Narrowing a big race field with Proform

February 4th, 2020

By looking to the past we can see the future, if like me you like to examine the demographics of previous winners to narrow a big saturday or festival handicap to a more manageable subset of runners who fit the trend of former victors and go from there, then Proform has many tools that are extremely useful in that process to a fraction what it would be without.


Where to start?

Well firstly getting the right race you wish to examine, by using perhaps the name (be aware there may be changes over the years or some might have similar names due to sponsors of the whole card). If the name does not give you what you need use the race class, course it is run at and the distance it is run over combined as a key should give you what you are looking for 95% of the time. We are often looking at valuable contests so the race value is also a potential filter in the event of their being a consolation race as in the case of Stewards’ Cup for example.

On the more higher profile events you hear or read things reported such as “this mostly goes to an unexposed sort” or “4 of the last 5 winners ran at the Cheltenham or Newbury last time out”.

It is good to absorb these and use the Proform System Builder lens to then test their validity and significance. ‘No four year old has won this previously’ and only 4 had attempted in the decade is far less convincing than - ‘no winner ran over 2m1f or shorter on their most recent run’ from 50 previous qualifiers.

To start, use some of the major relevant filters, filtering the full population of previous race runners each time. When you are satisfied then add all these in at the same time and continue to query and remove anything that appears to hold water and some sort of eye catching strike rate (lack thereof), % of rivals beaten or winners versus market expected wins scores.

Be very careful to not strike off potential qualifiers on metrics just because they show no previous wins/places, as the saying goes “absence of evidence is not the same as evidence of absence”.

Conversely be prepared to recognise for example that the 2017 winner was an outlier and not make accommodations to fit him in to the general trend to give the feeling that this system has found 10/10 winners from last decade, this back fitting is something I have fallen victim to before and may result in no meaningful narrowing of the field and reduced profitability in the larger scheme of big race narrowing as a whole.

Just to note here that one race on its own may not give you enough data, but you can make decisions based on similar contests as a whole. Such as, horses who ran in a chase last time out have a win strike rate of only 2.5% in class 1 handicap hurdles, we could see that in this case it falls well below the market expected win rate, 8 winners compared to an expected 14. As opposed to one race and them being 0/5, it doesn’t make you as sure of your decision to omit or include.


For example purposes let us take the steps through next Saturday’s Betfair Hurdle at Newbury

Use filters of: Hurdle only,Distance <3740, class 1 only, Month February, Course = Newbury (ignore contests at other venues as in 2019).

Firstly here is a full list of what we are dealing with. This race was abandonded in 2009 and moved to Ascot in 2019, that has left us with 11 runnings to analyse.

The trend suggests that this goes to a more unexposed or improving sort, five and six year olds do represent the lion's share of runners but account for all 11 winners, the rest are 0/80.

But by the same token it appears that their recent experience will back this up, most ran in a non handicap contest last time, and 7 winners were having their first handicap run and none had more than 6 handicap hurdle runs previously.

While some contests may go to more battled hardened horses who had been present in many of the major handicaps that season and rise to the top eventually, the Betfair Hurdle is unexposed horses who have been in the frame on recent starts. The rather more exposed but still only 6yo Heathcote in 2007 the exception. First 3 last time out 10/116 compared with 1/110 for 4th or worse on their most recent run.

Very few runners come from anything but a hurdle contest but Recession Proof and Violet Dancer both running in a low grade All Weather contest probably due to circumstances of cold snaps at the time in respective years. Though they had also fulfilled the criteria of placing or winning on their previous hurdle run.

While Zarkandar came straight here on the back of a juvenile grade 1 win at Aintree the previous spring many of the runners have had the benefit of a more recent outing. There is no guiding factor as to relevance of days since last run with winners at both ends of the scale from fresh to busy. The previously mentioned outlier Heathcote also being the one with 4 or more runs in the last 3 months.

It is one of the highest handicap hurdle prizes in the UK for the season at c.£87k to the winner therefore you would expect some high class runners to be present.

We will examine where in the line up winners generally come from.


We see from the graphic that it agrees with the previous narrative that improving sorts who are not fully exposed are the ones with the best record, the highest concentration of red is somewhere in the middle of the official rating ranked on the day, of courses number of runners on the day and the dominance of a top weight in the ratings affect this information but as a general guide it is useful.


Bringing this all together to guide us for what a Betfair Hurdle winner looks like.

Age 5 or 6
Less than 5 Handicap Hurdle Runs and 10 or fewer hurdle runs
Didn’t run over fences before
Finished 1st, 2nd or 3rd last time out
3 or fewer runs in the last 3 months

That would have left you with 60 runners and 10 winners from 11 contests staged at Newbury (trends hold for 2019 Ascot running) the one missed being in 2007 so since the Betfair sponsorship began this race carries a strong homogeneity among its winners.


From the early entries for Saturday’s contest it already reduces the field to 9 from 30.

In my experience for certain races where the trends are strong and over half the field can be eliminated this is extremely useful and profitable, you also have opportunities not just for backing but for laying, dutching and trading that you can explore with the Proform data.