Track variant question...

[thaskalos]

In Aqueduct's 6th race tomorrow, the #7 horse (Gatto Nero) and the #8 horse (Amply Rewarded) both raced last out on February 5th...but in different races; the 1st and the 6th race respectively.

Both these races were run in exactly the same final time (1:13.6)...and both the #7 and the #8 finished in 3rd place -- beaten by 3.75 lengths.

And yet...the Beyer figures for their respective races are markedly different. The 7 has been given a Beyer of 52...while the 8's performance has been downgraded to a Beyer of 43...a difference of almost 4 lengths.

I don't calculate variants for Aqueduct's winter meet, so I am wondering; is this a case of the track getting much faster as the day went on...or are the Beyer figures off the mark on this one?

[ntheiroff]

Andy uses the "projection method" which aligns the horses performance to a previous performance. And yes he could also have split the variant.

My guess is the first option.

[gm10]

Andy uses the "projection method" which aligns the horses performance to a previous performance. And yes he could also have split the variant.

My guess is the first option.

Even if he uses a projection method, it should still be a DAILY track variant, an average of (projected - actual) for that day. It must either a split DTV, or he's adjusting the numbers on a per race basis so that there's consistency in them.

[cj]

Even if he uses a projection method, it should still be a DAILY track variant, an average of (projected - actual) for that day. It must either a split DTV, or he's adjusting the numbers on a per race basis so that there's consistency in them.
Most likely it is a split variant. He doesn't adjust races on a per race basis unless the time of a race is just totally out of line with all the others. This does happen often enough, especially at smaller tracks, but no so much at NYRA tracks. Trust published times at your own peril.

[PhantomOnTour]

I didn't split my variant on 5Feb at Aqu. They ran 8 sprints that day and 1 route. My final time sprint deviations from par were:

R1: s5 (23.7-47.9-113.6)
R2: s2
R3: s6
R4: f6
R5: f5
R6: s2 (23.0-47.8-113.7)
R7: f3
R8: s2

Yes, the first 3 sprints were a bit slower than the last 5 sprints but not a big enough difference for me to split my variant. The races in question (R1 and R6) both got basically the same pace and final time fig from me.

[BIG49010]

Bris has them one point apart, I didn't split my variant either. It will be interesting to see the results.

[sjk]

I've got the 7 four points faster than the 8 that day.

[Big Bill]

Bris has them one point apart, I didn't split my variant either. It will be interesting to see the results.

Which BRIS product provides a track variant?

Big Bill

[BIG49010]

Which BRIS product provides a track variant?

Big Bill

BRIS doesn't provide variant, I do my own variant, just commenting that it didn't appear BRIS made any adjustment.

[cj]

BRIS doesn't provide variant, I do my own variant, just commenting that it didn't appear BRIS made any adjustment.

BRIS never splits variants.

[mistergee]

would you trust a days variants when there is a fast 5 and 6 race and a slow 5 and 6 race anyway?

[The Hawk]

BRIS never splits variants.

A big reason why their figures aren't worth much.

[Mr Saratoga]

Beyer, and Bris figs are only a tool. Most people put way too much faith, and cash into them.

[Robert Goren]

If you ever try to figure these thing yourself, you will find they vary greatly from race to race. Some time you think you see a pattern head in one direction as you get deeper the card, but it still could be random. if you split a card, you should have reason other than just numbers tell you to do it like a drying track. Somebody wrote a whole book on being fooled by randomness. If going to make your own numbers you should probably read it. I am not not much of a believer in changing variants unless there is some reason to. That includes from day to day as well. If it is hot and dry every day all week, the variant for all the whole week should be the same. If you get a goofball number in the middle of that kind of week, it is in all likelihood an outlier. The day in question is a strange day. I would not put any faith in any number I got from that day. The nice thing about doing your own numbers is that you know which days have numbers you can trust and which days they are not worth the paper they written on.

[Elliott Sidewater]

good post Robert - I think you hit the high points. Personally, I would not have split the variant on the day in question unless I knew something more; about a change in weather, that the track was rolled or harrowed heavily between races, it started raining - something other than just running more races over the surface. Sometimes at Philadelphia Park in the past, when I was doing my own figures, I found that the first race of the day, particularly on weekdays, had suspect timings. It's as if their teletiming system hadn't been set or calibrated properly, and you'd see weird stuff like a maiden claimer running 21 flat for the first quarter mile. One time, strong wind gusts came up while a field was walking to post and that one race got a 9 tick adjustment from me at six and a half furlongs. Real world events held outdoors don't conform to anyone's preconceived idea of order. There is an element of randomness there that defies rational explanation some of the time, but if it's small enough, chaos won't matter that much. If you're good enough to make the right adjustments for the right reasons, there's money there. If you're doing this three times a week at one track, it's way too much. To my way of thinking, constructing a great set of figures is part art and part science.

[raybo]

Which BRIS product provides a track variant?

Big Bill

Bris PPs and MCP data files have the DRF track variant included.

[Big Bill]

Bris PPs and MCP data files have the DRF track variant included.

Ray,

I know that BRIS Pace and Speed figures include track variants, but I was quite sure that the variants are not shown separately in any of the BRIS products. That's the reason I asked the poster.

Big Bill

[cj]

Ray,

I know that BRIS Pace and Speed figures include track variants, but I was quite sure that the variants are not shown separately in any of the BRIS products. That's the reason I asked the poster.

