Posts

Musselburgh Draw & Pace Bias

Draw and Pace at Musselburgh

For this article we are back across the border to analyse draw and pace data from Musselburgh racecourse, writes Dave Renham. To help me with this piece I have used some of the tools available on the Geegeez website, those being the Draw Analyser, the Pace Analyser and the Query Tool.

I will be looking at race data going back to 2009 as my starting point but, as before, I will examine a more recent data set in detail, too (2015 to 2019), where appropriate. The focusas with all the other articles in the series, is on handicap races with eight or more runners.

Musselburgh Course Constitution

Musselburgh is a right-handed course roughly ten furlongs in circumference, with no notable gradients, and is generally considered to be fair. The 5f sprint trip is raced on a straight track while 7f races and above take place on the round course. (There are no 6f races).

 

Musselburgh 5f Draw Bias (8+ runner handicaps)

Since 2009 there have been 218 qualifying races over the past 11 seasons, a significant sample, and here are the draw splits: 

The general perception I think is that horses drawn next to the stands’ rail (high) have an advantage. There is a kink in the straight track after two furlongs and, in theory, that should aid those runners drawn high. However, the stats for 8+ runner handicaps do not especially back that up, such horses winning only as much as middle draws, and neither group performing distinctly better than low starting stalls. Now a look at the A/E values:

Middle draws seem to offer better value than higher draws despite their similar win percentages. This does imply, albeit only slightly, that maybe higher draws are slightly overbet due to the perception of draw bias.

However, when the field size increases a slight bias does start to appear. In handicap races of 11 or runners (90 races) we get the following splits:

Thus, in bigger fields, horses drawn out wider (lower stall numbers) definitely start to struggle. The A/E values back this up too.

Again middle draws offer the best value out of the three draw thirds.

Ground conditions do not appear to make any difference to the draw so let us move on to to looking at each draw position broken down by individual stall number.

For this distance I have needed to change the way I collate the data. The reason for this is that the higher draws are positioned next to the rail so in many respects analysing individual stall positions in the ‘normal’ way becomes irrelevant. What I mean by this is, that stall 8 could be drawn next to the rail (in an 8-runner race), but in a 17-runner race stall 8 is actually ten stalls away from the rail. Hence I am using a trick that Nick Mordin used many years ago in his book Betting For Living when he flipped the draw. I am reversing the draw figures if you like and looking at them in their relation to their position near to the stands’ rail. I still used the Geegeez Query Tool to give me the relevant data, but it took me more time to adjust and sort out the final figures:

These stats indicate there may be a slight stands’ rail bias as horses drawn 2, 3 or 5 stalls from the rail are all in profit. Also the each way percentages for those drawn within five of the rail are all over 30%. Having said that, it is not something that one could be too confident about. What I would be more confident in is that horses drawn ten berths or wider from the stands’ rail look at a disadvantage. This correlates with the 11+ runner draw splits mentioned earlier.

Onto a more recent data set looking at the past five seasons (2015-2019). Here are the draw splits for the 100 races that have occurred during this time frame.

No surprises here with an even looking split.

The A/E values correlate with long term figures shared earlier:

Again middle draws have offered the best value.

 

Time for the 5 year stats for individual draw positions with the same twist as discussed earlier (draw positions effectively reversed):

The slight rail bias that was mooted earlier is not displayed with this more recent data set. However, as you would have probably expected the stats indicate that horses drawn ten or further from the stands’ rail remain at a clear disadvantage.

Musselburgh 5f Pace Bias (8+ Runner Handicaps)

Let us look at pace and running styles now. The overall figures (2009-19) are thus:

As is often the case, front runners enjoy a decent edge – as 5f biases go it is around the overall UK course average. Hold up horses have a poor record and look best avoided unless the pace is likely to be frenetic.

The front running bias does seem to strengthen slightly the firmer the going. The stats for qualifying races on going described as good to firm or firmer is as follows:

Improvements in strike rate, A/E value and IV; also the each way placed percentage increases too.

In terms of field size there is no clear change in front running bias.

Finally in this five furlong section a look at draw / pace (running style) combinations for front runners over this minimum distance. Remember this is looking at which third of the draw is responsible for the early leader of the race (in % terms):

Your first 30 days for just £1

Higher draws get the lead more often than any other third. You would expect this as they are drawn closest to the rail. I must admit that I had expected the high draw percentage to be a bit nearer to 50%.

