The reference in the after-race footnote of my post yesterday to Lee's way of drawing up an initial shortlist has attracted some interest, and there is no question that when applied to yesterday's race it was more successful than applying my understanding of how VDW drew up his initial shortlist. So I thought I'd discuss them today.
Lee's approach was presented on the Gummy VDW thread on 18/07/04 by reference to the 3.40 at the Curragh that day, as follows:
"The first 5 in the forecast:
Naahy *
Abunawwas *
Latino Magic *
Millennium Force *
One More Round
Trade Fair
* - Three most consistent
The 3 coming from the highest class race last time out:
Sea Dart
Avorado
Latino Magic
Sea Dart and Avorado can be disregarded, as both have showed nothing. Note Latino Magic is the only one to appear in both lists.
Indeed I have him coming out best of all when rating the above horses, and for me he comes out the most likely winner."
This approach to some extent matches that VDW showed in his discussion of the 1988 Mackeson won by Pegwell Bay, when he considered (a) those horses with the three highest ability ratings from the three most consistent ratings and (b) the horses coming from the three highest class races last time out. Both (a) and (b) gave him four horses with Pegwell Bay the only one to feature in both, so seven horses in total.
My own understanding derives the consistent horses shown in the Pegwell Bay table - six, of which five were what I refer to as "automatic" consistent horses and one was a "discretionary" consistent horse. (The * against Gee-A's 11 in the reprint of the Pegwell Bay article is a palpable error: it is not in the original and is plainly incorrect in the reprinted version.)
Lee's approach makes a lot of sense, in that it melds the horses with the three lowest consistency totals from the first five/six in the betting forecast - the "automatic" consistent horses - with horses that may be less consistent (in the sense of having higher last three race placing totals) but come from the highest class races last time out of all the runners. However, it can I think be demonstrated not to be the approach VDW used in what I refer to as his main method.
If we take all VDW's selections except for those such as Soaf, Aldaniti, Saher and the "Sporting Chronicle selection box method" (where VDW was clearly demonstrating alternative ways of "narrowing the field") or from other methods such as the "handicap hurdles" and the "best bet/next best" ones, ie those selections found by what I refer to as the main method, we have just over eighty. Of these, 90% came from the "automatic" consistent horses, and would have been found both by Lee's approach and by mine. Eight came from outside the "automatic" consistent horses: Crown Matrimonial, Desert Orchid (13/12/86), First Division, Love From Verona, Prominent King, Righthand Man, Roushayd and Son Of Love.
All eight are included in the initial subset of consistent horses by adding the "discretionary" consistent horses to the "automatic" ones. Only half are certainly found by Lee's approach of adding horses with one of the three highest class last time out races to the "automatic" ones (Crown Matrimonial, Desert Orchid, Roushayd and Son Of Love). Some sort of argument can be set up to take two of the other four out of the picture. If one believes that, in the 1978 Erin, Decent Fellow's consistency total should be counted as 10 rather than the 7 shown in VDW's table, Prominent King would be an "automatic" consistent horse and thus found by both approaches. One can also argue that First Division was not really a selection but used as a means of making a narrow but important point. But neither Love From Verona nor Righthand Man is captured by the Lee approach and the strong indication in the respective texts is that both were bets for VDW. Neither had one of the three lowest last three race placing totals from the first five and six, respectively, of the betting forecast (we know what forecast VDW used for the Righthand Man race and Love From Verona was well down the field in the Mail, Life and Chronicle forecasts), and neither came from one of the three highest class last time out races of the fields concerned (Love From Verona came from the 10th highest class race of the field and Righthand Man came from the 8th highest).
In short, while Lee's interesting approach may for all I know include more eventual winners in his first subset than the "automatic" plus "discretionary" consistent horse approach I use, the evidence is that the latter finds subsets that include all the relevant VDW selections while the former misses a small number.
Finally, it is worth adding that the approach I've referred to as Lee's was the one he demonstrated some six years ago. He may well still use it or he may have developed his approach further.
