A new way of looking at charts

[Valuist]

Sometimes I wonder if I'm the only one who thinks the DRF would be better off printing their charts differently. Instead of seeing horse A is 2nd by 1.5 lengths in front of horse B at the second call, why not just print the lengths behind the leader at each call? Its a bit of work, but is the equivalent of "seeing the race from overhead". Here's how the chart for Sunday's Garden City Stakes would look:

numbers are lengths behind at 1/2 mile, 2nd call and finish

Indy 500 9.0 6.5 0.0
Dimitrova 4.5 2.5 4.0
Campsie Flls 4.5 2.5 6.25
Baie 6.5 4.0 7.25
Acago 1.5 1.0 7.25
Andover Ldy 7.0 5.5 10.75
Sand Sprngs 0.0 0.0 12.25
Pttnvlle Ldy 2.0 3.0 39.25

A quick look and you can see how badly the speed tired. This is especially helpful is races at "about" distances on the turf which see very bizarre fractional times or soft turf races in which a :50 flat half may very well be a very fast pace.

[andicap]

Subscribers to HDW get charts that look exactly like that -- and you're right -- it's a lot better IMHO.

(although the DRF charts are much more complete than HDW's in most other respects -- HDW doesn't list owner, complete conditions, winner's breeding, date. track and final position of prior race.)

[kenwoodallpromos]

They probably just type as they see the race, froim the leader to the back for the race as a whole. PP's show the race from the styandpoint of the hose being studied. IMO the industry likes to make things confusing and varied!

[kenwoodallpromos]

If it was too easy too many winning picks would make payouts way too low!

[GameTheory]

Beaten lengths are actually an artificial construct. If the finish position in the chart looks like:

Pos / Lengths
1 - 2
2 - 1
3 - 4
4 - 1
5


Beaten lengths for the 5th place horse would be 2+1+4+1 = 8. Does that mean that the 5th place horse was 8 lengths back when the winner crossed the finish line? No. It doesn't even mean he was in 5th place at the time! (Although he probably was.)

The above example is really saying, "The first place horse finished two lengths ahead of the horse immediately behind him at the moment he finished [which horse was immediately behind him at the time is undefined]. The second place horse finished one length ahead of the horse immediately behind him, etc."


To make this clear, let's take a more extreme situation. Imagine a horse wins by 50 lengths. The chart will say 1 - 50 for the winner, and the place horse will be 2 - x. Does that mean it was the eventual place horse that was 50 lengths back? No. The eventual place horse could well have been in last place at the time the winner crossed the line and just starting his big closing move. We can only guess where the place horse was. We only know the winner was 50 lengths ahead of SOME HORSE behind him.

So you see, adding up beaten lengths to get a "snapshot" picture of the race is creating a snapshot that may not have ever occured in reality. Beaten lengths are really units of time, not distance. (Horse 1 finished x units of time ahead of horse 2; horse 2 finished y units of time ahead of horse 3, etc.) Unfortunately (or fortunately, depending on how you look at it), how much time they represent is undefined and changes from race to race (and even horse to horse).

[hurrikane]

Game,
interesting perspective on the charts.
I"m wondering if you can integrate the essoteric 'race shapes' into some kind of handicapping tool?
Is that where you are looking?
Also, should probably ask on your thread but I don't seem to be able to get the exotic pays in the parser. Particulary p4,6,8
thanks


Ken..
why would moring works be more meaningful than charts of actual races run. You diminish their importance but tout the morning works as gospel.

Ken,
really, just seriously curious about your interpretation.

Thanks

[GameTheory]

Originally posted by hurrikane@HTR
Game,
interesting perspective on the charts.
I"m wondering if you can integrate the essoteric 'race shapes' into some kind of handicapping tool?
Is that where you are looking?I use race shapes somewhat. That type of analysis seems to have limited value -- for me, anyway.

Regarding the charts, I don't really see it as "my perspective" -- it's a fact. The positions/lengths in the charts for each point of call represent the spacing & time of the horses passing a single point, not a snapshot of where all the horses were placed at a single instant. We tend to think of it as a snapshot, which works ok because it is close enough, but that ain't really what it is.




Also, should probably ask on your thread but I don't seem to be able to get the exotic pays in the parser. Particulary p4,6,8
The exotics work as far as I know, but they are handled in sort of strange way that might not be obvious. None of the names of the exotics are built-in to the parser as predefined fields because there are so many different exotics and they come up with new & special ones sometimes. So you have to loop through the "wager:X:wager type" field to find the one you're looking for, or just enumerate a bunch of slots for them like in the xrd.tpl sample. (The X in this case refers to exotic #X, not horse #X.) Email me & I'll explain how to do it, and pick out specific ones.

