Some questions on fig methodology...
Jerry, you've been pointing out what you see as implausible #s on Rags leading up to the BC. Whether you're right about those specific figs or not, I'm not in position to comment since I'm on sabbatical from serious horseplaying -- the BC was only the second race card I've handicapped all year. (maybe I should do that every year – less fun, but the ROI would be wonderful)
Anyway, your arguments against those Rag figs sounded reasonable enough to honestly give me pause, but then came a race like the Distaff, won by Shug's filly by an acre, a few races after a colt broke a rein and was pulled up. Got me thinking...
Pleasant Home surely got a new top, probably a big new top, no? Or did the field collapse behind her? Both a rest-of-field collapse and a big new top? How can we tell? On Rags, almost the whole field had iffy to bad patterns, but the winner was forward moving in general. Problem (for me) was that she had short rest coming off a top – I didn't play her.
OK, back to your fig methodology... Imagine for a minute that Pleasant Home broke a rein coming out of the gate and was pulled up. If the rest of the field ran poorly, as the runaway winner may make things appear on your figs, how would you have known without the runaway winner?
You argued in your pre-BC posts something like \'how could a horse get a worse figure destroying a top GR I field than someone else did running second in a lowly allowance field?\' Well, if Shug\'s filly were pulled up early in the race, that would have still left 12 of America's top fillies & mares, all pointed for the richest race of the season... so wouldn't you have been in the position of needing to infer some change of track speed so that the "winner" (in real life, the filly who actually ran 2nd) would get a better fig -- the decent fig it takes to defeat a field of America's top females? Or would the absence of Pleasant Home not made any difference in the # you assigned for others in the race?
I'm genuinely interested in your response – I know these issues aren't easy, you've thought a lot about this, you score plenty using your figs, and I'm sure there are plenty of times you are right about changing track speed. Slides, I'd bet happen often. But jump shifts without weather/maintenance shifts... well, let's see your response.
Jerry, sorry you had a rough day – and I'm glad at least some of your customers had a good one. At least on BC days, there's more than enough in the pools for all of us.
Best,
SP
Steve-- this is the right kind of question. It is easier to show this stuff than to explain it, which is why I often post the races (like the Jockey Club), so let\'s wait until I post the figures for the day to have the specific discussion. But a couple of points:
1-- Very few of the significantly changed variants involve pulling one race loose, as opposed to a slide, unless there are different circumstances-- 1 and 2 turn races, sealed track, watering (or not watering), runup differences, etc. Having said that, you don\'t always know the different circumstances-- especially if you don\'t believe they make a difference, and don\'t look for them. I posted about this after the Belmont this year in response to a post of Len\'s.
2-- When I pull a race loose, it\'s not good enough to know that it is obviously wrong to do it with the surrounding races. It has to be clear what the RIGHT correction is. If it\'s not clear, I leave a box, no figure. If it\'s pretty clear but there is a chance it\'s wrong, I do a figure, but review the race later when the horses have run back.
3-- To really KNOW whether the Ragozin figures in question were implausible, you would have to see his sheets for those races, with the figures they assigned-- for ALL of them. If you saw his BC sheets, you saw what he assigned several of the Gold Cup horses-- let\'s see what those sheets look like if he posts BC sheets with figures done.
4-- Here\'s the flip side of the example you gave. Let\'s say a card is cancelled after one or two races. You would of course have to make figures based on those races alone. But if they ran the rest of the card, Len is saying that would be wrong-- you would have to use the later races an make an average, even if it\'s in conflict with the conclusions you would draw FROM THOSE HORSES THEMSELVES.
5-- Your description of my comments (GI vs allowance) were an oversimplification, the same one that was made by Michael and others. If you read my responses to them from a few days ago (especially my post on the \"One More Pop Quiz\" string at 7:12 p.m.), you\'ll see what I really was saying. Short version-- it depends on ALL the horses in those specific races (see 3, above).
6-- What those who have not made projection figures don\'t get (especially if they have read Beyer or some of the earlier books which push the use of pars) is that it\'s not just about the winners. When you have a tight data base using weight and ground you use lots of horses within each race to make your figures.
7-- But yes, one of the tough questions you face is what to do with a runaway winner-- do you give him a big one, or just assume the others collapsed and he ran as usual. That came up with the first big one War Emblem ever ran-- the race before the Illinois Derby. We were the only ones to give him the huge jump. It comes up a lot. But with stake horses it\'s reasonably easy.
More after I do the day.
Jerry,
\"6-- What those who have not made projection figures don\'t get (especially if they have read Beyer or some of the earlier books which push the use of pars) is that it\'s not just about the winners. When you have a tight data base using weight and ground you use lots of horses within each race to make your figures.\"
This is really not true. Beyer never advocated making projection variants based on just the winner. He uses all the horses in the race also. His references to using PARs were mostly limited to the first few weeks you start making figures at a track until you have enough data to make projections.
The question of tweaking figures to keep them in line with expected PARs over time is another issue, but I haven\'t been able to get the full answer to that.