Handicapping :- the WTR ratings rework &
Questions Arising

Article 13_002_1
20th January, 2013

1. Introduction - the WTR Recalibration

The World Thoroughbred Rankings (WTR) bring together the best Thoroughbred racehorses from the major racing nations in a single ratings list. Interim ratings are published through a year. Then, at an end of year meeting, usually in Hong Kong, official handicappers from the nations meet to agree the final annual ratings of the best horses. The body that runs this process in the International Federation of Horseracing Authorities (IFHA) and the handicappers from the British Horseracing Authority (BHA) represent Britain at the IFHA.

The rating of the top horses in each year has been going on since 1977 and originally just a European venture. The process has developed over the years with more nations involved and different organisations having control of the process. The early years of the ratings saw far more horses rated 120+ than since the late 1990s. The IFHA had not sought to 'fix' this issue since the International Ratings of the top horses are rarely used to assign race weights and are of 'academic interest' to a large extent.

In recent years there has been more interest in the general racing public over where current star racehorses fit into the historical levels. This reached a high point in 2012 with Frankel touted as the 'greatest ever'. This meant that the IFHA now had to address the issue of the early years of International Ratings being too high compared to later years. BHA 's Phil Smith has said in interview recently that "..the time had come to grasp the nettle.". Which implies that the issue had been left unresolved for some period of years.

BHA handicapping head Phil Smith appears to have been given the task by the IFHA committee to review the issue and come up with a paper to detail how the issue could be resolved. He produced a draft paper in 2012 which was reviewed by the committee. They basically accepted the proposed solution but asked for the methods given in it to be reviewed for statistical correctness by an expert. An Professor in Economics from Ireland, with links to the racing industry, assessed the paper from Smith and suggested some changes which were implemented.

The methods agreed where then used to recalibrate the international ratings from 1977 forward so that they produced a more directly comparable set. The outcome of this process was announced along with the end of year ratings for 2012. The BHA then ran a presentation on the changes in ratings for top British trained horses on January 15th, 2013 and released a press statement. The IFHA published a paper which gave more details on the methods used to recalibrate the ratings.

This article has two main purposes :-

This article is in great part a response to the coverage of the WTR recalibration by the Racing Media in Britain. With a few honourable exceptions this has been of a lamentably ill-informed and biased nature. Section 2 of this article gives a flavour of the types of responses an interested follower of horseracing in Britain would have been exposed to by the 'Authority Figures'. The Pundits who purport to speak knowledgeable about Racing here and set the agenda and standards for others to aspire to.

- International Federation of Horseracing Authorities :- Link

- IFHA's World Thoroughbred Ranking Index Page :- Link

- World Thoroughbred Rankings (WTR) produced by Committee within IFHA :

- Quote from IFHA WTR page = "The World Thoroughbred Rankings are the IFHA's official end of year assessment of the top thoroughbred racehorses. The Rankings comprise all horses which have run during the calendar year, and which have been rated at 115 or above by the World Thoroughbred Rankings Conference. In addition to the annual official listing , World Thoroughbred Rankings (WTR) will also be published on an interim basis during the year in each case encompassing the leading racehorses in the world over the preceding six month period . This interim publication of WTR was formerly known as "World's Leading Horses"."

- Summary info about the IFHA's WTR Committee = "The World Thoroughbred Rankings are compiled by the World Rankings Supervisory Committee and published by the International Federation of Horseracing Authorities (IFHA). For further details on the World Thoroughbred Rankings, please contact: Nigel Gray, co-chairman, World Rankings Supervisory Committee Head of Handicapping and Race Planning, Hong Kong Jockey Club Garry O'Gorman, co-chairman, World Rankings Supervisory Committee Senior Flat Handicapper, Irish Turf Club Tom Robbins, chairman, North American Ratings Committee Executive Vice President, Racing & Industry Relations, Racing Secretary, Del Mar Thoroughbred Club"

- Note that this is NOT a BHA organisation and the BHA Handicappers are part of the Committee that meets at the end of year to finalise the year end rankings. The decision to re-work the historical data was a NOT a "BHA decision" and they contributed under the committee framework.

