Dancing Star and David Probert win the Qatar Stewards Cup at Goodwood. 30/7/2016 Pic Steve Davies/Racingfotos.com

Past Pace as a Predictor of Future Performance, Part 2

This is a follow up piece to the article I shared with readers earlier this month, writes Dave Renham. In that article I highlighted any horse aged four or older, that in the 2019 flat season ran at least ten times in sprint handicaps (5-6f). This gave me 303 individual horses to review from which I started to examine their individual run style data for each race they ran in 2019. This included all their races, both handicap and non-handicap, and over any distance; and it accounted for over 4,000 races.

Run style (pace) data on Geegeez is available for every single race, both flat and National Hunt, and is split into four sections:

Led – the horse or horses that take the early lead;

Prominent – horses that lay up close to the pace just behind the leader(s);

Your first 30 days for just £1

Mid Division – horses that race between the middle of the pack but in front of the rear ‘quarter’;

Held up – horses that are held up at or near the back of the field.

Points are assigned to each running style with leaders getting 4, prominent 3, mid division 2, and hold up horses 1.

My aim for the first article was to try and determine whether recent pace data was a good predictor of future run style. I think the numbers in part one proved that beyond reasonable doubt. In this article I am going to dig still deeper and share more pace stats with you.


Last four starts

Geegeez pace maps provide pace / running style data for each horse for up to their last four UK/Irish runs. Full career run style records can be found in the Full Form section of the cards. As I have chosen to examine horses aged four or older that have raced numerous times already in their career, the data collected always contains their pace scores from all of their previous four runs, thus the pace score totals for each horse can vary from 16 (four races where they have led early) to 4 (four races where they were held up).

Let us first look at how likely a horse was to lead if they had led early in at least one of their last four starts; comparing this with horses that didn’t lead in any of their last four starts.


As the graph shows there is a huge difference between the figures. Horses that have led early at least once in their last four starts, went onto lead 26.7% of the time next time out; those horses that had not led in any of their last four starts took the lead just 6.3% of the time on their next start.

It is clear therefore that a horse that has not led in any of its recent races is unlikely to lead early in its next race: approximately one horse in 16; whereas horses that have led at least once in their four previous races have a slightly better than 1-in-4 chance of leading early next time.

Let us now analyse how likely horses are to lead on their next start when we compare how frequently they led in their most recent four races.


I have tilted the graph horizontally just to mix things up a bit! There is essentially good correlation here – the ‘led in all 4’ percentage is slightly lower than the ‘led in 3 of last 4’, but that is probably down to the fact that the ‘all 4’ sample was relatively small and possibly slightly skewed.

All in all, horses that have led consistently in their most recent runs will lead more often than horses who have led less regularly, who in turn will lead more than horses that have not led recently.


Those that had led exactly once LTO

Before moving on I want to take a quick look at horses that led in just one of their last four starts, the reason being that I wanted to see if the position of that run made a difference. Here are my findings this time in tabular form:

This was a slightly disappointing set of results in truth as I was hoping to see a greater difference between the top percentage in the table and the bottom. However, there is still reasonable correlation, with those who led most recently more likely – relatively, at least – to lead again this time.


Horses with pace scores totalling 13 or 14 in their last four runs

When discussing running styles / pace and, in particular, when looking at the last four runs for a particular horse, it is perhaps easier to think of it in numerical terms following the 4, 3, 2, 1 points system allocated by Geegeez. I looked at horses gaining 15 or 16 points in their most recent four starts in the first article, and below are data for horses that have scored either 13 or 14 points in total. These are the possible combinations that produce 13 or 14 points.

Each set of four pace scores do not necessarily occur in the order shown above: a 4,4,4,2 pace combination, for instance, could occur in four different ways in terms of the order of pace styles in the last four races:

There is nothing mind-blowingly significant about this, I just felt it important to clarify that there are different orders of the same combination.

I wanted to understand how likely a horse with 13 or 14 points was to lead on its next start when we compare the number of 4s (number of times it led early) in its last four runs.

Horses that have led just once in their last four starts (one 4) would have any combination  of 4,3,3,3; those who have led twice (two 4s) could have either the 4,4,3,3 combination or the 4,4,3,2 combination. Different combinations of 4,4,4,2 and 4,4,4,1 would have seen a horse lead three times in their last four races (three 4s).


This graph illustrates what one would hope – the more 4s (early leads) in their last four runs, the more likely they are to lead in their next race. Roughly 45% of horses that had different combinations of 4,4,4,2 or 4,4,4,1 led in their next race.


Horses with pace scores totalling 11 or 12 in their last four runs

Next, using the same idea, we will look at total pace scores of 11 or 12 achieved in the last four races.

To save time I am not going to go through all the possible combinations of 11 and 12, although it is possible to create these totals with no 4s (e.g. 3,3,3,3), one 4 (e.g. 4,3,3,2) or two 4s (e.g. 4,4,2,2).

Again, I am looking to see how likely a horse with these points totals was to lead on their next start when comparing the number of 4s in their last four runs. First let us look at the data simply comparing zero 4s in the last four runs to at least one 4:

There is a significant difference here for horses with an 11 or 12 points pace total. Horses that had led at least once are far more likely to lead next time when compared with horses that have failed to lead in any of their last four starts. Now let’s compare zero 4s with one 4 and with two 4s:


Horses with 11 or 12 points go on to lead next time 27.7% of the time if they have led in 2 of those last four starts (two 4s). In turn those who led once (one 4) go onto lead 17.7% of the time. As we know from the previous table those who have not led in any of their last four starts (zero 4s) have gone onto lead 11.4%.

