Punting Angles: Chelmsford City Racecourse Part 2

In the previous article I focused on some angles for playing the polytrack at the Essex course at Chelmsford City, writes Jon Shenton. To be brutally honest, keeping the word count down to something sensible proved impossible and stumps were drawn as the light was fading late in the evening.

However, after a short break it’s time to pad up again, get back to the crease and finish building this meaningful innings. If you missed Part 1, or want to revisit it, you can do so here.

First up today, let’s look at some stallion data.

Stallion performance at Chelmsford

Using geegeez.co.uk’s Query Tool and evaluating all runs at Chelmsford with SP’s of 20/1 or shorter we get the following list of stallions with A/E values of greater than 1.00 (where they have had 100 runners or more).  The data is sorted in descending A/E order.


These articles have already discussed the merits of Lope De Vega progeny on all-weather surfaces, especially at Gosforth Park, Newcastle. That stallion also has a perfectly respectable Chelmsford record. Analysing “Lope” runners by race distance at Chelmsford gives the following picture:


There appears to be a distinct variance in performance between races of a mile or shorter and those longer than the 8-furlong trip. His progeny’s record beyond a mile is 4 wins from 33 whilst the numbers at up to a mile show a highly competent 18/72.


It’s not the most conclusive, or robust, angle in the portfolio but is worth tracking as it may develop into something a little more solid over time. If you have time, do re-visit Lope De Vega at Newcastle (see article link above), the stats are stronger for that course.

Top of the table is Medicean, so it would be impolite to move on without further reference to his progeny.  Again, here are the numbers based on race distance in the table below:


Like Lope De Vega, there is a split at around the mile distance: 6/47 at the longer trips and 20/92 over shorter.

Medicean retired from stud duties a couple of years ago so this angle has a limited shelf life, in truth it is probably reducing in relevance already. However, there are still winners to be had (Sharp Operator went in on the 24th September for example). It’s one to keep an eye on, rather than build as a cornerstone of a punting portfolio. Interesting yes, unmissable no.


5 furlong races

Let’s go back to the specific race distances, starting with the fast and furious five-furlong burn ups. The course map illustrates how they break near the bend at the end of the back straight.

Like some other courses I’ve evaluated in this series, the Chelmsford five has all the hallmarks of suiting a low drawn early pace speed merchant.

Evaluating in more detail using the tried and (semi) tested approach from part 1 sheds light on the hypothesis.


For those not familiar with the layout;  the table is a combination of draw bias in the left hand box (using the draw analyser IV3 numbers) and the Pace profile (Pace Analyser with IV) consolidated on one table on the right hand side, by number of runners in a given race.

Your first 30 days for just £1

For more detail on the numbers and what they mean I noticed Matt had addressed this particular subject in his “Silly Question Friday part 2” post, which you can find here.

The tables above cover all races over five furlongs at Chelmsford on Standard or Standard/Slow surfaces (very small number of events on the latter going) and relate to the actual stall position, not the drawn stall number (this simply adjusts for non-runners). It’s quite helpful that the maximum field size at this trip is 12 meaning there is a bit less eye-bleeding data manipulation to get through (secretly enjoyed!).

First impressions are that the bias is less apparent than I was expecting. In my mind I expected to see a sea of green to the left on the draw table (good) and an expansive pit of red on the right (bad).   Whilst there is undoubtedly a tendency towards those drawn on the inside, with stall 1 looking very healthy, it’s far from a binary profile. Plenty of animals are prevailing from wider stall positions. That said, the outside two stalls marginally underperform in almost all field sizes.

Pace, however, is much more clear cut. Shifting our gaze to the table on the right, we can see early leaders are universally green in nature with IV performance of a minimum of 1.5 in all cases. To be clear for those still not au fait with Impact Value (IV), that means early leaders are at least one-and-a-half times more likely to win than horses adopting other run styles.

Prominent runners fare reasonably well but those raced more steadily through the early stages generally have it all to do at the sharp end.

The main inference from these data, in reasonably strong terms, is that pace is of greater importance than stall position at five furlongs.

The best / easiest way of performing a quick check-in to see if this holds true is to use the heat map on draw analyser. In this case below I’ve taken data for field sizes of 8 and 9 (illustrating IV). However, it is straightforward to check other field sizes using the tool. As always drop me a note in the comments or on twitter if you need any guidance.

The exact numbers are always interesting; however, the colour coding shows you really what you need to know. The map does show that a low draw is perhaps more forgiving if an early leading position is not secured, but there is no doubt overall that ‘(early) pace wins the race’.

Whilst all of this is nice and makes perfect sense there is another side to the coin: the value side.

My pre-conceived belief was that low draws would be where the action is. When I wager at Chelmsford this is ingrained in my psyche and is always the first thing I look for. Whether this has been picked up through media talk, using Geegeez, or typically what I’ve seen at other tracks I’m not sure. But if I believed it, I surely can’t be the only one?

If I’m not alone then it’s highly possible that a low draw at Chelmo is in danger of being overbet. If the claim that pace is more important than draw holds true then maybe wider drawn, pacy animals are a great betting opportunity. Yes, sure, winners are more likely to be unearthed from lower stall positions, but perhaps the value is elsewhere with the market underestimating higher gate numbers.

The most effective way to check in the toolkit is to repeat the table format, but this time using the A/E number (again, details of A/E, Actual / Expected, can be found here). As with IV, the higher the number the better, with 1.00 being par performance (in a perfect world with no over-round for the bookies).


Interesting? The picture is choppy for sure, mainly due to the small datasets derived by analysing each stall position based on field size (manifesting a few zeroes, for example).

However, I’m confident that there is a greener hue to the right side of the table than the left; maybe not rainforest green but certainly including tinges of Kermit in comparison to the Bert-and-Ernie-like yellowness of the left-hand side.

This table is effectively confirming that the low stalls are broadly over-bet.

Taking stall one as an individual case study, in the first table in this section this berth has an IV3 of 1.28. It’s not a perfect measure but it sufficiently makes the point that winners are quite likely to originate from the inside box when compared to the average. The A/E comparison scores for stall one are all below that level (illustrated by the blue dotted outline), in some cases significantly.

The bottom line is that by backing trap one blindly in five-furlong races winners should be plentiful but cash will probably be conceded: the market has sussed it already.

To re-enforce / labour the point, below is a cut from the draw analyser which splits the draw into low/mid/high segments in field sizes of 8+.

The image confirms the assertion, namely generic low draws have an IV of 1.04 but an A/E of only 0.73. Conversely, high stall positions struggle in relative terms with IV (0.8) but have a higher A/E at 0.88.

However, when considering run style, we can see that those which led early – especially from wider out – have been very profitable to follow. Indeed, breaking fast from a wide draw may enable a horse to cut the first corner and carry more speed into that turn.


What does that mean? Simply that value can be found in the wider stall positions when there is early pace thereabouts.

In conclusion, with regards to the Chelmsford five-furlong range:

  • Finding the early leader (or at least a horse that is prominent) is the key factor in establishing a likely winner of the race
  • A lower drawn horse is more likely to prevail over the distance; however, there is evidence that the market overcompensates for the low draw.
  • A horse drawn in a middle to high stall is more likely to generate a long-term profit, especially if able to show early speed.


6 furlong races

Moving up by a furlong to the three quarters of a mile trip, runners start well down the back straight, thus giving jockeys and horses more time and room to sort out their positions before the bend. The maximum field size over this distance is 14. However, there have only been 31 races with a combined 13 or 14 runners so I’m going to leave these on the bench for the data analysis.


The table shows that, arguably, the bias over six towards low stalls is stronger than that over five. Most of the stall 1 and 2 data is green in nature, indicating that winners are more likely to originate from those positions than anywhere else. This holds true particularly well where there are 9 or fewer horses taking part.

Where there are ten or more participants the picture is less clear. It may be related to sample size (22 races with 12 runners compared to say 60 with 8 entrants), or it may be related to greater scope for congestion; but there isn’t anything too obvious – in my mind at least – to explain why the larger field sizes shape differently.

One thing that is not open to question is the effect of pace on the outcome of six-furlong races: yet again, being at the front end early pays handsomely.

Based on both the draw and pace details you’d expect a low drawn trailblazer to be of primary interest and, whilst that is true, as with the minimum trip pace seems to be the kingmaker. The heat map below shows IV performance for field sizes of 8-10, and is unambiguous in terms of how most winners race.


In this case the low drawn early speed combination appears to be almost unbeatable, but the enduring message is that if a runner is held up, dropped in or generally in the hustle and bustle of midfield it’s a big ask to pass the speed horses.

The same assertion made for five furlongs about lower draws being overbet could hold true over this course and distance too. I did repeat the full table treatment but for the sake of brevity here is the same broad-brush view covering all field sizes that have been analysed.

It’s the same story again, low draw equals higher overall probability of winning (IV 1.18) but the A/E doesn’t match it at 0.81.  But overlaying pace onto the equation is the route to profit, especially away from the ‘obvious’ inside berths.

A footnote on Pace

This may or not be of interest to some of you but it’s worthy of inclusion in my view.  When I started working with pace there was something gnawing away that didn’t sit right with me.

It comes down to the fact that any given race there is only one horse that is tagged as led/leader and there can be several tagged as held up, mid-division or prominent.

It is logical horses that also-rans are far more likely to be contained within the held up (or mid-division) classification. They start near the back and stay near the back! Could it be, then, that these no-hopers skew the data for the off-the-pace categories and in fact a quality hold up horse has the same chance of winning as a quality front-running animal?

To scratch this itch, analysing the performance of favourites by how they are ridden is a logical method. And what better way to do it than by evaluating the five- and six-furlong races at Chelmsford contained in this article?

The below table shows the performance of favourites in sprint races by early run style:

The pace aspect holds up! Leader favourites outperform the market with an A/E of 1.27 and a whopping IV of 3.89! Those market toppers which are dropped in have an A/E of just 0.62, which equates to a negative 40% ROI. Ouch. These data satisfactorily allayed my own curiosity and fears, anyway!

That about wraps up this Chelmsford two-parter covering as it has trainers, sires, and delving into races over 5, 6, 7 and 8 furlongs in fine detail. I hope you’ve found at least of something of use.

The regularity of racing at the all-weather tracks means data are more readily available than their turf counterparts and I’d fully recommend the geeky/curious amongst you to get stuck in to analysing racing on the artificial surfaces as a starting point.

  • JS

Punting Angles: Chelmsford City Racecourse Part 1

After a short summer break recharging the batteries (in theory) it’s time to get back to work and begin preparations for the onset of winter, writes Jon Shenton. That doesn’t mean National Hunt yet, I’m afraid. Rather, we’re going to get stuck into the polytrack of Chelmsford, hopefully stealing a march by doing some early research before the real all-weather schedule starts to kick in.

Chelmsford Overview

‘Chelmo’ has been a fixture of the racing calendar from 2015, ignoring its brief prior incarnation as Great Leighs in 2008, and is widely known for offering impressive prize money. In 2018 £5.2m was shared across 63 fixtures according to the official website. That, as well as the track’s proximity to the Newmarket training centre, has arguably led to a better quality of racing on this artificial surface than any of the others.

The track constitution is illustrated in the course map below. It is just about a mile in circumference and the turns are relatively broad and sweeping in nature, sufficiently so to develop a turf track to sit inside the current AW oval. There are chutes for the seven- and eight-furlong starts, more of which later.


Chelmsford Top Trainers

Before checking out specific race distances, we’ll adopt our usual tactic of scanning the trainer ranks for potential profit.

Usually in this series of articles data relates from 2012 to present day. However, as Chelmsford has only been up and running for four years, there are obviously less data to go on in terms of overall duration. However, that is more than compensated for by the fact that in its brief existence there have been over 17,000 runners at the track. To put that into perspective there have been fewer than 7,000 runners at Epsom from 2009 to date. All data in this article covers racing up to Friday 30th August 2019.


Trainer Performance

Using geegeez.co.uk’s Query Tool, the below table shows all trainers with an A/E performance of greater than 1.00, concentrating only on runners sent off at 20/1 or shorter; and there needs to be a minimum of 100 runners for a trainer to qualify.



The top pair of Charlie Wallis and Derek Shaw are certainly of interest, perhaps David Simcock too.  Aside from that at this helicopter level, there isn’t too much to get excited about.


Charlie Wallis

Wallis’s stable is based in Essex so the relatively high volume of runners at his local track makes sense. It’s noteworthy that the yard has a real all-weather specialism, with over 70% of their total runners appearing on artificial surfaces. Being a relatively new team (training since 2015), this may change as the operation develops and progresses. Until then, runners from the team are well worth monitoring here.

Analysing Wallis animals by the distance at which they have competed results in the following splits:


Sprinting is obviously a key focus. A large proportion of runs, wins and returns have been sourced over the 5- and 6-furlong ranges. For angle building I’m only interested in these short distances although you could easily argue that the sample sizes over further are insignificant and, in time, they may show similar performance to the sprints. That might be the case but I’m happy to stick with the larger samples up to three-quarters of a mile.

Wallis over 5&6 at Chelmsford puts up some nice numbers without too many more filters. If I were being a perfectionist, it’s preferable that one of his has had a recent run. Using horseracebase.com to drill down further, the yard has never had a winner (on any course) when a horse has been off the track for more than 90 days and, ideally, a run in the last 45 days would be optimal for this angle.


SUGGESTION: Back Charlie Wallis runners at 20/1 or shorter over 5 and 6 furlongs at Chelmsford, [Optional, exclude horses that have not run in the last 45 days.]


Derek Shaw

Moving on, Derek Shaw is another cornerstone of UK all-weather racing and, much like Wallis, a similar proportion (70% or thereabouts) of the yard’s activity is focused on the ‘sand’.

Checking the performance of his 248 runners by SP provides something on which to chew. The data below are for horses running at Chelmsford with an SP of 12/1 or bigger.

Your first 30 days for just £1


With only five winners from 81 bets and a strike rate of a smidgen over 6% in my view it’s marginal whether it would be worth fishing in that pool long term. However, taking all Shaw Chelmsford runners at 11/1 or shorter we build a potentially compelling picture when further analysing by race class:


It’s crystal clear that there is a division between performance in classes 2 and 3, and in classes 4-7. I think it’s not unreasonable to assert that the Shaw string wouldn’t contain the best raw materials in terms of racing talent with which to work. Perhaps some of the better class races are just a notch too high for the animals at Shaw’s disposal.


SUGGESTION: As always, simplicity is best and that’ll do for me, back Derek Shaw at Chelmsford in Class 4-7 races where the SP is 11/1 or shorter.


Without too much delving, those are two straightforward angles to file away in your QT Angles for use over the main all-weather season and beyond.


Ian Williams

One other worth bringing to your attention, though just bubbling under the 100-runner level (with 96), is Ian Williams. I don’t propose to go into detail here, but his numbers are worth keeping in mind (and perhaps researching yourself if you have the time).


Short Priced / Fancied Runners at Chelmsford

As you may have noted from previous columns, I’ve started to get a bit of a taste for angles focusing on shorter price runners. The table below simply illustrates the record of trainers where runners have an SP of 5/1 or shorter (50 runs minimum).


Obviously, there is some duplication with the trainer data presented earlier; Wallis and Shaw predictably are prominent (Williams too). Of the others, at first glance Messrs. Dwyer, Tate and Easterby appear to be potentially worthy of shortlisting when the cash is down. No doubt that some of these could stand alone as angles. However, before piling in it would be highly advisable to check consistency of performance. Based on samples of this relatively small magnitude it is perfectly plausible that the inclusion on this table is attributable to a golden year or two.


Specific Chelmsford Race Distance Analyses

One mile races

One advantage of the all-weather is that we can almost take one of the key variables in racing out of the equation. Changes in underfoot conditions are less prevalent, though weather variance can affect the surface more than the official going relates; so, coupled with the abundance of meetings on the AW tracks, there is nearly always a rich source of data regarding pace and draw to delve into.  Virtually all races at Chelmsford are on Standard, however, there are a handful of Standard/Slow contests included in the analysis from this point onwards.

Our first zoom into the profile of a specific trip is over the mile. If we refresh our memories from the course map, the start is located in a chute and there is approximately a furlong and a half of racing prior to the first left hand bend, where the runners join the main track. That does not give much time to secure a good early position, and being trapped deep on that first bend is a realistic danger.  In other words, there is a whiff of low draw bias about the set up here, especially in bigger fields.

I’ve compiled the draw and pace data and attempted to consolidate it in a single table. At first glance it may appear complicated, but hopefully with a small bit of explaining will be quite simple.


The table is basically a mash-up of draw bias (using the draw analyser IV3 numbers) and the pace profile (Pace Analyser with IV) consolidated on one table by number of runners.

A quick refresher of what IV3 means: it is simply the average Impact Value of a stall and its nearest neighbours. For instance, the IV3 of stall six would be the average IV of stalls 5, 6 and 7

The numbers are one thing, and the colours are another, but what does it mean and how can the insight be used to optimise our chances in finding potential winners?

Broadly speaking, the greenish tinged numbers represent good performance with the red ones conversely not so good (as Sven might say).

Without doubt, there are more green shades in the lower box numbers, indicating the expected low draw bias. This appears to hold true for all field sizes too: an inside draw is a positive when assessing the merits of an individual horse.

Moving across to the pace box to the right, the green numbers are concentrated around the leading and prominent runners.  At first glance it looks like an early-to-the-lead horse is the most desirable.  On closer inspection, however, we see that a prominent runner is arguably as valuable in terms of winning potential for most field sizes. The deep green relating to leaders in 13/14 runner races fields (data based on small samples of 24 and 20 respectively) gives a possible visual skew to the data.  What is in no doubt is that being up with the speed is highly desirable and, related, hold up horses generally have it to do.

Low daws are good, and early speed is good, but what happens when they are combined? That’s where our old friend the draw/pace heatmap (found at the bottom of the DRAW tab on flat race cards) can offer some valuable insight.

Evaluating races where the number of runners is between 7 and 10 inclusive over a mile (chosen as they are the most common field sizes so sample size is larger) and consolidating in the heat map (IV) we get the following composition.