[hurrikane]

thanks GT
will email.

[Tom]

I thought that at the moment the winner hit the wire, they took a photo of the field and figured out the beaten lengths, but your point is valid-how many times has a horse won by 5-6-7 lengths and the place and show horse change positions after the winner has crossed? Sooooo, nothing is really that accurate.
So why is everyone so concerned about timing to the 100's of a second?

[Zaf]

Originally posted by Tom
I thought that at the moment the winner hit the wire, they took a photo of the field and figured out the beaten lengths, but your point is valid-how many times has a horse won by 5-6-7 lengths and the place and show horse change positions after the winner has crossed? Sooooo, nothing is really that accurate.
So why is everyone so concerned about timing to the 100's of a second?

They do this for Quarter Horses, very precise.

But for thoroughbreds its all guestimates. It can be done for thoroughbreds, but it won't be done in the near future. And for that matter they won't upgrade the tote (past - posting system) or lower our takeout. Even the freakin greedy racinos won't lower the takeout.

Damn they got us by the balls, and we will keep playing.

See you at the contest Tom.

ZAFONIC

[GameTheory]

For all the calls except the finish, they eyeball it or refer to video.

The photo finish camera for the finish uses an aperature which is a very thin vertical strip -- it only photographs the finish LINE, that's it -- just a few millimeters wide. But the film stays constantly moving (it doesn't take frames like a movie camera) and captures the horses over time as they cross over the line. This is why the horses look so strange in photo finishes -- they are often in impossible positions because the picture was built up from hundreds of tiny slivers. They are also stretched out -- there is a chart to correspond distance on the film with real distance to convert to lengths. If the camera is running at the wrong speed, the lengths will be off as well.

[sjk]

GameTheory,

For the purpose of making figures, the lengths need to be turned back into time. It would be helpful to better understand how fast they run the film.

How do they adjust for the fact that the horses are moving much faster at the finish of some races (high caliber sprint) than in others (low caliber route)?

[GameTheory]

Originally posted by sjk
GameTheory,

For the purpose of making figures, the lengths need to be turned back into time. It would be helpful to better understand how fast they run the film.

How do they adjust for the fact that the horses are moving much faster at the finish of some races (high caliber sprint) than in others (low caliber route)?

Yes, beaten lengths don't really make sense until you turn them into units of time. But I think people are too mezmorized by time. I am surprised by how many winners you can pick just by using position & beaten lengths data alone. The horses are running against each other, after all, not the clock. I make distance/track/variant adjusted figures with an automated but complex process. They are about as good as the BRIS final time figures I guess. But using them alone, I can only pick about 3% more winners than by looking at position/beaten lengths.


I don't see why they'd need to adjust the speed of the camera for different races, but I have no idea. The question has been brought up before -- do they actually use the distance between the horses measured on film as the lengths, or do they measure the time between horses, and then convert that to lengths with an arbitrary rule like 1 length = 1/5 second. I can't find enough consistency in the chart data (at a single track) to believe the latter, but it is unclear exactly how it is done...

[Valuist]

The point here isn't whether or not horses are mischarted by a quarter or even half a length; its that this is a great way to see the pace shape of a race, especially when the raw times don't make sense or when races are run at "about" distances. I'll take my subjective pace judgements any day over the BRIS pace ratings.

[takeout]

Originally posted by GameTheory
Beaten lengths are actually an artificial construct.
Very interesting. I never thought of it that way before.

Valuist,
I have purchased text charts from Trackmaster in the past and notice that they do theirs like past-performance lines, i.e., lengths behind the leader instead of ahead of the next horse.

[Dave Schwartz]

Henry Kuck wrote about this like 12-15 years ago.

When one considers that the race does not "freeze" as the winner crosses the finish line it makes one consider the beaten lengths differently. What it actually says to me is that only horses that are battling to the wire have even remotely accurate lengths-to-times conversions.

That is why our paceline "adjustments" include some (optional) maximum lengths back adjustments.

Perserverance becomes an issue.


Regards,
Dave Schwartz

[Suff]

Originally posted by GameTheory
. They are also stretched out -- there is a chart to correspond distance on the film with real distance to convert to lengths. If the camera is running at the wrong speed, the lengths will be off as well.

