I made use of mail order from a UK outfit that to obtain titles published over there. One of them was "Betting For A Living"!
I just read (on Kindle) Dave Nevison's first, "A Bloody Good Winner" - that had a nice Beyer "$50,000 Year At The Races" feel. I enjoy his RacingTV work.
Winning Without Thinking - was his magnus opus imo - a summation,i.e. his life's work, of everything he struggled to convey and theorise in his previous books. More in the realm of the possibilities of idea generation within horse racing data than actual prev. practical books such as Betting For A Living, Mordin On Time & The Winning Look. That book, The Winning Look showed his expertise on paddock handicapping and not many books in that sphere, yes Ledbetter & Ainslie's old classic The Body Language of Horses, Australian Geoffrey Hutson's two books Watching Racehorses and Watching More Racehorses and also a Joe Tackach book if i remember right. Mordin's stands up with the best of them.
A good few year's back on this forum there were links usually on the homepage to a series of articles Joe Tackach done - are they still hidden somewhere on this site. Collecting articles on "physicality handicapping" and "paddock review" for a planned book from somebody else, expected in 2022. Would appreciate it if somebody could provide a link if still available. TIA
A good few year's back on this forum there were links usually on the homepage to a series of articles Joe Tackach done - are they still hidden somewhere on this site. Collecting articles on "physicality handicapping" and "paddock review" for a planned book from somebody else, expected in 2022. Would appreciate it if somebody could provide a link if still available. TIA
I remember those, but I don't know where they were. Takich also had an excellent video. He had a lot of great info to share.
__________________
Ranch West
Equine Performance Analyst, Quick Grid Software
A good few year's back on this forum there were links usually on the homepage to a series of articles Joe Tackach done - are they still hidden somewhere on this site. Collecting articles on "physicality handicapping" and "paddock review" for a planned book from somebody else, expected in 2022. Would appreciate it if somebody could provide a link if still available. TIA
I never created ratings exactly as presented in the book, but conceptually my thinking is very similar. I would add that since that book was written, we now also have high quality publicly available pace figures from TimeformUS. So when thinking about how competitively the race was run, you can also look at the figures and not just watch the race.
__________________
"Unlearning is the highest form of learning"
Nick Mordin's adapted Quinn's "Class of The Field" type ratings for UK racing
Going through this old thread and having read and own both Quinn's "Class Of The Field"(Quinn was really the expert on Class in the US) and UK Handicapping author Nick Mordin's first book "Betting For A Living" i just remembered that Mordin attempted to recreate a similar type of Rating system adapted for UK/Irish Racing - obviously we (UK/IRE) had no sectional times then so they were based on actual "Handicapping" principles , ie Margins beaten, weight carried etc but they could be recreated now that we have GPS sectionals across the board at over the 40+ tracks now in the UK , 20+ in Ireland.
This was Mordin's definitions of what Quinn called the (CQ) - Competitive Quality Scale. Anybody that has read Quinn's original book will notice right away the similarities even though the two types of designations are for different jurisdictions.
Mordin tested Quinns theory that the higher the Race Class, the higher the number of lead changes therefore the higher number of challenges a horse could withstand by Class - at the time our Official Class System were based on lettering with a min-max Official British Handicapping Authority Rating between each Class band - it has now changed to a larger spread numerical system but here was the results (admittedly from quite a small sample) back then.
Attached was the Ratings table he used for winners for the various designations by Class and (CQ). He also further used some rules based decision making based on Quinn's "The Handicappers Condition Book" adapted to UK/IRE racing.Back then i used them as "Ability Pars" to make Time adjusted Class Ratings with standard times and variants using Final Times and they worked very well, at least better then just using the Official BHA Class Rating and Mordin had the decency to give credit to James Quinn for the original idea.
James Willoughby's Class Ratings for 2yo's using iteration.
Obviously processes have got more sophisticated now but here's a method that employs 4 different concepts designed by UK technical horse racing analyst James Willoughby for analysing 2yo's.
1) The method employs Head to Head to type analysis where a "win/loss" is determined by finishing position , ie 1<2<3<4 etc such as this similar to Normalised Finishing Position except that it is cumulative as the season progresses. So a horse by the end of the season may show 42 wins/12 losses. These are shown as "Odds Ratios" so the Odds Ratio would simply be 42/12 = a Head to Head Win/Loss Rating of 3.5. A Laplacian prior of 1 win and 1 Loss is added to the horses very first start.
2) The method employs a "Race Strength" factor simply by calculating the geometric mean of every horses "Odds Ratio" in the race - in this process a Laplacian prior of 5 wins and 5 losses (Odds Ratio of 1.00) is added to each horses cumulative H2H Win/Loss Record - this has shown to provide better results.
3) Iteration - it employs the algorithm designed by Dr David Hunter in 2004 which is an MM (Minorisation-Maximisation) algorithm and used a lot in Bradley-Terry type Models
4) It offers a "Closed Form Solution" as once the 2yo Populations ratings cannot be iterated anymore that is the process finished (No rating can be changed with it's final iteration) - a Horses "Highest Rating" during the process is it's rating to take as it maximises its product of wins and losses with "race strength" - typical they are scaled 1 to ~50, although it is better to use at least 3 figures after the decimal point of the base rating for "precision" this excludes out "ties", "joints" etc - from that a regression could be run against a horses best Beyer figure to find the regression equation to normalise the two. Tests i have ran show it correlates to around 0.89 with an ELO type rating system (which is pretty damn good) and the only variables are finishing position optimised as H2H wins/losses as odds ratios which then derives the race strength factor - no weights, margins beaten, times, etc.
To derive H2H Wins/Losses in a race (see Pic 1)
Later after a few races because the H2H Win/Loss totals are cumulative the horses might look like this going into a new race see (Pic 2)
"Race Strength" is derived by the race geometric mean of the race odds ratios - during this process we attach 5 wins/5 losses (Odds ratio of 1.00) to each horses cumulative H2H win/loss record - then take the geometric mean - Rating is derived by Race GeoMean*new Odds Ratio - see (Pic 3)
For a more fuller explanation and explanation much better than i can here's James article on it - he gives a list of the top 2yo's ratings by this method that ran in 2020, a very large majority of whom turned out to be Group/Listed Class
As "shippers" are a big concept in your jurisdiction , this may be of interest.
Here James uses the famous Google "PageRank" algorithm to measure the quality of a track