Best Contender Qualifiers

[traynor]

What criteria have you found to be the most accurate in distinguishing between contenders and non-contenders? Not the same as "eliminations" but rather a means to distinguish entries with a serious chance to be close to the winner at the wire as opposed to those who are likely only to win if major catastrophe befalls all the other entries.

I have been using a set of criteria in building models, and "analyzing the analysis" indicates there are substantial differences between tracks (just as there is with most other things). Specifically, a "set of rules" applicable at Track A may be useful, but applying that same "set of rules" to Track B might be deceptive (to put it mildly).

Examples (from the test utility I am currently building):

Btva Pace Contenders AllPosts 0.90 790 Winners in 875 Races
Btva Pace NonContenders AllPosts 0.10 85 Winners in 875 Races

MOH Pace Contenders AllPosts 0.72 472 Winners in 657 Races
MOH Pace NonContenders AllPosts 0.28 185 Winners in 657 Races

Translated (briefly): Using a specific set of criteria to qualify contenders includes 90% of the winners at Batavia. The same set of criteria used to qualify contenders at Mohawk misses badly--28% of the winners would be categorized as "non-contenders."

I bet a LOT of exactas and trifectas (usually key entry + whatever others are most likely to be in contention near the wire). I assume I am not the only one. Including or discarding the appropriate entries in such wagers is critical.

Any suggestions, ideas, thoughts?

[Ray2000]

What criteria have you found to be the most accurate in distinguishing between contenders and non-contenders?

Any suggestions, ideas, thoughts?

Maybe not most accurate but still very important.

Average of last 6 starts for position gained or lost from post to the half pole.

Don't use breaks or qualifiers.

Must account for 2nd tier starts. (1st trailer = imaginary post 1½)

Allow for agressive driver changes.

Examples:
Draws Post 2, sits 5th at half pole = -3

Draws Post 8, sits 2nd at half pole = +6

Pace not considered for this criteria. (used in stretch performance or Form)

[traynor]

Maybe not most accurate but still very important.

Average of last 6 starts for position gained or lost from post to the half pole.

Don't use breaks or qualifiers.

Must account for 2nd tier starts. (1st trailer = imaginary post 1½)

Allow for agressive driver changes.

Examples:
Draws Post 2, sits 5th at half pole = -3

Draws Post 8, sits 2nd at half pole = +6

Pace not considered for this criteria. (used in stretch performance or Form)
Thanks! That is a whole different way of looking at the process than the one I use (which is what I am really looking for). I will look at it from that angle and see how it goes.

I should have explained the "why." I don't handicap races in the "conventional" manner. Rather, I apply a model pattern to all entries, and those that fit are the selections (that is a greatly simplified version, but much less verbose and boring than the actual processes). It works fine, but I have been noticing areas at some tracks in which winners seem to "come out of nowhere" (or at least so it appears to me). That may be because I am looking too much at recent form as part of the qualifying process, and results could be improved by using other qualifiers.

I have actually just started looking at the "contender/non-contender" process, because I don't normally consider it in handicapping. I am building a search utility to isolate contender/non-contender patterns on a track by track basis, and will post any results that I think might be helpful.

The figures for Batavia and Mohawk in the first post use a (very) simple format that is probably a carryover from the days when I supported my thoroughbred wagering "career" with funding obtained from scanning harness programs tacked on the wall of the Niagara Falls OTB--average 30 seconds or less per race. Basically, I only considered the last two races, and ignored any entry that had not been 3rd or better or within 2 lengths of the leader at some point in at least one of those two races. At Batavia, 90% of the winners are included in that initial process, and it frequently reduces the race to two or three "contenders" that can then be separated using other techniques. Crude, but effective. Especially at Batavia and Buffalo.

However, to use a phrase common in business, I think I may be leaving too much money on the table by avoiding races in which most--if not all--of the entries lack any useful (or readily identifiable) indications of current form. That would be foolish.

[traynor]

Contenders qualified by 3rd or better, leading or within 2 lengths of leader at least one point in one of the two most recent races. Only additional factor that improved results was odds in today's race 3/1 or less. Other than today's odds, additional qualifiers (including close at multiple calls) REDUCED overall accuracy, including all the "tried and true" nonsense like dropping in class, trainer changes, driver changes, etc., etc., ad nauseum.