Heat Map of mile races at Chelmsford with 7-10 runners inclusive using IV


The heat map paints a very clear picture:

  • Low draws are desirable irrespective of run style
  • For those drawn in the middle, a prominent or leading style is preferable
  • For those drawn high, a front running style is the only favoured approach 


7 Furlong races

Before wrapping up, from reviewing the course map I thought it may be interesting to use the same approach over the seven-furlong trip. The hypothesis is that a low draw may be of less relevance as horses and riders have a full three furlongs to get a position before the first turn. Thus, it ought to be possible to negate the risk of being trapped out wide and, therefore, potentially ease the sort of draw bias seen at the mile distance.


Alas, the hypothesis doesn’t hold true as the data indicate that there is still a form of bias towards lower stalls when viewing through the prism of IV3. That said, the draw doesn’t appear to have too much effect until field sizes of nine or more are experienced. In the broadest terms stalls 1-6 seem to be better off than stalls 7 and above in almost any circumstances.

For larger fields of 13 and 14 runners there appears to be a strong bias to the lower numbers although, again, sample sizes are smaller. Usually that can be attributed to getting out of the gates and securing good track position early on, ordinarily up with the speed and avoiding hazards in running brought about by a congested field.

Again, a quick check of the heat map can help:

Heat Map of 7f races at Chelmsford with 13-14 runners inclusive (IV)


This view is only comprised of 41 races but it’s clear that a horse in a low stall has a stronger hand to play than its wider-drawn competitors in the biggest field sizes. If that same low drawn horse leads it has an IV of 3.81 which means, it’s nearly 4 times as likely to prevail as the average!

In the second part of this Chelmsford epic, I’ll cover sires, jockeys, the fate of favourites, as well as the impact of draw and pace on 5- and 6-furlong races.

Until then, thanks for reading.

- Jon S

Punting Angles: Haydock Racecourse

For this edition of Punting Angles, I’m going to concentrate on the enigmatic Haydock Park, writes Jon Shenton. Whilst the course is home to both National Hunt and Flat racing, it is the latter that I’ll be evaluating in this edition given the time of year. For whatever reason, it’s one of those tracks that seems difficult to read, racing developing on both rails and atop a seemingly unique range of underfoot conditions, "Haydock Soft and Heavy" almost becoming an official going description in its own right. The track is synonymous with horses slogging through bottomless ground in pursuit of glory.


Haydock Going

Although there isn’t too much of punting value in it, I still felt it would be of interest to benchmark how Haydock shaped compared to the rest of the UK in terms of the official going for races over the last five years.

The below graph illustrates the ground conditions for UK flat races.


Immediately, the orange bars relate that Haydock proportionately has much more racing on the easier side of good: it’s nearly four times as likely to race on heavy than the UK flat average too.

Knowing that may not directly help in the pursuit of profit; however, searching for mud-larks or horses whose sires loved sploshing around in the deep ground may be a pragmatic activity in preparation for a wet Haydock meeting.


Haydock Course Constitution

What about the course itself? The map below illustrates a tight loop with a straight up-to-6-furlong run. Based entirely on the map alone, curiosity is piqued with regards the draw for the 7-furlong trip. The journey from the stalls to the first bend appears to be an exceptionally short one implying that a wide berth could be a problem.



Haydock Trainer Angles

However, before checking that out and getting into individual race specifics, let’s first take our usual glance into the world of trainer performance. Using Query Tool, I'm evaluating all runners from 2012 with an SP of 20/1 or shorter. A minimum of 50 runs is required at an A/E of greater than 1 to secure a position in the top trainer table below.



There is much upon which to mull here.

Despite the relatively small number of runners, Hugo Palmer's figures offer something to satisfy even the most voracious appetites in angle finding.

Keeping it simple by analysing the yard's runners by market strength, the following split in performance is observed.


There is a very clear dichotomy here in terms of a Palmer runner which has been supported versus one that has not. Notably, with a place record of 70% on those runners sent off at 11/2 or shorter it has been even more profitable to back each way, albeit to double the stake.

Of course, on a sample size like this, it only takes a single 20/1 winner on the next day to create a profitable “unfancied” segment. The cliché of fine margins applies without doubt and there is a chance (probably a 20/1 one) that I'll endure an element of regret as a Palmer animal bolts up at a big price here in the near future. However, with my cool data-driven mind I’d rather sup regularly in moderation than binge excessively once in a blue moon. Albeit there is a place for both, I’m sure!

Suggestion: back Hugo Palmer runners at Haydock where they’re supported in to 11/2 or shorter


Second stop on the trainer list is the seemingly ubiquitous Tom Dascombe who has a prolific volume of runners at the track.  It’s impressive in terms of pure scale but to beat the market with an A/E of 1.15 over so many runners is quite a rare feat.

Looking a little more closely, there is a noteworthy split along race class lines.

Your first 30 days for just £1


The competitive stuff of class 1 and 2 racing is obviously more of a challenge and it is likely that these races attract more top yards and animals thus making it tougher for the Dascombe runners to prevail even with home advantage. Whilst the performance of his charges is perfectly respectable in those upper echelons, I’m happy to pass them over in punting angle terms.

There is also a clear distinction between relative performance in handicap / non-handicap races in the Class 3 to 5 races, as can be seen below:


The table is unambiguous: handicap racing is where it’s at for Tom Dascombe's Haydock horses.

Suggestion: back Tom Dascombe Handicap runners at 20/1 or shorter in Class 3, 4 and 5 races


[As a footnote, it’s not unheard of that the yard fires in the odd big-priced scorer at the track. If you’re inclined to play with fire, there are worse places to go than a Haydock Dascombe runner to satisfy those punting pyromaniac tendencies. In C3-C5 races he has had 6 winners from 65 at SP odds of 20/1 or greater across all race types, which is somewhere between great and probably unsustainable!]


Ordinarily sticking with a deep dive in to the record of two trainers from the top table would be enough, but seeing John Gosden playing a prominent role means it would be negligent to let him pass by without further evaluation.

A good starting point with Gosden, Stoute, Charlie Appleby and the like is always to check performance by distance given their predominate modus operandi is to dominate the middle-distance division. Checking Gosden entrants at Haydock by race length gives the below breakdown:


Sure enough, sprint race performance is less compelling. Plenty of winners, yes, strong IV, definitely; but finding winners doesn’t always mean value.  Despite a strike rate of over 20% the return over 5, 6 and 7 furlongs is not a positive one. To eke out profit it appears as though an even healthier strike rate of over 30% is required. Happily, this has been apparent at distances of a mile or greater: 31 winners from 93 runs, a nice round third. A 68% ROI is not to be sniffed at either.

Suggestion: back John Gosden runners ay Haydock at 20/1 or shorter from races of 1 mile or greater in distance

Haydock by race distance

Moving on from the trio of trainers, let us now evaluate the shape of races at the track.

7 furlong races

As you will recall, I was particularly interested in the 7-furlong trip based on the course map. The left-hand bend which is seemingly close to the start could result in some interesting pace and draw snippets, with an expectation that low draws close to the rail should have the best of it.

Using the Draw Analyser Tool and combining the heat maps for draw position and field size for Impact Value (how likely a horse will win from that position with 1.00 being par) we get the below data to evaluate relating to the whole spectrum of ground conditions.



Conclusions are probably less obvious in the data above than they have been in previous articles. However, there are still some noteworthy and useful outputs to consider.

Firstly, evaluating hold up runners, there is a big stripe of red (Red Stripe?) confirming that it’s a tough gig for a horse that’s held up to win from a high draw. This also applies to middle stall positions in larger sized fields (see the top of the dotted black box).

In general terms it is spoken often that a high stall position is less relevant for animals who race by stalking from the rear as they can drop in and wait. However, these data show that adopting the waiting tactic from wide over seven at Haydock is not generally a good plan. This is surprising, or at least mildly counter-intuitive, as the home straight is over half a mile in length, giving hold-up horses plenty of time to wind up and make their move. It is hard to argue with the facts, though.

The data also appear to indicate that it’s not a productive strategy to try and secure an early leading or prominent position from a wider draw in middle- to large-sized fields (the red coloured zeroes on the table). Through watching race replays this starts to make sense. Horses sharply away from wider stalls have only a small run to get across to the rail / near the front prior to the the bend; if they fail to get a front berth quickly enough, they face the issue of being trapped wide and covering significant additional ground.

Even if they get out apace, given the proximity of the bend, all it takes is for one or two from inside stalls to be away well and it’s difficult for the wider drawn speedster. As the turn develops, our wide trailblazer has the choice of burning through more juice to get to the front or travelling further: both potentially terminal to the chances of him or her winning the race.

When investigating this, I expected a variance based on ground conditions. Surprisingly though, and broadly speaking, the conclusions work in all racing circumstances. If anything, high draws have it even tougher as the going deteriorates though data is quite sparse from which to draw anything more than lukewarm conclusions.

Low drawn and/or early pacers look to be generally the best bet on a consistent basis (the blue dotted boxes), as expected. Although, like a lot of other courses, it appears as though early speed is of more importance than draw, unless the horse has a starting position in the proverbial car park in a big field, as already mentioned.


Haydock Straight Track races

As noted previously the straight track at Haydock stretches for trips up to six furlongs in length.  Evaluating the draw using the analyser tool again for all ground conditions and number of runners for the straight track paints the following picture:

The data displayed relates to the IV3 calculation. I’ve picked this to try and smooth out variance and noise to help illustrate a general pattern. A definition of IV3 is simply an average Impact Value of a stall and its nearest neighbours. For instance, the IV3 of stall six would be the average IV of stalls 5, 6 and 7.

There is an indication that, irrespective of the number of runners, there is a definite bias to horses with higher stall numbers. The red and amber colours represent draw positions which are less likely to house the victorious horses. Green is good and it would be anticipated that based on historical performance horses that prevail are more likely to start from these stalls.

There is a line displaying the straightforward mean average (AVG) for each stall position which reminds me of the pH chart from my gruelling chemistry lessons in times gone by (even though it’s the wrong colour). Anyway, progressing from low to high the hue gravitates from an angry dark orange to a lovely tree-hugging green, showing that, on balance, low stalls are not the place to be drawn.

With that intel in the bank we can add a bit of spice by applying some complementary pace data.

Taking the minimum trip alone in the first instance.


The data is split by field sizes, and the information shown represents the IV data for each run style by underfoot conditions. The column entitled "races" simply represents the number of races that are included in each line of data. You can then draw your own conclusions based on sample size.

The field size 12-16+ (data on the left) has a limited sample size contained therein so firm conclusions are not sensible. Having said that, on a sounder surface it appears to be a challenge to win from the very rear; instead front runners boss proceedings. Whilst on the sludge of a Haydock soft or heavyTM surface, hold up (and mid division) run styles seem to be the way to go.

Is there something in that? Maybe, but I’m not sure as we only have eight races to go on for those relatively deep ground conditions. Sectional timing may help us understand these things one day as it may be that too fast a pace has collapsed setting things up for the closers on these small samples.  As things stand, it’s a bit of a leap of faith to assume it’s a true representation. Nonetheless, it is interesting.

The smaller field size data contradicts it, sadly. Broadly speaking, it pays to be on the pace over the five furlong range, where the prospect of less pace contention and, therefore, an ability to rate energy more efficiently is manifest.

Moving onto 6-furlong races, we get the following:


There is a little more data to go at over this distance. For all field sizes, horses that lead or are prominent early are most likely to win across all surface conditions. In larger fields there is an absolute bias to the lead horse.


Closing thoughts

In conclusion, when evaluating a Haydock race on the straight track the first item to look for is early speed, the second item should be a high-ish stall number. If both those boxes are ticked then it could be a good play, depending on the animal, of course.

A Dascombe or fancied Palmer runner on the straight track with early speed and a high draw would be a very exciting prospect, until it misses the kick anyway!

 - JS

Punting Angles: Beverley Racecourse

Now the 2019 Royal Ascot carnival is confined to history, my attention once again turns to finding some interesting insights on some of the UK’s less glitzy racecourses (which, in their defence, is all of them!), writes Jon Shenton.

Ascot may well be a full bodied, world-class (and expensive) Michelin star racing experience, but sometimes a hearty pub meal and a couple of pints hit the spot like nothing else.

So, clumsy metaphor out of the way, for this edition of punting angles, I’ll be focusing on the picturesque Yorkshire track of Beverley.

I enjoy this particular northern circuit. Fast and furious large field sprints spring to mind, as well as a large cast list of trainers, jockeys and owners which should lead to some reasonable angles and opportunities.

Let’s first take a look at the course map:


The diagram illustrates a couple of seemingly devilishly tight turns.  There is also a stiff uphill finish, the little red triangle pointing upwards indicating the highest point on the course is shortly after the finishing post, the lowest point being diagonally opposite. Thus, the final two and a half furlongs are a gradual climb, testing stamina as well as speed over shorter distances. Horses competing at the 5-furlong trip face an uphill task literally almost every step of the way.


Beverley trainers

Before analysing specifics regarding race distance profiles, a customary evaluation of trainer performance is in order. My starting filter is that all races from 2012 are included, but only where the runner SP is 20/1 or shorter.  50 runners are required to qualify in the table.


To be perfectly honest there doesn’t appear to be too much on here to get overly excited about.  The duo at the top of the table, plus potentially David O’Meara are probably the ones to focus on, if any.

Taking Richard Guest as an example, there is a definite point of interest from a punting perspective.  The Wetherby based operation has had only one solitary winner where there has been a SP of between 11/1 and 20/1 as the numbers below display:


The 10/1 or shorter row is a clear indication that market support for a Guest representative is a significant factor in assessing the likely performance of a stable runner.

Not much more to say on that, in truth.

However, whilst I was evaluating runners at the skinny end of the market, I noticed something that I think is worth bringing to your attention. Step forward, Mrs Ann Duffield. Hers is a yard I haven’t really taken a great deal of notice of previously from a data perspective: despite being a regular on the circuit I’ve never established anything robust in stats terms relating to runners from this stable.


The table above displays Duffield runners at the track, segmenting them between fancied and less well fancied runners. The delta between the two is noteworthy: just two winners from 127 runners at 6/1 or greater with a painful A/E of 0.25, IV 0.26 and a bankrupting ROI of -61%. Compare this to a strike rate of 35%, 54% ROI and IV of 3.05 for the shorter-priced entries and it’s looks like it’s potential party time when Duffield horses are towards the top of the market.

We can go slightly further:


Your first 30 days for just £1

This table shows the 11/2 or shorter SP data by odds rank, i.e. the position of the horse in the market, 1 being favourite. It may be an arbitrary point, but it certainly appears as though there is a differential between the horses residing in the top two of the market and ones further down the pecking order.


I’ve earmarked an alert in my portfolio to track any Duffield shorty that is at the top 2 of the market at Beverley and is less than 11/2 in price.


I’m learning that this game is all about constant evolution, by nature I generally search for horses at the more speculative end of the market. However, the more data I crunch the more I’m learning to appreciate these sorts of shorter priced opportunities. They offer balance and, in a world of risk and reward, such angles can keep the wheels turning when the more ambitious plays are stuck in their inevitable ruts. At least in theory, anyway.

There are other trainers (Brian Ellison, Michael Bell to name but two) where market support appears to be of significance. The table below is for your reference and contains the Beverley A/E performance for each trainer for each odds bracket. To qualify, a minimum of 25 runners in the 6/1 or less category are required. These data hopefully show how I stumbled onto the Duffield angle.    This approach will, I think, become a staple of how I evaluate trainer and market support in the future.


That’s enough about trainers, maybe a bit too much in fact! Turning the focus, now, to a couple of the specific race distances that the course hosts during the summer months.


Beverley Five Furlongs

Over the minimum trip of five furlongs, races start from a chute beyond the home turn at the bottom of the home straight. It is not around a bend as such, but there is a pronounced dogleg to the right at about halfway and a general curvature in that same direction for much of the trip. This ordinarily would point to a low draw bias, as the low stall numbers are situated towards the far rail, therefore offering the shortest route to the finishing post.

As a result, it would make sense that low draw numbers generally prevail in 5f contests at Beverley given that topology of the trip. The numbers confirm the theory.

Draw bias (IV) at Beverley for races at 5f by field size on Good to Firm, Good and Good to soft ground


Using IV for races with ground conditions of good to firm, good and good to soft, the above table certainly points towards a low draw as being the place to be; or perhaps more accurately a high draw is the place to avoid.

There seems to be an indication of a low draw becoming more advantageous as the number of runners increases. This certainly makes some sense, any advantage from the general curvature of the track from a low draw should increase as the physical distance in starting position becomes greater between the low and high wings.

There is always (or should be) a companion piece when analysing the draw, namely pace. We already know that early speed in a general advantage in these sharp sprint races from Dave Renham’s excellent series on early speed.

Again, Impact Value (IV, or how often something happens in relation to its peer group, where 1 is ‘normal’ and the further away from 1 is better or worse) is my weapon of choice. The visual below is an attempt at recreating the heat map within the draw analyser but for multiple field sizes in the same table. Its content again covers the more rank and file ground conditions from good to firm through to good to soft. [When the ground is soft or heavy, the draw bias at Beverley can reverse with runners often making a beeline for the near side running rail].


To my eye, early speed is important almost irrespective of stall position. It does reaffirm that the larger the volume of competitors the more challenging it is to prevail form a high draw. The big fat zero for Led, High Draw and a field size of 14+ relates to 10 runners, 0 wins and only 2 places.   Not big numbers but a nil is a nil. More importantly, logic supports the notion that these runners are significantly inconvenienced by race conditions.

Prominent runners do remain competitive throughout, perhaps with a notable bias to the lower side of the draw in the medium and large fields. Horses with mid-division and held up run styles face both a literal and metaphorical uphill battle and a lot to overcome.

Of course, nothing is impossible, and any horse can win any race, as our editor is always reminding us! Even the red ‘danger zones’ in the table are generally populated with numbers above zero, meaning at least some winners are found even in these relatively hostile environments. It’s about playing the percentages, however, and hopefully by using data such as these, small incremental improvements can be attained to improve long term results.


Beverley 7.5-furlong races

I’m acutely aware that there is a danger of sounding like a broken record here.  However, the adage of pace wins the race is seldom more apt than in relation to events over the 7.5-furlong distance at Beverley.  The actual official distance is 7f and 96 yards so do bear that in mind as it can be advertised as a plain old seven. Those extra 90-something yards can be of critical importance, especially with the stiff uphill finish coming into play.