- Link to IFHA WTR Recalibration Paper (January 2013) :- Link

- Co-Chairman's Preface to the WTR 2012 Recalibration:- Link

- BHA WTR Press Release Page (Jan 15th):- Link

Staged interview between Master of Ceremonies Mike Vince and Phil Smith from the January 15th BHA Presentation:- Link.

Staged interview between Mike Vince and Chairman of the IFHA WTR Committee Chairman Garry O'Gorman (Ireland's Official Handicapper) from the January 15th BHA Presentation:- Link.

- Phil Smith Blog on BHA website covering the WTR Recalibration (Jan 22nd):- Link

Back to Top of Page

2. TV, Papers & Pundits Response

- TV Response - C4 on Sat pm Jan 19th, ATR on BHA Presentation Day & then on Sunday Forum Jan 20th

The TV coverage was especially poor with B2yoR being able to vouch for the lack of quality on the programmes above. The analysis and punditry seemed to have the following common problems :-

- Pundits being asked about handicapping who have little or no knowledge about it. A number seem to be in the set that believes rating horses is a mystical 'art' performed by geeks who have little relevance to the real racing follower and audience. This is one of the Fundamental problems with British Punditry. A good working knowledge of handicapping, how the BHA implement it, how ratings relate to horse 'Class Levels' and how that it visible in races should be a foundation for any pundit. The lack of this showing up in the inability of most pundits to 'stick a number' on anything related to racing including performances and class types of horses. Instead using vague, and ultimately meaningless, phrases such as 'nice type', 'bolted home', 'could be anything', 'one to follow', and so forth.

- Pundits and commentators who had not read the IFHA's recalibration paper. They were therefore pontificating about a subject they had not addressed even if they had the abilities to do so adequately.

- Pundits falling back on an excuse that the WTR and International Ratings in general were a sideshow and irrelevant. Mick Fitzgerald dismissing the issue as a "matter of opinions" and Richard Hoiles as "all subjective" on the C4 Saturday afternoon programme. Neither showed any evidence of having read the IFHA paper but were being asked for an authoritative answer before the largest audience for Racing that week. Arch narcissist McCririck chose to call the re-working "a fix" [shifting Frankel to clear top rated] on ATR's 'Sunday Forum' programme but then tried to say this is what the public would be thinking. He did not appear to have read the IFHA paper and it was not clear he had anything to say about the issue beyond his usual ploy of choosing one issue, relevant or not, to jump up and down about. On the same programme Peter Thomas, lounging around in his usual cannot-be-bothered manner, called the recalibration "all nonsense". Again apparently a Pundit dismissing the whole issue when he had not read the source documents. ATR's 'Sunday Forum' has, unfortunately, dropped almost entirely into the Two-Blokes-in-a-Pub approach of ill informed people sitting around saying definitive things about areas where little or no preparatory work has been undertaken.

- [The WTR and the IFHA recalibration are not subjective or based solely on opinions. They are implementations of a set of simple algorithms and processes. Any worthwhile Pundit should be able to build on their own knowledge to analyse the IFHA approach and point out problems, areas needing more work and so on].

- Pundits following up dismissing the IFHA ratings by an appeal to really subjective issues to rank horses. Their own impressions of how the horse raced and the effect it had on them. The races it had won and the nebulous 'Body Of Work' (BoW) approach. The BoW defence having become very popular since Sea The Stars' race style and lack of a 4yo campaign held his rating down so that even the BHA's Phil Smith seems to think it is a ratings maker.

- Pundits focussing only on the Frankel element, and perhaps Dancing Brave & Shergar depending upon the Pundit's age, in the story. The lack of any other knowledge ensuring they could only deal with the story in a superficial manner.

