Track Master DTV Quality Control
 

How can the same day, same track, same distance, and basically same race type have 2 different daily track variants ?

These two horses ran last on August 9th. They are running against each other tomorrow in race 6 at Parx. The race is a $25.000 Starter Allowance for non winners of 2 life who have run in a race for a claiming price of $25,000 or less.

What goes on here ?
Looking back to August 9th The track was fast I have never seen this before. especially on a fast track at the same distance.

Miss Molly won race 1 The DTV was -9
Joyful Thing won race 8 The DTV was -1

Attachment 38270

Attachment 38271

closer look
 

There were 4 routes run at Parx on August 9th

Race 1, 4, 6 and 8 all were at 1 mile


Attachment 38273



It appears that the DTV for this day was -1

Here are 2 horses who won or ran in race 4 and race 6

The winner of race 4 was Tale of Peace. She has not raced since .
However Small Secret has. She next raced on 8/23
Her PP's from that race say she ran on a -1 DTV on 8/9/14

Attachment 38274

In race 6 the winner was Calypso Girl She next raced on 8/31
She too gets a -1 DTV for 8/9/14

Attachment 38272

billv just a suggestion "don't use DTV" in the configure tab. just the SR.
i m 'caping PID tonite i'll try it.

http://www.trackmaster.com/track/info/tchdesc.htm

I searched for "Daily Track variant" on TrackMaster.com, and in the description for their "Results Charts" Track Variant is:

TRACK VARIANTS - A highly accurate measurement of the track surface for each race.

So, it says they are giving the Variant for each race. In your first example:
Miss Molly won race 1 The DTV was -9
Joyful Thing won race 8 The DTV was -1

Race 8 was at the end of the day so the variant shows the track was getting slower as the day progressed.

What do you think Bill?

true this has been discussed and written about in several H'capping books.
i once asked "how can there be a track bias when the track is "raked" after each race?" well steve davidowitz answered but i forget what he said:o

TRACK VARIANTS - A highly accurate measurement of the track surface for each race.

RACE Variant
 

So It appears what we see as a heading on RDSS should be RV for RACE Variant

Here is another horse from race 1 It's Purple Rain

We see a - 9 RACE Variant

Attachment 38280

But for race 2 and 3 4 6 and 8 we see a -1

Attachment 38278

Attachment 38279

Attachment 38277

Nice work detective, that is a big swing between two races, something isn’t right there.

It is not unusual for the bias to change over the course of a day, Woodbine gets faster as the day moves on. You would have to monitor what type of track maintenance is going on, whether its gallop mastered, cultivated, harrowing, or roto-tilled, and how many times throughout the card. I know Woodbine generally gallop masters the poly track every two races.

Bill,

Handicapper's Data Warehouse has the same variant for both horses from my Aodds 98 Super Gold.Both horses have a '29' variant...



Ray

Bill, and everyone,

I just spent tome time talking with Craig Walker from TrackMaster, refreshing my understanding of how TrackMaster DTVs (or RVs!) are calculated.

DTVs for different surfaces are always different (Turf DTVs will always be different from Main Track).

DTV for a given surface will be the same for all races, regardless of distance, EXCEPT:

1. when a certain number of the horses in the same race all run final time Speed Ratings (SR) which are significantly faster than their expected SR. Maybe 1 or 2 horses running so much faster might happen, but several of all horses in a race doing so most likely indicates an event impacting track speed for that race (weather, e.g. grooming). Thus, the normal DTV is modified for that race to bring the SR in line with what those horses would be expected to run, based on a calculation involving their SR over the past 6 months of their histories.

2. When the Official Track Condition changes from previous Condition, the DTV for that race will likely also change - e.g. from Fast to Muddy.

All these DTV calculations are done algorithmically, not by human intervention, for consistency and predictability.

So, For PRX August 9, the Dirt Track DTVs were:
Race 1 = -9
Races 2 - 8 = -1

The Weather was listed as Clear and the Track as Fast for all 8 races.

Thus from point #1 above, since the DTV of Race 1 was calculated as 8 points FASTER than all the rest of the day's races (-9 versus -1) it means that more than a few of the horses in that race ran faster than their past 6 months' SR indicated they should probably run. Thus their final time SR should be smoothed out to their likely normal ability by recording that some event related to that race (the first one on the day) enabled the horses to run faster than their norm. So the DTV was increased FASTER than the DTV shown for all the rest of the races that day. Horses in the rest of the races on the Dirt that day all ran consistently and the DTV was shown as a consistent -1.

If you want to cross-check a day's set of DTVs from the TrackMaster Result Charts for any other card (say, 1 or 2 cards?), just let me know. If this does not relate specifically to Older Sartin programs, perhaps we could put that in a different Forum.

