• Hi Guest Just in case you were not aware I wanted to highlight that you can now get a free 7 day trial of Horseracebase here.
    We have a lot of members who are existing users of Horseracebase so help is always available if needed, as well as dedicated section of the fourm here.
    Best Wishes
    AR

Bad News For Bower ?

MySportsAI

MySportsAI

Yearling
Perhaps this question has been answered somewhere buried amongst the threads on Bower but I had to answer it myself having looked at Bowers ideas some 20 years ago and found little of value. My question is that if Bowers narrowing the field has worth then by applying it to create the top half of a field you should see from blindly betting the top half an improvement in amount lost in the pound or in other words ROI% than you would normally expect to see. I ran some code to check this on handicaps form 2021 to 2025. I selected only runners of 10 or more and races which had a mid point OR horse within 2lbs of the mean OR. I then ranked the races on OR and BFSP and added the two rankings together to create a 'BowerRating'. The lower the rating the better and I then divided the field into the top half. The number of bets blindly backing this top half was 46,165 and the ROI% to BFSP was -2.81% after 2% comm which is pretty much what you would expect from blind picks. Happy to be corrected on anything I have got wrong or suggestions to check other variations
 
I modified the code to add the OR difference to the BFSP and this is a little more promising losing 1.5% on turnover which to flat stakes IS better than pin sticking or random selections. In reply to Chesham can you explain how finding winners is likely to find profit, my perspective is that we need to find value and raw process i first tested was not finding value the second method does help in this regard but I want to check my code again
 
What I did next was to load the resulting file up into MySportsAI because I wanted to see if Machine Learning could find a more effective way of combining the OR difference value and the BFSP value, if there is any interest I can write this up. By the way the addition of these two values is OR difference + the raw BFSP as opposed to the difference between the fav BFSP and each horse, is this correct or is it the BFSP difference rather like with the OR
 
Hi all, with the filters that I use with my program I have 3 picks that I look for which have a good return for the tome I have been doing Bowers, they are, bronze, silver, and gold picks.these so far are showing a good return
6
🥇 GOLD FILTER — 7-12f · 8 runners · 3/1 odds · Balanced · Class 6 — historically 53% win rate, +110% ROI

5
🥉 BRONZE FILTER — 7-12f · 7 runners · 5/2 odds — historically 39% win rate, +54% ROI

As I'm on phone and not the computer can't put the silver up yet but will when I get on it.
This is early days and no doubt will change later but hopefully the updates I do to keep it fresh will help and hopefully improve.
I have 4 bronze running today and one gold and will post them later, just a point I only back it if it looks right to me so sometimes I may miss 1 or 2.
I no there is no method or system that is full proof but if it works for now I'm happy
 
After a recheck on my code I found that adding the OR difference to the BFSP returned -2.09% on the top half of the field and using BFSP difference ie the difference between the lowest BFSP and each other runner made no difference to this figure. Backing all the bottom half of the BowerRanking returned -3.65% so there is some gains however if you had simply backed all the horse in top half according to BFSP you would have only lost 1.6%. It seems to me that this process may sound clever but is in actual fact bringing nothing to the table
 
Sorry MySportsAI MySportsAI

My bad , I thought that I was showing how Bowers has been adapted. But I now understand that you are looking at it from purely a coding exercise . I have deleted my posts 👍
 
Thanks, what I am trying to do is investigate whether the Bowers formula actually brings anything to the table and so far it appears to bring very little
 
Hi MySportsAI MySportsAI,

Just to confirm that it's the points difference from top OR + Odds = Bower rating. Not sure if you were doing it that way from your posts or not.

Furthermore, it is not strictly the top half of a field. The elimination starts at the bottom (points) and then stops when, close to the halfway point, there is a sufficient points gap between the two.

Regards
 
Yes on the first question Leodis and I will check the half way point on ratings rather than ranking and get back but going out now
 
maybe the secret and i honestly dont know as not something i ever tried but is get all the qualifiers in top half then only bet horses in that half who where beat less than a certain amount of lengths in say there last race but would likely be impossible to check back MySportsAI MySportsAI but could be checked going forward.
 
I'm late on to the Bowers scene overall, but noticed different variations on Bowers, written in old raceform clippings, when compared to the smartsig magazine articles. Both versions give food for thought. The Balanced or Not Balanced discussion wasn't in his early writings either, but he did mention focussing on the better class handicap races. £5k+ equivalent to £15k+ nowadays, I'd settled somewhere in between, maybe £10k+ handicaps.

Most recent ones were OR diff+Latest Live Odds, taking the latest show before the off, Bill Benter-esque style. The short-list would then be studied further, not all backed or dutched.

Earlier clippings, intimated using the RP Ratings instead of the OR, and the RP Betting Forecast instead of the Live Odds. nagwa nagwa can take the credit for digging those ones out. Maybe Bowers updated his thinking, although both are interesting to explore.
 
You must log in or register to reply here.
Back
Top