To get a feel for the track I find it’s a sensible idea to sit back, perhaps with a cold Peroni (or other suitable equally enjoyable beverage) and take in a few race replays. The course map shows those tight turns but by perusing visual evidence it’s much easier to comprehend and, ultimately, to bring data to life.

Even without the support of stats this trip has all the hallmarks of being a front runner’s playground.   Happily, this can be checked using Query Tool to confirm the hypothesis or disprove it.

I’ve used QT in this case (as opposed to the pace analyser) as I want to compare our subject matter course with other tracks in the UK.


The table above contains data relating to the fate of front-running animals at a 7f trip (to the nearest furlong). The query filters are simply, all races from the 2014 season up to 7th June 2019, Distance 7f and a Pace Score of 4 for the runner to denote front running status; and I’ve sorted by win percent.

In the UK, only Chester has a bias towards pace greater than Beverley at this distance.  It’s a great benchmark as we all know the benefit of early speed at the Roodee. For Beverley to be in the same ballpark is a pleasant and potentially useful surprise. The trailblazers have a very strong record with close to 30% maintaining their advantage at the line, 56% hitting the frame, a very high A/E of 1.61 and a super high IV of 2.68. That’s a rock-solid foundation to build upon.

Expanding on this a little, the numbers in the green and red table below represent the overall pace profile of the 7f trips of the courses in question.



It’s very interesting to note that not only does Beverley have a confirmed and pronounced front running bias, a prominent running style also scores well in comparison to the other tracks.

In simple terms, it’s more important at Beverley to pick a horse with a prominent or pace-setting profile than virtually everywhere else (at this distance). Even if a horse doesn’t lead, the closer it is to the front of the pack the better.

The corollary of this is that hold up horses have a very moderate record over the course and distance (A/E 0.45), the poorest of all the listed tracks.

Field size is worth consideration when analysing front runners. It’s logical to assume (and an obvious point to make) that it’s easier to get a lead in a field of 4 than of 14 for example. The graph below shows the A/E performance of leaders at Beverley over the 7.5 furlongs range based on the number of competing horses.


I’ve excluded a handful of data in the graph related to races with 4, 15 and 16 runners across a total of 7 events, as it’s not helpful given sample sizes are extremely small.

The performance line tracks upwards, demonstrating that A/E improves as the number of runners increases. This works from a sense point of view as horses racing off the pace have a huge challenge to overcome, and simply, there is more of them in bigger fields. An abundant volume of runners means less racing room, so picking up and sweeping by the field is a big ask with a relatively short straight of only 2-and-a-half furlongs: advantage front runners. Importantly, with A/E being a measure of implied profitability, these data show that if you can consistently predict the front runner(s) in larger fields at this course and distance there will be due reward.

That’s all for this edition of Punting Angles. Hopefully there’s plenty to put to work in your own Beverley betting, and don’t forget that the tools here on geegeez.co.uk – especially Draw, Pace, and Query Tool – can give you this sort of leg up at any track you care to look into.

Please feel free to drop me a line with your suggestions, questions or comments. I’d love to hear from you.

  • JS


Punting Angles: Goodwood Racecourse

With the Goodwood May festival upon us it seems as good a time as any to apply a little focus to the Sussex track, writes Jon Shenton. The hope is we'll discover a few snippets of info along the way to boost our chances of a profit at the course over the rest of the summer and beyond.

Racing has been an integral part of Goodwood history since 1802. The track is synonymous with the Glorious Goodwood festival with its three Group 1 races taking a prominent position in the racing calendar. There’s much more to racing at the course than that shiny centrepiece, however, and there are plenty of other meetings and notable races to enjoy throughout the flat season.

As usual, let’s start with the training performances at the track.

Goodwood trainers

The table below shows the trainers who have high-quality records at Goodwood. The data within relate to runners with a maximum SP of 20/1, and are sorted by descending A/E and include all races from 2012 onwards. To qualify the yard must have had a minimum of 50 runners during the period.


For those of you who have been following these articles over the recent past, you may remember the Mark Johnston editions. I don’t want to trample over old ground but suffice to say Johnston handicappers who have had a recent run (last 25 days) are a serious proposition. Whilst this angle wasn’t as lucrative in 2018 as previous years the performance has been consistently impressive over time. Those are horses to keep on your side. A link to that article is here.

The other notable name to take from the trainer table is William Haggas. It’s the Impact Value of 2.52 which immediately draws the eye, especially on what is a healthy relative volume of runners. As you might expect with stats such as these there is a general all-round excellence contained within.  The only significant and justifiable enhancement I could establish is associated to the age restrictions of races.

The data show, amongst other things, that his record with 2yo's and in 4yo+ races is perfectly respectable. It isn’t, however, quite as sharp as the races containing horses aged 3. It’s low volume stuff though, with the place performance consistent across all  groupings indicating that I might be looking for something that is not there. As a result, I’m not convinced that there is an angle beyond keeping Haggas horses firmly in your cross-hairs at Goodwood.

Moving on, those yards which currently have less than desirable records at the track we get the following picture:

I was very surprised that Saeed bin Suroor is top (or bottom) of the pile with an A/E of just 0.52. Messrs Fahey, Varian, Hannon and Balding are all on the list too with Fahey’s runners returning a strike rate of a meagre 6.6%. Perhaps these yards are too proficient to stay on this cold list indefinitely and the cream will rise to the top in due course. Here and now, though, the numbers demand we proceed with caution.

To complete the trainer view, the table below contains the best trainers (in terms of A/E) for Goodwood during the month of May, perhaps offering a couple of clues for the next few days. Not that I’d advise anyone to back runners from these yards blindly but there are some impressive numbers in here, Roger Charlton most notably.  That said, there is a danger of the data reverting towards the mean based on such small volumes.




In terms of the pilots, the data below show all active riders with an A/E of greater than 1.00. William Buick probably has the stand-out record. Again, all round excellence means dedicated deeper focus angles are difficult to find.

The deadly duo of Norton and Fanning have a very close association to the Goodwood-friendly Johnston yard. Therefore, it would be reasonably logical to assume that their records could be attributable to the trainer connection. The intel below shows that whilst that is undoubtedly true, when they are jocked up on rides for other trainer, performance remains largely in line.


Whilst this may be of limited interest in isolation, I think it may lead towards a question of pace. In general, Fanning and Norton are considered to be enterprising riders at the front of a race. Perhaps they prosper at the track irrespective of who is employing them because of their propensity to effectively judge pace from the front? More on that shortly.


Straight track pointers

Your first 30 days for just £1

As the course map below illustrates, Goodwood has a complex array of starting points, routes and undulations.  The least confusing element is perhaps the confirmation on the map that there is a straight track for races up to 6 furlongs in distance.


Before searching for clues on how best to tackle the straight course, it must be noted that analysing the factors of pace and draw (like I’m going to here) in a broad way is a challenge. There are several variables that need due consideration, field size and ground conditions being the primary drivers of variance in determining how the race unfolds from a pace and draw perspective.

Fields here can range from 2 to 20-something, and underfoot conditions obviously can vary meaning that many multiple permutations can exist. All the same, there is merit in attempting to decode the data.



First let's look at the draw.

Using the draw analyser tool in Geegeez Gold the table below shows the performance of horses, by draw segment and based on the number of runners in the race, using Impact Value. I’ve only analysed races with six or more runners and I've used the actual drawn position (i.e. accounting for the effect of non-runners) rather than the race card drawn number.


The data covers all race ground from Firm through to Soft.  As noted in the above paragraph going conditions can have a significant impact on draw stats. However, in the case of Goodwood it’s fair to assert that the numbers on display are reasonably representative of the whole spectrum of ground challenges faced by the animals.

Here is a graphical representation of the very same data.


I include this as I think it illustrates a clear picture: horses that are drawn in lower or middle stalls are far more likely to prevail than horses drawn in high stall numbers on average. This applies to all nearly all field sizes (apart from arguably in 8-10 runner fields where the delta appears marginal) and to both 5- and 6-furlong distances.

The red line (representing those animals with a high draw) deteriorates the larger the field in general terms, especially if the race comprises of 11 or more participants.

The highest drawn are stationed on the stand side rail, nearest the cameras, the numbers thus progressively moving lower towards the centre and beyond to the far side. Racing usually develops between the middle and that stand side rail as a few horses generally tack across in that direction.

A rail is often an asset to have nearby but for this track it appears to be far from the case. Let’s complement this with a sprinkling of pace data using the Pace Analyser tool in Geegeez Gold.

The table below is based on the same conditions as the draw data above:


It is perhaps unsurprising that being on the speed early is an advantage over the sprint distances.

Putting both pace and draw together you’d expect a low/middle draw with a prominent or front running run style to be optimal. We can validate this by checking the draw/pace heat map (in Geegeez' Draw Analyser).

This picture covers  5- and 6-furlong races, on Firm through to Soft where there is a field size of 9 through to 12.

Interestingly, it appears as though a high draw is acceptable if the horse can zip out of the gates and secure an early lead. It could be claimed, using this data, that pace is of more importance than draw. High drawn horses who get to the front are 1.43 times more likely to win than the average in spite of the ostensibly challenging stall position.

That makes sense: racing room can be at a premium at Goodwood and it’s very feasible that horses get boxed in, especially in a big field. Those high drawn animals can have nowhere to go if horses congregate and the race develops on or around that rail or side of the track. The jockeys starting their journey from the low and middle numbered stalls should have more options to avoid trouble in running; unless of course the field sizes are so large that the low numbered stalls are situated on the far side rail as in, for example, the Stewards' Cup.

A heat map taking account of field sizes of 14 or more confirms the thinking:


In large fields even prominent racers struggle to get the run of the race from a high stall position, probably due to the relative lack of options in running. Horses drawn low retain a degree of flexibility in how they approach the race and can win from off the pace. Now all that remains is to find the right horse that this might apply to on race day!


Round course and longer distances

The 7-furlong trip has just shy of a quarter of a mile from the stalls to a tight right-hand bend into the straight. Most races develop on the far rail, the opposite to the straight track races.

Again, early speed holds sway. Attaining good track position at the bend is clearly of primary importance. Evaluating the draw for the trip over seven using the graphical format (below) shows the significance of stall position.

Whilst it’s reasonable to say that low draws generally have an advantage it only appears to become a concerted one in double digit field sizes. In these larger fields low drawn speed merchants around the bend are very much of primary interest!

In smaller fields pace is still an advantage but, naturally enough, draw appears to be less relevant. Like the straight course, high draws are perfectly fine if you think your horse can get to the front early and control the fractions. In basic terms, if you can pick the leader early in the race consistently over seven furlongs at Goodwood you will have a strong hand to play over time. The same principles apply over the mile too.


Distances greater than a mile

The races between nine furlongs and two miles are represented from a pace angle in the data presented below. There is perhaps a marginal preference for front running speed in general apart from the shorter relative distances (9 & 10f) where early speed is a significant advantage.

It’s repeating the same message: the major takeaway from the data is the reinforced view that it is  difficult for hold up horses to win in larger fields.  That makes perfect sense given the tight and undulating nature of a track where hard luck stories seem commonplace. Let’s hope that you’re not on  one that falls out of the stalls!


That’s it for another edition, I hope you find things of interest in the above and I’ll certainly be watching Goodwood races with a keener eye than usual over the next few days and months. Good luck!

- Jon Shenton

Punting Angles: Windsor Racecourse

I was toying with a couple of subjects on which to base the content of this article, writes Jon Shenton, when I read a highly enjoyable edition of David Probert’s thoughts on this very site, which you can find here.

My eye was drawn to the section on Windsor.  It’s a course that historically I’ve had a patchy record when betting, so, inspired by Mr Probert’s words, I elected to dive into the deep Windsor / Thames waters to try and find some data driven treasure.

The course is a staple of the flat season: with its regular Monday slot in the calendar it forms an important part of the campaign due to the high volume of meetings and central location for many trainers. For the army of 9-to-5'ers like myself it also offers a chance to kick back and watch some racing due to the usual evening nature of the meet.

The fate of the favourites

Whilst researching, the first factor that made me sit up and take notice was the performance of the jollies. Using geegeez Query Tool (odds SP rank = 1, races from 2012 season onwards) the below table shows the tracks with the best performance by favourites ranked by A/E.

The numbers are certainly of interest regarding the Berkshire course. Surprisingly, backing the horse (or horses if joint favourites) at the top of the market in every single race during the last seven years would have returned a 4% profit, with close to a 37% strike rate.

It’s one thing understanding that favourites generally do well but why is that the case? Sadly, I have no answer, only several hypotheses. The evening nature of the meeting gives punters longer to study? The clientele who frequent the meeting are casual punters who back proportionately more outsiders, therefore boosting the value at the shorter end of the market? The “vibes” behind the fancied runners are stronger at the track? Or maybe the nature of the track plays to form more than others. It’s difficult to pinpoint specifics; the volume of data does make it hard to ignore however.

Before delving into further detail (and as a bit of public service) here is the view of UK tracks which have the worst performing jollies over the same time period.


Earlier I raised a potential hypothesis in relation to the time of the meeting being a factor, more study time effectively meaning the market becomes more efficient. It doesn’t feel like it could be a credible factor? Well, data talk! If we take our info and analyse it by meeting time, there is perhaps a surprising outcome (using horseracebase for this aspect).

Favourites prevail 4.5% more often at evening meetings than at afternoon fixtures! An 8.3% ROI has been attained during the later meetings, opposed to small loss during the earlier timeslots.

In truth, I’m not sure what to do with this intel, and it could obviously be mere happenstance, but thought it to be a worthwhile detour and if you have any theories please do share them in the comments. I did check other courses with a significant number of night meetings and the difference was certainly less noticeable than for this track. Perhaps this is evidence of potentially reading too much into data with no sound reason behind it.

Parking the time of day theme, for now at least, and getting back to evaluating the market leaders in Windsor races, I next assessed the age of the protagonists. This starts to paint a picture of where additional focus may be a rewarding exercise.


Two-year-old favourites at Windsor

The younger end of the age spectrum appears to be the area to concentrate on; it’s certainly where there is a greater demonstrable value. Again, we’re into conjecture about why that may be the case but, equally, the data are clear and compelling.

Starting with the 2YO group first, a logical extension would be to analyse the data based on the experience of the horse through checking the number of previous runs.


According to the info a first time out horse which is sent off favourite is worth taking on, generally speaking. The numbers are undoubtedly inferior to the animals that have at least a modicum of racecourse experience. Removing the debut runners, we’re left with 120 wins from 223 runs, A/E of 1.22 and a return of 24.4% on funds invested.


The table above shows the consistent out-turn of this angle. It’s not my usual hunting ground but I have to say I’m extremely interested in seeing how this one pans out over the glorious British Summer.


Your first 30 days for just £1

Suggestion: back all 2YO favourites at Windsor if they have had previous racecourse experience


Before moving on, there is one potential fly in the ointment. Knowing which horse will be at the head of the market at post time is not an exact science. Realistically, having to take the rough with the smooth (unless you have the luxury to back at the last possible moment) will be the nature of an angle such as this. In other words, there will inevitably be a few winners missed. However, particularly with BOG, there will be plenty of returns at prices better than SP, or at least you’d hope so. Swings and roundabouts.


Three-year-old favourites at Windsor

Moving on to the 3YO Classic generation, there is a clear distinction between handicap and non-handicap races.

That’s a stellar win rate for the non-handicaps at around 44%. However, market expectations are higher leading to an A/E performance of just 1.02 and there is only a single point of SP profit from the 206 runners. As a result, it’s a tough gig attempting to find value there, albeit winners will be plentiful. The market appears to be exceptionally efficient.

However, the handicaps offer a degree of hope, an A/E of 1.09 and 14% return are reasonable if sustainable.

It seems highly plausible that an unexposed 3YO favourite against potential older rivals in a 3YO+ race may perform better than a favourite running against the same unexposed rivals in a 3YO only race. Therefore, a check on the performance by age restriction of the race would make sense.

Sure enough, the numbers support this theory. Again, a potentially reasonable angle with a high strike rate and, in the context of betting on horses, relatively low risk. I’m less sure about the 3YO only races but it’s a matter of personal taste.


Suggestion: back 3YO favourites at Windsor in 3YO+ Handicap races


Windsor Trainers

Departing from the favourite theme there are several stables that seem to be synonymous with strong Windsor form.

The data above show the top eight yards using runners with a maximum SP of 20/1. The A/E is ahead of the market for all of them at greater than 1.  Having said that, and frankly speaking, the only two that really appeal in terms of further analysis are Ed Walker and Roger Varian. The IV’s (Impact Value, a measure of how much more often than the group - trainers at Windsor, in this case - perform as a whole, where a figure greater than 1 is better than standard) for both are very strong. The rest of the cast are probably worth another check at some stage but time (and word count) precludes such deliberations today.

Firstly, there is something rather remarkable regarding Ed Walker’s runners:

That’s a striking difference in performance based on SP. Not a single winner at 13/2 or greater from 43 darts thrown.

To check if this is happenstance or a general trait of the stable it’s best to compare the performance by SP for all runners (not just Windsor) for the yard. Evaluating all runs from Walker;

  • 3 victories have been notched from a total of 342 attempts at 18/1 or greater (A/E of 0.27)
  • 31/839 at 8/1 or larger (A/E 0.59).

These numbers indicate that the Ed Walker stable tends to know what chance its runners have, and should be noted for the “don’t back without support” list.

Thus, in general a supported runner representing Ed Walker at Windsor is a serious proposition (this article and data does not include the most recent winner from the yard on Monday 29th April, He’s Amazing at 5/1).


Suggestion: back Ed Walker horses running at Windsor with an SP of 6/1 or less


Secondly, Roger Varian is clearly an elite trainer whose horses often appear to be in the winner's enclosure on a hazy summer evening. His performance is solid all-round, arguably aside from his 2YO’s who seem a little under-powered at 2/14.  Taking those out of the equation we again have a similar story to Walker in terms of supported animals delivering much better performance than the relatively neglected entrants.

Same rules apply...


Suggestion: back Roger Varian horses aged 3 or greater running at Windsor with an SP of 6/1 or less


There is a cautionary note however in the case of Varian: 2018 returned one victor from nine attempts, a much lower number of winners and runners than previous years. It may be that Roger V has targeted other races/courses in recent times so monitoring of the situation will be required. 


