Quantifying post position

[Lou G]

I've come to the point in a program I'm developing where I'd like to include the effect of post position (both past and today's) in my code. Beyond knowing that early horses are disadvantaged by outside posts and closers by inside ones, I really don't have the first clue how to incorporate this into code. Do any of you who program care to share your thoughts on how to make a number out of this?

Best to all,
Lou G

[acorn54]

i know that in winning at the races book by william quirin there is an appendix on just this factor
he illustrates linear regression analysis using finish position in relation to the the post position the horse had
hope this helps
acorn

[cj]

Wow Lou, that is a tough one.

For example, obviously the outside post in a route race would be at a big disadvantage at a lot of tracks. But, we've all seen the 10 horse that clears the field and moves to the rail before they've run half a furlong. Do you really want to adjust for a post in this situation? Post can certainly affect the trip, but it doesn't have to for sure.

How do you adjust for the speed horse on the rail that breaks a length slow and gets boxed in, virtually eliminating any chance he may have had?

I'm not saying it can't be done, but its not something I'd want to tackle on a programming level.

[keilan]

I've come to the point in a program I'm developing where I'd like to include the effect of post position (both past and today's) in my code. Beyond knowing that early horses are disadvantaged by outside posts and closers by inside ones, I really don't have the first clue how to incorporate this into code. Do any of you who program care to share your thoughts on how to make a number out of this?

Best to all,
Lou G

Why would early horses be disadvantaged by an outside post. In many instances that's the place to be. And there's nothing wrong with closers drawing an inside post. Your comments are true at some tracks/distances/surfaces but those are the exceptions imo

[Turfday]

I do not write code, but the software programmer for our site is an expert. As an Equibase partner, we have every single race run in North America and Canada in our database since Jan. 1, 1995.

We have post position stats for every track in North America and Canada broken down by the three important factors:

1) Surface (turf or dirt)

2) Field Size

3) Distance

We do not LUMP field sizes together and do not lump distances together. We believe that the dynamics of a race change as field sizes increase and decrease, just as we believe you can't clump all turf routes or dirt routes together.

Aside from the raw statistics available, we also use a formula called Betting Value Averages, which uses field size, closing odds, finish position and a "rolling average" timeline which gives more credit for recent trends, but incorporates all of the data for each individual post, field size and distance.

Without spending hours and hours of research, for example, I can tell you that at many tracks the outside posts in dirt route races are not NEARLY as bad as the public perceives and in fact, have higher win percentages and based on my BVA's are actually more productive.

I can also tell you, as odd as this may seem, that posts five and six at the specific distance of SIX furlongs, generally UNDERPERFORM all of the other posts at that distance. This was from a huge sample size of many individual tracks around the country that I looked at about a year ago. I surmised that since they were the middle posts (and thought to be benign by the crowd), horses breaking from the inside and outside were more likely to "sandwich" the post five and six breakers, from the inside out and outside in, or catching them pressured from close quarters between horses.

For the most part, drawing at outside post position in a turf race is worse than drawing on on the dirt.

[banacek]

Why would early horses be disadvantaged by an outside post.
As someone who learned to handicap at bullring tracks and still does some (e.g. Hastings), it can be a big deal. There are (one-dimensional) speed horses at Hastings every day that I can throw out just because they've got one horse inside them with as good of early speed. The horse tries for the lead and pushed the pace alongside, can't get it and is toast. But if he was on the inside of the other early speed horse, it is often completely different. He might be a single for me!

The funny thing is I didn't think it mattered at bigger tracks, and it certainly doesn't matter as much, but through the years I pay attention to it there too. A horse pushes the pace 2 or 3 wide all the way (and around the turn where it really hurts) and runs out of gas. If he would have been on the rail, it can often be a different story. He goes on my watch list right away.

[andicap]

What I think would be productive is knowing how favorites -- or maybe stronger favorites, those under 2-1 in an 8 horse field say -- did under various circumstances.
This would allow you to handicap the crowd which is pretty damn important.
I'd love to know that at Belmont (hypothetically) the favorites win only 21% at post 5 at 7f sprints. Obviously you'd need a big database going back a few years to get enough sampling. And even then it would be tough considering you have to take field sizes into account. Maybe you could group certain field sizes to get a bigger sample.

