Archive study on horses that have run 3 pt or more tops in final derby preps

[Wrongly]

Small Timer

You can download the data as an excel spreadsheet and then sort on any column you like.

What quickly appears to me is the top 5 fastest TG figs each year account for 10 of those 16 winners or roughly 63%, another 12 finished second or 3rd so that\'s 28 of 48 or 58% ITM.

Interesting to me is the center of distribution in the pedigree only 11 of those 48 where 1.0 or higher.  There are always exceptions to any rule but lower CD tend to win a higher percentage of distance races.  Dr. Roman has some charts on his web site.

[ajkreider]

This isn\'t a pedigree issue, and I\'m not sure if it is worth noting.  But it appears you shouldn\'t play a horse for a big jump up relative to his peers. By the data, it looks like a horse that had a last out route figure that ranked outside the top ten was unlikely make it into the Derby super.  

19/64 did it, which actually seems okay.  But 9 of those 19 were on wet tracks.  Two others were jump ups from poly races.  Since 1998, when the Derby was run on a fast track, at least three of the four superfecta horses had a last out top ten route fig - with the exception of Barbaro\'s year, when it was 2/4 (including Barbaro).  

Edit:  it was 2/4 in Giacomo\'s year.

[kmart4503]

I emailed Steve Roman to confirm and below is his response in regards to the data I used.

The archival Derby data is unchanged.  The figures are those on the day of each year\'s race.

Regards,
Steve

[TGJB]

Good, thanks.

Only problem with that table is Real Quiet is still ahead of Victory Gallop.

[smalltimer]

Thanks wrongly, I would have never figured that out on my own.  I was trying to get someone else to do the work for me since I bought a new computer yesterday and have installed lots of software thus far, excel/word notwithstanding.
Not to waste your time, but I\'ve done a spreadsheet or two since I was doing reel to reel computers in the Air Force in 1971.
I think sorting the winners into one group is the fast to finding any pertinent similarities.
Have a great one fellas!

[pres711]

Jim, I don\'t post often here but always enjoy you thoughtful analysis.  Have been a long time user and fervent derby fan.  Saw my first live derby in 1972 and got hooked.  made 30 years without a miss till age, work and time caught up with me to settle on living room downs.  anyway, I was prepared to drop WS on bounce theory looking for more value.  I\'m sitting on a major three fig futures bet on CC at 31-1 so I have value there.  I\'m also live to GAR, VIT and Sam for smaller plays.  Don\'t usually play so many but in Feb all of them looked pretty live.  I still think that CC is going to take something special to be beaten this year.  they have not yet asked him to run.  Good luck, bring them home safely.

[louisvilleguy1111]

I am really starting to think those future wagers on General A Rod are looking better and better every day!

[phil23]

Ok, I used the Difference between their Derby fig and their Previous Route Best AND their Dosage. I broke it down by two groups: Dosages < 3 (which is what that Romans guy claims means they are more likely to be routers, vs Dosages >= 3 (more likely to be sprinter types.

For Dosage <3 (sample size 171)
Tops: 9%
Pair: 21%
Off: 21 %
X: 49%

For Dosage >=3 (sample size 127)
Tops: 11%
Pair: 24%
Off: 19%
X: 46%

I just used a couple of simple COUNTIF statements to breakout the numbers from the columns in excel. Someone maybe can check my numbers to ensure I didn\'t screw up?

Less than 3 might seem a random dividing point to pick but (apparently) by definition, that is the dividing line for routers and sprinters, so, that\'s what used. The fact that VG, BODI, POINT GIVEN, TAKE CHARGE INDY, ORB all fall into the sprinter category seems...insane?...but what are you gonna do.

Anyway, based on this quick look at it, 30% of the horses supposedly predisposed to ROUTE Pair or Top, vs 35% of the horses supposedly predisposed to sprint.  In other words using dosage to divide them, the sprinters actually did better...which does not exactly give one great hope that this (dosage index) is a great way to identify horses who will perform better than average at 10f in the derby.

[TGJB]

Interesting. Basically, based only on that criteria, there\'s random correlation, which is information (though clearly not conclusive). That\'s at least a start towards a meaningful conversation about distance ability.

[pres711]

I always thought that the dosage for distance was <4, would that make a difference?

[phil23]

http://chef-de-race.com/dosage/drf_series/original_drf_pt2.htm

He says at the bottom of the page that <3 is where \"stamina is significant.\"

Also, if you run it >=4 there\'s only about 30 samples in that one.

[TGJB]

Can you run 3 groups, one in the middle range and one on each end?

[phil23]

For Dosage <2.00 (80 sample size)

Top 11%
Pr 21%
Off 18%
X 50%

For Dosage >=2.00 and <3.50 (149 sample size)

Top 8%
Pr 21%
Off 26%
X 46%

For Dosage >=3.50 (68 sample size)

Top 12%
Pr 25%
Off 12%
X 51%

[TGJB]

Romans won\'t like it.

[kmart4503]

I ran the same with slightly different groups and below are my results.


For Dosage <=2.1 (96 sample size)

Top 11%
Pr 20%
Off 19%
X 50%

For Dosage >2.1 and <=3 (110 sample size)

Top 6%
Pr 24%
Off 22%
X 48%

For Dosage >3 (91 sample size)

Top 12%
Pr 25%
Off 20%
X 43%