Intersting stuff. This Chart....is it Standardized? Meaning a high enough % of charters use it or its standards? And if so...where can I see one? Or can you recall from memory the general jist of it?

[takeout]

Isn't Equibase supposedly working on some kind of new charting system that would make callers obsolete?

[GameTheory]

All my info is from reading and from asking questions. That's all I know.

Suff, you mean you didn't arrange a tour/demonstration of the photo finish system at Saratoga? Put it on the list for next year. I would suggest those of you who are on-track or have connections to go talk to the people that run the equipment and see if anything has changed recently and to get the nitty-gritty details. If you find anything out, report back here. The equipment may or not be standardized at this point. In past years I know it was not, and smaller tracks are less likely to upgrade quickly. This is the same basic system that they use for other kinds of races -- cycling, track & field, etc.


The new system, if ever implemented, would be GPS based and would give the exact position of every horse at any arbitrary point throughout the race. How they would package that data could be a problem as that is a massive amount of data for just a single race. Likely they would just make charts like now, except with perfectly accurate data.

[grjr]

If you had a GPS based system wouldn't you be able to download the info and plug it into a simulator and be able to "watch" the entire race? It could be not only the times at certain points but also the distance from the rail and such. That would be pretty neat.

jr

[GameTheory]

Originally posted by grjr
If you had a GPS based system wouldn't you be able to download the info and plug it into a simulator and be able to "watch" the entire race? It could be not only the times at certain points but also the distance from the rail and such. That would be pretty neat.

jr

Yes, that's right, but that is a large amount of data per race that someone has to store & serve up. i.e. more $$$ to do so -- we'll probably have to pay a premium for that, at least for a while.

[DJofSD]

Yes, the perfect chart would be based upon three units of measurements: time, distance from the rail and location around the track.

Look at it this way, what if you had a GPS unit on the nose of each and every horse. Make it a transponder or recording unit. You'd capture x, y, z and t - but who cares how much the head bobs up and down during the running of the race, so drop the z. Give me those two space co ordinates along with the corresponding time and you don't need to use beaten lengths. And even if you had only 1 second resolution, you'd still have a heck of a lot more data and more accurate data than you have now.

Just image the amount of data you'd have for a typical 10 race PP with a full field of horses!

DJofSD

[BillW]

Careful what you wish for guys. If all data was perfect it would be close to impossible to win at this game.

Bill

[takeout]

Originally posted by BillW
Careful what you wish for guys. If all data was perfect it would be close to impossible to win at this game.

That may be right but I would love to see it put to the test. Would probably have to include accurate drug and vet data in with that as well as a host of other stuff that we are not now privy to.

It occurs to me that if the good folks in MD haven't seen fit to give Secretariat his rightful Preakness time in 30 years, that we don't stand too much of a chance of getting data that's too accurate for day-to-day racing. :D

[GameTheory]

Originally posted by BillW
Careful what you wish for guys. If all data was perfect it would be close to impossible to win at this game.


I don't believe for a second that it is lack of accurate data that holds people back. However, certain types of handicappers will always be hurt when new data becomes easily available -- those who were working hard on their own to assemble it. So trip & pace handicappers who would most crave a GPS rendering of each race the most would be actually be the most hurt by it becoming available.

When modern speed figures appeared in the Form for everyone without them having to work for it -- who did it hurt? Speed handicappers.

But in the pari-mutuel system, the closing of one avenue means the opening of another.

[Valuist]

He's right. There's plenty of info available now. Trip and pace handicapping angles are still profitable but using GPS would only cut into those profits.

I remember the good old days, when decent home made speed figures were all one needed. Like them or not, both the Beyer and Equibase figures are far more accurate than the old speed rating and variants used by the Form.

[VetScratch]

Originally posted by takeout
Isn't Equibase supposedly working on some kind of new charting system that would make callers obsolete?
I had heard about this project at EquiBase and assumed the first visible output is represented by the "NEW" instant charts that are available for some tracks via BRIS SuperTote.

For example, these so-called instant charts are currently available for AP but not for BEL, so I figured AP was one of the early tracks in the EquiBase testing and rollout program. By instant, I mean SuperTote seems to offer fully descriptive and complete EquiBase charts within 20-40 minutes of each finish.

If this is the "new" EquiBase chart system, one has to wonder how the new charts will compare for accuracy with the old ones, whether the instant charts are subject to reviews/changes before they get archived as "official" charts, and to what degree the process has been truly automated.