This is yet another clear example that more 4s in recent runs really does positively impact the chances of a horse leading next time.


Horses with pace scores totalling 9 or 10 in their last four runs

Here is a table comparing zero 4s versus one or more 4s for horses with pace totals or 9 or 10. For the record there is only one combination where two 4s would occur with a score of 10 (4,4,1,1), and it is not possible with a score of 9.

Once again the number of 4s in the last four starts does make a difference in terms of the chance of the horse leading next time out. The more 4s, the more likely they will lead again.


Last six runs

To finish I wanted to dig a little bit deeper still and look at the last six runs rather than the last four. What I have done is to create pace averages over those six runs. Again the maximum average would be 4 (six 4s) and the minimum average 1 (six 1s). I have split the averages into groups to see if the horses with higher averages are more likely to lead next time than lower ones. Here are my findings:

This table shows perfectly what I had hoped it would: horses with the highest pace averages over the last six runs lead more often than the rest, with excellent correlation between the decreasing averages and the decreasing percentages.

I further calculated the average pace figure for each group on their next start. In other words I added up all their pace scores on the next start and divided by the number of races/runs. Again we have excellent correlation as this graph shows:

Horses that have a pace average of 3.50 to 4.00 for their last six starts have produced a pace average of 3.26 on their next start. Those averaging 1.00 to 1.49 yield a much lower next time out average of just 1.53. Again, there is excellent correlation between the latest six-run pace average and running/pace style next time out.



In this article I have focused on the pre-race prospects of finding the percentage chance of horses that will lead early. The rationale is, I hope, obvious: we already know that such horses have the potential to secure us a profit if we can consistently predict which one is going to take the early lead in a sprint handicap. However, the final table, below, looks at the chance of being held up next time using our last six race pace averages, just to prove this approach helps us to predict hold up horses, too!

The table again correlates beautifully, this time in reverse, with the lowest six race pace averages having by far the highest percentage of hold up horses next time.

As we know, Geegeez provides the last four pace figures with a last four race average; for those keen to dig further, using the last six races (found in a horse’s Full Form) seems to work equally well.

As a final note, below is a 'cut out and keep' reference to the four-race data in these two articles which should prove very useful for those who bet in older horse sprint handicaps.

And finally finally, the percentage chance of each run style based on a horse's last four pace scores and the number of times it led:

I hope you'll find that useful.

- DR

Other Recent Posts by This Author:

Your first 30 days for just £1
14 replies
  1. sondrio2
    sondrio2 says:

    Wow, brilliant article, love this thanks for doing so much work, pace is starting to take a big part of my research so this helps enormously.

  2. Rotund legend
    Rotund legend says:

    Thanks for pulling this article together, which I did find interesting. My betting interest on the flat and all weather is in races up to and including one mile.
    A thought I had was, in terms of horses that rank 3.5 to 4 is: what is the performance of a horse that fits this category when stepping up in distance?
    So, for instance if a horse leads over 5 furlongs and then steps up to six – I presume it still leads ? But does it perform to the same level in terms of finishing position by doing so?

    Thanks Martin

  3. Monsieurbernie
    Monsieurbernie says:

    pace is important but i think alot of this data can be wasted if the aw track surface is riding different to usual… say the past 12 months at kempton have suited horses that are coming from the back for instance. The past 5 years worth of data will just throw us off and be backing ‘led’ horses when infact it pays to back hold up horses.

    how can we figure out what the tracks are racing like on the aw? you just get the old standard to slow … that means nothing , without watching every race for the past 6 months we wont know or does it change from week to week depending on how far they have had the tractor compressing the surface!?

    • Matt Bisogno
      Matt Bisogno says:

      These data relate to turf and AW races, Bernie.

      Regarding going on AW, standard to slow is deep; after it has been raining, tracks are often speed-favouring.


  4. Fatcat1
    Fatcat1 says:

    Another good article Dave, thanks. As a matter of interest, in todays 4.20 Chel 6f maiden sprint with 6 runners. There was a 1/2 shot who had achieved nothing in several outings and a horse called POWER OVER ME who was the only Prominent runner (no leads in race) and it was same rated fig of 8 , as was favourite but it’s price was 33/1 and it could be backed with 3 places on Betfair Played a small ew. POW!! won at 22/1. Just goes to show, Pace gives a great edge in these short races.

    • Monsieurbernie
      Monsieurbernie says:

      i backed that! but only in a lucky 63, daft bet really… the rest lost as the jockeys cottoned on you needed to lead… so the ones thereafter struggled to lead that i put in the bet!

  5. Antony
    Antony says:

    Your right Bernie I found a couple year’s ago the perceived pace at Chelmsford was not working historical data is right to a degree but things change maybe for a short period then revert back to historical I found the best way to check pace bias is query tool over the shorter date ranges but undoubtedly pace is important but you have got to be on the right side of it and query tool allows you to be sunchu

    • Monsieurbernie
      Monsieurbernie says:

      yes… this historical data was not really any good with the ever changing aw surfaces… like you have have just started recently to go back two years… might be more of benefit or even a shorter period?? but then sample sizes are getting smaller

  6. gshields
    gshields says:

    These 2 part articles on pace will forever be in the Geegeez Blog Hall of Fame. Absolutely first class. If Query Tool eventually expands to allow us to play around with total pace scores (1-16) ….and their outcomes , we could be on our way to Nirvana. Many thanks for such great insight Dave.

Comments are closed.