These values are current as of today. I am recoding the parsing utility to categorize by track size and post position (inner vs outer), and generate a generic set of values. I will post results when available.

Btva Pace Contenders AllPosts 0.90 790 Winners in 875 Races
Btva Pace NonContenders AllPosts 0.10 85 Winners in 875 Races

Fhld Pace Contenders AllPosts 0.86 469 Winners in 548 Races
Fhld Pace NonContenders AllPosts 0.14 79 Winners in 548 Races

Har Pace Contenders AllPosts 0.88 399 Winners in 456 Races
Har Pace NonContenders AllPosts 0.13 57 Winners in 456 Races

Mea Pace Contenders AllPosts 0.80 620 Winners in 774 Races
Mea Pace NonContenders AllPosts 0.20 154 Winners in 774 Races
Mea Trot Contenders AllPosts 0.78 638 Winners in 813 Races
Mea Trot NonContenders AllPosts 0.22 175 Winners in 813 Races

MOH Pace Contenders AllPosts 0.72 472 Winners in 657 Races
MOH Pace NonContenders AllPosts 0.28 185 Winners in 657 Races

MR Pace Contenders AllPosts 0.82 727 Winners in 891 Races
MR Pace NonContenders AllPosts 0.18 164 Winners in 891 Races

Nfld Pace Contenders AllPosts 0.79 603 Winners in 762 Races
Nfld Pace NonContenders AllPosts 0.21 159 Winners in 762 Races
Nfld Trot Contenders AllPosts 0.78 493 Winners in 632 Races
Nfld Trot NonContenders AllPosts 0.22 139 Winners in 632 Races

Nor Pace Contenders AllPosts 0.80 347 Winners in 432 Races
Nor Pace NonContenders AllPosts 0.20 85 Winners in 432 Races

PcD Pace Contenders AllPosts 0.80 609 Winners in 759 Races
PcD Pace NonContenders AllPosts 0.20 150 Winners in 759 Races
PcD Trot Contenders AllPosts 0.72 364 Winners in 505 Races
PcD Trot NonContenders AllPosts 0.28 141 Winners in 505 Races

Phi Pace Contenders AllPosts 0.82 674 Winners in 825 Races
Phi Pace NonContenders AllPosts 0.18 151 Winners in 825 Races

PRc Pace Contenders AllPosts 0.84 386 Winners in 459 Races
PRc Pace NonContenders AllPosts 0.16 73 Winners in 459 Races

Scar Pace Contenders AllPosts 0.86 438 Winners in 512 Races
Scar Pace NonContenders AllPosts 0.14 74 Winners in 512 Races

Stga Pace Contenders AllPosts 0.80 582 Winners in 732 Races
Stga Pace NonContenders AllPosts 0.20 150 Winners in 732 Races
Stga Trot Contenders AllPosts 0.80 408 Winners in 511 Races
Stga Trot NonContenders AllPosts 0.20 103 Winners in 511 Races

YR Pace Contenders AllPosts 0.89 756 Winners in 851 Races
YR Pace NonContenders AllPosts 0.11 95 Winners in 851 Races

[traynor]

Exactly the same criteria as used for win position, in this example, for PLACE position. Data is incomplete. When I have more time I will layer by inner/outer posts for both win and place values and post it.. I have NOT tested whether or not additional factors would improve the PLACE values. So much to do, so little time ...

Btva Pace Contenders AllPosts 0.91 926 Placers in 1015 Races
Btva Pace NonContenders AllPosts 0.09 89 Placers in 1015 Races

Fhld Pace Contenders AllPosts 0.82 469 Placers in 573 Races
Fhld Pace NonContenders AllPosts 0.18 104 Placers in 573 Races

FLMD Pace Contenders AllPosts 0.81 424 Placers in 523 Races
FLMD Pace NonContenders AllPosts 0.19 99 Placers in 523 Races

Har Pace Contenders AllPosts 0.88 431 Placers in 490 Races
Har Pace NonContenders AllPosts 0.12 59 Placers in 490 Races

