Breeders' Cup Day

[kev]

Ok thanks.

[TGJB]

Jimbo-- by the time the day gets to me, it\'s in the form of a \"rundown\"-- we have input all the data, and the computer has spit out a winning figure for each race based on a \"mechanical\" variant (formula driven average for the day, sometimes turns out to be close and sometimes not), and a figure for every horse based on their relationship to the winning figure, due to beaten lengths, ground, and weight. I then make \"corrections\"-- which basically means decide what figures should be assigned to each race (as opposed to each horse-- you can only adjust a race as a whole, or I really would be doing the nonsense that Friedman claims I am of \"giving horses what I want\").

In order to see how I did a day, you don\'t really need to see what the overall variant was (in this cast about minus 6 from a totally arbitrary \"par\", before my adjustments). You just need to see how I did the races relative to each other. Adjustments of a point or so in either direction from where you think the day is (even assuming you think it did not change) are nothing-- the underlying data is not nearly accurate enough to support the position that what you have come up with is exact race to race (wind readings, wind formula, slight changes in track speed, let alone the significant ones I talk about in \"Changing Track Speeds\").

Except, of course, if you are a religious fundamentalist:

\"Now I am ready to change the variant to fine-tune the number based on an analytical look at each horse\'s development-- but if I change one figure, I must equally change them all. If I want to change Holy Bull\'s provisional rating for a 1 1/4 mile race from 5 1/4 to the more likely 4, I must subtract 1 1/4 from every other horse-- including sprinters-- running on the dirt that day. I can\'t tidy up the horses\' lines by subtracting a point here and adding a point there\".

--- Len Ragozin (\"with Len Friedman\"), \"The Odds Must Be Crazy\", page 53.

That is absolutely true. But only if you assume that the time you are using is exact, track speed is a constant (see \"Changing Track Speeds\" in the archives), that the relationship between distances is a constant (ditto, also note changing runups), that your wind information is accurate (it\'s an estimate, usually done before or after a race, not during), and your wind formula accurate (another estimate based on looking at the AVERAGE effect of wind over lots of races, and doesn\'t take into account the currents that are created when wind hits a big building at different angles).

Other than that it\'s fine.

[miff]

\"In order to see how I did a day, you don\'t really need to see what the overall variant was (in this cast about minus 6 from a totally arbitrary \"par\", before my adjustments\"


Jerry,

Can you briefly elaborate on \"minus 6\" Does that mean dirt races were being run 6 lengths faster than the totally arbitrary par? or was the variant 6 points lower than par?Thanks


Mike

[TGJB]

Miff-- 6 lengths slower than the arbitrary par, based on the \"mechanical\" variant. Since my adjustments ranged from minus 4 to plus 1 (roughly), the day ranges from minus 10 to minus 5. Most tracks these days run maybe minus 3 to about minus 20 (*#%@ Keeneland). They are much slower than they were when we started 20 years ago.

By the way, we do that mechanical variant different than anyone else, as far as I know. Most others use claiming pars-- we do it by using a formula based on the past figures of the horses running that day, just to get us started. On a lot of days there are not many claimers (sometimes not any, like BC day), and this gives us something that gets us in the ballpark, so the person doing the day doesn\'t have to juggle big numbers in his head (like minus 117) when looking at each individual horse.

This day should give everyone a good chance to see what happens if you don\'t break races out. Do you really want to have the first race going even (3 points) slower? Would you want to have to have all those fillies behind PH running even (4 points) worse? Or, would you want to have Stevie et. al. running faster?

Hell, you don\'t even have to think about it. All you have to do is wait for Friedman to post Ragozin\'s BC figures. And that won\'t even be the best part-- unless they do something really crazy, like break their own rules. We\'ll see.

[miff]

Thanks Jerry,

I always have difficulty with the notion of what makes \"sense\" based on PREVIOUS efforts.Although it is highly unlikely that a majority of GR 1 runners would X in a race, (but it seems to have happened) then thats what I would always prefer to see, instead of figs adjusted to what makes \"sense\" based on prior efforts.You know as well as I that sometimes they just don\'t run.

I\'m not saying that with regard to your BC figs which I feel are excellent and in line with what I thought you would give and how the runners performed.Only Pleasant Home is a bit tough for me to accept knowing his previous plodding, ground loss style, although she did make a \"huge\"run thru the lane.