Windsor Jockeys

Reading though Mr Probert’s blog it certainly helps pinpoint the value of having the right pilot on board. I have very few jockey angles, my primary belief being that factors such as horse and trainer ability (or patterns) have much more weight in assessing the likely outcome of a race. However, perhaps there is untapped potential to consider here.

The table below shows a Famous Five using the same criteria (races from 2012 at 20/1 SP or less), again sorted by A/E.


The top two, Harry Bentley and Andrea Atzeni, stand head and shoulders apart, the IV’s are impressive but so are all the other numbers. Atzeni is allied to the Varian operation which accounts for 16 of his 44 wins (51 runs, IV 3.22).

Finding variables to sharpen the focus is difficult though given that so many of the inputs to a horse's performance are non-jockey related. You could argue that a pilot’s performance could vary depending on ground, race distance, pace, and/or number of runners, but Atzeni is strong across all factors. There is no real angle beyond taking an Atzeni-ridden runner very seriously.

With regards to Harry Bentley, it’s more straightforward to find opportunities. The below graph offers interesting insight, namely that races over shorter distances are less productive, at least in terms of A/E and IV.

At races over the minimum trip through to 6 furlongs Bentley is only 4/49 with an A/E of less than 0.5. The results at distances at a mile or greater are a polar opposite: 22/59, A/E 1.93, IV 3.44 and ROI of 113%

That will do as a nice micro to test.


Suggestion: back horses where Harry Bentley is the jockey at Windsor at distances of 8 furlongs or greater (SP 20/1 or less)


That about wraps it up for this edition. I plan to dive into some other UK courses over the coming weeks and months which I hope will be of interest (and potentially utility). I did plan to evaluate pace/draw, course experience and other factors but lost the battle with time on this occasion.

- Jon Shenton

A correlation between trainer and sire?

stallion DUBAWI at Dalham Hall Stud, Newmarket
31 Jan 2019 - Pic Steven Cargill / Racingfotos.com

A couple of weeks ago, I began thinking about this article and was trying to work out what I was going to research in terms of content and subject matter, writes Jon Shenton. For quite a while I’ve felt that it’d be interesting to evaluate prolific sires whose progeny train on (or otherwise) from the age of two to three. I started work by “messing around” on Query Tool (I believe that’s the most accurate description of the activity) trying to determine whether the data were material enough to construct an article from.

As I explored, a part of the data jumped out at me, exhibiting something that I hadn’t really considered previously. I wasn’t quite sure if it made sense, and to some extent I’m still not, but the more I think about it the more I think there is a possibility that it is underwritten by a degree of discernible logic.

I’m talking about the so far unconsidered and unheralded angle of trainer / sire combos!

Before you start sniggering, let’s work this through. Many of us consider trainer/jockey combinations a cornerstone of punting. Nothing wrong with that, I love a good old review of TJ combo data as much as the next man or woman.  But why would trainer / stallion be any less relevant? After all, the bloodline is essentially the raw material that a trainer crafts, shapes, develops and races. You could argue that if a trainer has expertise with a specific bloodline or stallion, they may have an edge on knowing how to get the best from those related animals; or that prior success leads to the acquisition of similarly bred horses.

I realise this may sound fanciful, but for now we ought to suspend any such reservations and see what can be unearthed.

The data utilised in the article are restricted to horses running at the ages of two or three, the rationale being that generally racecourse evidence will take precedence over breeding as a primary indicator of performance in older horses. Of course, there may also be merit in looking at the more experienced cohort but that’s for another day.

Starting with a broad-brush the below table shows simply the stallions that are ubiquitous in flat turf racing. The dozen detailed in the table below have accounted for over 24,000 runners since 2012.

There is one name on the list that stands out like my thumb after a rare and clumsy day of DIY, Dubawi. The Darley Stud resident checks in with a current price tag of a cool £250,000 per covering, which is the third most expensive in the world (behind Galileo and Deep Impact). His progeny reads like a hall of fame with a monumental 38 Group 1 wins. Current leading lights include Too Darn Hot and Quorto, ably supported by Benbatl, Wild Illusion, and Kitesurf in terms of recent G1 victories.

The graph above is very simply the strike rate of Dubawi stock in comparison to the other prolific stallions contained in the table.  Performance is significantly higher than the others, nearly 5% ahead of his nearest rival, Sharmadal.

If you’d backed every 2&3-year-old Dubawi runner over the last 7 summers at SP you would have yourself a tidy 7% profit which, quite frankly, seems ridiculous. It casts any cash ISA onto the naughty step in comparison (1.45% easy access, 2.3% fixed were the best I could find if you’re interested!).  Incidentally, just for the avoidance of doubt, the small print states that in no way am I advocating a Dubawi investment fund as a tax-free alternative to an ISA...

In terms of TS combo the table below shows Dubawi’s by trainer:


No huge surprises in where the majority of Dubawi offspring end up plying their trade. The boys in blue take the lion’s share of the animals, particularly Charlie Appleby.

There is without doubt promising data contained therein. As you’d expect, an elite stallion’s progeny in the hands of elite trainers results in elite level performance (aside from perhaps the Varian operation based on this intel). The non-Godolphin duo of John Gosden (who does inform buying decisions for Godolphin) and Sir Michael Stoute stand out a little though. Key data items of A/E, win% and ROI all point towards further investigation.


John Gosden / Dubawi TS combo

Your first 30 days for just £1

The baseline performance of 44 wins from 148 runs, A/E of 1.15 and a SP ROI of 42% may well be good enough, but it’s always worth examining in more detail to establish if we can find any legitimate enhancement.

A logical starting point of refinement with this data is by horse age. We might expect with Gosden better 3YO numbers than the 2YO equivalent, given his general focus on the Classic generation. The numbers in the table below support that.

At first glance it would seem to make sense to sharpen the focus by building the angle around the 3-year-old’s only, that would be a very reasonable approach.  However, in this case there may be another factor that could be of equal, if not greater, importance.

We know that in general terms Gosden is not a renowned trainer of sprinters so it would make sense to check the profile of these 2 and 3-year-old animals against race distance.

The graph above shows the average strike rate by distance of Dubawi progeny (orange line), which is pretty consistent across all distances. A Gosden-trained Dubawi, however, appears to underperform versus the average at distances up to a mile; but, from that 8-furlong distance up to 12.5 furlongs, Gosden strike rates are roughly 10-15% greater than the overall Dubawi benchmark (the unhelpful spike of 100% for 9-9.5f is derived from a perfect 1/1 record at that distance so can be discounted as meaningful data).

Taking only these races of a mile up to an extended mile and a half the table below illustrates the annual performance:


For me, using distance rather than age is a better approach in this case. It’s a subjective call though. Gosden’s mastery of middle distances feels instinctively more important than the age of the horse but it could certainly be cut both ways.

I’ve included both factors individually in this article to try and share some of the thought processes and considerations I use when inching towards data-driven conclusions. It’s seldom that I take a linear approach to building an angle and I always check several factors to try to work out what is driving it and then make a call accordingly.

As a final call-back here are the 108 runners across both age bandings at those key distances of 8-12.5 furlongs. Clearly Johnny G Is peaking with the classic generation but to throw out all 2YO’s based on the higher-level data could be an erroneous conclusion to reach, especially on a small sample.


Back John Gosden 2 or 3-year-old runners sired by Dubawi at race distances of 8-12.5 furlongs inclusive, or all 3YO Gosden Dubawi runners if preferred


Sir Michael Stoute/Dubawi TS Combo

Secondly, let’s address the subject of the TS Combo of Stoute/Dubawi (so much for my usual focus on the smaller underdog operations!). With 15 Classic wins and 10 Champion Trainer awards there aren’t too many unknowns when evaluating Sir Michael’s imperious record on turf.

Starting with horse age again makes complete sense. It’s probably edging into cliché territory regarding Stoute’s famously patient approach to the training of thoroughbreds, but it would be expected that the older grouping demonstrates stronger performance than the juveniles.

Sure enough, in this case it certainly appears as though the numbers support the view that a slow and steady progression is the preferred mode of the stable. Unlike with Gosden, I’m not motivated to search for something in the 2YO grouping. 3 wins from 25 tells its own story and as a result I’m content to leave a Stoute juvenile alone.

Again, SMS is generally not a trainer associated with the sprint division.  Splitting Dubawi 3YO progeny performance paints the picture below.

The data contained in the red box are of great interest: 39 runs, 19 wins, A/E 2.04 and ROI of 94% to SP, albeit on a small sample size. This could be construed as a nano-angle as micro seems too grand for it in terms of scale, but the statistical basis of it is sound enough.

In very specific terms Stoute had three 2YO Dubawi offspring making their racecourse debuts in 2018: Karnavaal, Calculation and Vivionn. I doubt very much that we’ll see huge prices on them when they race at three years of age, but they’re in my tracker and if they run between 10-12.5 furlongs during 2019 they’ll certainly be of interest.

Back Sir Michael Stoute 3YO Dubawi runners between 10 and 12.5 furlongs inclusive


So, where does that leaves us in terms of trainer/sire combinations? In Formula One, the best cars generally are piloted by the best drivers; as a result they win with almost tedious regularity. I suppose calling out Gosden/Stoute & Dubawi is not hugely left-field and maybe isn’t dissimilar to asserting that Mercedes/Hamilton will perform well in the predictable world of F1. Not that backing that F1 combo will lead to a profitable outcome, their dominance heavily factored into the market.

I feel the data are very interesting and that there is at least a modicum of logic supporting this route in. After all, as stated earlier, the raw materials of the racing game are provided by the bloodlines of the animals, trainers can only work with what’s there already and if premium stables are working with exceptional bloodstock that surely must be a positive? The only real surprise may be that their appears some value in the market for these high-profile trainers and benchmark stallions, at least hitherto.

Finally, by way of wrapping up the TS Combo subject I’d like to share a bigger data table showing the raw info of the most productive combinations in terms of A/E from 2012 onwards (75 runs to qualify). I did check the stand out Dalgleish/Kodiac axis but the number of qualifiers in 2018 was low (only 2) so it’s one to keep a watching brief on. Clearly there is further research to get stuck into if you are so inclined – do share if you go down that route, and thanks!

- Jon Shenton






Early Season Turf Pointers

April trainers

With Cheltenham now a fading speck on the horizon our next scheduled stop is the cavalry charge of the Lincoln in only a few days time, writes Jon Shenton. The shackles of winter are off (hopefully), Spring has sprung, and the flat turf season is well and truly on the way.

It’s without doubt my favourite time of year, certainly in terms of the racing calendar. The promise of the long, warm summer nights and a plethora of punting challenges stokes the fires like no other.

Conventional wisdom is that bettors should tread very carefully in the opening few weeks of the season whilst form-lines are built. Whilst that might be true to an extent if you’re a pure race reader it is certainly of less relevance to the data driven approach that I primarily use.

Horses having long absences, an array of new talent on show and highly variable underfoot conditions all contribute to devilishly difficult puzzles. Data can be your friend and ally under these circumstances and it can give you an edge on the general population.

A sensible point to start would be evaluating trainer angles for April performance.

The below table shows the April numbers, sorted by A/E and only including the usual SP of 20/1 or shorter animals. All races since 2012 are analysed.


One can clearly delve into any of these further. It’s certainly of interest that the highly populated Fahey yard is profitable over a high volume of runners. The same applies to Gosden, O’Meara, Appleby (Charlie), Haggas and Beckett. If they’re delivering runners to the track in April, then these data give a degree of confidence that they are likely to be competitive.

In pole position, however, is the veteran trainer Mick Easterby. He will be 88 years-old at the end of this month! If at a similar age I’m lucky enough to be around, I’d be hugely disappointed to be still working (understatement!) so it surely shows the enthusiasm he has for the game. Those rich experiences over the years certainly seem to have been put to good use in getting the yard's runners blasting out of the stalls early.

The April output is impressive with an A/E of 1.61, a nice strike rate (19%) and an ROI of 41% is more than welcome.

Evaluating performance against SP there is no winner at 18/1 or 20/1 from 26 attempts so from an angle point of view I’m going to exclude those personally. I do realise entirely that this may be folly, mathematically you’d only expect 1-ish winner from 26 attempts at those odds. But given the number of angles I operate and the relatively high number of daily bets I’m always happy to be more selective and potentially leave a winner or two on the bench.

Taking the 16/1 (SP) or shorter only it leaves 129 runs, of which 123 are in handicaps of some description. The remaining half-dozen non-handicappers have failed to register a single win. It’s clearly a yard focussed more on handicap racing so I’m happy to trim the angle accordingly again.

I also want to understand if April performance is uncharacteristically positive against the rest of the year. It could be that the basis of this angle applies to other months.

The graph below effectively puts the notion of strong other periods of the year to bed. It overwhelmingly illustrates the peak month for Easterby is April, with spikes in both win and placed rates in the month. It’s generally downhill from there as the season progresses.


Finally, to understand the consistency of the potential angle, a check of performance by year is helpful. Doing so we get the following split:


29 wins from 123 runs, 1.79 A/E with a 78% ROI. That’ll do for me. With no fallow year since 2014 this goes into my active angles as one to follow. Ordinarily these should go through a bit of testing before committing, but where’s the fun in that? I’ll be live with this in April, trying to get early prices. A high volume, small stakes approach mitigates the risk to some degree and enhances the entertainment value exponentially!

Back Mick Easterby in April handicaps at 16/1 or less on turf

Your first 30 days for just £1


Working down the list sequentially, the second-best performer in terms of A/E is John Quinn. The Yorkshire stable is a powerhouse of racing in the North. Around two thirds of his April runs are on relatively local Yorkshire tracks.
Starting with the April performance vs. rest of year this time we have the following by win strike rate:


On the chart I have marked the April data point with a red circle. Like Mick Easterby, it is clearly a landmark month for the stable.

A point of note, the March number is only representative of a handful of runners (15), and the same applies to November’s apparently phenomenal peak (17) so it’s easy ignore these months given the paucity of data.

Also, like the Easterby angle there is no winner at 18/1 to 20/1 so a small snip to the criteria to only take account of SP’s 16/1 or shorter is my personal choice. Looking at the annual performance there are two poor yyear, 2013 and 2014, which weirdly are also the same as Easterby. It might be that those were particularly cold or wet springs, leaving the horses a little short in their work, though that is no more than conjecture.



I’ve poked around looking for other trends or items of note with these data. In truth though, nothing stands out and there is usually little point in forcing it, such efforts usually leading to at least a degree of backfitting. Simple is best.

Back John Quinn runners at 16/1 or less on turf in April


Maiden & Novices

The onset of a new season means an absolute battalion of untried, untested and unraced 2YO’s will all hit the track for the first time. Like a lot of readers I don’t generally play in this type of race. Paddock judging is out personally, aside from worldly insight such as “that’s a big horse” and “that one looks a bit fired up” I have nothing to offer in this field, though I very much respect those who can read the confirmation, maturity and fitness of these babies. I have limited sources (i.e. none) of yard and course chat so the only thing in my armoury is my old mate, data.

From 2012 to date there have been no less than 14,911 horses making their racetrack debuts on turf as two-year olds in maiden or novice races. Changes to the novice programme in 2017 do make individual analyses on Maidens or Novice races more difficult on a like for like basis which is the reason that I’ve compiled them together.

This time I’m going to evaluate yards with a high number of runners, searching for the good and the not so good. The relatively massive table below shows first time out trainer performance in maiden and novices from 2012 onwards. I have elected to leave an SP filter out of the equation for this data set. The logic behind that is with debutants you could argue that the market is more likely to get it wrong and big priced winners could be more prevalent. This may or may not be true but that is the rationale for leaving the data as “pure” as possible.


As you might expect, there are some wild variations in performance. Firstly, the ones to potentially avoid, out at least around which to be wary.

Messrs Bell, Stoute and Easterby (Tim not Mick!) have a quite frankly appalling record under these conditions. In fact, the volume of combined winners is of such paucity that I can add it up confidently in my head without consulting any technology.

41 wins from 743 runners (I did have to check the runner number with a calculator). A strike rate of just 5.6%, with a combined loss of about 46% in terms of ROI. Good luck with that!

Of course, we know that SMS famously nurtures his charges along at a careful pace, so it makes complete sense for him to be here. The others are possibly more surprising. Geegeez Gold is of huge assistance in alerting you to these red flags on the trainer icon on the racecard, showing FTO performance of that trainer for the last two years.

Back to the macro-level data in the table relating to the last 6 years. The only trainers eking out a profit in the list are John Gosden and Andrew Balding. Gosden has the most impressive strike rate, 18.6%, on the table too. I must confess, I did find this a tad surprising so with a degree of curiosity I investigated it further.

Zooming in on monthly performance is logical in my mind. The early season calendar is rife with sprints. Short distance blasts are not something you’d ordinarily associate Johnny G with so might expect performance to be less positive early in the season in maidens/novices;

Sure enough, volume of runners, strike rate and ROI all improve as we  move into and through through the hot summer (ha ha). Indeed, Too Darn Hot (August), Cracksman (October) and Coronet (September) all prevailed on their debut run in recent years.

In general terms you might think that Gosden’s strong hand of 2YO’s will be focussed towards the future, and specifically their 3YO campaigns. In fact, it’s quite common that he waits until his charges are three before giving them their first run: La Ti Dar is perhaps the best recent case in point.

To be honest, despite knowing all this there is not enough here to generate a sufficiently strong angle for me. I have evaluated race class, sex of horse and a number of other variables but there is nothing of huge significance. That said, I’d always be very mindful of a Gosden debutant once we get beyond the summer solstice and maybe play on that basis, but it’s certainly not for me in terms of a discreet “system” to run with.

Given the sheer heft of runners (633) and the worthy A/E attainment (0.99) it would be slightly remiss not to comment on the Fahey operation a bit further. In a similar way to Gosden it’s hard to find a robust angle to recommend although there are some clues and pointers worth drawing out.

Firstly, the earlier in the season the better as the graph illustrates, April and May are very strong in comparison to the rest of the year.


There is also interest when evaluating at the SP’s of all the stable's Maiden and Novice runs. The line graph below illustrates the cumulative profit or loss position by SP. In basic terms it shows that it is most profitable if Fahey’s first time out animals have been backed to 4/1 or shorter. Virtually every banding bigger than that is loss making.


