Using PRBs to Analyse Run Style Bias

In this article I am going to crunch some data connected with the run style of horses.

When I use the term ‘run style’ I am talking about the position a horse takes up early in the race, within the first furlong. I will be splitting these early positions up into four groups that are used by https://www.geegeez.co.uk – this site has plenty of run style data to get your teeth into as a researcher or a punter.

Geegeez splits the run style data into four distinct groups which are based on the race reading comments found in the racing press. Here is how they split the runners up:

Led – as the word ‘led’ suggests, this is the horse that take up a position at the front. Occasionally you can get two or more horses vying for the early lead and in that case, I would count them all as leaders. These runners are often referred to as ‘front runners’.

Prominent – horses that race up with the early pace just behind the leader(s). Race readers will use comments such as ‘prominent’, ‘close up’, ‘chased leaders’, ‘tracked leaders’, etc.

Mid Division – horses that race around the middle of the field.

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

As I have mentioned before, run style is one of my preferred areas of research – possibly my favourite area.

The reason for this is that I think there is still an edge to be found as a good percentage of punters still do not take run style into account. Front runners under certain circumstances have an edge over other run styles and this occurs in both flat and National Hunt racing.

This edge can be very significant as has been seen in other articles I have written on the subject and hopefully this article will show that once again.

I will be using run style data going back to 2009 and will be concentrating on 5-furlong handicap races with 7 or more runners. 

As I have mentioned in previous pieces handicap races tend to be better to use than non handicaps for this type of research due to their more competitive nature. Regular readers of mine will know I take the same approach when I do draw research sticking to handicap races only.

In this article I am going to examine 5f handicap course data sharing the PRB run style splits for each course. PRB stands for ‘Percentage of Rivals Beaten’ – this is one of many horse racing metrics I use to help me try and measure horse performance. I like this particular metric because it considers all the runners in a race.

This means we can be more confident in its findings than some other measures of performance, especially if the race sample size is on the small side. As far as PRBs are concerned the average figure should be 0.50 or 50%. Hence given a level playing field the four run style groups should all hit this mark. 

I have always been a fan of handicap races, and the shorter the better. Races over the minimum trip of 5 furlongs gives a bigger advantage to front runners than flat races over longer distances.

In addition to this, some courses offer a bigger advantage to front runners than others as you will soon see.

In the table below, I am going to share all the 5f handicap PRB data (7 or more runners) for each turf course in the UK. The courses are listed in alphabetical order:

Course	Led	Prominent	Mid Division	Held Up
Ascot	0.56	0.48	0.50	0.50
Ayr	0.62	0.53	0.48	0.45
Bath	0.55	0.52	0.49	0.46
Beverley	0.60	0.51	0.51	0.45
Brighton	0.60	0.51	0.47	0.48
Carlisle	0.56	0.51	0.50	0.49
Catterick	0.63	0.55	0.50	0.39
Chepstow	0.60	0.56	0.48	0.40
Chester	0.65	0.58	0.47	0.39
Doncaster	0.51	0.51	0.51	0.48
Epsom	0.63	0.50	0.50	0.42
Ffos Las	0.46	0.52	0.51	0.51
Goodwood	0.58	0.53	0.49	0.46
Hamilton Park	0.61	0.51	0.50	0.45
Haydock Park	0.56	0.51	0.48	0.49
Leicester	0.58	0.51	0.55	0.43
Lingfield Park	0.66	0.55	0.47	0.42
Musselburgh	0.58	0.55	0.49	0.44
Newbury	0.60	0.53	0.52	0.43
Newmarket	0.55	0.55	0.44	0.46
Nottingham	0.60	0.51	0.48	0.47
Pontefract	0.66	0.51	0.46	0.45
Redcar	0.63	0.53	0.49	0.42
Ripon	0.66	0.56	0.50	0.37
Salisbury	0.60	0.56	0.45	0.42
Sandown Park	0.54	0.50	0.53	0.47
Thirsk	0.63	0.53	0.50	0.42
Wetherby	0.52	0.57	0.44	0.44
Windsor	0.62	0.53	0.50	0.44
Yarmouth	0.56	0.51	0.50	0.47
York	0.62	0.54	0.48	0.44

As you can see only one course, Ffos Las, has a ‘led’ figure below 0.50. Meanwhile 18 courses had a ‘led’ figure of 0.60 or above.