- Presenters interviewing connections of the famous horses involved to ask for their views. None of these people are handicappers and presumably had not read the IFHA paper. Which lead to a number of emotional and gut reaction responses as the people stood up for their horses. Walter Swinburn for Shergar & Guy Harwood for Dancing Brave as examples. Pundits struggling to say anything worthwhile would then repeat the emotional responses of the connections as if they had some relevance to the real issue. Which centres around the validity of handicapping & statistical methods and not emotional responses.

- [Interviews with Phil Smith - References?]

- Newspapers and Racing related Websites :-

A selection of links below and all seem to follow a similar format of majoring on Frankel and then quoting some of the IFHA or BHA's representatives at the January 15th presentation. Little questioning or analysis of the IFHA approach evident and a number of the articles appear to be written by the Press Association and syndicated.

Some interesting snippets in the Greg Wood, Will Hayler & Marcus Armytage reports and all three seem to have been at the presentation. But, all three basically question the relevance of handicapping and ratings. Hayler noting there was a thorough 26 page IFHA document to read but he had got bored so had given up with it part way through. Armytage spends time complaining about Phil Smith using a lot of slides and stats to explain the issue at the presentation. Then adding a few quotes but really only showing interest in the Frankel part of the story. Wood is usually more thoughtful and swapped some Tweets with Willoughby & Rowlands about the recalibration implementation. But his piece is standard stuff and invokes Orwell in characterising the IFHA approach, questions the worth of ratings and then descends into using the phrase 'number juggling' at the end. Oh, dear.

- BBC Website Article:- Link

- Greg Wood article in the Grauniad:- Link

- Racing Plus article:- Link

- Racing UK article:- Link

- Will Hayler Sporting LIfe website article:- Link

- Marcus Armytage Telegraph article:- Link

- Twitter & Timeform, ex-Timeform related

- Timeform = Head of Development Simon Rowlands put a response on the Timeform Website :- Link

- Simon Rowlands on Twitter :-

15th Jan = "Too kind. However, according to P Smith, all a computer can help with is working out pounds per lengths..."
17th Jan = "Good point about the increase in horse pop. Logically the top 57 should be getting higher if the base of pyramid getting wider."

- James Willoughby on Twitter :-

15th Jan = "You cannot slice the top 57 elements off different distributions, normalize the sub population and then compare. Nonsense"
15th Jan = "Ratings have meaning only within the context of the universal distribution. You have to go back and parameterize properly"
15th Jan = "Presentation by @rowleyfile on ratings = 140, presentation by BHA = 106$. Exercise in playing with numbers not good enough in modern world"
17th Jan = "I asked EXACTLY same question to Phil Smith. He just didn't get it" [Response to Rowlands et al asking WIlloughby about the increased horse population meaning MORE horses should rate 130+ (or whatever level)].

- Response in other Countries?

[References & links to responses in other countries? Because the recalibration affects the early years when Britain, Ireland & France were involved is the response limited to those countries? Any downgrading for Italy & Germany? Major racing Nations like Australia & the US cover the reclibration at all? Has Ireland responded to Vincent O'Brien stars like Alleged being downgraded?]

Back to Top of Page

3. Questions on the Rework

- Why the Slippage between 1977 and 1990s?

- What processes are in place now to ensure no further slippage (or Uppage)?

- Why was the recalibration process and results announced now?

- Why are there not an increased number of Horses rated 120+ Worldwide with a larger population of horses rated?

- The Paper says it does not want to say the previous Handicappers were 'Wrong' but then lists a number of issues with the previous methods. Why were these methods not faulty?

- The paper notes that horses are much more likely to be downgraded, at the IFHA's year end workshop, that in early years. The reason for this apparently being that the present 15 handicappers present are more likely to downgrade than the total of 3 in the earliest meetings. If all meetings were implementing a simple set of algorithms consistently then the 15 versus 3 logic is faulty. How much 'Touch & Feel' and 'Art' is/was involved in the process of agreeing year end ratings? How much is simply stated algorithms & formulas?

