As evidenced from recent races discussed here, it is usually not difficult to sort out the consistent horses, the probables, the probables with form and then the class/form horse, but only in a small minority of cases does the resultant class/form horse have all the characteristics of a "winner in the race". And when one considers handicaps, it is not surprising that many winners are not among the horses highlighted by the basic VDW numerics. For every handicap winner with the 1/1/1 last three race placings of a Sunset Cristo, there are numerous winners with a profile more like that of Roushayd - 3/4/6 - or even less attractive viewed solely from the placings perspective.
In one of his Gummy contributions Lee explained how he addressed this problem - by viewing the consistent horses as what I refer to as the "automatic" consistent horses (those with one of the three lowest placings totals from within the first six on the forecast), and adding to them horses with one of the three highest last time out placings. That is, as VDW did with Warner For Leisure, Gee-A and Smart Tar in the 1988 Mackeson example. I think Lee assumes that is how Roushayd came into consideration (he was not an "automatic" consistent horse), and indeed in one post he referred to Roushayd as not a consistent horse.
I think the balance of evidence from the VDW examples is against Lee here, and prefer my solution that in a significant proportion of races, there are "discretionary" consistent horses, identified by objective rules, to be added to the "automatic" ones (and anyone interested in exploring the two solutions should start with Prominent King and then consider Righthand Man as well as others). But even working with the definition of consistent horses I do, the inescapable fact is that many handicaps are won by horses not identified as consistent in the obvious, numerical, sense and the question arises is there a way of capturing any of those which, but for the initial placings total, do have all the characteristics of a "winner in the race". That is the task I think Lee has addressed, and to which I have been giving a lot of attention this season, albeit with as yet only partial success.
Today's 3.45 provides a good race to illustrate my current approach to trying to go beyond the classic handicap class/form horse like Sunset Cristo and identify the Roushayds which have less obvious form. The race works out very straightforwardly from the perspective of the main method, with three "automatic" consistent horses - Misplaced Fortune, Secret Witness and Piazza San Pietro - and no "discretionary" ones. So no need to deploy the probables method of rating. All three are in my view form horses, but they are 14th, 15th and 16th in the ability rating ranking (and with such exposed horses I am uninclined to give much weight to the time-rating ancillary rating), and none comes close to having the profile of an "80%" horse. That is not, of course, to say that none of them can win: despite having quite a few runs to their credit two are 4yos and could yet show further improvement and even older horses have been know to improve. But none comes close to being a Sunset Cristo type.
To those three, if he was applying the approach he showed on the Gummy forum, Lee would add Flipando, Hitchens, Midnight Martini and Tombi as coming from the three highest class ratings races last time out. I am currently working rather differently, paying attention to horses from the top five of the ability rating ranking and even closer attention to those from the top five which are also in the first six of the forecast and have already won races of comparable class to that under consideration. In the 3.45 that means Hitchens and Lowdown get first attention, and then Abraham Lincoln, Flipando, Joseph Henry and Midnight Martini.
It would take too long to discuss each, but the questions one is addressing are the same each time: first, can one reasonably suppose there is unexposed form in the horse's profile (ie form that does not show up via the placings totals), and second is there reason for supposing that today is the trainer's target for the horse.
Of the six I've listed, Hitchens strikes me as the most likely, for the following reasons:
1) he is in the first six of the forecast and has won higher class races than today's 117 (a 312 and a 130);
2) his profile since the more recent of these wins, on 26/09/09, shows consistency;
3) his last run suggests he is close to being in good form and the conditions today, when considered in the context of his recent profile, suggest it is plausible for him to win.
Taking (2) first, all Hitchens' wins to date have been over 6f, and the admittedly fairly limited evidence suggests 5f is too short and over 6f too far. Thus I am concerned only with his handicap runs over 6f since 26/09/10. There have been five such runs, and looking at them in the context both of the race on 26/09/09 and today's gives the following picture (AOR = average OR of the field):
26/09/09: class 130, AOR 90.4, off 95, won
10/10/09: class 259, AOR 95.0, off 98, beaten 2.3l
23/10/09: class 117, AOR 89.9, off 99, beaten 3.8l
24/07/10: class 324, AOR 96.2, off 98, beaten 2.9l
31/07/10: class 623, AOR 101.1, off 99, beaten 7.0l
14/08/10: class 405, AOR 96.3, off 97, beaten 2.1l
05/09/10: class 117, AOR 90.8, off 96
The above suggests that in races of the class in which he has been running, Hitchens struggles off marks much above 95, but except when pitched in way too deep, (as in the class 623/AOR 101.1) he has run consistently well. In general, the picture is of the nearer the class is to 130/AOR 90.4 and his OR to 95, the better he does, and the further away the worse.
If we focus on his runs this year, ignoring the 6.5f opener at the Curragh, probably over further than ideal, they were all 6f races in class terms well above his win in class 130/AOR 90.4. Leaving to one side the class 623, we see two decent runs, the latter arguably showing improvement:
24/07/10: class 324, AOR 96.2, off 98, beaten 2.9l
14/08/10: class 405, AOR 96.3, off 97, beaten 2.1l,
ie the 14/08/10 race was of higher class on both measures, and even allowing for the fact he ran off 1lb less in the relative weights, an arguably improved performance. And as Hitchens' history suggests he is at his best on good/firm, the going on 24/07/10, his run on 14/08/10, on good/soft, may have been even better than the numbers indicate.
Today the trip is right, and the going at present shown as good/firm. Logically a horse which was beaten only 2.1l off 97 in a class 405/AOR 96.3 would seem to have a good chance in a class 117/AOR 90.8 off 96. In other words, a plausible interpretation of the data is that Hitchens has shown improvement over his relevant runs this season and is now poised to win a race of essentially the same class, off essentially the same mark, as per his last win.
Judged by ability ratings, there are two better horses in the race - Midnight Martini and Flipando - but a similar look at their recent profiles leads me to conclude that thet are less likely to win today than Hitchens, and I can't see a convincing reason why Lowdown, equal third rated on ability with Hitchens, should reverse placings from their runs in the Great St Wilfred on 14 August. In short, on the approach to trying to identify winners with unexposed form (in VDW's sense of that term) Hitchens looks plausible, and I'd rather back him than whichever of the three probables with form one regards as the class/form horse.
However, this is very much work in progress and I am far from confident that I have yet identified the full characteristics of such winners and certainly I don't yet have enough examples to be able to put a warranted probability figure against Hitchen's name