FWIW, I have been spending a lot of my programming time on RDSS integrating TrackMaster XML Result Charts with the RDSS database. Shortly after each race (~20 minutes) the Chart is available to download and integrate with all the other info for that race, so the DTVs are visible as the day progresses. You can correlate the DTV with what you are seeing race by race by way of weather and track maintenance (and thus have confidence in those numbers). (You can also use the Results to maintain realtime Track profiles of 'what's winning', as well as maintain Decision Models and update Betting records - but that's a different discussion).

I hope that helps!

Ted

Interesting, thanks for that Ted.

How a track gets groomed before the racing day starts can have an impact on the bias. For example, after the morning workouts at Woodbine, the track is generally Gallop Mastered, which means they are raking down 1½” in depth. Once a week they Cultivate the track after the morning workouts, which means they are raking down 4” in depth, or they might Roto-till the track which is 5” down in depth. The deeper they go the softer the track becomes, thus having a bias of slow or fast depending on the surface and track. As the surface gets raced on it gets compacted back down, and the type of grooming in between races will change the surface’ consistency. Was the track sealed the day before? If so, what was done to the track this morning? Then you factor in the sun, cloud cover, temperature, humidity, precipitation before/during/after, and you have another set of variables.

What lacks in the industry is transparency, how many tracks are forthcoming with their track maintenance programs?

factor analysis
 

Thanks Ted and Hoof

This is from the factor analysis manual

Attachment 38283

August 9th was a Saturday

Parx runs live racing Saturday Sunday Monday and Tuesdays
unless there is a Monday holiday, In that case they may also run on Weds.
They may card a race on Friday if its a big race weekend like the triple crown races or the breeders cup ,

When they came back to race on Saturday August 9th, they had not raced
since Tuesday.

I want to look closer at that first race on the 9th I find it odd that a MC 10 for 3+ F
( bottom of the class ladder ) would produce so many fast times

does not seem to fit
 

Hi Ted

I have been looking this over. The explanation you got from track master
may be good in general but I hope they take a closer look and see that there is something odd here.

First of all This race is very aberrant but just for 1 horse.
Look how Miss Molly blew away the field

Attachment 38286

Pars

Based on my Parx database par for Maiden $10,000 Females 3 or 4 and older
is 24.6 49.3 115.2 and 142.6

Miss Molly alone exceeded par by running on a big lead and earning a
92 Final fraction ( phase 1 ) and a 138.7 final time .

No other horse in the race ran to par.

Here are the Phase 1 readouts for the previous 15 winners
of MC 10000 for older females only

Look how aberrant the readout are compared to the others.
Miss molly ( listed last ) ran its LP and TPR and 3rd fraction energy
Very different than the other 14 winners

Attachment 38287



Lastly who else in the race ran faster than expected ?
Using Valm the par POR for these 15 races would be a 63 speed rating

Besides Miss Molly, Who else exceeds par ? only the place horse is close with a 59

When the PP's come out for each of the other horses from this 8/9/14
race this is how Valm will rate them I used a flat 16 DTV since they all ran on the same track anyway

My theory is the race variant was set too fast only because 1 horse blew away the field

I could very well be wrong but I don't see any reason to rate this race 8 points faster than every other race that day.

Attachment 38289

Miss Molly "Line 2"
 

Miss Molly ran today 9/23/14 in race 5
Looking back at my previous post
Valm calculated the final time of the 8/9/14 race to be 82

Track Master has given the race a 85

Attachment 38299


Miss Molly ran early but faded in the first race follow her maiden win
but as this chart shows that race was a way to high class turf experiment
( because of the wild 3rd fraction) they tried her on the turf

Today Miss Molly returned and ran in a CL 10 Non winner of 2 life
I have tracked where the previous 15 MC 10 3 + F winners next ran follow their maiden win. In the other 15 races for 3 + F maidens This class level is seems the correct place.


Attachment 38298

The problem is the 8/9/14 race is so aberrant. This further supports my
feeling that the -9 is not true

In today's race 9/23 race 5 Miss Molly has found her proper level
but look how she ran Very Early with a Early % of over 70
Her Phase 1 numbers for today's race were EP 80 LP a very week 68.5
That is 23.5 points off the "92" she earned on a easy uncontested early lead on 8/9/14.