- The paper notes that the internal consistency of early years was faulty due to issues such as 'leapfrogging' and promoting horses above the 120 entrance level before it was lowered to get them listed. Why did the recalibration not address these issues?

- Why were current handicapping techniques not applied to all of the earlier years to produce re-worked poputlation ratings?

- What evidence is there that the approach implemented of taking slices off the horse population in each year and then normalising these sub-groups relative to each, within set tolerances, is valid?

- What comments did the Statistical Advisor make on the first draft approach? Did he raise any issues with the final approach?

- What was the reason for the decision to apply a rule that no two consecutive years could have a larger downgrade in the second year (Monotonicity)? Was this is recommendation from the Statistical Advisor employed?

- Do all the members of the IFHA use similar Handicapping methods? How many use Manual Handicapping solely and do any implement some form of Computer Handicapping?

- The IFHA paper says that increased travel for horses to compete abroad made International handicaps harder to set. Why did the increase in collateral form available across national boundaries not improve the situation?

- The paper further states that the "huge increase" in Pattern Races across the world has diluted competition. What has the IFHA and the International Handicapping community done to raise this issue with the appropriate bodies to reverse this trend?

- Does the BHA think it should have done anything different in the presentation on January 15th? Should there have been a 'Recalibration for Dummies' type document to support the more data laden offerings?

- Have the BHA team been surprised, in good or bad ways, by the response to the recalibration?

Back to Top of Page

4. General Handicapping Questions for the BHA

- Does the BHA use Manual or Computer Handicapping? Is any form of Computer Handicapping used?

- Are there any BHA plans to implement any, or any further, forms of Computer Handicapping?

- Is the identification of Under/Over rated races as collateral form develops a Manual or Computer process? [The 2011 Word document linked to below suggests this is a 'Human' process looking at a screen with the five previous and five later ratings for each runner in a race displayed. Why could this process not be done continuously by a computer program which then flagged up 'suspect' races, for example. Not 'Computer Handicapping' but using a more consistent approach and lessening the human inutuiton issues.]

- Does the BHA feel that splitting the handicappers by Race Distance causes any issues in being able to take a 'whole population' view on handicapping? Does it affect the decisions on Computer Handicapping use?

- The recalibration document cited handicapping to an expected 'Norm' in an average year leading to over-rating and 'Top Down Handicapping'? How is this problem handled now? How does the use of Average figures for Group & Listed interact with this issue? How does the use of these figures by the Pattern Race Committess afect the handicappers?

- Does the BHA use any 'Race Standardisation' techniques in their handicapping?

- Given that the split of Race Rating Downgrades to Upgrades approaches 2:1, for Maiden races, as highlighted in their document, what proposals are there to bring the original ratings given closer to the true levels?

- The the majority of races run in Britain are not run at a 'True' pace, i.e. slower than the ideal to demonstrate the horses' full abilities. What processes do the BHA use used to Project out to final rating for horses involved in a compacted finishing positions caused by thg slow paced races?

- How are Population Medians used in the Handicapping process?


Back to Top of Page

5. Links and References

- BHA Website Page outlining what the Handicapping Teams does:- Link

- MS Word document on BHA Website detailing what the Handicapping Teams does:- Link

- BHA Handicappers' Weekly Blog:- Link. Recommended because includes useful insights into how the handicappers deal with current races and horses and highlights the methods in use. Also interesting to see how they respond robustly to complaints from connections of horses and also to common beliefs within the racing community which are not based on any research. The BHA handicappers do seem to use small sample sizes a lot when doing some light research to 'disprove' these beliefs.

- Historical Editorial Document from John Whitley's 'Racing Research' Website:- Link. Good summary of how to build Collateral Form and Time based Handicaps. Also good historical background on the difference between "Computer Handicapping" & "Manual Handicapping".

Back to Top of Page

© British2yo Racing. All rights reserved.