Mike

[TGJB]

Mike-- The concept of the projection method (or any method) of making figures is that we use the horses\' previous figures to make today\'s-- we don\'t have a machine that gives us an \"objective\" measure of \"resiliency\", to quote Friedman\'s claims of what his figures represent. As such, there is no basis for deciding that a whole field of horses \"X\"d, except by tying the race to others. And as the science shows, (\"Changing Track Speeds\"), aside from common sense and experience, that\'s not a safe thing to do.

The PH race was obviously one of the reasons (the main one) why I said doing the day was brutal. SP had the questions right-- if the winner wasn\'t in the race, or you didn\'t have any surrounding races, how would you do it? Probably differently-- we go with the data we have. But there was no way you could make the case of that race going slow enough to tie it with the rest of the day-- even if you didn\'t know about the track maintenance. It would have given all the rest of them simply horrible figures-- and it was beyond unlikely mathematically that so many would do so on the same day. George took somebody on the other board through something similar recently, but try this-- look at all the fillies in the Distaff, count the number of times they have run this year, and the number of times they have run more than 3 points worse than their top. Then take that percentage, and work out the chance of ALL of them doing it at the same time. Let\'s say it\'s 30%-- the chance of 3 doing it is 30% of 30% of 30%, or 2.7%. And that\'s just 3 horses.

If I didn\'t have any other data to work with-- like, if it had been 2 turns, and the only one of the day--and PH had not been in the race, I would have given Ashado and the rest better numbers. Even with the additional data, if there was a way to do it that made complete sense for the rest of them I would have done it that way, even if it meant giving PH a minus 8. But there was not, and the more you took off, the further it got from the surrounding data, which is not conclusive, but yet another piece of evidence. So you put it together the best you can, all things considered, and find the most likely scenario. It\'s both art and science-- and has nothing to do with dogma.

TGJB,

Given everything you said about the PH race, let\'s suppose \"theoretically\" she wasn\'t in the race and you assigned the race a different figure. In both cases, \"as is\" and \"theoretcially\" would you keep this all (the fact that it was a tough day) in mind when these horse came back and you had to try to square another race in the future.  

[miff]

Jerry,

I do understand, but you must admit that PH\'s fig is only 1w re ground and to my way of interpreting your figs, that is a monster performance by a heretofore rather nice/decent horse.I\'m not saying it\'s wrong, just tough to swallow, and I want to see her next performance.

Mike

[TGJB]

CH-- it\'s easier to answer this in the abstract. When there is significant doubt about a figure, I do review it again, starting about a month later, when the horses start running back. In this case, I would not pay much attention if PH went back, because she figures to. If she ran another big one-- a pair, or a little better or worse, I would pay attention. But it\'s all the other horses I would be looking at.

TGJB,

>But it\'s all the other horses I would be looking at.<

OK. I agree. I wasn\'t clear. PH could run practically anything in her next start and it might not indicate much about this last figure unless perhaps if she paired it.  It\'s the rest that are important.

[miff]

Getting senile. I BELIEVE that someone mentioned to me yesterday at the track that PH was being bred and if so we\'ll never know. Anyone hear anything about PH\'s future.

[TGJB]

Correction-- I spoke without looking when I answered a question yesterday. Not that it really matters, but before corrections, we started the BC day at minus 8, not minus 6.

[marcus]

Plus PH ran well 2nd out and did not come forward after that for a while as also was the case 1st time she ran the 1 3/4 as a 4 yo and that was only about a 3.25 move , now off the that 5 pt top it remains to be seen but I would venture to guess that the BC effort took alot out of her and she probably should be retired for breeding purpose\'s now . In my opinion , the way she ran that number is supported by her pattern but the top itself is unsupported so a bounce next , at the very least , is likely with this one   .

[BitPlayer]

TGJB -

I\'m curious about the adjustments for races 2 through 4.  If I understand your process correctly, the wind adjustment was already in the numbers before you made the adjustments listed in your post.  That being the case, and with all three races being at the same distance over the same surface without any water being added, it looks odd that the adjustments for races 2 and 4 are the same (representing no change in track speed, I think), but the adjustment for the intervening race (the Juvenile Fillies) is 1.5 points different.  What do you think accounts for the difference?  Gusting or swirling winds?  Pace?

(Respect omitted per your request.)

BitPlayer

[TGJB]

Bit-- there certainly are slight wind issues race to race. But there are a zillion other things it could be-- see \"Changing Track Speeds\" in the archive section.