The three highest figures for these early leaders (all 0.66) have occurred at Lingfield, Pontefract and Ripon.

I would now like to share some additional findings about these three courses over this 5f trip.

Lingfield – this minimum trip at Lingfield on the turf is not a distance that is raced a lot due to the fact that Lingfield also have races on the all weather at different times during the turf season.

However, when they do, horses that take an early lead tend to have a huge edge. The stats suggest that a front runner is around eight times more likely to win than a hold up horse. Now of course you will get more hold up horses per race than front runners, but even so these numbers help to show the true strength of this front running bias.

Prominent racers also have a much better record than hold ups so when analysing these races, I would focus on potential front runners and those that are likely to track the pace.

Pontefract – Over 53% of all front runners have either won or been placed – that is a huge figure.

If you had been able to correctly predict the front runner pre-race you would have secured a profit of £111.96 to £1 level stakes equating to a return of just over 64 pence in the £. You would have obtained a fair profit also if backing the front runner each way.

It is also worth noting that it is easier to lead at Pontefract from a low to middle draw due to the turning nature of this track and higher draw front runners are not only rarer, but their win strike rate is lower than their better drawn counterparts.

Ripon – The first stat to note is that Ripon has the lowest course PRB for hold up horses at 0.37 coupled with the joint highest for early leaders.

Hence one could argue that this C & D has the strongest front running bias over 5f in the UK.

The ground conditions though are very important as far as the bias is concerned. On good or firmer ground the bias is far stronger than it is on easier ground. The comparison for different ground conditions is shown in the table below:

Going	Led	Prominent	Mid Division	Held Up
Good or firmer	0.69	0.56	0.49	0.36
Good to soft or softer	0.56	0.55	0.52	0.42

On a softer surface front runners still do best, but only just. Under such conditions there is not much in it between front runners, prominent racers and horses that race mid division.

Hold up horses are still at a big disadvantage though. Hence keep a close eye on the going and the weather when there is a qualifying 5f handicap at the track this year.

Finally, the front running bias strengthens when the field size increases. In races of 11 or more runners the PRB for front runners stands at an impressive 0.74. I want to share the PRB figures for Irish courses now.

There are far fewer Irish courses that race over the minimum trip of 5f, only seven courses in fact, and here are the splits:

Course	Led	Prominent	Mid Division	Held Up
Bellewstown	0.66	0.59	0.46	0.39
Cork	0.71	0.53	0.42	0.45
Curragh	0.64	0.53	0.47	0.46
Down Royal	0.60	0.53	0.51	0.44
Naas	0.67	0.54	0.46	0.45
Navan	0.60	0.54	0.49	0.44
Tipperary	0.64	0.55	0.48	0.41

What immediately stands out is the strength of the front running bias across all seven tracks. All PRBs for front runners hit 0.60 or above. Let me again focus in on the three with the highest ‘led’ PRBs.

Bellewstown – Of the 58 qualifying races front runners have won 18 which equates to 31% of all races. Prominent racers have won 28 of the races from a much bigger pool of runners.

In races over this C & D I would put a line through any horses that are likely NOT to race up with or close to the pace.

Cork – Cork has the highest PRB figure of all the courses for front runners standing at 0.71. This figure increases on good going or firmer to 0.75. When considering all going conditions if you had managed to correctly predict the front runner pre-race you would have seen a return of over 100 pence in the £. Oh, for a crystal ball being in tip-top working order!

Naas – Naas follows the pattern of some other courses I have noted in this piece in terms of the fact that the front running bias increases on better ground. On good or firmer ground front runners have produced a PRB figure of 0.70.

The only issue is that there are not too many 5f handicaps each season at each of these three tracks. You could expect five or six races at each course this season so keep a close eye out for them.

So, there you have it. Ignore run style bias over 5 furlongs at your peril. Indeed, before I wind this up, it should be noted that 5f non handicaps also give a huge edge to front runners.

The problem I find with these races though is twofold. Firstly, it is harder to predict the likely front runner due to the horses often having limited past run style data to go on.

Secondly, even if you do have some concrete past horse run style data, if the likely front runner is outclassed, then it is likely to find it difficult to dominate from the front, let alone stay out in front for the full race distance. 

Before I finish, in a future piece I hope to examine the data for 7+ runner handicaps over 6 furlongs. As a rule, the front running bias is less strong, but it is still potent at some courses. It is time for me to do some more digging …….

David Renham