Churchill-Ellis Park

[iamt]

I do agree that the main cause here is just a cluster of bad luck, but we are at the extreme end that at this point.

If you accept CDI have no control over what is going on (ie can't make it safer), if they remain at Churchill they are left to hope for a statitical correction. The upside of that is actually quite low, peope will continue to talk about the breakdowns early in the meet, while the downside of even a 'normal' number of accidents over the rest is much larger.

[o_crunk]

FYI

Equibase has updated their RDS system to close the CD meet on 6/4 and they've put a place holder for the track_id ELP for a meet that begins 6/10 and ends 7/2. I guess we will see Monday whether is this going to be what they do - transmit cards as ELP for the races at ELP under the CD meet umbrella.

[classhandicapper]

Originally Posted by Jeff P

Andy,

Yes.

There IS a reason for renaming/branding a meet.

And there ARE a lot of moving parts.

If a track wants to brand their meet and make it stick out by giving it a catch phrase name:

GO for it!

But keep in mind the data is one of those moving parts too.

Imo, when a track asks Equibase to brand an upcoming meet a certain way:

They should also ask Equibase to not break the data.

And Equibase should strongly (and quietly behind the scenes) suggest to track management that they be allowed to get the data right if that doesn't happen.

Imo, that means:

The rebranded track name can and should be front and center when Entries are drawn, on the front page of the DRF, on Track Programs, on TV, and on Track Video.

But the data itself (data file names, result file names, running lines, workouts, and compiled stats) should bear/be based on a permanent track code where the races are run.

The only valid reason to change an existing track code would be if a track is torn down, the track surface(s) are removed, or moved to a slightly different location, and then rebuilt (such as Belterra which was essentially built on the same site where River Downs used to be.)

Horse Racing is a data driven business.

The first order of business:

Get the data right.


-jp
.

+1.

Paper and pencil handicappers and apparently the industry itself didn’t have a grasp on why this would be an issue for database/spreadsheet handicappers, but given the complaints after “BAQ” you would think they’d understand the inconvenience by now. It’s not the end of the world, but why should I have to spend time changing data in a bunch of tables that should be correct from the start.

[MJC922]

Originally Posted by o_crunk

FYI

Equibase has updated their RDS system to close the CD meet on 6/4 and they've put a place holder for the track_id ELP for a meet that begins 6/10 and ends 7/2. I guess we will see Monday whether is this going to be what they do - transmit cards as ELP for the races at ELP under the CD meet umbrella.

Thanks much for the info, I'm always grateful to have people like you and CJ who communicate these things. I'm supposedly now on an email list for notifications of the sort, this after missing the first three months of charts from that fairmount to fanduel debacle.

'Outer' course bit me a couple of years back too. For way too long of a time I was caught calculating these as dirt races until a customer was kind enough to report the races had been on turf. Big mistake. Yes I know I should've been a lot more in touch than I was but really do we need an outer anything? Unless there are three dirt courses or three turf courses then no, not IMO. Two courses anywhere just tag one as the inner and let the other be the main. Oh well, chalk it up to the stuff that happens over the years.

[classhandicapper]

Originally Posted by o_crunk

You would never alter the source of the data for storing it as-is. For whatever the need is, and the need may change, you store it as it was transmitted from the "source of truth", in this case Equibase and transform it as needed. In this case, some clients wanted BAQ, some wanted AQU. Some wanted BAQ display but AQU historical stats. You get it. Meanwhile, the races actually took place at AQU.

It is non-issue for me. Still doesn't make it right. Right is better than wrong. BAQ was wrong on almost every data level from the single source of truth.

You and Jeff are obviously correct.

The only reason I can think of why a user might want to store it as BAQ is because some trainer or other patterns are seasonal. If it was stored as AQU you might miss things that are normally more likely to occur seasonally at Belmont or BAQ. Other than something like that, I don’t get it.

[MJC922]

Originally Posted by classhandicapper

You and Jeff are obviously correct.

The only reason I can think of why a user might want to store it as BAQ is because some trainer or other patterns are seasonal. If it was stored as AQU you might miss things that are normally more likely to occur at Belmont or BAQ. Other than something like that, I don’t get it.

Exactly, as usual you understand the nuances, that's why I haven't bothered changing it (for the trainer stuff). The downside to that is of course for the first season there's nothing stats-wise, unless you change it for just the first year and then roll out of it later. That's an option, I just haven't bothered since nobody is really subscribing to my stuff for trainer-related info anyway. It's more about the class ratings and form cycles. I must say Twinspires has some good trainer stats available on their site for each race. For now I prefer Amwager for my tote monitoring though.

[o_crunk]

Originally Posted by MJC922

Thanks much for the info, I'm always grateful to have people like you and CJ who communicate these things. I'm supposedly now on an email list for notifications of the sort, this after missing the first three months of charts from that fairmount to fanduel debacle.

'Outer' course bit me a couple of years back too. For way too long of a time I was caught calculating these as dirt races until a customer was kind enough to report the races had been on turf. Big mistake. Yes I know I should've been a lot more in touch than I was but really do we need an outer anything? Unless there are three dirt courses or three turf courses then no, not IMO. Two courses anywhere just tag one as the inner and let the other be the main. Oh well, chalk it up to the stuff that happens over the years.

I have been developing off the RDS system for over 10 years now. It's a totally fine system and I'd guess the initial developers of it some 25+ years ago had both the domain knowledge and deep knowledge of how to build a robust relational database system.

The changes made to it since then all seem to be driven by those with specific domain knowledge about racing and absolutely no respect for how relational database systems should work. It is very frustrating since developing off of these systems is not cheap - both from the perspective of the cost of acquiring the data and building the racing knowledge to do so. You might be a black belt SQL slicer and dicer but if you have no racing knowledge, you are hopeless.