Backing all 4/1 or shorter runners would result in a £26 profit to a £1 level stake (represented by the green arrow on the graph), whereas backing all 9/2 or greater would return a £97 loss (red arrow on the graph). We know two things about Fahey Maiden and Novice performance. Firstly, April and May performance is good. Secondly, horses at 4/1 or shorter are profitable. So, if we take April/May runners at 4/1 or shorter at SP I’d be optimistic we’ll find a reasonable angle. The table below gives us our answer:

There we have it. A small number of prospective bets, and at 4/1 or shorter it should be relatively low risk if unspectacular. It’s not really my sort of usual angle or bet (I tend to favour Hollywood odds long shots) but if you are inclined to have a bet in a maiden and novice race a short priced backed Fahey charge in the spring wouldn't be a bad place to start.

Back Richard Fahey First time out horses at 4/1 or shorter in Maiden/Novice races in April and May


I hope in the above I've offered a few potential pointers for success at the start of the British flat turf season. Do feel free to play around with Query Tool on some of the other names in the big tables, and leave a comment if you find anything of note.

- Jon Shenton

Jon Shenton: The Spring Trainers

Over the past week or two I’ve noticed as I’ve left work for the commute home that it’s getting lighter, writes Jon Shenton. The seemingly everlasting spell of only being at home in eternal darkness is coming to an end: spring is on the way!

It also means that the flat turf season is looming into view and I used the equine flu downtime to do a bit of much needed spring cleaning on my flat systems portfolio. More of that to come over the course of the next few months. Until then, let’s have a brief delve into a few thoughts and potential patterns regarding the spring-time National Hunt game.

Before I start in earnest, it does need noting that the relatively dry winter, coupled with equine flu, means that seasonal based angles may not be as relevant as trainer plans may have been shifted through necessity. Sadly, we’ll only know in a few months, but I still think there is strong merit in finding training operations which typically go through the gears over the next few weeks.

In the spirit of not changing the habit of a lifetime, I’ll kick things off with a massive data table, containing the trinity of A/E, Win(SR)% and ROI.


What are we looking at here then? Essentially, this is a table contrasting March-May trainer performance with all other times of the year in terms of A/E, SR and ROI. My usual 20/1 cut off applies (that is, I am ignoring all runners >20/1), and the trainers illustrated all have seemingly strong spring records, certainly in comparison to their performance across other times of the year.

The data are sorted by variance in A/E performance between the March-May period and the other months. I think by showing this contrast it helps highlight those yards truly peaking at this time of year.


Matt Sheppard

For example, Matt Sheppard sticks out like the proverbial sore thumb, with an A/E variance of 0.75, a SR variance of 14.5% and a ROI variance of 91%. At a high level, absolutely everything about these data point towards a promising angle, including the fact that it is another under the radar yard which always comes as a nice bonus.

Here is Sheppard's record from 2012 onwards, by year.


Not many negatives there!

However, there are always a few simple checks to apply, especially on a seasonal angle. A good starting point to achieve that is to check the performance by individual month, hopefully validating that our spring record is superior to other periods.

By splitting out by month we’ve found something potentially of interest. Firstly, the A/E in February is touching 1.8. This was not part of our original data set. The A/E in May is also less impressive (although still very reasonable) in relative terms at 0.97. The conclusion from this is that the radar needs a slight adjustment, potentially shifting to the left of the calendar a tad. Taking account of the February data makes sense, perhaps at the expense of jettisoning May. I’m not willing to expunge that month yet, as there may still be merit in drilling down into the data further.

First port of call for the zoom is to analyse Sheppard’s record by race type.


Striking difference?  I think so.  Look at that chase record in comparison to the other NH race codes. There’s a very strong 37% strike rate, A/E of 2.15 and a return on investment of very nearly 129%. At SP, too, with Betfair SP returns better by another 20 points!

Incidentally the May record for chasers is five wins from 26, A/E of 1.25 and ROI of 38%. Moderate in comparison to Feb, Mar and April, maybe, but still bankable.

The table below shows the February, March, April numbers only, leaving it as your call on whether to include the month of May.

Your first 30 days for just £1


Suggestion: Back Matt Sheppard Chase runners in Feb, Mar, April (and optionally May) at 20/1 or less


Chris Down

It is going to be very hard to follow that angle with something quite as compelling. Still, nothing ventured, nothing gained, eh? Let’s establish a better understanding of the data relating to Chris Down's Cullompton (Devon) yard.

Whilst checking a few contextual pieces of info I noticed something worthy of attention. The graph below shows the volume of runners from the yard by month. That clearly looks like a trainer that gears up for the months of April and May in particular, potentially keeping their powder dry over the January and February period where they have relatively few runners.


There clearly could be a multitude of reasons for this. One logical possible may be that in general horses trained by Mr Down prefer quicker ground than the typical NH animal, hence a focus on spring-time runners.

The table below shows all yard runs since 2012 (at 20/1 or less SP, irrespective of month). Interestingly, 56% of all runners were on ground officially described as good or quicker. This compares to 42% across the whole National Hunt population.

We can look at this in two ways. Firstly, this yard purely gears up for the spring. Secondly, and alternatively, this yard could simply be waiting patiently for good ground and the fact it is spring is merely incidental.

It is probably of limited relevance what the key reason is as the data are clear. However, it’s always of interest and importance to have an underlying logic when selecting or deselecting variables in your search for profitable punting angles. In this case there is a clear demarcation between Good/Good to Firm results and those on the more traditional ground conditions of National Hunt racing; but not enough for an angle.

So, back to the original spring data. The table below clearly illustrates a yard that focuses on competing over the smaller National Hunt obstacles in the main: this is in keeping, perhaps, with horses who prefer faster ground.

NH Flat and chase runners are low in volume and no firm conclusions could ever be drawn on such a small sample size. However, what we do know with certainty is that hurdle performance is strong. We don’t know for sure that NH Flat or chase form would follow suit if we had more data so for that reason I’m happy to exclude them.

There is another factor worth consideration, and it can be of interest with the smaller, arguably more regionally focused trainers such as the Down operation. The table below indicates the results of the yard by general area within the UK.

Evidently, this is a trainer who prefers to keep things close to his Devon base. Performance on the tracks determined as being in the South West is markedly more impressive than the rest of the country. Approximately two thirds of runners are saddled in this region, too, which of course makes sense from a transportation perspective.

For the record those courses are Cheltenham, Exeter, Newton Abbot, Taunton and Wincanton. Although data points are limited, the record in “central west England” (which is to say, Hereford and Ludlow) and Wales (Bangor, Chepstow, Ffos Las) appears to hold up too which makes sense given their proximity to the Devon base of the yard. Whichever way we think about it, runners from Chris Down's stable travelling away from its South West base perform poorly in comparison.


Suggestion: Back Chris Down Runners over hurdles at 20/1 or less where they are running in the South West, Wales or Central West England.


For completeness and perhaps your own manual adjustments here is a table of performance by specific track.


Mark Walford

Finally for today, we head to the dual-purpose yard of Mark Walford. Based in York there are approximately 30 horses in training currently across both flat and national hunt racing. At face value, performance on the level is a tick or two behind the performance over obstacles with roughly double the volume of winners in the National Hunt sphere from a broadly similar number of runs.


Adopting the monthly split approach illustrates peak performance in April through to June (arguably with an interesting period in the early throes of the NH season in Oct/Nov). This means another radar refinement: effectively ditching March and adding June to the equation.

The official going spread is again of interest:


That’s consistent enough and, angle wise, it doesn’t really change things. I’ve included the table as, like the Down yard, there’s a notably high proportion of runners on good ground, doubtless caused in part by the months in question. I did check if the yard's record on good or quicker held up all year round to see if that was an angle worth progressing but, in truth, it didn’t. In terms of volume of runners, winners, and A/E, the core months of April, May and June are where it’s at!

The data also show us (table below) that the Walford yard is one where horses appear to be better for a recent run.  So, in terms of angle I’m going to tighten it by selecting only runners who have had an outing in the last two months / 60 days.


Suggestion: Back Mark Walford runners at 20/1 or less in April, May and June where the horse has run in the last 60 days


That's all for this edition. There are a few potentially interesting angles to be found when evaluating some of the other trainers on the main list for spring form published at the top of the article. At the very least I think all of these yards can have a small upgrade associated with any of their runners during this part of the season.

- Jon Shenton

Jon Shenton: Heading Further Leftfield

This is the second of a two part article. In the first piece, which can be found here, I highlighted a number of training operations worthy of a second glance if their horses ran well enough last time to finish in a placed position.

This follow-up will explore the other side of the equation: trying to find those trainers whose horses perform well after a moderate or disappointing run, specifically having finished unplaced last time out (LTO).


The Leftfielders

You may recall I named this group of trainers the “leftfielders” for what are obvious reasons, hopefully. In basic terms, these yards appear to deliver a winner seemingly irrespective of whether the horses ran well last time out or not. In the same format as that last article, the table below covers National Hunt handicap runs from 1st Jan 2012 where horses have had a starting price of 20/1 or shorter.



The data is split by trainer, showing performance by horses that placed last time out (left hand block) and horses that were unplaced LTO (right hand block). This table of leftfielders is sorted by the variance in performance between the two separate datasets. In this case it’s showing the yards who are exceeding market expectation with their runners who were unplaced in their last race.

And here is the same info represented graphically (using A/E).



I find some of this data striking. All these yards outperform market expectation with runners that had a moderate/unplaced last run. That’s certainly interesting, however, the real insight for me is the variance in comparison with their in-form runners.

Perhaps it’s labouring the point but if we take Joanne Foster’s data, the A/E of her placed LTO animals is 0.55 (based on market evaluation of 8 wins from 86 runs) and the A/E of her unplaced equivalent is 1.48 (based on 26 wins from 157).  That's a variance of 0.93 in A/E performance between the two datasets. The graph shows the scale in difference with the Keevil, Stronge, Coltherd, Dalgleish and Spearing yards all showing notable deltas between their placed & unplaced performers.

Another way of appraising the data is by analysing strike rate. The graph below illustrates the same dataset but by winning percentage across the placed/not placed variants.



Here we can see that not only do these trainers beat the market with their unplaced last time entries, but their strike rates are at least comparable with their more in-form (as measured by last day placing) charges. In other words, last time out finishing position is less relevant for these yards. Taking the Foster yard again, a horse placing LTO prevails in its next race 9.3% of the time, compared to 16.56% for those unplaced last time. Whichever way you look at it, that’s a big difference and should be ripe for punting angles.

Let’s zoom in on the Joanne Foster data as a case study. The Yorkshire-based stable is another small, focused outfit. Exactly the sort of yard that I like to follow. Having had a total of 39 winners from 407 runs in National Hunt racing (since January 2012) the ability to track the stable runners and performance closely should be attainable. Incidentally, nearly a quarter of the yard's runs have been at Sedgefield (15 wins from 97 with a ROI of 0.1% at starting price) so entries at that track from Foster should always be noted.

Returning to the placed/not placed data, there are three aspects of the Foster yard that I think can be drilled down upon to make the healthy starting point of her overall unplaced LTO data (27 wins from 156, A.E 1.48, ROI 33% to SP) potentially even more profitable.  They are starting price, going and days since last run.

Your first 30 days for just £1

There are three data tables below showing the splits:

The reason that I’ve shown the three separately with the same data (the 27 wins from 146 runs) is that if we'd gone through sequentially it could seem like a heavily back-fitted approach. By keeping the larger data set and analysing the variables against that broader info we can be slightly more confident that the data is indicative of a general trend in performance.

The numbers outlined by the red border in the tables show moderate performance, which I’m happy enough to remove from the equation. Horses returned at an SP of 17/2 or greater have a 5-from-90 record losing 19% of funds overall; while horses running on “non-winter ground” have a 5-from-56 record, and those having had more than 45 days off the track are a wretched 1-from-32.  All of these 'chuck outs' are rational and explainable too.

Removing the red border runners gives us:



A very small micro angle – but a high strike rate (39%), A/E at 2.53 and 166% ROI. One of the reasons I’ve laid out the components is that the angle is probably strong enough for you to choose which elements or not you’re comfortable including or excluding. For me, I’m going for it, this is in my systems portfolio until further notice!

Back Joanne Foster runners at 8/1 or less, on Good to Soft or softer where the horse has run within the last 45 days


I could go into more detail on any of the trainers on the master table at the start of the article hunting for micro angles! After all, there are some massive names on there (Gordon Elliott, and Anthony Honeyball of this parish, to name but two). However, the key message is that all these yards have a propensity to deliver irrespective of a low-key last run by one of their charges.

When trying to unpick a handicap race we all like the comfort blanket of a nice recent run for our potential wagering proposition. But over the past year or so my mentality is changing / has changed and I take my comfort from data such as this instead.

Finding a horse from one of these yards that I like and which bombed LTO gives a ripple of excitement, after all I think I know something that the punting masses generally don’t. Naturally, it doesn’t always work and clearly an unplaced run LTO can be a symptom of a horse with limitations or being out of form rather than an off day. Nobody said it was easy!


Like Shelling P's

As I’ve been writing this, looking out of the window on another frosty Leicestershire morning, something else has occurred to me. As an extension of the theme I’ve been wondering if there are some of the training establishments whose horses bounce back specifically from being pulled up in their last run.

My hypothesis would be that some trainers could offer instructions to the jockey to pull up more readily than others (it could of course be the jockey too), or that they are more inclined still to run a horse when the weather conditions have gone against the horse, or, well, for any number of other reasons. Regardless of reason or theory, it’s worth a quick look.



Now isn’t that interesting?  The table shows trainers where a horse was sent off at an SP of 20/1 or less and had pulled up in its previous run. To qualify for the table an A/E of greater than 1.00 is required.

There are some National Hunt behemoths on here: Nicholls, Henderson, Tizzard and Hobbs for example. It does make sense to an extent. They generally house animals with greater talent so you’d expect them to be more capable of bouncing back after a bad run in comparison to other stables. It does appear, though, that the market might not always fully factor that in. That’s perhaps understandable; after all backing a horse with a P as its latest entry on the form-line requires a degree of bravery. However, in general terms there will be value plays where these trainers listed in the table are in the game.

The clear leader here is David Pipe so let's have a look at his data in more detail.

31 wins from 161 runs, A/E 1.75 and ROI of 62%.

We could leave it there, that’s strong!  But let’s check a bit closer anyway.

Here is the distribution by starting price.


A bias towards the shorter end of prices here, it appears as though we are on relatively safe territory ignoring all runners at 12/1 or greater, where a combined record of 3-from-65 and a £20 loss to a £1 level stake is persuasive enough for me. That leaves the following consistent performance over several years (notwithstanding the solitary runner in 2019 so far).



And Just like that another micro is added into my portfolio:

Back David Pipe runners sent off at 11/1 or less and which were pulled up in their previous run

Sometimes it can be that easy finding angles, other times I can sit for hours searching high and low for snippets worth exploring.  I like this one though and am looking forward to seeing how it pans out over the coming weeks, months and maybe years!

That just about wraps up this edition. A final word, though. If you have any theories or hypotheses that you think might be worthwhile looking into please drop me a line, either via the comments below or on twitter

I can’t promise I’ll investigate everything but I’m sure there is some gold worth panning for in the collective minds of geegeez.co.uk subscribers / readers.

Speak soon!

Jon Shenton

Jon Shenton: From one place to another?

I’ve been meaning to check this out for ages, writes Jon Shenton. With a bit of downtime over the Christmas and New Year period (if you can have downtime with three kids) I finally got around to pulling my thoughts together to test a few theories with some lovely data.

I’ve always been curious about whether there are some trainers where we can get a better view of a horse’s likely performance based on its form in recent runs.

In fact, I have touched on it in some of my earlier pieces on this site. Using data such as the LTO Winner element of the Trainer Snippets report here on geegeez.co.uk we know that some yards are more adept at backing up a win with another victory. In this article, then, I will broaden the search a little to placed form, rather than evaluating last time out winners only.

Let’s get straight into having a look at a few numbers.

The data below shows all National Hunt Handicap runs at 20/1 or less since 1st Jan 2012 (up to 11th Jan 2019) split by whether the horse placed last time out (LTO). I’ve selected handicaps only largely to avoid the inconsistencies associated with horses on their first few runs. There may be an argument to expand this, but one has to start somewhere.

The key here is that A/E and ROI are virtually the same across both datasets. From a punting perspective you’d lose just short of 15% of your cash backing all horses that placed LTO, and a similar level of damage would be inflicted if you only backed horses that did not trouble the judge during their last venture.

The main difference is strike rate, a horse “in form” based on last run prevails 16.66% of the time as opposed to 11.44% if it was off the board that last day.

Broadly speaking, a winner that placed last time out returns a starting price of about 4/1 on average.  One that didn’t place the last day has an SP of 13/2 or thereabouts.  Considering both aspects perform consistently in A/E and ROI terms the market appears to be getting it right.

Let us now drill down somewhat, by yard, in search of trainers worth tracking in specific circumstances.

To do that I’ve pulled together a fairly large master data table showing performance with horses that placed last time, and data showing where the horse hasn’t hit the frame in their last race, and merged it together by trainer to try paint a picture. Still with me? Good!

To make sense of it all I’ve categorised things into three areas of interest and will discuss each in turn.  The categories are as follows:

  1. The Outperformers: trainers who generally are beating the market irrespective of what their charges did last time out
  2. The Rollers: these guys appear to roll when their animals place in their previous run and follow up with another good run
  3. The Leftfielders: these yards appear to deliver a winner seemingly irrespective of whether they run well last time out or not

In this article, I'll review the first two categories, starting with...

The Outperformers

The first theme or thread represents general outperformers, which is to say handicap excellence all round.

Just to explain the tables and data in more detail first.  The “block” to the left shows the trainer performance for horses that placed last time out in terms of runs, wins, actual vs, expected and return on investment.  The “block” on the right shows the performance of all runners that failed to make the frame in their last appearance.

The column on the far right (A/E var PL – NP) shows the difference in A/E performance between horses placed and unplaced last time out.

This should be an elite list, and for that reason it’s small in size. These yards have an A/E of 1.00 or greater irrespective of horses placing or not last time, with a positive ROI across the board to boot. To qualify they needed to have saddled a minimum of 50 runners for each aspect at an SP of 20/1 or less.

