I noticed the following post on the successor-to-Gummy forum and thought it worth reproducing here:
"My view on why the VDW writings remain unresolved after 30 years is that the writer tries to mix oil and water.
He introduces the reason why a horse wins the race then tries to resolve that with a statistical approach when in reality one has no relation whatsoever to the other.
VDW reminds me of one of those pioneer scientists from history, who having made an astute observation, ultimately drew the wrong conclusions from it.
An example that illustrates this contention is 'malaria', a disease from which I have had much inconvienience.
Early scientists noted that this disease was common to warm, humid climates and consequently, but wrongly assumed that the disease was airborne, baptising it with the name we know today. As a result of this erroneous conclusion dwellers of these climates spent their lives walking about with masks over their mouths in the vain hope that they would avoid contamination.
It never occured to the scientists that the mosquito needed a hot and humid climate to proliferate and was the vector responsible for the transmission of the disease.
So is VDW's mistake.
He spotted that it was the class horse that usually won the race, providing that it was fit and 'on', but he chose to look for the 'class' using statistics.
It may well be that the class horse is in the first six in the betting but it doesn't have to be. Just because bats and butterflies have wings doesn't mean to say they are birds. The class horse has class, its position in the betting is totally irrelevant, just as its last three finishing positions are.
A horse has a career, that career has to be developed by training, a racehorse doesn't just exist as a ready wrapped, immutable machine. When the trainer is happy that he has got as far as he can with it he will then look for an opportunity to win with it. He will look to find a race where the class of his horse is going to be better than its rivals. Statistics have nothing to do with this decision making process though they may arise as a consequence of it.
It is no use looking for class from predefined areas that in reality have nothing to do with the intrinsic nature of class.
It is no use hoping that by applying rules, such as 'must have shown improved form in its last three runs', that you are wielding the net that will catch the fish when the true nature of class demonstrates that it cannot be constrained by such bonds.
To get anywhere with these methods a student must first understand VDW's mistake."
I was a fellow member with JIB of a private forum for some years, and he posted extensively on the Gummy forum and elsewhere, but do not recall having seen the above post before. Having now seen it, I think it is very insightful and largely correct.
I agree with John's point that there are two different matters involved - what he refers to as the oil and the water - but I am not yet sure whether John is right to see the mixing of the two as a mistake, or whether it was a conscious strategy on VDW's part.
At various places in his writings VDW makes comments which suggest the latter as a possibility - a strategy. For example, in his letter to Tony Peach of February 1996, reprinted in "Systems In My Racing", VDW wrote:
"When I first began to write for Sports Forum [the letters section of the Sporting Chronicle Handicap Book] it was clear that to splash the whole lot in front of your readers would be a pointless exercise and only by adding bits as time went by could it be hoped a doubtful, critical and sometimes abrasive readership would eventually see the light.
I had intended to give away everything in due time, but you will recall telling me you had decided to call a halt to discussions of my methods in your column. This was fine by me, but only a fraction had been revealed at that time."
What he had "revealed" was much, though not all, of what I think of as his main method, set out most fully in the March 1981 article, and that is in large measure what John describes as a "statistical approach": first five/six in the betting, the three lowest consistency totals, the four highest ability ratings etc. all, we can reasonably assume, backed up by statistical findings of the kind outlined in that article. There is of course more to the method, the much discussed issue of "form" in particular, but it would be fair to say that it is in large part numerical, based on statistics.
That, it seems to me, invites two lines of inquiry. First, does the method work? It is quite clear that there is no agreement on what the method comprises, so anyone answering the question can only do so on the basis of his or her interpretation. My answer, on the basis of my interpretation set out on this blog, is yes it does, it but generates (at least in Flat handicaps) very few confident bets. When one arises, it normally wins but looking at, typically, one or two handicaps a day when there is one or more of class 3 or above, and some class 4s, I don't find many bets a season. And anyone who thinks that that indicates that the interpretation must therefore be incorrect would do well to re-read the March 1981 article: five selections made for 7 March 1981 but only two were bets. A class/form horse in the large majority of races, but few with all the "winner in the race" characteristics.
The second line of inquiry is whether what VDW presented - this essentially statistically-based method with its variants in terms of the initial narrowing of the field - was, as he wrote in his February 1996 letter, just "a fraction" of what he knew, or was it in reality all in the sense that he had gone no further than methods based on statistics. If the former, then I see the main method as a kind of basic introduction to the real business, finding the class/form horse in a race not by the numerical techniques but by the kind of approach John touches upon and which he exemplified in some of his forum posts.
I don't know when John wrote the post reproduced above, but he knows a great deal more about the true VDW story than most, and from what he knows he has reached his own conclusion about VDW the man and given them on the Gummy forum. We now know rather more than we did when John posted on Gummy about the discovery of VDW's true identity, and I have to say that what I have learnt since supports John's conclusion. But even with this additional knowledge there still remains, for me, a percentage of uncertainty which, as things currently stand, may not be resolved in VDW's lifetime. I think the percentage is low and that John's view - "mistake" nt strategy is very likely to be proven correct.
If we take John's view as correct - that VDW hit on the fact that "it was the class horse that usually won the race, providing that it was fit and 'on'" - but had the wrong explanation - the task is clear: to find a more satisfactory way of identifying the class/form horse in a race. More satisfactory, that is, from an intellectual perspective and because, freed from VDW's numerical limitations, it will find more betting situations. (A simple example of this latter point: VDW indicated in the March 1981 article that we should concern ourselves with horses with one of the four highest ability ratings in the field. In practice, we discover by studying his examples, he sometimes relaxed that slightly - but not much. Yet in practice horses which are the class/form horses in their races and have all the other attributes of a VDW main method bet except a high enough ability rating ranking often win. Similarly, horses with all the "winner in the race" attributes except a low enough last three race placings total often win.
The potential task might be described another way. On 17 January 1981 VDW wrote: "readers who fully understood my previous letters will know it is the balance between class, form and the other factors [especially capability] which shows the good things". The main method is a way of finding some of those "good things" - the likes of Little Owl, Sunset Cristo, Wing And A Prayer, etc. Another way would be to do what John suggests and look at the underlying realities rather than numerical simplifications of statistical findings. That would mean freeing oneself from such concerns as positions in betting forecasts, consistency totals and ability rating rankings and concentrating on (a) finding better ways of assessing relative ability - ways of identifying what John terms the "intrinsic nature of class" - and (b) what in my view is VDW's main contribution, his undertanding of "form". If one could do that, then the fact that the class/form horse in the race so identified had a relatively high consistency total, or a relatively low ability rating would, as John suggests, be neither here nor there.
Did VDW in reality work in this latter way, and give us the main method as a first step on what he hoped would be a voyage of discovery for his readers? I'd like to think so, but as indicated above I doubt it. But although a point of interest, it is irrelevant from a practical perspective as we are absolutely free to try to work in the way John rightly suggests is intellectually more satisfactory.
