from :

http://www.gummyracing.org/




Last week we examined how the relationship between a horse's energy output and its maximum speed can cloud the conventional weight-handcapping equations. In other words, you cannot be sure that weight will have the effect that theory prescribes.
In response to the same observation, some have looked to eliminate weight from their speed figures entirely, in order to produce a supposedly purer measure of a horse's ability against the clock. Could this approach lead to more insightful figures, or is it unscietific folly?
The most famous proponent of disregarding weight is the US speed figures guru Andrew Beyer, whose writing has influeced many others to tread the same path. He does not include weight in his calculations because he believes its effect is difficult to quantify and generally overstated by mainstream handicapping.
The notion of removing weight from the process of compiling speed figures seems heterodox when first considered. Anybody who has been nurtured on British racing has come to accept that weight does indeed affect race horse performance. Moreover, as demonstrated last week, the traditionalists have pure science on their side, at least.
Can the US experience be translated across the Pond? Not without qualification. It must be remembered that the range with which Beyer and his compatriots have to deal with is much less than in Britain.
US trainers baulk at their horses carrying 126lbs
(9st-4lbs) in top class handicaps, while rarely is a horse saddled with less than 116lbs (8st-4lb). If US handicappers regularly encountered horses winning a 0-70 handicap off 10st (140lb), then running in a 0-90 with 8st-8lb (120lb) it is unlikely that they would discard so entusiastically.
There are also important differences in course format and the pace of races between the US and Britain, Also, Beyer figures are not pure expressions of speed, but ratings using finishing times as a starting point. When a race produces a time which is not consistant with logic, Beyer and his associates sometimes employ a going allowance unique to the aberrant race in order to artificially create a figure. In some instances, the apparent inconsistency is readily explained by the effect of weight.
One of the best known examples of Beyer's fudging concerned Fantastic Light. In his prep race for the Breeders Cup Turf of 2000, the superb Godolphin runner Fantastic Light encountered a slow pace in the Man O' War Stakes at Belmont and ran the 11 furlongs in 2min 17.4. Earlier on that clear day, a minor race over 10 furlongs on the same course was run in 2min 4.3 In order to run the equivalent time, it should taken F.L. no more than 12.5 seconds to complete the extra furlong, yet he took over 13 and was somehow awarded a Beyer figure 15 points higher. There is absolutely no justification for this trickery if you are producing speed figures which are supposedely based on finishing time alone.
Beyer rejects the notion of class yet, ironically, it is because his figures are manipulated to reflect class that they work so well. The pillar of his work is that finishing time is more important than class, but he is completely wrong. The point is that class is an ethereal, incalcuable quality whereas finishing time - basically an approximation of class - is readily quantifiable and therefore it is easier to determine its objective worth.
In British racing, the higher you go up the class scale, the less likely it is that the final time of a race will be a good guide to the class of the main protagonists. The reason for this is simple: a high percentage of good horses possess acceleration and are ridden to conserve energy, which results in a tendency for steadily-run races. Most moderate horses, by contrast, cannot quicken and have to be ridden forcefully to maximise their chance, which leads to more truly-run races.
Good races are still run in faster times than lesser races, but not by as much as they should be. For instance, a horse carrying 9st in a class E 0-70 handicap-and therefore running off a BHB mark of 56- should, in theory, run about 5sec slower per mile than a Group 1 winner rated BHB 125 carrying the same weight.
(a difference of 69lb equates to 28.75 lengths at 2.4lb per length, and a length corresponds to roughly 0.18sec at a mile: 28.75 lengths x 0.18 sec/length = 5.15sec)
In reality, the average Class E race is run just under 3.3secs slower per mile than the average Group 1 race, according to the Rac/Post database. This is much less than prescribed by theory, and the difference is that Class E races are generally run at a truer pace.
Consider a typical example of 2 races over a mile on the same card, a Group 1 race and a Class E. (The ground is fast, and all other things are equal including weight carried) On average, the Group 1 race will be run 3.3secs faster, which to those doing ratings including weight on a parallel scale to the BHB handicapper, corresponds to 44lb.
The winner of the handicap duly gets a speed figure in keeping with its BHB rating, but the Group 1 race is going to come out nowhere near 125.
To those whom weight is not a consideration, a system of par times, according to the class of race, is the most common way of scaling their figures. So the time of the Group 1 race will come out true to their expectations, the diference of 3.3secs between the 2 races in the example above being typical. When the Group 1 winner comes out again, those using weight in their speed figures might be sceptical of the form, whereas the 'par times' man will be optimistic.
As can be seen from the examples given, the pure mathematical calculations of handicapping can often be confused by other external influences such as pace. The different handicapping methods will usually result in varying figures for the same race, and it may be only the manner in which a race is run that determines which method seems more sound and accurate on which occasion.