If we take the Kenneth Slack yard, horses which placed last time won 23 of their next starts from 86 runs. That was right on expectation against the market (A/E 1.01) whilst recording a small profit of about 7%.

His unplaced last time runners were 20 from 79 in terms of wins in the review period, but this time A/E is 1.35 and ROI is close to 50%.

A Slack runner based on these numbers is always worth a second glance.  However, it seems as though the horses’ last race performances may be irrelevant to how well they’re likely to run this time. The win strike rates are very similar, too (c.25-26%), but because the market seems to be factoring in the last run form there is more value in backing the runners who did not place on their last outing.

Your first 30 days for just £1

All of the above five trainers are Outperformers, consistently beating the market irrespective of whether their horses’ previous run indicates they’re in form – as evidenced by a placing – or not. At the top of the tree is Tom Lacey, though the market is now catching up with his progressive stable. As things stand backing all Lacey’s runners blind is a still a rewarding enterprise… just! The bird, however, may have flown.

Interestingly, and perhaps surprisingly, Nicky Richards is outperforming expectation across nearly 800 runners. That merits further investigation another time but, for now, following this yard closely is the recommendation.


The Rollers

Just by way of a quick refresh, a Roller is a trainer who performs strongly where their horses hit the frame in their most recent start, in comparison to the ones that were unplaced.

The table below shows trainers with 50 or more runners in each category (placed or not last time out). Data is sorted in order of the variance in A/E performance between the placed and not placed LTO (the far right column).

Some of these variances are mighty in scale.

For those that are more visually stimulated, the graph below shows the same data as the table in terms of A/E performance by trainer. The blue blocks represent horses placed last time, the orange line representing those that did not. It shows the range of difference in performance by trainer, the size of variance getting smaller as we work left to right across the graph.

Right at the top we have James Evans (H J Evans http://www.hjamesevans.co.uk), a relatively modest (in terms of size) operation based 20 or so miles away from Cheltenham.

Modest in size maybe, but if an Evans horse placed LTO they enjoyed a 21 from 69 win rate, A/E of 1.67 and ROI of 112%

Compare this to those unplaced in their previous outing:

12/146, A/E 0.71 and ROI -24%.

That’s a stark difference potentially illustrating that this yard takes its time to get horses ready but, once firing, they look worthy of support.

Drilling a little deeper, the James Evans yard in its entirety has only ever had one winner at greater than 14/1 (UK NH racing) so we can refine slightly.

I’m happy to leave it there, this angle has been profitable every year apart from one so that will go into my systems to follow library from now on, a simple angle:

Back James Evans runners with an SP of 14/1 or less that placed LTO


Another trainer with a stark difference in performance based on last time out placing is Caroline Fryer (17/56 placed LTO 28% SR vs 6/78, 8% SR not placed LTO http://www.carolinefryerracing.co.uk). Hers is another team with a select number of horses running under rules. On closer inspection there is a slightly different story to tell here, however.

Of those 17 wins from last time placers, ten relate to a single horse, Riddlestown. Now a 12-year-old his last run was on Boxing Day. It may be that age is catching up with him but he’s still going into my geegeez tracker. When he’s had a competitive run (placed) in the recent past (30 days or sooner) he is 10/31 so I’ll be keeping an eye out for an uptick in performance. Riddlestown has an entry at Leicester on 22nd January, so I will be watching closely for signs of improvement with a view to potentially following him next time.

Every trainer on that list above merits further analysis, especially those who fly under the radar of the general masses (which I think is the case for most of them).

A yard I’ve followed with some success in the past is that of Iain Jardine (http://iainjardineracing.com), though admittedly more on the flat where Jardine has a strong record in staying races. That is particularly true where his horses are fit (had a recent run) and have been backed (10/1 or less). So it’s not a huge surprise to see him on this list with his NH runners.

Taking his ‘placed in previous run’ data for National Hunt (30/106) and refining I’d expect a similar pattern to his flat runners to be apparent: a recent run and relatively well supported in the market.

I’m not going to show the data table (you’ll have to trust me!) but there isn’t a winner in his dataset sent off at greater than 10/1, albeit only from 13 runs.  A similar pattern to the flat runners so I’m going to ignore those for now, though I am mindful of the small sample size.

The table below shows the 10/1 or shorter NH placed LTO runners by days since last run.

I’ve used coloured font illustrate some of the subjectivities when dealing with data.

The Green section is easy, it clearly shows that if a Jardine horse sent off at 10/1 or shorter was placed in a very recent run (20 days or less) it is worth following.

The other sections are less clear, the red font illustrates those horses that have been off track for more than 60 days: one winner from only 7 runners sent off at 10/1 or shorter, so inconsequential. Personally, considering the overall profile of the data I’m happy to leave a short-priced Jardine placed last time out runner alone if it’s not been since within a two-month period.

As you may have speculated already the amber section is most open to interpretation, with ten wins from 44 runners and a very small SP profit to a level stake. I think in this case I’d probably leave alone, primarily to limit the number of bets and to keep a more focussed portfolio. However, with exchanges and BOG it may be that backing these ‘three weeks to two months off’ runners will turn a healthy position.

Iain Jardine NH Handicap runners with an SP of 10/1 or less that have run in the last 20 days, placed LTO


This article and the data referenced within is broad brush stuff. For example, in the high-level data there is no consideration of factors such as when the horse’s last run was, which grade, distance, class, etc. Form must be more relevant when those things are positively aligned with today’s task.

Next time I’ll look at the “Leftfielders”.  There is a raft of interesting stuff to discuss with these yards.  Word count and a determination on my part to try and be a bit more succinct means I can’t do it justice today! Until next time, happy punting.

 - Jon Shenton





Newcastle Punting Pointers: The Angles of the North?

In my last article I accidentally stumbled into unpicking all-weather course form and the relative importance of it at each track in the UK, writes Jon Shenton.  It wasn’t my intention to evaluate anything in that area; however, when exploring a vast ocean of data sometimes you end up going where the wind takes you as thoughts develop around the words and numbers on paper.

If you didn’t read that article (link here), a tentative conclusion was that a previous course win was more indicative of a likely follow up victory at Newcastle than any other all-weather surface.

Until that moment I had very limited interest in the tapeta at Gosforth Park as a vehicle for punting. Aside from the odd evening “leisure” bet, almost universally doomed to failure, I’ve hitherto watched perplexed from afar.

To my untrained eye the victors seemed to be pattern-less in terms of my usual all weather starting points of pace and draw, the immaturity of the surface also meaning data regarding trainer, sire and anything else you can think of is less reliable from which to build even vague conclusions.

So, when the intel from that last article showed Newcastle in a favourable light it got me thinking: it was time to have a proper delve into the delights of the northeast venue.

There is a uniqueness regarding Newcastle compared with its AW cousins. Namely, that it has a straight mile. Apart from Newcastle’s five to eight-furlong and Southwell’s five furlong straight track all other races on AW in Britain and Ireland are contested around a bend.   This could be of potential interest for a number of reasons. As I’ve already alluded to, the usual staples of pace and draw could be less important without a turn than we see at other AW venues?


Why could pace and draw be less relevant at the Newcastle straight track?

To start with, on a straight track all of the horses compete over exactly the same distance.  This is not the case when racing around a bend where inside draws have a shorter distance to run.

Imagine an Olympic 200m final around a tight bend where all athletes start at the same point.  Not even a peak Usain Bolt content on fried chicken could overcome a lane 8 / car park draw unless he was running against people like me.

As well as the emphasis of draw on a turning track the AW tracks typically are tight in nature, with shortish home straights. This often leads to greater extremes of front runner bias. Thus finding a competitive front running animal with an inside draw on the AW is usually a compelling wagering proposition.

All of this points to Newcastle being a fairer test over the straight track than the other artificial surfaces racing around a turn. Fairer for the runners and riders, but trickier for punters?

Could it be that factors such as course form, pedigree or trainer angles play a more significant part in determining the outcome of a race? That’s all supposition at the moment, but let’s dive into it.

A continuation of the course form theme seems like as good a place to start as anywhere.


Course Form at Newcastle

A quick refresher, the graph below is from the previous article, it shows the adjusted strike rates at UK all-weather courses split by horses’ number of previous wins at the same track.  The adjusted view was to reflect/standardise the effect of different field sizes. In other words, a race at Newcastle should be harder to win as the average number of runners is 10.9 whilst at Chelmsford it’s a more meagre 9.05.



The green line to a clear degree illustrates that Newcastle previous winners have a higher probability of a future win than course winners at other AW tracks.

By adopting the same method but splitting the Newcastle races by straight/round track performance we hopefully will find something of interest. Firstly, we need to take field sizes into account.  Straight track races are popular with an average of 11.28 entrants per event, 1.11 more than the round track average field size of 10.17.


NEWCASTLE Avg Field Multiplier
5-8F STRAIGHT 11.28 1.16
8.5F+ ROUND 10.17 1.05
OVERALL 10.90 1.12


Using the same format, the graph below shows the profile of previous course winners’ strike rates by distance of race.



I think this is quite insightful.  There appears to be an indication that previous course form is more valuable in predicting a winner over the straight track of 5 to 8 furlongs, than it is over the longer trip.

Now the volume of runners is quite small, particularly on the round course where two or more previous course wins are concerned but there is definitely enough to upgrade a previous course win on the straight track in comparison.


Pace on the straight track

Your first 30 days for just £1

Lesson number 1 in Geegeez.co.uk land is that pace is a game changer in punting life.  It’s certainly been a key component in my improvement in race reading and is just about the first thing I look at when trying to evaluate any equine contest.

We’ve already generated the supposition that front running pace bias may not be as important at Newcastle as it is on the other UK all-weather tracks due to the fairer nature of the straight; but do the numbers back that up?


Well, yes. The above chart is eye-opening. It illustrates the Actual/Expected performance by pace score for each of the all-weather tracks in the UK. The data covers all 3YO+ and 4YO+ handicaps and all races up to 8f in distance.

You can see the old adage of “pace wins the race” is pronounced across all of the tracks apart from Newcastle.

The blob annotated with “a” above shows the fate of hold up horses on the straight track at Newcastle. There is clear daylight between their performances when compared with late runners at every other track. In fact, horses that are held up actually fare well even in comparison to their front running rivals at the track. Certainly, trailblazers are not the be all and end all that they can be on some tracks, as the blob “b” illustrates.  Both front runners and hold up horses have an identical A/E performance of 0.99.

Lumping in all races from the minimum trip to the mile is potentially dangerous and clearly analysis by specific trip length may lead to slightly different and more solid conclusions.  However, in terms of proving a point that race profiles are different on the straight Newcastle track to the typical AW ones I think this does enough. The bottom line is don’t be put off by a horse stalking from the back of the pack at Gosforth Park.



Hopefully it’s reasonably understandable but evaluating full draw implications of a straight vs. round track is a tough ask for an article of this length given the variables in distance, race type, number of runners and the like.

That said, by way of a quick guide, below is a broad-brush summary of Newcastle draw performance.  It only considers handicap races of 10-12 runners.  It’s also important to note I’m using actual draw position (i.e. accounting for non-runners), not racecard draw number.



What does the above tell us? In truth, not a great deal! Maybe, just maybe, there is a hint of bias towards the wings of the track, especially for races over 5, 6 and 8 furlongs. Sometimes this makes sense as races develop against a rail and perhaps that is what is at play here.  But… for no obvious reason the 7f distance contradicts other distance data by suggesting there is a hint of middle track bias. In conclusion, it’s all pretty marginal and if you find the right horse, with the right profile, the draw at this course appears to be less relevant than most in terms of stall position.



It’s quite early to draw meaningful conclusions on stallions to follow at Newcastle but the below table shows some potentially emerging talent.

It’s derived from geegeez.co.uk’s Query Tool and illustrates all runners at 20/1 or less; and to qualify for the table an A/E of 1.25 is required, as well as a 10% ROI.


The volumes are generally too thin to draw firm conclusions and build bankable, watertight angles, especially as some of the performance will be driven by individual animals repeatedly winning. Even so, it’s a good list to keep in mind to help generate a shortlist when evaluating a race, particularly when form, or course form, is at a premium.

There is merit in just pulling out a couple to discuss briefly. The most successful couple of sires on the Newcastle all weather, in terms of winner numbers, are the renowned Sea The Stars and the progressive Lope De Vega.


Sea The Stars

Firstly, Sea The Stars… His progeny’s 15 wins are comprised of 13 individual horses.   John Gosden’s Champion Stayer, Stradivarius, is the most illustrious, having recorded his first success (on his third run) at Newcastle, over the straight mile. That is the very same course and distance that stablemate Enable made her debut on, incidentally. Clearly, Johnny G likes to blood a top class type on the tapeta here.

In fact, it’s not uncommon for horses of real quality to get an early spin on the Gosforth Park sands. The apparent level playing field of the track is a feature which attracts some of the elite stables to test their youngsters at a formative stage of their careers.

Reverting to Sea The Stars, below shows his progeny runs by race code.


Not bad all round but there is a clear distinction between AW and Turf data. If we zoom in a little further and evaluate the performance by the UK’s different AW venues, we get the following.


Here we see that Newcastle is driving the superb AW performance. Yes, Southwell, Chelmsford and Lingfield all show promise and we should take note of the offspring of Sea The Stars when they run at those venues. But Newcastle is where it’s at.


Lope De Vega

Lope De Vega was campaigned exclusively in France under the tutelage of Andre Fabre and doesn’t on the face of it have a particularly strong all-weather pedigree. However, much like Sea The Stars, his progeny has performed generally better on the artificial surfaces, in win strike rate and profit/loss terms at least.

Newcastle performance is strong (see the table below), but so too are Chelmsford and Wolverhampton. The surface at the midlands track is also Tapeta so that makes some sort of sense (albeit that it was polytrack until 2014). If you delve into sire records, quite frequently a good Newcastle record can be indicative of a better than average Wolves one and vice versa. The Chelmsford one is harder to explain, though it may be simply that Lope De Vega is a top class sire all round.



If we take that trio of courses and check P&L performance over different trips, we can see below that Lope De Vega offspring are less productive over 5 and 6 furlongs than other distances.



So, I think we have a potentially nice micro here: Lope De Vega progeny, 20/1 or shorter, 7-14f at Chelmsford, Wolves or Newcastle. 27.6% strike rate, 52% ROI to level stakes with strong A/E and IV numbers. The table below shows the precise numbers.

All Lope De Vega at 20/1 or less, 7-14f at Chelmsford/Wolves/Newcastle:



A final word on the trainers who have taken to Newcastle’s newish surface, the above table shows those yards who have had 25+ runs, an A/E of 1.00 or above and an ROI of 10%+.


Before I talk about the table a couple of mentions for trainers not on the list. As stated earlier a number of elite trainers use Newcastle as a proving ground for their potential stable stars. John Gosden has had 75 runners (at 20/1 or less) including Enable, Without Parole and Stradivarius. Sadly though, and for obvious reasons, these are all quite well found in the market. Hugo Palmer is another who is inclined to send runners north as part of their education and development, but without profitable import for punters.

To those actually in the table, where there is a mix of northern track specialists and selective southern raiders. Sir Mark Prescott and William Haggas both clearly send animals up to the north-east that have a fair chance, and it is somewhat surprising to see these practitioners showing a level stakes profit. Moreover, as their strike rates at 31% and 38%, and related Impact Value numbers of 3.02 and 3.53, demonstrate, they’ll keep you in the game more readily than most.

The more local names of Menzies, Tate, Whitaker, Bethell and so on are all worth tagging too, although with only a handful of winners I wouldn’t necessarily generate micro angles to follow until there is a greater body of evidence.

Good luck, thanks for reading, and a happy new year to you all.

 - Jon Shenton

Southwell Statistics: Horses for Courses?

There are few racing betting mediums as divisive as Southwell All-Weather, writes Jon Shenton. I know people who barely acknowledge its existence, and yet, in the other camp, are people like me: I absolutely love it with every fibre(sand) of my being!

Indeed, I love winter all weather racing, full stop. It’s probably as a result of me getting some (well earned) gardening leave from January to March 2017, when I really started to immerse myself in the world of racing. Those halcyon days of studying my new toys (Geegeez Gold being the main one) in the morning and watching the racing in the afternoon on ATR will live long in the memory.  I was drawn to Southwell because it seemed a bit easier to navigate than the complicated world of National Hunt racing.  No vagueness on ground, no fences or hurdles to consider, and a whole stack of course form to evaluate.  Perhaps some moderate, relative early success helped too.

Whilst it may not be to the taste of everyone, supporters assert that the deep, stamina sapping test provided by the track offers a unique challenge and adds to the rich tapestry of UK racing. Arguably, it serves as an outlet for horses to show their ability who aren’t ordinarily suited to other racing surfaces.

It also has the important attraction of familiar names returning year after year, which as we know isn’t then norm for the racing on the level. It may be a stretch to claim superstar status for the main protagonists, but there are legends such as La Estrella (16 wins from 21 runs at the course) and General Tufto, who has run no less than 125 times on the fibresand over the last 10 years. 125 times and still counting!

To be honest, that’s even more frequently than I’ve attempted to explain odds and probability to my poor, not really interested, long suffering and very tolerant partner. Yes, on occasion I’m surprised I have one too! Anyway, let’s crack on. What follows are a few thoughts and insights which I hope will inform your Southwell wagering hereafter.


USA Bred horses at Southwell

It’s relatively well documented that horses with a pedigree originating from the good ‘ole US of A are worth consideration on the fibresand.  There is certainly logic in this given the perceived proximity between the Southwell surface and the dirt tracks of America.  The table below illustrates track runners by sire origin, for all races in 2012 onwards (three major countries only included)


Origin of Stallion Runs Wins Win% P/L (SP) Place% ROI (SP) A/E
(GB) 7881 892 11.3 -1815.6 29.3 -23.0 0.86
(IRE) 5919 700 11.8 -1327.8 30.1 -22.4 0.85
(USA) 1248 225 18.0 155.6 38.8 12.5 1.03

Southwell (AW) runners by country of origin, 1st Jan 2012 to present


The picture is pretty clear: US-bred horses outperform their UK- and Irish-bred counterparts significantly, winning more often (18%), beating market expectations (1.03) and returning a profit (12.5%) at SP.