The draw/run style heat map, sorted by Percentage of Rivals Beaten, again points to early leaders from a pace perspective and middle to high from a draw perspective. (Any score above 0.55 implies a bias to that section, below 0.45 a bias against that section).

 

To conclude, in terms of the draw, higher draws are at a disadvantage as the field size gets bigger, with draws ten or further away from the rail having a particularly poor record. Pace wise, front runners have the edge and this seems to strengthen on firmer ground.

 

Musselburgh 7f Draw Bias (8+ runner handicaps)

As mentioned, there are no six furlong races at Musselburgh, so the next distance we'll review takes in the round course and the seven furlong (seven-eighths of a mile) range. The 7 furlong trip has had 189 qualifying races from 2009 to 2019 which is another decent sample. Here are the draw splits:

The 7f trip sees low draws start closest to the inside rail. However, this does not appear to give them any concrete advantage.

Let’s look at the A/E values to see if they correlate with the draw percentages:

Similar A/E values offering no real edge.

Drilling into the stats when the going gets softer there is a suggestion that low draws have an advantage. The problem is that there have only been 19 races on soft or heavy ground. Having said that, 12 races have been won by low-drawn runners compared with just two for higher-drawn horses. The placed stats strongly favour lower draws, too, under such conditions, but 19 races is far too small a sample to take at face value.

Time to look at what the individual draw positions offer over the 11-year period between ’09 and ’19. We can view these in the normal way:

Nothing particularly significant here as one might expect looking at the other draw data. However, draws 1 and 2 clearly have the best placed strike rates which is interesting.

On that theme you could have made a 36 point profit backing the two lowest draws in one point reverse forecasts over the 189 races. There were enough winning bets to create a small profit. For tricast fans, perming the three lowest draws in full cover tricasts would have yielded a huge profit of just under 3600 points! There were only five winning tricasts, though, and the profit basically relied on one monster payout.

Onto the last five seasons for 7f handicaps at Musselburgh. There have been 94 qualifying races since 2015, with the draw splits as follows:

These are similar figures to the longer term ones. Higher draws have performed slightly worse in the last five years but it is likely not statistically significant.

Onto the A/E values (2015-2019):

Middle draws have been the best value of the three draw thirds in the last five seasons. However, there is no edge to really take advantage of.

Now a look at the individual draw figures for this latest 5-year period:

Again nothing clear cut and ultimately 7f races offer little interest for the draw punter (despite those aforementioned forecast and tricast figures). The PRB3 data - a rolling three-stall average of percentage of rivals beaten - suggests that the course constitution does slightly favour inner-drawn horses, though this has so far yet to manifest itself in bottom line profit. Nevertheless, it is worth being aware of.

I will be looking closely at any future races on softer ground, though, as it is possible that there could be a low bias under those conditions. Here is the same view, but on soft or heavy going:

 

Moving on the seven-furlong handicap pace data, here are the overall pace figures going back to 2009:

This makes much better reading and front runners have a very strong edge, even more so than over 5f. More recent data offers a similar picture so this is a bias that we must try and use to our advantage.

This front running edge looks to be stronger as the ground starts to soften. On good to soft or softer there have been 47 races giving the following splits:

There also seems to be a slight increase in front running bias when the field size grows. In races of 11 or more runners, front runners win 21% of the time with an A/E value of 1.85; in races of 8 to 10 runners the strike rate is still 21% but the A/E value drops to 1.50. It should be noted that mathematically it is harder to win in bigger fields so even though both win percentages are at 21%, it is clear that in effect front runners have been more successful in bigger field races.

Let us now look at the draw / pace (running style) combinations for front runners over 7f.

Lower drawn horses get to the early lead more often – they are positioned closest to the inside rail so this is what we should expect. Having said that I would have expected a higher figure than 40%.

The draw / run style heat map offers a perfect diffusion of green to dark orange when viewed on PRB; this is normally a strong indication of a repeatable bias:

To conclude, over 7f the draw in general is extremely fair, but possibly lower draws have an edge in soft or heavy conditions. Pace wise, however, front runners have a bankable edge in all conditions which seems to increase on good to soft or softer going.