You would want to have a running total as well to see if the trend has obviated in the last year. Maybe over 3 years there is a strong trend but the people has caught on and it's not as strong this year as it has been. Or there was some weird bias this year due to the weather/humidity, new track superintendent, etc.

Conversely at which post positions did horses overperform, i.e., the public consistently underbets their real chances due to some unwarranted prejudice about a post position.

I'd also try and keep track by running style as well, if possible. But this might be slicing the onion way too thinly. Maybe favorites who are early speed horses win much more than favorites who are stalkers or closers at various post positions.
but when you start to consider the various field sizes and such you really wouldn't get much of a sample I guess.

[Lou G]

CJ - Yes, I know it's a bear and may not even be possible. If my application was database oriented I could just toss PP into the hopper and let the chips fall wherever.

Keilan - we're both half right. I just had a look at the PP report provided in Equisim which breaks PP out by running style and the numbers for E and S runners show a steady decline as they move out from the rail. The thing about the inside being bad for late runners (told to me by a retired jockey years ago at AP) is that if the field bunches up rather than stringing out the horse can end up having to either go very wide to circle the field or try to find room through traffic. His opinion, not mine. All I personally know about horses is that they are very tall, have a leg at each corner and dislike me on sight...

Thanks for your thoughts, guys - it may turn out to not be doable but I'm trying to cover all bases.

Lou G

[kenwoodallpromos]

You may want to do some more independent study of the effects yourself so tailor your code for your own purposes.
See what races are affected by any horse that can clear and how much energy is expended.
I can often eliminate a contender who cannot win on the outside, and can find an overlay with a horse going from outside to inside.

[Dave Schwartz]

My experience indicates that PP should be cross-referenced with running style (i.e. E, EP, P, etc.)

I also find that field size (as mentioned by Turfday) is very important. Being outside in a 6-horse field maybe a positive while being outside in a 12-horse field may be negative.

Personally, I would build such a model "dynamically." That is, based upon today's field size, number of front runners, and whatever else you feel is important.


Regards,
Dave Schwartz

[Overlay]

With reference to Quirin's findings, in six-furlong races at tracks which were one mile in circumference, post positions 7-12 all had impact values below 1.00. Those horses starting from the most outside position (regardless of field size) had a cumulative impact value of .74. The only statistically significant result which qualified as an independent variable was a 1.22 impact value for horses starting from post position 1. At the finish, the average horse breaking from post 12 ended up one length behind the average horse breaking from post 1. As far as early speed, there were no significant findings, since horses that did not have early speed starting from inside posts tended to get pinched back, accounting for an average difference of only one-half length between post positions 1 and 12 by the first turn. In route races of 1 to 1-1/8 miles run around two turns on mile tracks, post positions 5-12 all had impact values below 1.00. The only significant result was again a positive one for post 1, with an impact value of 1.44. The average horse breaking from post 12 ended up 2-1/2 lengths behind the average horse breaking from post 1 in a 1-1/16-mile race. (1-1/2 lengths of this amount was lost in in the run around the clubhouse turn to the first-call position.) From the standpoint of early speed, horses starting in post position 1 had a 47% chance of running first, second, or third at the first call; 2-3 had a 40% chance; 4-9 had a 33% chance; and 10-12 had a 25% chance. At tracks less than one mile in circumference, post positions 1-4 in short one-turn sprints had a cumulative impact value of 1.16. Horses breaking from posts 1-5 in two-turn sprints (5-1/2 to 7-1/2 furlongs) had a cumulative impact value of 1.13. In routes at smaller tracks, post position 1 had an impact value of 1.29, post positions 2-7 averaged 1.01, and post positions 9-12 averaged .79. In two-turn turf routes of 1 to 1-1/8 miles, post positions 1-3 had a value of 1.20; 4-6 had a value of .97; 7-9 had a value of .90; and 10-12 had a value of .60. Within these overall averages, however, there was wide track-to-track variance, based on the distinctive contours of many turf courses.