WHEN publishing the official marks for Cheltenham’s handicaps last week, Phil Smith went to the trouble of issuing a statement outlining the reasons why BHA handicappers keep their own performance figures for Irish jumps races, stating that there was “no semblance of any anti-Irish bias”, using the collective strike-rates for Irish versus British runners in jumps handicaps - viz. 11% v 10% for all handicaps, and 5.2% vs 4.8% for what Smith called the “four major festivals”, including Cheltenham and Aintree in the spring.

Smith uses the above figures to justify his handling of Irish runners in handicaps whereas his critics, notably Kevin Blake (late of this parish) use that very justification to conclude that he is deliberately biased against the Irish by artificially depressing their expected rate of success.

The argument Blake makes is that “An Irish trainer is only likely to bring one of their horses to Great Britain to run in a handicap hurdle or chase when they believe that not only is their horse in top form, but that they are progressive enough to go close,” and therefore such runners would be expected to outperform local handicappers “given that this group contains every poorly handicapped and out-of-form performer in the country.”

There is a valid point being made here in that for ALL handicaps run in Britain, the subset which are travelling from Ireland are more likely to represent live chances for the reasons stated.

On the other hand, Blake conveniently ignores the numbers for the major meetings. To suggest the relative underperformance of British horses at Cheltenham is in any way down to them being poorly handicapped and/or out of form is subjective and misleading.

That’s without latching onto the use of the term “badly handicapped” in relation to British runners. If Irish horses are winning more races because the opposition is badly handicapped, then Irish horses are ipso facto well handicapped, at least in relative terms, and let’s be honest, handicap marks are never going to be an absolute measure of ability, but a relative measure within a population of horses.

In similar fashion, pointing out the average discrepancy between different trainers’ handicappers is not proof of an unfair approach without further evidence.

But this isn’t intended to be a defence of Phil Smith against Kevin Blake. There is a lack of genuinely robust analysis in either the BHA’s justification of their approach or in Blake’s rebuttal, and that’s my real gripe.

GUIDING PRINCIPLES

If we are really going to put the issue of perceived bias to bed, we need to be able to establish some guiding principles which are accepted by all, and then we need to find a way of analysing the results on an ongoing basis to ensure that fairness is being achieved.

There are some “rule-of-thumb” methods of determining whether BHA marks are ostensibly fair. One is to compare them not just with Irish marks, but also with those produced by independent ratings organisations such as Timeform, who have no reason to fudge their own ratings - and yet there has been very little light (and far too much heat) shed on the issue by ignoring this approach. Phil Smith has attracted criticism by using ante-post betting markets to justify individual handicapping decisions, but while it’s a dubious way of making a point, there is a perfectly good argument that the betting market on the day of race is a useful yardstick of handicapping efficiency, with market strength and liquidity ensuring largely representative odds.

Aiming to achieve parity in win ratio is a broad, but flawed approach, and there are reasons why certain subsections of the horse population perform worse in handicaps than others, particularly those who are fully exposed and beyond full maturity.

Measuring the effectiveness of a handicapping approach by a blanket analysis of winners allows for no nuance in identifying such anomalies. Stripping out obvious outliers and analysing results for all finishers based on percentage of rivals beaten or publishing impact values based on both win and place figures would produce more robust statistics, and separating such results into sub-categories by code and distance would also enable better understanding.

As an example, if Irish handicap hurdlers achieve a 10% strike-rate and chasers win absolutely nothing, then as a population they are winning the right number of races, but the figures show a worrying imbalance. That analysis may well be going on, but by failing to publish it, Smith and his colleagues are underselling the process.

On the other hand, if that analysis isn’t taking place, why the hell not? Insisting on relevant information such as this (and more fundamentally the BHA marks for Irish horses), being published and subjected to ongoing scrutiny is what we should be pushing for, rather than digging trenches for a phony war.