Some Stats

[pawras]

I have my own DB of past races, I generate my own kpi’s etc etc.

Anyway I’ve also loaded up the data that’s available from data.betfair.com so as to enhance my analysis, as such I have my version of bsp where by it’s the last price traded before the off, plus also record the price traded at around 700am.

As such I did a bit of analysis to compare isp against my version of bsp and the 700am price and came up with the following. Note the betfair prices have had no commission deducted.

See what you think….

 

[Desert Orchid]

Congratulations, pawras, on such an interesting first post.

What would you deduce from the stats?

 

[Danny]

Good to see you over here Mr.P

 

[pawras]

Congratulations, pawras, on such an interesting first post.

What would you deduce from the stats?

On average in stakes races the difference between bsp and isp for the winner once you take commission into account is fag paper thin.
On average winners shorten on betfair from earlies to bsp by about 7%.
If you’re not trading etc then back early with BOG if possible.


Danny’s interest is primarily big field races so I produced this for him.
The amount that winners tends to shorten on betfair for AW races is quite dramatic , but that is because there are only a very small number of big field races on AW.

 

[Desert Orchid]

If you’re not trading etc then back early with BOG if possible.

Excellent. That indeed is my MO and Slim Chance has told me more than once it's the right way for me to operate given I pretty much need to get all my bets on by about 10.30am.

 

[pawras]

How 'statistically strong' is the OR broken down by GB and IRE, race type and hcap vs stks?

So if you blindly backed the stand alone top OR in each race since 2001 what would you get?

 

[Danny]

Mate your memory is better than Colssuss's if you remembered you'd posted this 7 years ago.

 

[pawras]

Mate your memory is better than Colssuss's if you remembered you'd posted this 7 years ago.

There's no escape for you from my 'minging stats'....

 

[pawras]

Here's the PRB2 (percentage runners beaten squared) data for all trainers with runners between 20250211 and 20250216 based on data up to and including 20250208

I show it at different timespans to give a view of whether a trainer is in or out of form compared to their long term.

Percentage runners beaten is more granular than just simple strike rates.

It's too big for a screenshot and you can't attach spreadsheets here so have attached it as pdf and have highlighted the stand out ones that seem in or out of form.

 

[pawras]

Next time I look at the code for that I might add a new col showing how many runners they over the next week.

I have the same report but for jockeys but didn't have time to fk around creating pdfs and posting that one

 

[pawras]

Here's the jockey version of the report

 

[pawras]

As before here's my trainer PRB2 ratings covering different time periods for all trainers with runners between 25/02/25 and 02/03/25 using data up to and including 22/02/25. Allows you to see if a trainer is in or out of form. The last col "count of horses in last 365 days" as it's a distinct count it gives an indicative view of yard size.
Has some more time periods as it was requested by someone elsewhere.
I've highlighted the ones I view as particularly in or out of form.
If you don't know about PRB2 then an internet search will soon help

 

[pawras]

Following my post here

Thread 'Professional Punter Shane Reville'

Mar 20, 2025

I caught up with Shane as we embark on a new flat turf season, he's ditched the jumps to concentrate on flat racing encompassing Hong Kong too.

Simon Nott Zooms-In #10 Professional Punter Shane Reville

Irish based professional punter Shane Reville talks to Simon Nott about how he has decided to just concentrate on flat racing in Ireland and is also exploring the potential of betting on Homg Kong racing.

www.simonnott.co.uk

• Simon Nott
• Replies: 4
• Forum: General Racing Topics

• 

I felt inspired to do the stats to compare in the UK , IRE and HK , backing the stand alone top OR and my trainer PRB2 rating based off their last 365 days.

 

[pawras]

Being top OR seems more relevant in the UK than IRE & not really worth thinking about for hcaps in HK.

The results look interesting for the trainer prb2 rating in IRE chase hcap races.Only a 5% loss at isp over a big data set with no other considerations.

 

[Grey]

Thanks for sharing this. I'd like to be sure I'm understanding the data correctly, so please forgive my questions.

The left hand tables show the overall results of betting at level stakes on the highest rated horse in various types of races in the three countries?

And what exactly are the right hand tables showing? The results of betting at level stakes on any runners from in-form stables?

And in the pdf file, that last column showing numbers of horses per trainer: is it the number of horses that ran on the track in their name or the number registered as horses in training in their yard?

 

[pawras]

Thanks for sharing this. I'd like to be sure I'm understanding the data correctly, so please forgive my questions.

The left hand tables show the overall results of betting at level stakes on the highest rated horse in various types of races in the three countries?

And what exactly are the right hand tables showing? The results of betting at level stakes on any runners from in-form stables?

And in the pdf file, that last column showing numbers of horses per trainer: is it the number of horses that ran on the track in their name or the number registered as horses in training in their yard?

>>The left hand tables show the overall results of betting at level stakes on the highest rated horse in various types of races in the three countries?
Yes, blindly backing the stand alone top OR (i.e. if the top is 140 and more than one runner has that then no bet) to level stakes at ISP


>>And what exactly are the right hand tables showing? The results of betting at level stakes on any runners from in-form stables?
I'm into percentage runners beaten squared prb2 as a kpi (key performance indictator) for trainers by aggregating it up for set time period, in this case over the last 365 days prior to a race i.e "Trainer prb2 rating from last 365 days - backing any horse stand alone top rated"
This gives a more granular view of the trainer's performance than just wins vs losses or even wins & places vs losses.
So for any given race I have the trainer's prb2 rating derived from the 365 days prior to that race and I'm showing the results of blindly backing the stand alone top rated trainer prb2 at £1 level stakes to ISP.
Then by doing that for the same time period, race types and country as for the stand alone top OR it's showing a side by side comparison of the two quite different metrics.


>>And in the pdf file, that last column showing numbers of horses per trainer: is it the number of horses that ran on the track in their name or the number registered as horses in training in their yard?
The pdf file posted on 25/02/25 showed for all trainers with runners in the forth coming races, all courses, between 25/02/25 and 02/03/25, their calculated prb2 ratings for different time periods , the last 365 days being one of them, so as to get a view of their overall performance compared to other trainers with runners in that time period and whether they were particularly in or out of form.
The last column "count of horses in last 365 days" is a count of the distinct number of horses a trainer has sent out over the last 365 days (distinct as in if horseA runs 10 times it still only counts as 1) , this gives an indicative view (i.e. estimate not definitive) of the size of their yard.