Having said that it’s not ‘backing blind’ territory in my opinion, especially given the fact there are some exceptionally big priced winners in the sample. The biggest of all was a 100/1 shot, the Derek Shaw-trained Hammer Gun, who is definitely worth putting in the tracker for future Southwell entries as we will see shortly.

The Hammer bolted up in that particular race and, if you’re going for the Hail Mary play, I can think of worse places to do it than backing a US-bred runner at Southwell who is unproven on the surface.

As USA horses have a positive record at the track it would make some sense for American stallions to have similarly favourable numbers.

The below table shows sire records at Southwell for the same period.  This time I’ve only considered runners with a maximum SP of 20/1.  The usual reasons apply: I’m looking for angles which will return with a modicum of regularity.  Whilst there can be value at larger prices if you look hard and wait long enough, it’s not a game I want to play, or perhaps I can’t afford too long between drinks.  20/1 works for me, I know some of you prefer shorter. If you do, the data is there in the Geegeez Query Tool – go play!


Stallion Runs Wins Win% P/L (SP) Place% ROI (SP) A/E
Key Of Luck (USA) 67 15 22.4 21.8 47.8 32.5 1.39
Dubawi (IRE) 79 26 32.9 64.2 45.6 81.2 1.34
Ballet Master (USA) 53 9 17.0 1.3 43.4 2.4 1.30
Poets Voice (GB) 53 11 20.8 35.2 54.7 66.4 1.26
Refuse To Bend (IRE) 64 16 25.0 -5.2 40.6 -8.2 1.23
Speightstown (USA) 85 20 23.5 8.2 51.8 9.6 1.21
Street Cry (IRE) 129 35 27.1 40.5 50.4 31.4 1.21
Invincible Spirit (IRE) 130 32 24.6 22.0 37.7 16.9 1.20
Showcasing (GB) 55 10 18.2 24.0 45.5 43.6 1.12
Captain Gerrard (IRE) 105 21 20.0 54.3 40.0 51.7 1.10
Clodovil (IRE) 58 11 19.0 3.8 34.5 6.5 1.10

Sire performance at Southwell (AW) 1st Jan 2012-present at 20/1 SP or less


The table of top Southwell AW stallions has smattering of USA sires on the list, no major shock there. And, in the case of Street Cry, he was raced on dirt in America and latterly Dubai, winning the Grade 1 Stephen Foster in US and the Grade 1 Dubai World Cup in Dubai.

Ordinarily I’d now be searching through these data, trying to find a few nice angles to share and adopt over the next few months. In general, though, Southwell is a different proposition. Angles still have relevance but the number of course specialist horses can paint a different picture. I’ve already referred to the fact that one of the joys of the track is the number of repeat runners. Taking the top of the stallion charts (Key of Luck) we can see where the problem lies in angle creation.


This graph shows all of Key of Luck’s runners by individual animal, illustrating runs (blue) and wins (orange). The conclusion rapidly emerges: all of Key of luck’s progeny wins have been delivered by three individual horses, with 14 of the 15 coming from The Lock Master and Serenity Now! Even the most prolific stallion on the list, Street Cry, sire of the Australian darling, Winx, has a third of his victories from just two horses, namely Tatting and Fluctuation.

Based on this I don’t feel like many genuine angle opportunities exist in sire data. The samples are too small and the number of progeny involved are insignificant in many cases. No, for me, finding the right individual horses is the key. Then tracking and following them closely can be a productive method with which to approach the fibresand puzzle. Having said that, any Key of Luck or Street Cry progeny running at Southwell are still of interest and I’ll be watching them all closely and backing where conditions appear to be right.


Southwell trainers

Like any track across the world there are handlers who seem to know what is required for the unique Southwell test. Using the same approach as the sire table above here is the equivalent view for trainers, again sorted by A/E.


Trainer Runs Wins Win% P/L (SP) Place% ROI (SP) A/E
Carroll, D 101 24 23.8 37.6 41.6 37.3 1.63
Fell, Roger 55 11 20.0 33.8 38.2 61.4 1.48
Bailey, A 84 17 20.2 59.8 40.5 71.1 1.29
Furtado, Ivan 60 12 20.0 18.0 33.3 30.0 1.28
Brown, D H 67 15 22.4 -6.8 43.3 -10.2 1.25
Nicholls, D 103 22 21.4 32.9 37.9 31.9 1.25
McCabe, A J 184 34 18.5 31.1 36.4 16.9 1.19
Dwyer, C A 82 19 23.2 14.3 53.7 17.4 1.16
Shaw, D 251 44 17.5 5.5 41.8 2.2 1.15
Burke, K R 118 28 23.7 28.7 40.7 24.3 1.13
Kirby, P A 60 11 18.3 -11.6 40.0 -19.3 1.11
Butler, John 75 19 25.3 9.8 45.3 13.1 1.10
Bowring, S R 209 33 15.8 -12.4 35.9 -5.9 1.09
Your first 30 days for just £1

Southwell (AW) runners by trainer from 1st Jan 2012 to present at 20/1 or less SP


As is becoming tradition it feels right to have a quick delve into the top name on the list, in this case Declan Carroll.  The Malton-based trainer sends a high proportion of runners to the Nottinghamshire track. Indeed, the only course that is frequented more by his team is Thirsk, relatively local to the outfit.

Again, like the Key of Luck data, on the face of it, it seems that backing the Carroll stable representatives blindly is a good idea. In truth, it might be: there is a healthy strike rate, fantastic A/E performance and a reasonable return on investment.  On closer inspection though, we run into a familiar theme.



This graphical representation shows Carroll horses that have had 3 or more runs on the fibresand track from 2012 onwards at an SP of 20/1 or less; we can see quite clearly that Monsieur Jimmy and Shearian with their six wins apiece (the orange line) account for over half of the trainer’s wins during the nearly six years analysed. I think this illustration reinforces the fact that successful horses generally return to the track time and time again.

In other words, there is a selection bias in these small samples. It’s a repeat of the sire analysis scenario, and again begs the same question: is it worth following specific yards on the fibresand, or is it worth following specific horses?

The answer is difficult, as are all such responses to small sample sizes skewed by individual elements. What is not in doubt specifically is that Carroll knows what it takes to nurture a successful Southwell career for a horse and, once he knows he has one with the right aptitude, he isn’t afraid to keep running them.

By way of example, let’s examine the record of Shearian at the course under the tutelage of Carroll (he was with Tracy Waggott previously).



Impressive stuff. In spite of a remarkable track record, Shearian still, however, went off at a price of 15/2 on the 12th November this year. This, despite him winning in his previous run over course and distance.  Granted, he hit the crossbar on this occasion, in a grade where he’d largely struggled, but netted the rebound three days later with a comfortable victory back in Class 6. However, considering his price shortened significantly on the 12th throughout the day the bet represented potentially great value.

That value was present due to his previous eight runs (four on tapeta, four of turf) being fairly unproductive. To Shearian followers that is absolutely of no consequence whatsoever: his lamentable record away from Southwell is 61 spins for just two wins, both as far back as 2013. The cynic in me would point towards a summer of official rating reduction in preparation for a bountiful winter campaign cruising around the Rolleston venue, his AW rating having reduced from 73 to 65 over the period in question.

I recognise that you can always find examples to fit any given narrative; however, it does seem that Southwell form offers more reliability for predicting future prospects at the track.


Horses for Southwell

I’d love to be able to statistically assert and prove that course form is more important at Southwell than most places and I think I can do that, at least partially.

The graph below is quite simple in what it’s trying to show but not so easy to explain.  It contains data for all AW runners, by track, from January 2012 for 3YO+ and 4YO+ handicaps only.  I’m selecting these age groups due to the likelihood of more horse runs, and logically more course form to check. It’s the journeyman (or woman) type of horse that I’m interested in here.

Anyway, the graph below shows win rate by how many victories a horse has had at the track previously:



The thick blue line represents Southwell.  What it depicts is that, compared to the other all-weather tracks of the UK, a previous course win means the horse is more likely to win again at the same track. Newcastle is an interesting newcomer, and runs it close, albeit sample sizes are tiny for the three and four previous wins data points for that course.

This statistical evidence is all well and good, but it still doesn’t quite sit right. That is due to the fact that field sizes could have a bearing on the data.  If we take the black line above (Kempton) we can see that it languishes at the bottom, or close to it, across all bandings.

The only reasons that can be the case are either that Kempton has larger field sizes, i.e. more horses running equals lower strike rates, after all only one horse can win (dead heats not withstanding); or because course form doesn’t stand up as well as elsewhere.  The fact Kempton is “poor” in all categories does point to it having a higher than average number of runners per qualifying race.  The table below confirms this, to some degree at least:


Track Average field size Multiplier
Southwell 9.15 0.94
Chelmsford 9.02 0.93
Wolves 9.66 1.00
Lingfield 9.21 0.95
Kempton 10.52 1.09
Newcastle 10.89 1.12
Overall 9.70 1.00

Average field sizes for AW races from 1st January 2012 onwards


Kempton does indeed have notably larger fields than the average AW line up. Interestingly, however, so does Newcastle, a potential course specialist track in the making. So what does this mean, and for what can we use it?

In the table it confirms that a win at Newcastle is harder to get than a win at Chelmsford, and indeed anywhere else in the UK all weather landscape, based purely on field size. To prevail at Newcastle a horse has to be the best of 10.89 animals on average. At Chelmo, the cream of the crop rises above 9.02 horses, a significant 1.87 (17%) fewer.

To try and obtain a like-for-like comparison of course form, effectively taking field size out of the equation, we have to boost Newcastle and Kempton performance to take account of the higher volume of runners per race. Conversely, we’ll be downgrading Chelmsford, Lingfield and Southwell accordingly by deploying the multiplier column in the table above.

It leaves the following picture:



These data appear to show that in the pursuit of finding winners previous course form is considerably more valuable on the tapeta of Newcastle than other all-weather courses.  The new surface at Gosforth Park is still relatively new having only been in place for racing for just over two years, so the picture may change over time; but the fact that all races at up to a mile are on a straight track is a notable difference from the remainder of the all-weather scene and may contribute to it emerging as a "specialists' track".

However, even with the adjusted numbers previous course form still holds up well in comparison for Southwell.

Of course, this doesn’t necessarily directly translate to profitable angles, as course specialists are often well found in the market after all. But using data intelligently to assist in constantly improving our race reading ability has to be a good thing. If we find a course specialist with a favourable looking setup in terms of pace and draw (for another article, or check out Dave Renham’s excellent general series), we’re looking at a bet on the assumption that the price is reasonable. 


A Dozen Fibresand Masters

Let’s wrap things up. Much of this article has referred to course form and the longevity of horses who run at Southwell on a repeated basis. The below table shows some of the stars who thunder around the Notts oval with regularity. Each has had at least one run at Southwell during the past 12 months, and the table is sorted by A/E, with a minimum of 10 runs required to qualify.


Horse Runs Wins Win% P/L (SP) Place% ROI (SP) A/E
Custard The Dragon 10 6 60.0 18.8 80.0 187.5 2.75
Hammer Gun  11 6 54.6 111.6 63.6 1014.4 2.64
Piazon 13 6 46.2 20.3 61.5 155.8 1.90
Luv U Whatever 21 9 42.9 15.6 81.0 74.3 1.41
Stand Guard 14 6 42.9 -3.7 71.4 -26.6 0.88
Captain Lars  15 5 33.3 5.3 33.3 35.5 1.37
Philba 12 4 33.3 5.5 66.7 45.8 1.87
Shearian 21 7 33.3 25.6 47.6 122.1 2.28
Razin Hell 22 7 31.8 26.3 59.1 119.6 1.61
Royal Marskell 16 5 31.3 21.6 50.0 134.9 1.89
Pearl Nation  13 4 30.8 -0.1 61.5 -0.9 1.14
Samtu  13 4 30.8 29.3 46.2 225.0 1.43


Record breaking Stand Guard has since retired and there may be one or two others who have hung up their racing shoes, but the list should still be broadly active and, hopefully, profitable. Piazon and the aforementioned Shearian have already got their 2018 winter campaigns off the mark and I’m sure some of the others will be troubling the judge in the coming months. I’ve got a keen eye on Hammer Gun, and Samtu if reverting back to the flat, in particular. Here’s to a productive Southwell campaign for us all and a bit of Hammer Time over the festive period!

 - Jon Shenton (@jonnyshents on twitter)







Early NH Season, Part 2

A few weeks ago, my last article focused on National Hunt trainers who fly out of the gates in the autumn, writes Jon Shenton.   When compiling data and researching angles for that edition there were a few other areas of interest which I’d like to touch on today.

A key aspect that was considered for the aforementioned piece was evaluating where trainers had a runner returning to the track after an absence of more than 180 days, or about 6 months.  The thinking is that some trainers will have horses wound up and ready to go after a summer absence, while others’ animals generally come on for a run, taking a long-term view of the season ahead.

The below graph shows the total volume of runners returning to the track after a layoff of that magnitude.  Clearly, now is a good time to dive into which trainers are ready to go or otherwise.  As can be seen, we are in peak season for long absence returners.

Graph illustrating number of horses returning to the track after a break of 181+ days, since 2010, by month


Bargepoles and Scary data

My general approach is to always try and provide a few pointers to find a reasonable return over the medium to long term.  However, there is definite value in identifying horses through which to strike a line: data for those inclined to lay in other words.

The first stop is what I’d uncharitably term a ‘bargepole list’. The table below comprises of trainer records in terms of horses making a reappearance after more than 180 days off the track.  50 runs is the minimum level for inclusion and I have sorted in reverse A/E, accounting for all runs from the start of 2010 onwards.


Trainer performance for all runners from 2010 where the horse last ran 181+ days previously

Trainer Runs Wins Win% P/L(SP) Place% ROI(SP) A/E
Jewell, Mrs L C 51 0 0.0 -51.0 5.9 -100.0 0
Menzies, Rebecca 52 0 0.0 -52.0 11.5 -100.0 0
Young, Mrs L J 62 0 0.0 -62.0 11.3 -100.0 0
Carroll, A W 81 1 1.2 -72.0 11.1 -88.9 0.17
Stephens, Robert 56 1 1.8 -39.0 17.9 -69.6 0.24
Newton-Smith, A M 51 1 2.0 -40.0 11.8 -78.4 0.25
Dennis, David 73 2 2.7 -61.8 12.3 -84.7 0.3
Wintle, A 57 1 1.8 -48.0 10.5 -84.2 0.31
Brennan, F J 55 1 1.8 -26.0 10.9 -47.3 0.34
Henderson, P 79 2 2.5 -63.0 12.7 -79.8 0.37
Dyson, Miss C 99 2 2.0 -71.0 9.1 -71.7 0.37
Easterby, T D 62 3 4.8 -36.3 22.6 -58.5 0.37
Thompson, V 53 1 1.9 -44.0 11.3 -83.0 0.38
Davison, Miss Z C 57 1 1.8 -36.0 10.5 -63.2 0.41
Normile, Mrs L B 67 1 1.5 -54.0 9.0 -80.6 0.43
Goldie, J S 68 3 4.4 -42.0 19.1 -61.8 0.44
Candlish, Jennie 129 5 3.9 -90.5 22.5 -70.2 0.47
Frost, J D 76 2 2.6 -37.0 7.9 -48.7 0.47
Bewley, G T 62 3 4.8 -42.8 27.4 -69.0 0.49


That’s a combined 30 wins from 1290 attempts with a A/E performance on average of 0.30.  Ordinarily I’d like to keep table data to a top 10 or so, but in this case, it felt a bit like a civic duty to share it all!

It goes without saying that if you’re backing a runner from these stables under these conditions that you need a very compelling reason to argue against the data. Obviously, it doesn’t mean that they can’t win – and horse can win any race – and, as ever, sample sizes are sub-optimal. Treating all of these stable runners with caution under these circumstances is advised.

The yards contained on the bargepole list are generally of the small/mid-range in terms of size.  Of greater interest may be to evaluate some of the household names of the game with the same conditions applied.  The table below contains larger outfits (100+ runs and not included in the first list above).  All have A/E rates of 0.8 or lower for horses where they are absent from competitive racing beyond the 180 days limit.


Trainer performance for all runners since 2010 where the horse last ran over 180 days previously (min. 100 runs at A/E less than 0.8)

Trainer Runs Wins Win% P/L(SP) Place% ROI(SP) A/E
Gordon, C 110 5 4.6 -48.1 18.2 -43.8 0.51
Webber, P R 188 8 4.3 -104.5 17.6 -55.6 0.51
Keighley, M 145 9 6.2 -61.6 22.1 -42.5 0.51
Dobbin, Mrs R 127 7 5.5 -83.0 21.3 -65.4 0.58
Williams, Ian 180 13 7.2 -72.5 25.0 -40.3 0.59
Hammond, Micky 110 5 4.6 -75.2 14.6 -68.3 0.59
Smith, Mrs S J 308 24 7.8 -104.2 26.3 -33.8 0.65
Russell, Lucinda V 332 27 8.1 -147.5 27.4 -44.4 0.66
Richards, N G 220 24 10.9 -68.5 35.0 -31.2 0.67
Hill, Lawney 120 10 8.3 -41.1 24.2 -34.3 0.67
Down, C J 106 4 3.8 3.5 17.9 3.3 0.67
Case, B I 102 6 5.9 -34.2 22.6 -33.5 0.7
Phillips, R T 115 5 4.4 -53.0 19.1 -46.1 0.71
Wade, J 166 10 6.0 -76.3 22.9 -45.9 0.72
Alexander, N W 169 12 7.1 -84.1 20.1 -49.7 0.73
Greatrex, W J 250 40 16.0 -106.7 37.2 -42.7 0.73
Wadham, Mrs L 118 13 11.0 -6.3 31.4 -5.3 0.74
Jefferson, J M* 163 20 12.3 -63.5 33.7 -39.0 0.75
Mullins, J W 175 11 6.3 -70.5 19.4 -40.3 0.76
Bailey, Caroline 100 7 7.0 -34.8 22.0 -34.8 0.77
Moore, G L 317 32 10.1 -144.1 24.6 -45.5 0.79
Dickin, R 116 7 6.0 -41.3 17.2 -35.6 0.79

*J M Jefferson yard now overseen by daughter, Ruth. It remains to be seen whether she adopts the same patient approach


A lot of these are undoubtedly considered elite level exponents of the training game.  They all will have short priced horses making their seasonal reappearance right about now.   Across the board the win strike rate is a moderate 8%.