Miss Molly is a early energy runner but that August 9th race final fraction
will always mess up your Sartin Program readouts

Attachment 38300

Par
 

Yesterday 9/22 race 9 was the next 8.0 3 + F MC10 race run since the
8/9/14 race

The winner Sugeily ran just about to par

Attachment 38301

Attachment 38302

Does not hold up
 

Hello Ted

I have bold faced something that does not sit well

First off If you have the Monmouth 8/22 card If you can checkout race 3
(I don't have the card) But I have the horse that finished 1st 3rd and 4th from the Parx 8/9/14 race
If you can let me know what speed rating horse 5 Sweetpuddin who placed) earned in the
8/9/14 race ? It does not look to good since Swweetpudding ran in the back of the pack at MTH too

According to what Trackmaster told you about their faster than in 6 months rule ( who thought this was a good idea ?) :rolleyes:
This does not appear to be a valid reason to give this race a fast adjustment

We see Miss Molly did earn a WILD 85 SR

I am curious to see how much faster Sweetpuddng ran in the 8/9/14 race
If she did not then nobody else in the race ran faster than their normal
speed rating ability

Here are the other horses in the race besides Sweetpuddin
These two horses who have run
back both get very low speed ratings of 12 and 14
First View who finished 5th and Stormy Norma who finished 6th have not raced since the 8/9/14 race


Attachment 38303

Attachment 38304


Attachment 38305

hello Bill v, please take some time and read this about TV. its long but interesting.
http://www.chef-de-race.com/pfs/variants.htm

Skimmed
 

Thanks Chef

I skimmed over the link. I will read it fully in a few minutes
In the mean time from what I skimmed I still feel, and will believe until I see
otherwise what speed rating was given to Sweeputtin (mth 8/20 race 3).

I am at this point convinced that some sloppy recording by trackmaster is still the issue.
I have tried my best to show that Miss Molly's race is not from a faster than the rest of the days races by 8 points track variant Its simply a case of a horse running a aberrant wicked fast 3rd fraction after getting a easy early lead

The Parx 3 year best, and track record for 1 mile was set by Peasant. 134.77

Attachment 38306

Peasant is ordinary bread and butter claimer.
Here are the RDSS readouts we saw before his record setting performance
on Jan 4 2104. The race was a 16,000 claimer for non winners of a race in 6 months.
The Trackmaster variant for that day was -29 or 29 fast ( frozen hard )? but unlike the trackmaster's alleged sloppy work on august 9th, every race on Jan 4th
earns the same -29 variant. All the winners times are faster than par.
Peasant earned a 104 speed rating. Its Phase 1 readouts
were EPR 95 LPR 88 = 183 He ran 7 early on the early late and 68.35 percent medium.

Rverge, that was a good read!
Thanks for the post!!
Mike

The methods of timing and chart calling are flawed. Whether the gate is placed properly or the timers are working correctly you just never know. When you add to this situation variant production that is computer generated, you have to expect these kinds of anomalies. Len Ragozin used to have his own timers at tracks around the country, particularly turf courses and time from the opening of the gate. So it is certainly acknowledged that horse racing is technology deficient and error prone.
Being forewarned is to be forearmed. You have to look at each horse past performances and note the lines that just don't belong. Horses can make major jumps in speed but it normally associated with maturation, change in surface or reaching fitness after a long layoff.
When you see a speed figure that is haywire don't use it. Bradshaw always spoke of preponderance of evidence. Make sure that the circumstances exist that could possibly support a big improvement in speed. Examine and evaluate each running line from gate to wire and convince yourself that the race was legitimate. The tracks have to incentive to insure fidelity in there reported times and charts. The bigger the prices the better in their view.

Bill, I'm travelling right now and will check that Result Chart and the other figures when I get home, on Monday.

Ted

Bruno Dejulio published and article today with a couple anecdotes of the Santa Anita dirt in the 90's. He talked about how the sun would shine only on the backstretch and not on the main stretch making the two sides of track almost completely different.

He mentioned an example in the mid 90's where the 4th race was a turf race, meaning there was an hour between dirt races (3rd and 5th races). The weather conditions apparently made the track very brittle and fast. The 5th race, 3 YO 25K claimers went 6f in 1:09:00- which is pretty fast, especially at that level. By the time races returned to the dirt again in the 7th race, the track had been watered and slowed down.

Equibase may not have perfect DTV figures, but they're likely doing an OK job. The fact that a track can change that much from hour to hour is ridiculous. It has to be extremely difficult to get an accurate track influence from day to day.

The best way to combat this IMO, is to beware of the flaws in the system. You can trust the adjustment and the speed figures, but look at the holistic picture.

I really like Mark's insight 2 posts back talking about taking a holistic look at a horse. If the last effort is a lifetime high speed #, make sure it's legitimate.

Where I get lost, is how does one determine whether the last effort was truly legitimate? I imagine quite a few short priced horses, with lifetime best speed #'s in the last race, lose everyday.