So who should I believe this time?

[takeout]

I mentioned in another thread that I've never been able to fathom why BRIS continues to use a DRF file for its PP data. Is this latest data discrepancy that I've noticed due to the same data passing through too many hands?

Same horse: (Banished Lover)

TSN has it as a claim
DRF (paper) has it as a claim
BRIS DRF file has it as a "Previously trained by"

[andicap]

When and where was it's last race? I can check the chart in Simulcast Weekly where it lists claims if it was at a major track.

[takeout]

It was claimed (I think?) on 8/13/03 at Del.

[takeout]

4th race

[takeout]

Andicap,

Any luck, or is Delaware not in the Charts Weekly? Sent an email to BRIS but haven't heard anything. I'm assuming it's a BRIS mistake only because I happened to see a paper DRF that, like TSN, had it as a claim.

Does anyone out there that has ever worked in the business know how these types of mistakes happen? Is it just a question of a wrong keystroke or something? I have been dealing with BRIS and TSN now for quite a few years and I have seen many differences between the two.

[BillW]

Takeout,

DRFSW shows no claim for that Horse on 8/13 at Del

Bill

[takeout]

Bill W,

Wow! That's interesting. The plot thickens. Do they show it as a "Previously trained by" like the BRIS file did?

[BillW]

Same Owner/Trainer on that date and the race before (Jul30? from memory). I have never seen charts explicitly indicate "previously traind by".

Bill

[takeout]

BillW,

DRFSW - Is that the Formulator stuff? The paper DRF has "Previouly trained bys" in it.

I just got a Trackmaster text result chart and it has it as a claim,

From - Lael Stables, trainer Michael Matz
To - Tradewinds Stable, trainer John Zimmerman

(In the PPs now, TSN & BRIS, it just has Zimmerman as the owner and trainer.)

So, what's to make of all this? What name did you see as owner/trainer?

[BillW]

Originally posted by takeout
BillW,


DRFSW - Is that the Formulator stuff? The paper DRF has "Previouly trained bys" in it.

So, what's to make of all this? What name did you see as owner/trainer?
takeout,

DRFSW - DRF Simulcast Weekly Charts

The Owner/Trainer listed was Lael and Matz.

Bill

[takeout]

Originally posted by BillW
takeout,
DRFSW - DRF Simulcast Weekly Charts

Of course. Sorry about that. Don't know what I was thinking. Chalk me up with a "well duh" for that one.

"The Owner/Trainer listed was Lael and Matz."

That would've been right, but down at the bottom of the chart there was nothing saying that the horse had been claimed by Zimmerman?

That's VERY weird if there's not a claim in the Simulcast Weekly chart because the paper DRF's PPs show a claim. Now how do you suppose THAT could happen? :confused:

[Brian Flewwelling]

My database based on Equibase Results Charts shows the Horse to have been claimed by John Charles Zimmerman (for Tradewinds Stable).

I find it curious that Equibase, the source of all current race data, was not mentioned in this tread. And it is free.

Fleww

[RonTiller]

takeout,

I am in the business, so I'll explain a few things, but I cannot diagnose the source of any specific errors with BRIS/DRF data.

BRIS has had a contract with the DRF for a long time, before Equibase was a twinkle in anybody's eye. As many of you remember, the chart callers worked for the DRF and the DRF had various employees at the track collecting data on workouts, charts, entries for upcoming races, etc. The DRF had THE database. Having a contract with the DRF, BRIS got data feeds from the DRF.

Equibase started up around 1990 and the 1st year of data they collected was in 1991. They had their own employees at the tracks and their own database, seperate from the DRF database.
This didn't sit well with many handicappers, who in essence were presented with 2 different versions of the races, including in a few cases, different versions of who the trainer was. The Equibase data was distinct from the DRF (and by association, BRIS) data.

When HDW started in 1994, we were in talks with both Equibase (the new kids on the block) and the DRF (the turf authority!). Each one provided data files, with which we could create databases to collect, analyze, make value added data (like stats or speed ratings) and sell the resulting products. We signed a contract with Equibase for various reasons, partly because we thought they would eventually win the data wars, but also partly because with Equibase, we got a continuous replication of their database. With the DRF, we didn't. Database professionals out there will understand the importance of replication (where all the Inserts, Updates and Deletes are to all of the tables are sent near real time to us).

At around the same time, TSN got a contract with Equibase, so now BRIS was covered by a DRF contract and TSN was covered by an Equibase contract (bet on both sides and no mattter who wins, you're covered). I presume TSN processes the database replication files as we do.

Wrench in the machine: The DRF, in the late 1990s, decides to get out of the data collection business and buy their data from Equibase, just like HDW, Trackmaster and TSN. Here is where I don't have direct knowledge, because I don't work at BRIS. BRIS has a contract with the DRF for data but the DRF is no longer collecting data; they are a customer just like the rest of us. Presumably, BRIS still gets data feeds from the DRF, but it is data that ultimately originated from Equibase, channelled through the DRF. TSN maintains a DIRECT connection to Equibase data because they have a contract with Equibase. I have no idea

To add another wrench, it is possible to have a contract with Equibase and not process the database replication feeds; that is, one could just store the flat file chart files that are available to all VARs in some tables and go from there. This would be equivalent to creating past performance tables from parsed out charts from the web. At HDW we replicate - I can't speak for any other VAR in the business. But if any of them only accumulate data, not replicate it, this could result in some data discrepencies also.

So why might there be differences between what Trackmaster (owned by Equibase), HDW (Equibase contract) and TSN (Equibase contract) have and what the DRF and BRIS have? There are kind of 2 data channels at work here. It is POSSIBLE that an error at the DRF data processing end could get into the DRF-to-BRIS data channel, propogating the error to BRIS, whereas the Equibase-to-EveroneElse data channel had the correct data. If both the DRF and BRIS share an error that the direct-from-Equibase VARs don't, this seems the likely culprit.

Let me say quite frankly that in addition to trapping actual data errors (impossible fractions, incorrect trainers, blinkers off when the horse doesn't show blinkers in the last race, and such), we worry about introducing our own errors in the mix That's why I take any error report that I get seriously - like doctors, our 1st responsibility should be to at least do no harm. We use our own data, so just from a purely selfish perspective, we want it as correct as possible. Alas, perfect we aren't.

As I mentioned, I can only speak for HDW. I'm not finger pointing at or being an apologist for other value added resellers, but perhaps these various data channels can help explain some data inconsistencies between resellers.

PS: The HIQ (horse in question) shows it was claimed in our copy of the Equibase database and like the other Equibase VARs, our reports and data files show a claim.

Ron Tiller
HDW

[takeout]

Ron Tiller,

Thank you for the very interesting post. I had always suspected as much and that had been my theory for some time now (2 databases in play with different info in each). I wrongly assumed years ago that when DRF became a customer of Equibase's that they would somehow match the two databases up back to the point where they both started collecting the data. This just seemed to be a logical step to me but I was wrong bigtime and mistakes in the DRF database can be found that go back well over ten years now for tracks that they pulled out on before and during the feud with EQ (like CT). All of these mistakes still reside in the BRIS archives and will surely never be changed.

[takeout]

A friend that uses the DRF pdfs has informed me that they do have the claim in them. This would appear to make it a BRIS mistake but doesn't explain why the DRF Simulcast Weekly doesn't show a claim in its result chart. ???

[RonTiller]

Maybe it would help to expand a little on what I know for sure and clarify what I don't know.

First, I don't know what happened to all the DRF data collected by the DRF prior to their getting out of the data collection business. Since Equibase has no Equibase collected past performance data prior to 1991 and the DRF has been running a long long time, the data they collected must be somewhere. After all, my copy of Seabiscuit has the complete DRF PPs for Seabiscuit in the back of the book; they're stored somewhere!

Second, I don't know how the DRF has handled the messy task of integrating the data they collected prior to 2000 with the Equibase collected and disseminated data from 2000 onwards. Like many other VARs, they have a replicated database of Equibase data back to 1991, which overlaps for almost 10 years data they ALREADY have themselves collected. I'm sure much thought and sweat (and probably a few tears) went into whatever decisions they ultimately made. Database design and data formatting issues could make any integration very messy indeed. Maybe somebody could corner Steve Crist at the Expo and ask him whose PP data is published for a 1997 PP: Equibase or the old DRF data. I suspect Equibase.

Third, I do know that there is one official source of data now: Equibase. Putting aside old PP lines for 9 year old geldings (and semantic quibbling over what 'official' officially means), races for 2 or 3 or 4 year olds have Equibase derived PP information; there is no equivalent DRF derived data with which it competes. Why would there be a difference in data among VARs in THESE races?

a. A few of our customers have pointed out that occasionally the DRF will have a workout published that we don't. I have researched each of these both in our copy of the database and in the data feed files that Equibase sends all the VARs. In each case I investigated, the DRF mystery workout was not in the Equibase data in our possession. Now, where did they come from? I thought we were using the same data. Are these occasional mystery workouts valid workouts that we don't have or are they invalid workouts that get erroneously added by the DRF? I just don't know. Regardless, the total quantity of these workouts is probably small compared to the total quantity of misreported workouts and falsely reported workouts (yes, at at least some tracks, it is possible to report a workout that didn't really happen in a time that didn't exist - the honor system and all).

b. VARs can round data differently. For times, HDW always uses hundredths, the number of significant digits that Equibase collects and stores times in (quarter horses are in thousandths). Thus, with a few exceptions, we disseminate exactly what Equibase has. The tradition of using fifths of a second requires either rounding or truncating the times. Truncating means 23.59 seconds gets disseminated as 23.4, the next lowest fifth interval. Rounding means 23.59 seconds gets disseminated as 23.6, the nearest fifth interval. I believe the DRF has historically truncated their times (I may be wrong here). Additionally, beaten lengths can vary slightly depending on how VARs might choose to round. 6.2 beaten lengths (say the horse came in 3rd, 6 lengths behind the horse in front of her, who was a neck behind the winner) could be truncated to 6 or rounded to 6 1/4.

c. VARs can modify data for error correction purposes. I can only speak for myself, but we do a check on all times and all intervals in all races and if a time is obviously wrong or an interval is impossible (e.g. the final furlong is run in 8.2 seconds), we blank it out. There is nothing good that comes from disseminating such an obviously bad time. Likewise, we will occasionally modify the jockey or trainer that Equibase sends when it is clear beyond all doubt that it is wrong. Note, this does not involve us changing Patrick Valenzuela to Mike Smith; it normally involves changing a newly created trainer to a well established trainer with an almost identical name. None of these modifications are more than occasional, but they can result in data differences.

d. VARs can make programming/database errors. While I personally know nothing of this (cough cough), people make mistakes. You do everything you can to validate data and double check results but sometimes errors happen. When they do, you fix them and double check anything else that might have similar problems and come clean with customers. Example: we make Ed Bain's Layoffs and Claims reports and a particularly perspicuous customer found a few errors in some of the statistics over the course of a year. He was right, the problem was fixed (ok, I created another problem fixing one of them) and customers were notified on the download site. Yes, I got a call from an unhappy customer who then (reasonably) questioned all the statistics, since I fessed up to this error. Nobody enjoys taking calls like that. The claimed/traner change issue that started this thread probably falls under this.

e. VARs can conglomerate data differently for statistics. An example from this very bbs is off track statistics. We use the Equibase definition but one could divide on/off track differently if one so desired. Another example is Distance statistics. Some may count 7.5f with 8, or 8 40yds with 8, and others may not.
Another example could be the counting of unofficial starts in a horse's performance statistics. When they are not counted, you get horses at Les Bois as 1st time starters with 12 UNofficial races in their PPs. We don't count UNofficial races in any stats but the experienced 1st time starter case still bothers me.

Fourth, I do know that Equibase error trapping has improved considerably from 5 or 6 years ago. Just keeping the trainer and jockey names straight there is a full time job. One of the advantages of a replicated database is that when they make changes to the jockey and trainer keys (perhaps they discover that a single trainer accidently got assigned 2 trainer keys, so they merge the 2 into a single trainer), the change is propogated to all the PPs in the replicated tables. It isn't reflected anywhere just accumulating PPs from charts. This can actually be another source of data divergence, if a VAR is only accumulating PPs and not replicating (I don't know what everybody does).

Fifth, I do know that we and others have worked with Equibase to identify errors (many identified by our customers), some more subtle than others and that once they are fixed by Equibase, they get propogated to any VAR doing replication. What happens at BRIS, with an indirect data link, I don't know. I don't presume that they, or anybody else in this business, just don't care. A big part of this is talking to the right person.

Well, I guess I've said enough.

Ron Tiller
HDW

[Tom]

They have problems with data. Check out the current discussion on the Derby List - somehting a simple as a horse's lifetime stats at a distance are apparently rocket science to DRF. As usual, the party line response down-played the errors. Intersting post about that and speed figures in general. A usually dull forum is running hot this week.

[takeout]

RonTiller,

Thanks for the insights. I'm going to print that one out and save it.

[Brian Flewwelling]

RonTiller

Thank you for two great posts.

I will add that i spotted a problem with two jockies and the same Name ... the Campbells at AP have the same initials. When i emailed Equibase they were all over it immediately. They negotiated with one of the Jocks to use a 'different' name.

So my data is useless for these lads prior to that date, but they were prompt.

Ron: I take it from you comment that all the pplines in your data for the races ridden by these jockies would be updated?

Fleww

[Figman]

Not quite sure if the DRF has these two jocks riding in New York State separated as well....R.Rojas at NYRA and R.Rojas at Finger Lakes are not the same jock. It's Raul at NYRA and Ruben at FL.

[RonTiller]

Brian,

You are correct about the jockeys. Here's the way it works. Equibase has at least one person (maybe more now or maybe it is a shared responsibility) who works to resolve name problems, for jockeys and trainers. Why are there name problems to begin with? Well, in the olden days, the tracks sent basically whatever name they wanted. Bill Jones trains a horse in race 6 yesterday; today, B. Jones trains the horse, tomorrow, W. H. Jones trains a horse. The trouble is, they are all the same guy.

Standardization of names was a real problem. There was in essence no official standardized name. Part of the problem was that some of the trainers themselves had no standardized name -that is, they had trainer licenses in multiple states but the names on the licenses were not identical from state to state. Same with the jockeys. This of course can cause problems if part of the jockey colony from Thistledown winters in Tampa and they don't have identical names on their licenses. It is not always easy, sitting in an office on Corporate Drive in Lexington, to determine if George Espinosa in Ohio and G. Espinoza in Florida are the same person. It involves quite a bit of detective work in some cases.

I know from personal experience that the people at Equibase have put in a lot of work to develop a system of standardization
and data entry at the track level that results in lots fewer name ambiguities, but some still slip through. My experience in the past has been that if I, as a VAR with Equibase, send them names I think are suspect, they promptly research them and let me know the results (sometimes I was right, sometimes I was wrong). I am happy to hear they deal with end customers in the same manner. This "negotiated with one of the Jocks to use a 'different' name" is great. As somebody who has spent hundreds of hours in the past fighting name ambiguities (and who hasn't?), it warms my heart to hear of things like this (yes, I need a life!)

To continue the story, once the corrected jockey or trainer is ammended for all applicable PP lines (this could go back13 years), all companies who replicate the data get update transactions to their copies of the database which keeps them in sync and which ensures their copies of all the tables have all the relevant updates. So yes, the next night when we run stats, the new jockey or trainer keys will be in place to yield stats based on the corrected data. This is why it is important to use a replicated database to calculate these stats rather than an accumulated data table; the accumulated data table does not reflect all the changes, merges and deletions of the jockeys and trainers. Owners and owner statistics are a completely different mess for a different set of reasons, which I can discuss if anybody is interested.

Incidentally, for spot play junkies, many years ago there used to be a C. Black and a Corey Black in the PP tables, roughly equally divided. One showed a healthy positive ROI and the other was soundly negative. We liked to speculate that somebody "in the know" sent in whichever version of his name matched his chances in the race, a signal to those who knew the code; in reality, pure random variation was the culprit. But I still have a warm place in my heart for conspiracy theories (you mean big oil companies, under the control of the masonic lodge, directed by the trilateral commission, DIDN'T hire the mafia to collude with the KGB to have Seabiscuit's entry mate run poorly in the Santa Anita Derby, ensuring a win by Seabiscuit (whose owner also hid a 200 mpg carburetor in Seabiscuit's stable)? I bet Dave Schwartz knows more about this than he lets on!).

Ron Tiller
HDW

[takeout]

Originally posted by RonTiller

Owners and owner statistics are a completely different mess for a different set of reasons, which I can discuss if anybody is interested.
Please do! Some of this stuff is finally starting to make some sense to me now.

One thing that I like that TSN *sometimes* does with owners is that they add the person's name after the farm/stable name.

TSN - Own: Windward Farm (E. H. Hackman)

DRF & BRIS would only have Windward Farm.

OTOH, TSN used to have last year's trainer record over top of this years. Something I found helpful. Then they pulled it. I realize that one man's treasure is another's trash, but it sometimes seems like resellers have a lot of info that they are holding back.

[Brian Flewwelling]

RonTiller
Yes, please continue. You are mostly confirming my guesses, but that is also useful. I assume they keep Jocks and Trainers info in a lookup table and access it with a key-code??


Originally posted by takeout

OTOH, TSN used to have last year's trainer record over top of this years. Something I found helpful. Then they pulled it. I realize that one man's treasure is another's trash, but it sometimes seems like resellers have a lot of info that they are holding back.

Maybe they dropped that info when the year was well along, and most trainers had sufficient data of this year.??? In which case it should return in Jan.

Fleww

[BillW]

Originally posted by takeout

OTOH, TSN used to have last year's trainer record over top of this years. Something I found helpful. Then they pulled it. I realize that one man's treasure is another's trash, but it sometimes seems like resellers have a lot of info that they are holding back.

Are you refering to the .pdf Past Performances? The .PCS data files contain both this year and last year (and also for the meet) trainer/jock records. Maybe I'm misunderstanding what you mean by "over top"?

Bill

[takeout]

BillW,

Yes, the pdf fifty-centers.

That's interesting because I don't see last year's trainer stats in the paper DRF or in the BRIS DRF $1 data file.

I'm "dimly recalling" and I hope I don't have my story wrong. I think it was somewhere in the early '90s when I noticed it and I *think* it was TSN.

I recall seeing what Brian Flewwelling mentioned about dropping last year's trainer info after the year was well along. I think that is what they are doing now. But this was different. They would leave the prior year's stats up there all year. It was helpful to me with shippers. I think they only did it for a couple of years.

[BillW]

takeout,

I have "Quick Play" .pdf PP's from TSN ... they only have this year/meet stats for both jock and trainer. but the data files (PCS) have both. I should mention that the info was just recently added to the .pcs files, maybe in the last few months.

Bill

[VetScratch]

When I extract from BRIS PP downloads, I ignore current and previous year statistics for jockeys/trainers because of the calendar year switch-overs. Here are the statistical categories that seem most meaningful IMHO:
Category Jockey Trainer
Last 365 Days Y Y
Last 90 Days Y Y
Last 14 Days Y Y
Current Meet To Date Y Y
Turf Last Two Years Y
Trainer/Jockey Combination (combined)
Jockey With ESP Running Style Y
Trainer Key Condition #1 Y
Trainer Key Condition #2 Y
Trainer Key Condition #3 Y
Trainer Key Condition #4 Y
Trainer Key Condition #5 Y
Trainer Key Condition #6 YThe Trainer Key Conditions (which are specific to today's horse and race conditions) are really helpful for anyone looking for promising trainer angles that may imply intentions.

[RonTiller]

OK, here we go.

First, I'd like to point out that I can't get into specific stuff here regarding the Equibase database or technical or contractual issues between us and them. Also, I am not a spokesperson for Equibase, BRIS or anybody else other than HDW.

Now, with regard to owners, as I said, this is a different mess. There is no real standardization of owner names and there probably never will be. There's just too many and too many TYPES of owners: corporations, individuals and partnerships (maybe there's another type I forgot). There is no good infrastructure in place to make sure Bill and Ted don't get reported as Ted and Bill at another track. I believe Equibase has made some progress on this front but a sure way to get a groan out of anybody is to bring up the subject of owner names (groan! - see, I told you).

Even more important, from the standpoint of compiling statistics on owners, is the issue of what constitutes the 'same' owner. Consider these 6 horses and their owners:

A: Bill
B: Bill and Ted (2 buddies pool their $ to buy a horse)
C: Ted and Alice and Frank and Susan (two married couples)
D: Doc Stables (Ted's tax write-off corporation)
E: Doc Stables and Fred (Fred is sharing a tax write-off)
F: Bill's wife Jill (she's on the paperwork for what else, tax reasons)

None of these are unusual examples in any way.

Do A and B have the same 'owner'? How about B and D?

If the answer is yes to both, then A and D have the same owner but completely different people actually 'own' the horses.

This complex network of inbred multiple partnerships, individual ownerships and corporate/organizational ownerships makes doing large scale owner statistics virtually impossible. I have seen more than one intrepid handicapper attempt this and they were all humbled by it.

This is not to say that on a smaller scale, somebody couldn't be tracking an owner/trainer combination - say Bob and Beverly Lewis with D. Wayne Lucas. I have spoken to individuals who do exactly this sort of thing and some claim to have huge success with obscure owners without having to get into the messy issues discussed.

On a large scale, breaking outside the easy cases like these or specialized knowledge cases of the true ownerphile (you know who you are), there is simply no single CORRECT answer and no non-ridiculously-labor-intensive answer even if there were. One can of course maintain that the CORRECT answer is each distinct case is a seperate owner with its own set of statistics - Bill is different than Bill and Ted for instance. This yields pretty much mush. I just won't do it.

We have had numerous requests for owner stats over the years and it just won't happen, for the above reasons. My hat is off to people who specialize in this really narrow band of handicapping statistics. You won't likely see any of the data companies encroaching on your turf.


I realize that one man's treasure is another's trash, but it sometimes seems like resellers have a lot of info that they are holding back.

We have so much information at HDW it staggers the mind. I mean it. Are we holding back? I guess, insofar as we don't disseminate every possible piece of data we calculate on every hard drive in the office. But 'holding back' has too much of a conspiritorial air about it. The free market determines most of these things. If enough people complained to TSN about such things, it might get added. But understand that every company and every software developer is innundated with suggestions and every one of them has to make decisions, based on their own vision (if they have one), practicalities, know-how and customer demand. (OK, no amount of complaining will induce me to put up half *ssed owner stats.)

I could produce a report that is 100 pages per race, for every race - the Mega Ultimate Thoroughbred Extravaganza (TM) - MUTE for short. The first reaction is: its too long (but we're not holding anything back). The second: if you just put this, this and this, sorted like this, I'd buy it (MUTE Lite?). But nobody else would, because they want other things sorted differently. Sally wants trainer stats going back 6 years exacly that are track specific; Dirk wants trainer stats for the last 30 days, combined; Beth wants 2nd after layoff stats but layoffs are defined as 30 days for clm 4000 and lower, 52 days for stakes horses (except grade 1 shipping from Europe who won one of last 2 races in last 69 days, unless trainer is one of top 5 at meet, then expand to 86 days); George thinks trainer stats are worthless (at least the way we do them) - the report needs instead decelerational uni-velocitous entropic thor-energy ratings.

I am being a bit facetious here but I am not exagerating much. And I am not mocking anybody; handicapping is a very very personal matter. That's why generic reports, no matter how many varieties you sell, won't make everybody happy (or maybe ANYBODY). Everybody needs to know different things to handicap a race. That's why so many handicapping programs and reports are available, why so many people jump from one program to another in search of a good fit, why so many people ultimately program their own stuff as the last resort for getting exactly what they want and need.

Ron Tiller
HDW

[takeout]

Originally posted by BillW
I should mention that the info was just recently added to the .pcs files, maybe in the last few months.

That's good news. Hopefully they will end up in the pdfs one of these days so that people like me (with no computer skills) can use them. :)

[takeout]

Ron Tiller,

Owners: I figured they would be a messed up deal (from my previous experience with all of the different and wrong trainer names) but I didn't realize they would have as many wrinkles to them as they do. Thank you for the examples and explanations.

Whenever I notice a discrepancy in the data from BRIS and/or DRF, as opposed to everyone else, the vast majority of the time it's BRIS and/or DRF that has it wrong. I always thought that it was because the data was being messed with too many times since coming from its original Equibase source. Too many cooks in the kitchen, so to speak. I don't know anything about databases and had never heard of the replication part of it. That sounds like it could indeed be the culprit or at least a big part of it.

I just stumbled across another example of the type of thing that I notice from time to time although this example was from a race that took place back on June 29th.

Saw a horse in the TSN PPs that won but was DQed and placed 8th (last). The comment said "unprepared start". I wanted to know more so I looked it up in the DRF .pdf result charts that I print out and save. (I only play one track - maybe that's why I notice so much of this stuff.) There was nothing in the chart about the horse being DQed. Nothing. A text result chart from Trackmaster (info straight from Equibase) showed the DQ.

[takeout]

I also notice that BRIS and TSN have different comments for this horse in their PPs. They both say "unprepared start," but underneath of that,

BRIS has- 06-28-03 Disqualified from purse money
TSN has - (Placed 8th through disqualification)

[RonTiller]

takeout,

I seem to write really long winded responses, so I'll be briefer.

We do get comment lines from Equibase but the DQ part is not part of the comment. How a vendor chooses to annotate a DQ is completely up to them. Same with a claim or trainer change or oddball things like 'ran for purse money only'.

Not only that, it is possible for a horse to be disqualified after the fact, due to something like a failed drug test (I'm not sure the official designation here is 'disqualified', but the result is the same). The horse may show a 1st place ORIGINAL finish, WPS mutuel prices and even an OFFICIAL finish of 1st place at the time of the race. Several days or weeks later, a failed drug test may DQ the horse to 9th for OFFICIAL finish but the horse still has an ORIGINAL finish of 1st and WPS mutuel prices attached to it for that race. Again, there is no 'correct' way to display this in a PP.

I guess I should clarify a point. VARs (Value Added Resellers) have contracts with Equibase to acquire data from Equibase, analyze and manipulate it, use it to produce NEW information (speed ratings, stats, power numbers, etc.), make products (whether that be pdf reports, comma delimited files or binary files for handicapping programs) and most importantly, sell those products to customers. We have a license to resell Equibase data and any data based on or derived from Equibase data to customers and there are of course licensing fees that all VARs have to pay to Equibase. I cannot elaborate further, except to say the licensing fees are substantial.