On a personal level, awareness of this data has resulted in a modification of my betting habits over the last few weeks.  Sure, sometimes using intel such as this will leave you kicking yourself as you leave a winner out but it’s all about getting a few more right than wrong in the long-term.


Your first 30 days for just £1

Winter Sunshine

Enough with the negativity. Let’s find a few rays of winter sunshine. Using the same 180 days off the track criteria with the addition of only considering runners at an SP of 20/1 or less (to prevent one or two big winners skewing the data) I’ve curated the following, more optimistic, data set.  This time I’ve sorted by ROI: bottom line profit is the ultimate goal after all. To qualify for the winter sunshine list at least 50 runs are required, a minimum of a 10% ROI at SP and a minimum of a 10%-win rate.


Trainer performance for all runners since 2010, 180+ days layoff, SP 20/1 or shorter

Trainers Runs Wins Win% P/L(SP) Place% ROI(SP) A/E
Bridgwater, D G 98 21 21.4 53.7 41.8 54.8 1.47
Easterby, M W 51 11 21.6 24.9 35.3 48.8 1.65
Hales, A M 74 12 16.2 34.0 35.1 46.0 1.42
Pauling, Ben 100 25 25.0 37.6 45.0 37.6 1.24
Honeyball, A J 119 26 21.9 41.2 43.7 34.6 1.14
Walford, Robert 59 10 17.0 18.9 30.5 32.0 1.33
Scott, J 118 20 17.0 30.4 39.8 25.8 1.23
Symonds, Tom 64 10 15.6 13.6 43.8 21.2 1.09
Williams, Evan 339 60 17.7 64.0 39.2 18.9 1.07
Scudamore, M J 74 11 14.9 13.3 35.1 17.9 1.27
Leech, Mrs S 74 10 13.5 12.8 27.0 17.2 1.14
OBrien, Fergal 189 38 20.1 29.7 42.3 15.7 1.12
Dartnall, V R A 119 19 16.0 15.8 40.3 13.2 1.09


A much more interesting set of results for backers, all pretty positive and all worth further investigation.  As usual it’d be remiss not to have a quick dive into the most profitable on the list, in this case the Cotswolds-based trainer, David Bridgwater.


David Bridgwater runners after a break of 180+ days, SP 20/1 or shorter by year

Year Runs Wins Win% P/L(SP) Place% ROI(SP) A/E
ALL 98 21 21.4 53.7 41.8 54.8 1.47
2018 7 2 28.6 19.0 71.4 271.4 2.41
2017 3 1 33.3 2.0 66.7 66.7 1.18
2016 13 1 7.7 -7.0 23.1 -53.9 0.71
2015 25 8 32.0 15.0 44.0 59.9 1.65
2014 15 2 13.3 0.5 40.0 3.3 1
2013 20 3 15.0 20.0 35.0 100.0 1.4
2012 6 1 16.7 -2.0 16.7 -33.3 1.23
2011 6 2 33.3 5.8 66.7 95.8 2.27
2010 3 1 33.3 0.5 66.7 16.7 1.89


Judged on this criterion, “Bridgie” has clearly peaked between 2013-2015 in terms of volume. However, he still appears to get his horses primed after a layoff these days, just in lower numbers.   Perhaps the increased activity during the peak years were as a result of his stable star The Giant Bolster finishing 2nd, 4th and 3rd in consecutive Gold Cup’s at Prestbury Park, thus raising the profile of the operation.  Delving slightly deeper into the data the performance is strong in the rank and file classes of NH racing (4 and 5), with 19 winners from 70 runs, ROI of 106% at SP. That’s probably an angle to keep in the back of your mind I suspect, rather than to follow blindly.

Picking another yard in a semi-random way (as I have an affinity for them) let’s check the Ben Pauling outfit. Willoughby Court signalled a change in fortunes with regard to my woeful Cheltenham Festival record back in 2017 and I’ve been following them ever since that momentous occasion.  The expanding yard is coming off the back of its most successful season and is clearly going in the right direction.

The beauty (or one of them) of evaluating data such as this is that it can act as a gateway into a deeper understanding of a trainer, generating a different angle or view to what was initially expected.  Let me illustrate:

Pauling’s 25 wins from 100 with a 37% ROI looks overwhelmingly positive (and it is), however, here is the breakdown by month 

Ben Pauling runners with 180+ off the track at SP of 20/1 or shorter by month

Month Runs Wins Win% P/L(SP) Place% ROI(SP) A/E
January 4 1 25.0 9.0 50.0 225.0 2.08
February 4 0 0.0 -4.0 0.0 -100.0 0
May 5 2 40.0 -1.0 40.0 -20.0 1.3
June 3 0 0.0 -3.0 66.7 -100.0 0
July 1 0 0.0 -1.0 0.0 -100.0 0
October 20 1 5.0 -14.0 35.0 -70.0 0.27
November 46 13 28.3 11.2 50.0 24.4 1.31
December 17 8 47.1 40.4 52.9 237.5 2.31


Look at October in relation to November and December.  They are pretty powerful numbers (small sample small-print applies).  In fact, they’re so powerful I have the strong inclination to check all of Pauling’s runners, irrespective of whether they’ve had over 180 days rest or not.  The graph below shows the split of profit and loss by month for all of the stable’s runners at 20/1 or shorter.


Ben Pauling P&L performance by month for all NH runners at 20/1 or shorter from 2010 onwards


The first thing to say is that the trend from the 180+ data is very much a representation of the whole yard’s performance.  Backing every Pauling entry during November and December appears to be a very promising area in which to potentially invest the kid’s university funds.  The whole stable appears to go into overdrive as we get towards the dying embers of the calendar year.

As a final and potentially arbitrary step, the Pauling record in Nov/Dec with fillies and mares is very poor with just one win from 28 runs.   Checking the overall year-round performance with the fairer sex there have been a skinny 5 wins from 68 runs, losing over 70% of funds invested.  As a result, I’ll happily exclude fillies and mares from the angle: training these has unique and different challenges, so exclusion can, I feel, be justified. That leaves the overall angle performance as per the table below.


Ben Pauling November/December male runners by year with, SP of 20/1 or shorter

Year Runs Wins Win% P/L (SP) Place% ROI(SP) A/E
ALL 205 56 27.3 104.7 46.8 51.1 1.35
2018 5 1 20.0 0.0 40.0 0.0 1.75
2017 72 18 25.0 14.2 47.2 19.7 1.17
2016 56 13 23.2 2.1 41.1 3.8 1.17
2015 36 15 41.7 54.3 52.8 150.8 1.72
2014 31 8 25.8 35.1 51.6 113.3 1.57
2013 5 1 20.0 -1.0 40.0 -20.0 1.56


In summary, backing Pauling male runners in November and December at 20/1 or shorter returns 51% to SP with a healthy strike rate of over 27%.  Maybe the market is catching up and pickings have certainly been slimmer over the past year or two.   Having said that, the yard is definitely still one to keep close to your thoughts as soon as we move into November.

Another trainer from the Winter Sunshine list, this time entirely based on volume, is Evan Williams.  The Vale of Glamorgan handler has delivered 50+ National Hunt winners every year since 2010 and is on track to do so again in 2018.

There is little doubt that this is an operation that gear themselves to getting horses out fresh and ready in October and November.   Using the P&L graph again, below is the distribution.

Evan Williams P&L performance by month for all runners 180+ days off the track, 20/1 or shorter, since 2010


A nice profit has been gleaned in the focus months; unlike Pauling, however, there are other potential periods of interest. Also, whilst the Pauling yard is historically flying with all runners in months 10 and 11 there is a clear distinction in Williams’ stable between fresh and already active animals.


Evan Williams Oct & Nov runners by month from 2010 by days since last run, SP 20/1 or shorter

Days since LR Runs Wins Win% P/L(SP) ROI
180 days or less 456 78 17.1% -84.0 -18.4
181 days or more 190 45 23.7% 84.4 44.4


As a result, we only want to consider the fresh horses from the yard, even though performance for the other horses is far from terrible.

If we want to sharpen up further, the trainer hasn’t had a victorious horse at odds of greater than 16/1 from 11 runs in this dataset.  There might be a big one out there though, as always, it’s personal choice in terms of appetite for risk and reward.

In summary, backing all Williams charges with over 180 days off the track in Oct/Nov at 16/1 or less would yield 53% at SP, delivering £95 profit from a £1 level stake.

I’m fully aware that October is in the rear-view mirror in 2018.  However, this year the stable was exceptionally quiet during the month in terms of qualifiers, finishing with a record of 0/5.  My guess would be that the exceptionally dry summer and autumn may be pushing this (and other) yard’s general routines back a few weeks, patiently waiting for winter ground.  If that is the case, then Team Williams may burst into life as the squad start hitting the course over the next few weeks.

Obviously, it doesn’t always work like that, it’s just part of the evolving punting puzzle that we all know and love.

Good luck!

 - Jon Shenton

NH Season Fast Starters

As I’ve alluded to in previous articles I would consider myself more of a flat game specialist, writes Jon Shenton.  However, with the onset of winter and the monumental battle of wills around when to put the heating on, perhaps you could argue that my timing is less than impeccable in terms of becoming a contributor to Geegeez.

Data are data, though – and in some ways the fact that I’m not invested so much in the history, the characters and the equine stars of the show arguably means I can be more objective about what I’m looking at.  In other words, the data can speak for themselves.  Every day is a school day and I’m hopeful that I can build some profitable and interesting angles to keep things ticking over during the cold, dark months when I’m wrapped in a blanket because I’m too tight to fire up the boiler!

In this article, I will try to unearth a bit of early season value with regard to the winter game.  That said, and as a starter concession, I still can’t work out officially when the National Hunt season starts.

As ever a reminder that analysing past performance is no guarantee of future spoils; but, as a minimum, it should help in generating ideas and approaches for evolve our knowledge and therefore our betting skill.

Let’s start with a broad-brush approach evaluating National Hunt runners by trainer during the months of October and November.  This time all the data have been crunched using the Query Tool on this very site, any runners on or after 7th October 2018 are not included.

All National Hunt runners with an SP of 20/1 or shorter by trainer in October and November from 2012 onwards

The table above displays trainers ordered by the best return on investment (ROI) at starting price (SP).  Encouragingly, there are nine of them returning over 10% without diving any deeper.

Top of the tree and first cab off the rank is Henry Oliver, the Worcestershire-based trainer who is returning a very substantial 61% over the period in question: it’s stating the completely obvious but that’s worth more than a quick glance.   First stop is to check the context of this apparent seasonal bounty, it may be that Mr Oliver is an all year-round cash cow.

All Henry Oliver National Hunt runners with an SP of 20/1 or less from 2012 onwards

If you backed every Oliver NH runner from January 2012 you would have a neat 5% return to SP with 90 winners from 534 bets.  Not quite ‘cash cow’ status but there are certainly worse ways to put your money on the line.  The below graph shows how the 26.7 points of profit is split by month.

Monthly P&L to a £1 level stake for all National Hunt runners at 20/1 or shorter from the Henry Oliver stable from 2012 onwards


First thing to note is that, like a number of NH trainers, the summer months are fallow for Oliver’s charges.  December aside, Oliver is operating at a profitable level over the winter months and I wouldn’t put you off tracking all stable runners over the core NH season so certainly a trainer to follow.

However, we started searching for early season value and clearly November sticks out like Brian Blessed playing hide and seek, returning 94% profit to ROI.  The 20% October ROI is worth noting, too.

Trying to dive deeper into those autumnal runners, evaluating variables such as obstacle type, race class, horse age or date of recent run doesn’t generate anything of real material value.   If you’re nit-picking, Oliver’s horses are 0/11 for runs greater in distance than 2m 6f in those months and 5/58 overall, something to keep an eye on.

The last metaphoric hurdle is to understand the consistency aspect of the performance.

The table below shows Oliver’s October/November runs by year.  Maybe a little streaky but scintillating performance in 2013, 2015, and in particular 2017, with a bit of a washout in 2016.  Only one losing year though (excluding 2018 thus far for hopefully obvious reasons) means that this is solid enough to go on the list!

All Henry Oliver National Hunt Oct/Nov runners with an SP of 20/1 or shorter from 2012 onwards

Suggestion: Back all Henry Oliver runners in October/November at 20/1 or less


Your first 30 days for just £1

The second luminary of the list is Fergal O’Brien, who quite simply has the best (in my opinion) and most entertaining twitter profile of all of the trainers, well worth a follow (@FOBracing) if you’re active on that medium. The stable contains relative household names such as Chase The Spud, Cap Soleil, and their first Grade 1 winner Poetic Rhythm to name but three of them.

There is no doubt the yard has impressive credentials and performance has been very strong over recent years.  If you backed every single stable runner at SP from January 2012 you’d walk away with 3.7% more cash than you invested.

I think there are angles aplenty when it comes to O’Brien, most of which are for another day but with specific reference to the early season view there are a couple of options to home in on for profit. The first is National Hunt race code

All Fergal O’Brien Oct/Nov National Hunt runners with an SP of 20/1 or shorter from 2012 onwards by race type

All profitable, which is nice. The pertinent angle for me though is his performance in bumpers, where O’Brien has nearly double the volume of winners than expected with a 186% return to boot.  Yes, the sample size is small, but within the data there are ten winners from horses making their debut (from 22), indicating that the yard gears up to get quality horses (or horses ready to win) out on the track in the months of October and November. Generally speaking, the later in October, the better as the record is 1/11 from the 1st-16th.

Profit in relation to hurdles and fences is quite small over those two months; however, if we zoom in a little closer there is a quite telling split in monthly performance, again it looks like the stable is peaking in November.

All Fergal O’Brien Oct/Nov Hurdle & Chase runners with an SP of 20/1 or shorter from 2012 onwards by race type

It’s not an absolute rule, and certainly doesn’t mean that a horse on the track on the 1st November is in different shape to one on 31st October, but it does indicate generally that as we start heading towards the big November Cheltenham meeting, the O’Brien yard picks up pace and is a definite one to follow closely.

Suggestion 1: Back all O’Brien NHF runners in late October/November at less than 20/1 SP

Suggestion 2: Back all O’Brien Chase and Hurdle runners in November at less than 20/1 SP


Moving to the trainer in the bronze medal position in the opening table, Harry Whittington: the Lambourn-based outfit is growing rapidly, currently housing nearly 50 horses with an increasing number of runners per year. I like these yards that are growing, it often means they’re on an upwards trajectory and are worth closer review.

First port of call is checking the race type in the table below, a small number of runners but the bumper aspect doesn’t look entirely compelling so I’m happy enough to exclude and keep a watching brief.

All Harry Whittington Oct/Nov National Hunt runners with an SP of 20/1 or shorter from 2012 onwards by race type

Again, evaluating the profile of Whittington’s hurdle and chase runners across the whole year gives an interesting picture in terms of P&L.  The graph below shows that very same P&L by month to a £1 level stake, it’s fair to say that Q4 looks quite compelling – another yard that’s fast out of the blocks for the new season.

Monthly P&L to a £1 level stake for all National Hunt runners at 20/1 or less from the Harry Whittington stable from 2012 onwards


If we analyse the October to December runs in terms of race class as a differentiator there is a further shard of light to assist profitable punting.

All Harry Whittington Oct-Dec Hurdle and Chase runners with an SP of 20/1 or shorter from 2012 onwards by race class


The basement C5 races are easy enough to ignore in punting terms, most of them crossing over with the NHF group we already discounted; the Class 1 & 2 are less straightforward, particularly if the yard’s expansion means they may be knocking on the door of the higher echelons of the racing ladder. Here and now I’d be inclined to back the C3 & C4 horses and track the C1/2 runners for signs of improvement or add to a shortlist to back on their relative merits.

Suggestion: Back all Harry Whittington’s October, November and December Chase/Hurdle runners at less than 20/1 in Class 3 or 4 races.


The final trainer I’m going to run through from the initial table is Venetia Williams, largely due to her volume of runners: to deliver a 17% ROI across 440 runners in the months of October/November from 2012 onwards is impressive and merits closer scrutiny.  That’s not to say all of the other trainers are not worthy of further investigation and I’d definitely be inclined to sharpen the focus on Messrs Pauling and Keighley in particular.  Have a play on QT yourself and maybe post anything of interest (or otherwise) in the comments below.

Returning to Venetia Williams, the Grand National-winning trainer has a profitable record during the months in question, but the below table tells a stark tale.  Clearly, Williams has a knack for getting her cavalry of chasers ready early in the season

All Venetia Williams Oct/Nov National Hunt runners with an SP of 20/1 or shorter from 2012 onwards by race type


Again, if we look specifically at the month, the record in November is much stronger than that of October.

Perusing the “Venetia” page at her website www.venetiawilliams.com the following sentence caught my eye:

“Since then Venetia's career has flourished. Never one to expose her horses to the high risk of summer ground, each year Venetia can be seen with the big Saturday winners during the core NH season”

There is a common belief that Williams’ runners love soft turf, and the statement above also seems to indicate a preference to avoiding the risks associated with summer ground.  On Geegeez we like facts to back up a theory, so the table below shows Venetia’s chase runners in November by official going.

All Venetia Williams Chase runners in November with an SP of 20/1 or shorter from 2012 onwards by official going


While there is confirmation that Williams’ runners prefer a softer surface, it is worth noting that the stereotyped ‘hock deep’ runner from this yard fares less well than those encountering merely ‘winter ground’, i.e. good to soft or soft.

There is one mild concern with the overall angle though, namely 2017 performance, showing a loss of 28%, this is also on the back of a moderate 2016.  It could be this angle has run its natural course, albeit I will be adding it to my own armoury this November.  Williams had a very quiet spell last winter, alluding to a potential problem in the yard so I’m just about happy enough to strike a line through 2017.  This is one for keen observation though.

All Venetia Williams Chase runners with an SP of 20/1 or shorter from 2012 onwards on good to soft, soft or heavy ground by year

Oh, and incidentally the Saturday assertion in the quoted sentence does have a degree of credence too.

Suggestion: Back all Venetia Williams November Chasers on Good to soft or softer ground with a 20/1 or less SP (with caution)

- Jon Shenton

Your first 30 days for just £1