I would point out the FP/FAN debacle occurred right in the middle of the meet. The specific geographical location for the races themselves did not change of course. The O course type is one of the few structural errors the initial developers made - surface vs course type. There was no way to know in 1991 that there would be other "main track" type "surfaces" that were not dirt at that time. The decisions made since then have all been Frankenstein'd onto the structure itself and never actually corrected at the schema level structure. Again, very frustrating that something this important is not driven by someone at Equibase who has both deep knowledge of relational systems *and* racing knowledge. It's always seemed very blase to me to make changes like this without considering the broader ramifications for the relational systems that power all the horse racing data products in the US.

[Tom]

Not sure if there were any more incidents, but scanning the charts, it appears all the horses finished the races. Keep a good thought.

With the move to Ellis, maybe Belmont will benifit with entries in the slew of good stakes races coming up. Be nice to see some full fields of top horses.

Now, will Ellis take the spot on America's Day at the Races?

[MJC922]

Originally Posted by o_crunk

I have been developing off the RDS system for over 10 years now. It's a totally fine system and I'd guess the initial developers of it some 25+ years ago had both the domain knowledge and deep knowledge of how to build a robust relational database system.

The changes made to it since then all seem to be driven by those with specific domain knowledge about racing and absolutely no respect for how relational database systems should work. It is very frustrating since developing off of these systems is not cheap - both from the perspective of the cost of acquiring the data and building the racing knowledge to do so. You might be a black belt SQL slicer and dicer but if you have no racing knowledge, you are hopeless.

I would point out the FP/FAN debacle occurred right in the middle of the meet. The specific geographical location for the races themselves did not change of course. The O course type is one of the few structural errors the initial developers made - surface vs course type. There was no way to know in 1991 that there would be other "main track" type "surfaces" that were not dirt at that time. The decisions made since then have all been Frankenstein'd onto the structure itself and never actually corrected at the schema level structure. Again, very frustrating that something this important is not driven by someone at Equibase who has both deep knowledge of relational systems *and* racing knowledge. It's always seemed very blase to me to make changes like this without considering the broader ramifications for the relational systems that power all the horse racing data products in the US.

It's wonderful you have a clue about the pain-points because their aren't many of 'us' 3rd party folks around. I developed off of this because the charts are my primary source data:
https://info.trackmaster.com/thoroug...bonlycomma.txt

As you can see there's still no 'outer' noted here and how long has it been? Ellis (Starr) there's one change you can crack the whip on @ TM


SURFACE

D D-Dirt
E D-Equitrack
F D-Dirt training
N D-Inner track
B T-Timber
C T-Downhill turf course
G T-Turf training
I T-Inner turf
J T-Jump race
M T-Hurdle
S T-Steeplechase
T T-Turf
U T-Hunt on turf
V T-Hunt Flat (Weatherbys only)

[PaceAdvantage]

The reactions here to CD compared to the reactions to LRL when they shut down racing for a few days back in April after some breakdowns are quite remarkable.

From what I gather, people were kinda losing their shit about the LRL situation...it was another overreaction and another "OMG, this doesn't look good at all for Maryland racing with the Preakness coming up..."

People were speculating back and forth what could be causing this...blah blah blah

But here, in this thread, we have people criticizing CD for overreacting to a STATISTICAL ANOMOLY (which is exactly what I said the LRL situation was...I believe I was criticized for being another uncaring asshole who wants to see horses die...or something like that).

Very weird.

[Maximillion]

Originally Posted by PaceAdvantage

The reactions here to CD compared to the reactions to LRL when they shut down racing for a few days back in April after some breakdowns are quite remarkable.

From what I gather, people were kinda losing their shit about the LRL situation...it was another overreaction and another "OMG, this doesn't look good at all for Maryland racing with the Preakness coming up..."

People were speculating back and forth what could be causing this...blah blah blah

But here, in this thread, we have people criticizing CD for overreacting to a STATISTICAL ANOMOLY (which is exactly what I said the LRL situation was...I believe I was criticized for being another uncaring asshole who wants to see horses die...or something like that).

Very weird.

It's probably an anomoly either way but the first thing to look at in both cases should be the horses themselves rather than the track imo.

[Augenj]

Originally Posted by Jeff P

I'm hoping Churchill track management decides not to mess up the data by having Equibase use track code CAE (Churchill At Ellis.)

Because the races will be run at Ellis:

My preference is that all data (data file names, result file names, running lines, workouts, and compiled stats) bear/be based on track code ELP.


-jp
.

Thank you for this.

[Augenj]

The elephant in the room is this unpopular subject of synthetic tracks.

https://www.courier-journal.com/stor...ks/1579341001/

[Thomas Roulston]

Note the factual error in this article:

https://defector.com/horse-racing-to...orm-blows-over

Prior to the policy changes, trainers had financial incentive to get horses into races regardless of their fitness, and they received at least some payout regardless where their horse finished. Under the new regime, only horses who finish in the top five places receive any payout.

Now, go to equibase and click on "Full Charts," then click on "Churchill Downs."

I would speculate that the chump who wrote this piece is not a horse racing fan.

[Augenj]

Originally Posted by Thomas Roulston

Note the factual error in this article:

https://defector.com/horse-racing-to...orm-blows-over

Prior to the policy changes, trainers had financial incentive to get horses into races regardless of their fitness, and they received at least some payout regardless where their horse finished. Under the new regime, only horses who finish in the top five places receive any payout.

Now, go to equibase and click on "Full Charts," then click on "Churchill Downs."

I would speculate that the chump who wrote this piece is not a horse racing fan.

I guess this makes the statistics wrong also?