What TSN or the DRF, Thorograph or HDW do with this data is largely up to all of us. There is no 'Handbook for the precise way to make and format an official PP'. The fact of the matter is, the DRF has been in the business forever and the DRF style PP has become a de facto industry standard. Trackmaster produces a DRF-style PP report; so do TSN, BRIS, ITS and others. This was always a conscious decision to create a DRF-style PP report. People like it; people expect it; people want it. The DRF adds something to improve their product. Some of the other VARs have been known to immediately modify THEIR OWN DRF-syle report to reflect the same DRF changes. I am not criticizing this. People want it.

Thorograph has an Equinbase contract, and look what they do. No DRF-style PPs last I looked. HDW makes PPs for our RSPos products, but I had the luxury of desgning a product we ourselves wanted. If a customer wants the weight carried listed on their PPs, they should NOT buy our product, because we left it off. A generic PP (not associated with a specific methodology like Thorograph or RsPos) normally has weight, because people want it. I guess we can be accused of holding back data, namely weight! I have had customers call who say they would buy our PPs if we added such and such, removed this and that, changed things here and there, and what they have in the end described is the DRF! I'd buy Thorograph sheets if they got rid of all the nonsense with plotting speed ratings and added running lines, comments, pace numbers, jockeys and trainer...oops, thats the DRF again.

I lost my train of thought, but I think you have the picture. BRIS and DRF data are joined at the hip; Equibase and TSN, HDW, ITS and Thorograph (Trackmaster too) are joined at the hip. We all have our own ways of doing things, formatting data, designing reports. We all make mistakes. We don't all have customer service departments in Italy or India though.

Rats, another really long winded response.

Ron Tiller
HDW

[takeout]

A few thoughts on parts of Ron Tiller's very interesting posts:

"First, I don't know what happened to all the DRF data collected by the DRF prior to their getting out of the data collection business. Since Equibase has no Equibase collected past performance data prior to 1991 and the DRF has been running a long long time, the data they collected must be somewhere. After all, my copy of Seabiscuit has the complete DRF PPs for Seabiscuit in the back of the book; they're stored somewhere!"

And that, I suppose, is what's keeping them in business. Equibase should've bought the old DRF database but either couldn't afford to, didn't have the vision to, or DRF wouldn't sell. Don't know if it would've been any bigger mess since '91 than we have now but it most certainly couldn't have been any worse and at the very least the BRIS archives wouldn't be full of the untold errors that have gone into them since '91.

"Maybe somebody could corner Steve Crist at the Expo and ask him whose PP data is published for a 1997 PP: Equibase or the old DRF data. I suspect Equibase."

Me too, except for maybe the trainer names. I think a lot of them were still wrong in '97. I would be very interested in hearing his answer.

The workout thing must be a real mess. I guess they're just phoning them in from all of these little training tracks now and we are just taking their word for it. I have seen the phantom works at DRF that you mentioned and have wondered about them also. Sometimes when I'm at the track I will have to get a work from the track program that is not listed in my PPs. Other times they will announce a work because it is not in the track program and I will have it in my PPs that I got well in advance of 48 hours before post. I have also occasionally seen what are almost certainly wrong workouts for horses in the PPs. (Horse that never left the east coast showing workouts on the west coast, and like that.) So much for "value-added". Indeed, most of the discrepancies that I come across are in the value-added stuff. I guess that would make sense though due to the increased handling of the data and probably the lack of replication for that kind of stuff.

"VARs can conglomerate data differently for statistics."

I wish there were some standardization in this area between the resellers. Fat chance that will happen. It would just show up more discrepancies and point out who was the tardiest in updating their stats. The different data conglomerations are definitely something to keep in mind if switching resellers.

Errors: That EQ-to-DRF-to-BRIS corridor seems to be a treacherous piece of highway. Some data that travels it is never seen again. :D