PDA

View Full Version : Brisnet running styles


raybo
10-02-2014, 12:39 PM
As a previous long time loyal Brisnet data file user, I have been trying to figure out exactly how Brisnet calculates their running styles for years now. Why? Because they work better with my method than those sources that only use horses' "good" races.

But, some of their assignments are quite puzzling. For instance, how does a horse with only 3 races, and the following 1st and 2nd call data, get assigned an "E" running style?

Start position, 1st call position, 2nd call position, 1st call B/Ls, 2nd call B/Ls.

Most recent race: 3, 3, 2, 1.00, 1.00
2nd race back: 3, 3, 4, 1.50, 3.50
3rd race back: 5, 4, 4, 2.50, 9.50

This same scenario could also be assigned an "E/P" and maybe even a "P" designation.

This is just one example of how confounding their criteria are for the various running style assignments they make (E, E/P, P, S, N/A).

So, if any of you have tried programming or otherwise calculating their running style assignments, without actually getting it from their data, and have "figured it out", please PM me. (I've read every article available regarding their criteria, but none of it works when programmed or calculated manually, because the criteria is not complete.

I normally would not post something like this, but since Brisnet does not offer an unlimited data downloads plan, even to their long time loyal customers thereby forcing them to seek other options, then I feel justified in making this post.

PA, feel free to delete this if you think it is inappropriate. I'm just tired of trying to figure it out myself. :bang:

DeltaLover
10-02-2014, 01:27 PM
I think the topic of this thread is interesting and a good subject to analyse

For starters, in this spreadsheet (http://themindofagambler.com/comparing_running_styles.xlsx) I present a comparison of Quirin Speed Points vs Bris Style.

The data is coming from a universe of 48,218 starters, accumulating the occurrences of each possible pair of QSP and BRS which (of course) appear to be correlated to some degree.

DJofSD
10-02-2014, 01:32 PM
I've always assumed those BRIS running styles were automatically assigned by some data analysis routine.

raybo
10-02-2014, 02:16 PM
I think the topic of this thread is interesting and a good subject to analyse

For starters, in this spreadsheet (http://themindofagambler.com/comparing_running_styles.xlsx) I present a comparison of Quirin Speed Points vs Bris Style.

The data is coming from a universe of 48,218 starters, accumulating the occurrences of each possible pair of QSP and BRS which (of course) appear to be correlated to some degree.

Thanks DL! From what I have read, Brisnet Quirin points use Quirin's original calculations and criteria. It is the running styles that I am having problems with, not the Q points. Their running styles documentation is quite general, and vague, in nature and "obviously" does not explain exactly all the criteria they use for running style assignments. Statements like "E horses like to be on the lead" are meaningless.

raybo
10-02-2014, 02:18 PM
I've always assumed those BRIS running styles were automatically assigned by some data analysis routine.

I am 99.9% sure that a computer program assigns their running styles. That is what I am trying to do also, but have been unsuccessful thus far.

I have written so many "what if" formulas for this stuff over the years that it would make your head swim. There must be a way to reverse engineer them, but I just don't know how to do that without using trial and error, and that has not worked.

GameTheory
10-02-2014, 03:16 PM
As a previous long time loyal Brisnet data file user, I have been trying to figure out exactly how Brisnet calculates their running styles for years now. Why? Because they work better with my method than those sources that only use horses' "good" races.

But, some of their assignments are quite puzzling. For instance, how does a horse with only 3 races, and the following 1st and 2nd call data, get assigned an "E" running style?

Start position, 1st call position, 2nd call position, 1st call B/Ls, 2nd call B/Ls.

Most recent race: 3, 3, 2, 1.00, 1.00
2nd race back: 3, 3, 4, 1.50, 3.50
3rd race back: 5, 4, 4, 2.50, 9.50Is this a real example? I think you still need to include the finish position/lengths to make any judgements...

raybo
10-02-2014, 03:23 PM
Is this a real example? I think you still need to include the finish position/lengths to make any judgements...

Yes, that was a real example. According to Brisnet, they don't consider finish positions or finish B/Ls, only 1st and 2nd call positions and B/Ls.

The method I use needs to know how a horse "normally" runs, not only how it runs when it has a "good" race ("good" normally means 1st, 2nd, or 3rd and B/Ls of 2 or less, and I never assume any horse is going to run his "good" race). When horses have only 1 or 2 of those races (or none), and they were in the distant past, they do me no good.

Capper Al
10-02-2014, 03:34 PM
I'm surprised to hear that BRIS running styles might differ from Quirin. