chicken or egg

[sammy10k]

I\'m trying to understand how DRF and Equibase come up with a lengths beaten number.

I\'ve seen finish line photos with a time strip up at the top.  I know they use one-fifth of a second is equal to one length.  So if a horse loses by 5 lengths according to them, are they saying he was one second behind the winner according to the time strip and we think that equals five lengths or are they saying that the horse lost by five lengths because that is what we measured?

[TGJB]

Photo finish cameras have always measured time, not distance, and convert it into \"lengths\" when they print it. As I understand it, most tracks have switched to super accurate digital devices these days. If it wasn\'t a weekend I would pass on your question to the guys at Equibase for a more complete answer-- next week, if I remember.

[sammy10k]

So if the winner of a six furlong race ran 1:09 and won by five lengths (according to the PPs) - the second place horse crossed the line in 1:10?

[TGJB]

They don\'t use 1/5= a length-- that was the old DRF \"speed figure\" formula. Don\'t remember what they do use, it\'s been a long time since I looked at that stuff, and I\'m used to working with lengths. I\'ll let you know after we talk to Equibase (and by the way, all data nowadays is theirs-- the DRF doesn\'t collect its own any more).

[sammy10k]

Ok thanks TGJB.  What I\'m really looking to do is glance at the results (for any distance) and be able to make a pretty accurate estimation of what time a non-winning horse ran.

[asfufh]

sam10k, I think the beaten lengths by Equibase have to be based on the difference between the horse\'s time at it reaches the finish line compared to the winner\'s time. Otherwise, a horse that is second when the winner reaches the finish line but gets passed by another horse to finish third would have less beaten lengths that the second horse. (after rereading this, now I\'m confused):>)Asfufh

[sammy10k]

asfufh - You are correct.  Finishing places are determined by the order of the horses when each hits the wire not the order they are in when the winner crosses the line.  Equibase knows the time for each horse but does not publish it - they use it to calculate lengths beaten.  What I want to know is how to take lengths beaten and go back to the actual time.

I contacted Equibase and they said there is no easy way to calculate the non-winner\'s times.  I take this to mean that the time/length they use varies based on distance or may even be more complicated by using the running speed of the horses to get a decent approximation of their time/length.

Jerry - I\'m curious if you heard anything different.  The way they are calculating this would be important to making figures because if they are using a constant time/length (even if it\'s changed for each distance) it would be misleading if non-winning horses were running faster or slower than that constant.  My guess is that they are adjusting the time/length based on the speed the race is run in (e.g. t/l would be different for a 1:33 mile vs a 1:38 mile) but I just wanted to find out for certain.

[TGJB]

Sammy-- have you ever posted here under another name? Hmmm...

The guy I need to talk to won\'t be in until tomorrow. I\'ll get back to you.

[sammy10k]

I don\'t recall ever posting here.  I\'ve been a casual reader of the boards for a long time but just got a login to post with.

Did you just read Andy Grove\'s Only the Paranoid Survive or something?

[TGJB]

No, it was because of the difference in apparent level of knowledge that appeared in your first post on this string, and that of your response to Asfuh, combined with the anonymous e-mail address.

Should be able to get you your answer tomorrow, although it figures to be a little busy with the Belmont day draw.

[sammy10k]

My initial question was very simple cause I thought the answer was simple.  When I was growing up my dad taught me to read the racing form and he said the way to convert lengths back to time behind was to use 1/5th of second per length - I assumed that was close to how it was done so I didn\'t get more specific with my question.

Later I emailed Equibase asking them my question of how to convert beaten lengths to actual finish times.  I figured they would say it\'s either 1/5th or 1/6th of a second per length.  The answer I actually got was that there was no simple way to do the conversion.  After hearing that, I knew the solution couldn\'t be using 1/5th or 1/6th.

At this point, what I assume they are doing is taking the finish times of non-winners, determining the rate of speed each ran at, using the rate to figure out how far they would have run when the winner crossed the wire, and then convert that distance into lengths.

[sammy10k]

BTW, if I\'m right about them doing it that way it\'s still pretty simple to figure out a horses finishing time as long as you know how many feet they use for a length.

[TGJB]

I just got done talking to two top people at Equibase (and by the way, whom did you e-mail, and who answered you)? They don\'t gather the information themselves, they buy the data from the photo-finish companies, and are going to talk to them and ask a whole list of questions I gave them. I am supposed to be getting the answers by early next week, and will give you the answers then.

I will say this-- a couple of months ago we saw a race where the lengths printed did not match up with what was printed in the chart (I was paying close attention because the horse in question was a Graph Racing horse). I called Equibase to have them check it, and they did, and published a correction. I said, look guys, we need to know this is not happening regularly. So they set up a system of spot checks, and they say that so far everything has matched up within a small margin of error.

More to follow when I get more.

[sammy10k]

I have an online account with Equibase and emailed a guy named Matthew that works in Ebusiness.  He asked someone in Operations my question and emailed me back saying that there is no simple answer.  I don\'t know who it was in Operations that he spoke with.

That\'s interesting that they buy the data from photo finish companies.  Hopefully they\'re all trained to read the and translate the photos the same way.  Obviously it would be better if Equibase could just buy the photos and have their own people read them but that would be more costly.

[TGJB]

They are not reading photos-- they are getting actual times of the horses, converting it into lengths, and giving that info to Equibase. More details next week.