Big Bill

These are two different things. The BRIS speed figures are not calculated using the DRF variant, but one devised by BRIS, at least that is what I think. It isn't included in any files, but you can figure it out if you take a few weeks of data files from around the country.

The DRF variant is the one printed next to the DRF Speed Rating in the DRF, at least I would assume as much. For example, you see 88-15, the 88 is the speed rating, the 15 the DRF variant. Again, this has nothing to do with the BRIS speed figures.

[cj]

Which BRIS product provides a track variant?

Big Bill

I would assume he meant the BRIS speed figures were one point apart, which would tell me the variant was the same for both races.

[raybo]

In the Brisnet PPs and MultiCaps data files, the DRF speed rating is located in fields 856 through 865, the DRF track variant is located in fields 866 through 875.

Yes, Brisnet's pace and speed figures include Brisnet's proprietary track variant, not the DRF's variant.

[cj]

would you trust a days variants when there is a fast 5 and 6 race and a slow 5 and 6 race anyway?

It depends. I'd probably trust it more than one that gave those races a 0 if it made no sense given the past history of the horses.

[Robert Fischer]

Andy uses the "projection method" which aligns the horses performance to a previous performance. And yes he could also have split the variant.

My guess is the first option.

bingo

[Robert Fischer]

It would be interesting to do a niche product like "TRIPLE CROWN 3YO DIVISION FIGS"

have


Beyer Methodology guy (projections)
Pars guy (pars based Kentucky Derby class of horse -or simply triple crown)
Comprehensive guy(trips,acute form swings, weather/trackcondition)
With a lot of time and work the above could come up with a quality product. The non-Beyer guys would have subjective stuff to do, and thus would need to do quality work.





Not really IMO a project worth it's time, but these type of hybrid methodology are an interesting niche product.

[gm10]

bingo

Regardless of whether you compare the times with par times, or compare the raw figures with projected figures, the DTV is still the average of those differences.

The likely explanation is that he split the performances in 2 groups and calculated two averages.

[Tread]

Anyone who thinks track variants should not change during the day should read/listen to the scientific evidence presented in the Horseplayer Expo presentation on the Thorograph website.

http://www.thorograph.com/archive/files/multimedia/Vegas2004/vegas.html

Unless you are really into flatly rejecting physics, weather, and common sense, it is completely ridiculous to argue that the moisture content of a track does not change during the day, especially when you consider the use of water trucks.

[cj]

Regardless of whether you compare the times with par times, or compare the raw figures with projected figures, the DTV is still the average of those differences.

The likely explanation is that he split the performances in 2 groups and calculated two averages.
Basically true, but I tend to weight those I have the most confidence projecting more heavily.

[gm10]

Basically true, but I tend to weight those I have the most confidence projecting more heavily.

It can be a weighted average as you say. It doesn't even have to be an average, you can also take the quickest race of the day. That isn't such a bad choice, especially in places where there a lot of falsely run races. It equates to saying ... "I don't know how fast the track was, but I know what how slow it wasn't".

[cj]

It can be a weighted average as you say. It doesn't even have to be an average, you can also take the quickest race of the day. That isn't such a bad choice, especially in places where there a lot of falsely run races. It equates to saying ... "I don't know how fast the track was, but I know what how slow it wasn't".

No argument with that, I do the same and ignore races where a projection is not really possible.

[Father Guy]

Only just joined the forum so apologies for revisiting an old topic but it’s an interesting one that will undoubtedly crop up again. Doesn’t the answer to this particular conundrum clearly lie in the in-race sectionals?

According to the scale I use the final time 6f going variant on Feb 5th was 136.5 – this is deduced from winning times after taking into account the run up, the abilities of the winning horses, weight for age and weight carried and then using the fastest 2 of those times compared to a universal 6f standard (not a class par). The going variant for all the 6f races, starting from race 1 was 155;141;145;141;138;155;135;142. The scale isn’t readily comparable with Beyer or Bris or Equibase but no matter as illustration only is the point of this post. Clearly the average for all the first 4 races is 145.5, the middle 4 races 145; and the final 4 races, which included the two fastest on the card, 143. So, if you use all-inclusive averages alone to assess variants, isn’t there then some suggestion in these figures that the track was speeding up through the day? Not at all. Not only is that method of assessing track variants flawed anyway because of the inclusion of skewing data, the in-race sectionals tell the truer story.

Calculating an in-race variant for the first three sectional timing points (after 2f, 4f, 5f) over the 6f course using the same method that provided the final figures above gives a fastest 2-race variant at each of those 3 points between 119 and 128; the discrepancy between the sectional variants and the final variant is explained by the faster speeds the horses are running within the race than at its end (early/mid race speed ratings produced are then adjusted proportionally so as to be on the same level as the final figures). What’s important in trying to nail the correct track variant here is the interpretation of the in-race variants and what the figures show is that the fastest race relatively of the 8 6f races after 2f, 4f and 5f was, on each occasion, race 2. And given that race 2 was a maiden claimer contested by exposed 4yo’s and older, that race by its own definition has to be the one, along with the overall fastest race, race 7, by which all other races on the card, and the track variant by association, must be referenced. There just seems to be no evidence at all, using all the times available and not just the winning ones, that the surface was any different by race 6 than it was for race 1.