As Game Theory suggested, PA members who are on-track regulars where the instant chart feature is up and running may be able to get details about what is really going on.

[VetScratch]

By new "instant" charts, I didn't mean the ultimate GPS system, but something must have changed since shortening publication deadlines is hardly a way to reduce labor intensity unless one is willing to sacrifice accuracy.

[sjk]

The Equibase charts have usually been posted by about 30 minutes after the race all along. I think the only thing which is new is the link to supertote.

[VetScratch]

Originally posted by sjk
The Equibase charts have usually been posted by about 30 minutes after the race all along. I think the only thing which is new is the link to supertote. Wow! I thought it was new because not all tracks show it yet on SuperTote. I can see how a good deal of the data and can be "generated" from pre-race information based on order-of-finish for the horses, but 30 minutes or less from finish to published distribution seems like a pretty tight deadline for composing the race commentary and putting together the intermediate calls. Are the intermediate calls for "instant" charts reviewed against video or simply based on notes from real-time observations?

[cj]

I agree the time is a little short, but what else do chart callers have to do between races?

[GameTheory]

The charts get changed all the time after they are posted. I always wait 24 hours or so to get them, and by then any changes in position/beaten lengths/comments (chart caller stuff) seems set in stone. It is possible those things are sometimes adjusted within an hour or two of posting, or after the card is over maybe a review is made.

But changes (mainly spelling corrections) in names may occur a day later, or even weeks later. Happens all the time. Very very rarely does any of the running line or timing info get changed. 80% of the corrections are to the names of the owners, generally not a big deal. The rest is for trainer & jockey names (usually missing or inverted initials), changes of sex of the horse (listed as colt originally when it should have said gelding, for instance), and also claim notices. If there is more than one claim slip put in for a horse it may take them a day or so to sort it out. DQ's based on drug tests may also appear weeks or months later.

There are no "unofficial" & "official" versions of the charts. They put them up, and then they're up. If they get changed later, then they're changed. But there is no formal review process by which they double-check everything. Corrections get made to the names because Equibase is alerted by the actual owners/trainers/jockeys (or someone else) of the errors. Now Equibase does of course have a list of standardized names in their db but sometimes they choose the wrong one initially (e.g. Jr instead of Sr), and there are also some bad ones in there. Many of the same errors are repeated over & over and then get corrected over & over.

But like I say, VERY rarely (maybe twice a year, all tracks) does chart caller data (position/beaten lengths/timing/comments stuff) get corrected more than 24 hours later. Ranchwest (I think) posted an example from Retama last year where a horse won by like 12 lengths and the chart had him winning by only 1 length or something -- that was one that was eventually corrected weeks or months later. (So if you see something like that, email Equibase and maybe they'll fix it.) But names of owners/trainers/jockeys (in that order of frequency) are corrected quite often days/weeks/months later, and Equibase relies on someone telling them about the problem.

By the way, all of this was told to me by an official at Equibase.

I assume many of the corrections actually come from HDW, because they go over everything with a fine-tooth comb. Equibase also told me that when changes are made, those updates are forwarded on to BRIS, HDW, DRF, etc. Some providers may actually directly use the Equibase database instead of making their own copy of it in their own format, so I'm not sure how that works...

[Valuist]

We had a great example today of how this can really work. Della Francesca, a shipper from Arlington, running in the Belmont BC on the turf.

Most handicappers may have looked at Della Francesca's last past performance line and concluded the horse ran near a slow pace and still couldn't get the job done. After all, the fractions said :25 :49 1/5 and 1:14 so it had to be slow, right? WRONG! Here's how the race charted out:

Honor in War 7.5 5.5 0
Better Talk Nw 9.0 7.0 nk
Mystery Giver 9.0 7.0 1.0
Dlla Frncesca 0.5 hd 1.25
Act of War 11.5 11.5 1.75
Shaanmer 7.0 6.0 6.25
Tijir 7.5 7.5 8.25
Candid Glen 5.5 4.0 8.50
Extra Check 0.0 0.0 12.50
Gretchns Str 6.0 4.5 16.50

The horse that Della Francesca dueled with, Extra Check, tired badly. 4 of the top 5 finishers came from fairly far back. In cases like these I make a check next to the horse who I feel was the best, and exerted the most energy, which in this case, was Della Francesca.

Also, James Quinn's graded stakes theory takes yet another hit as a one-time Grade 3 winner handles this supposedly Grade 2 field.