Mea Pace Contenders AllPosts 0.81 781 Placers in 969 Races
Mea Pace NonContenders AllPosts 0.19 188 Placers in 969 Races

MR Pace Contenders AllPosts 0.80 777 Placers in 971 Races
MR Pace NonContenders AllPosts 0.20 194 Placers in 971 Races

MOH Pace Contenders AllPosts 0.75 514 Placers in 685 Races
MOH Pace NonContenders AllPosts 0.25 171 Placers in 685 Races

Nfld Pace Contenders AllPosts 0.80 757 Placers in 952 Races
Nfld Pace NonContenders AllPosts 0.20 195 Placers in 952 Races

Nor Pace Contenders AllPosts 0.79 352 Placers in 444 Races
Nor Pace NonContenders AllPosts 0.21 92 Placers in 444 Races

Phi Pace Contenders AllPosts 0.80 747 Placers in 934 Races
Phi Pace NonContenders AllPosts 0.20 187 Placers in 934 Races

PcD Pace Contenders AllPosts 0.80 653 Placers in 815 Races
PcD Pace NonContenders AllPosts 0.20 162 Placers in 815 Races

PRc Pace Contenders AllPosts 0.84 403 Placers in 480 Races
PRc Pace NonContenders AllPosts 0.16 77 Placers in 480 Races

Stga Pace Contenders AllPosts 0.76 552 Placers in 728 Races
Stga Pace NonContenders AllPosts 0.24 176 Placers in 728 Races

Scar Pace Contenders AllPosts 0.81 424 Placers in 521 Races
Scar Pace NonContenders AllPosts 0.19 97 Placers in 521 Races

YR Pace Contenders AllPosts 0.87 929 Placers in 1064 Races
YR Pace NonContenders AllPosts 0.13 135 Placers in 1064 Races
' --------------------------------------------------------------
Mea Trot Contenders AllPosts 0.80 700 Placers in 874 Races
Mea Trot NonContenders AllPosts 0.20 174 Placers in 874 Races

Nfld Trot Contenders AllPosts 0.77 508 Placers in 656 Races
Nfld Trot NonContenders AllPosts 0.23 148 Placers in 656 Races

PcD Trot Contenders AllPosts 0.71 373 Placers in 529 Races
PcD Trot NonContenders AllPosts 0.29 156 Placers in 529 Races

Stga Trot Contenders AllPosts 0.73 385 Placers in 524 Races
Stga Trot NonContenders AllPosts 0.27 139 Placers in 524 Races

[willphorse]

Trackmaster Speed Ratings

Best last 3, as long as one other in last 6 races was within 3 points. If not, 2nd highest speed rating last 3 races is used. Ties go to 2nd highest SR last 6 races. If 4 SR's or less are showing in PP's, use highest SR, period.

👍 Very good contender breakdown.

[traynor]

Trackmaster Speed Ratings

Best last 3, as long as one other in last 6 races was within 3 points. If not, 2nd highest speed rating last 3 races is used. Ties go to 2nd highest SR last 6 races. If 4 SR's or less are showing in PP's, use highest SR, period.

👍 Very good contender breakdown.

I don't monitor Trackmaster Speed Ratings. Have you done any testing over time to determine the percentage of races in which the contenders you select using the Trackmaster Speed Ratings win compared to the number of races won by other entries?

[willphorse]

I have. It decreases incrementally. Based on my records (About the past year), I see that the top ranked horse wins about 40% of the time. Top 3 win 77% based on my records.

[Stick]

Will
Any problem with shippers using this method? I find the speed ratings to be much more accurate when using a rating from todays track.

[traynor]

I have. It decreases incrementally. Based on my records (About the past year), I see that the top ranked horse wins about 40% of the time. Top 3 win 77% based on my records.

That is a great way for people who don't code their own apps to qualify contenders. Thanks for suggesting it!

My (possible) obsession with "contenders" is in that area in which (most) others discard the entry as a "non-contender" for win. In your approach, about one-quarter of the winners are excluded. Many (if not most) of those excluded will head up very healthy exactas (as well as better-than-average win mutuels). I think it is an area of research with a lot of profit potential.