 

Musselburgh 1 Mile Pace Bias (8+ Runner Handicaps)

I will start our mile handicap analysis by looking at the 2009-2019 data - 90 races during this period have given the following draw splits:

There is no clear draw bias looking at these stats, but when you break the data down into halves, the bottom half of the draw won 61.1% of races to the top half figure of 38.9%. Hence a slightly lower draw seems preferable.

Let us break the mile draw data down by stall position:

Draws 1 and 2 both have decent A/E values and, breaking the data down further, stalls 1 to 4 have been 2.1 times more likely to win than draws 10 or higher (A/E values of 0.93 versus 0.66). Hence taking all things into account a lower draw seems preferable over a very high one, as reflected in the below IV3* chart:

*more information on IV3, and all of our metrics, can be found here.

 

The last five seasons have seen a fairly even split draw wise when splitting into thirds; draws 10 or higher have continued to struggle winning just twice from 37 runners. A look now at the pace findings for this 1 mile trip going back to 2009:

As with the two shorter distances, front runners have a definite advantage over a mile. This is one of the strongest mile pace biases in the country and it should also be noted that exactly half of all front runners went onto finish in the first three. Horses held up at the back early do not have a good record once again. The bias is consistent across all going and field sizes, although you could argue that in smaller fields (8-9 runners) it has been slightly less potent.

Finally in this section a look at which part of the draw gets to the lead first:

Although lower draws are positioned next to the rail, they do not get to the lead the most. This is probably due to the fact that there is nearly 4 furlongs until the first (and only) turn and wider drawn jockeys are keen to get a more expedient trip.

Again, we can see the golden triangle when looking at draw / run style in concert, though this time it more a 'led' bias, with a mark up for low drawn prominent and midfield (ground saving) racers.

 

As with 5f and 7f handicaps, over one mile the front running pace bias offers the most interest and it is a strong one. Draw wise I would always prefer lower draws over higher but all in all I don’t perceive it to be a significant factor.

 

Musselburgh 1 Mile 1 Furlong Pace Bias (8+ Runner Handicaps)

There have been only 44 qualifying races at this distance but some interesting findings: 

Higher draws seem to have an edge and the A/E values strongly correlate:

My concern with these figures is that they are not easy to explain – if low draws had this advantage I would assume there was an inside rail bias; with higher draws having the edge it makes virtually no sense. The most likely scenario is simply down to variance as the sample size is not that big in reality. However, it may be that jockeys are able to play more of a waiting game by dropping high-drawn horses in at the back of the pack. All may be revealed shortly!

Let’s break the data down by individual draws to see if that helps:

It is difficult to make much of this either – the unusually good stats for stall eight reinforces my belief that the draw splits cannot be relied upon.

Onto pace now, and below is performance by run style.

Once again at Musselburgh we have a decent front running bias and hold up horses have an even worse record than the three shorter distances, so bang goes that theory about why wide-drawn horses have fared best!

This is surprising as normally the longer the distance, the harder it is for early leaders to make all the running; likewise longer distances normally see a much higher percentage of wins for hold up horses.

To conclude, there is a strong pace bias for the fourth consecutive distance over 1m1f. The draw stats suggest a high draw bias; but, as stated earlier, I am struggling to rationalise this in the overall context, even though the PRB data support the win and place tables above. Weird!

 

Musselburgh 1 Mile 4 Furlong Pace Bias (8+ Runner Handicaps)

This is the longest distance I have looked at in any of the articles but I would like to share one set of stats. The draw is fairly even and, over 12 furlongs where they start just before the winning post and make a full loop of the track, I do not feel it is worth going into too much detail.

But pace wise we continue to see that front running edge, even over this relatively long trip. Here are the 2009-2019 stats, taken from the geegeez Pace Analyser:

The figures suggest that this may be the distance where the front running edge is at its strongest. This is very surprising given the distance we are talking about. Maybe it is down to the fact there is additional sharp bend soon after the start at 1m4f and front runners get more of an advantage going the shortest route into that turn.

 

Musselburgh Draw and Pace Bias Summary

Although there is little out of the ordinary in draw terms, Musselburgh is a course of real interest when viewed from a pace angle. Looking for potential front runners at all distances from 5f to 1m4f is definitely a strategy worth considering. The draw is generally not a major factor but there are subtleties that one needs to be aware of.

Thanks, as always, for reading, and good luck!

- DR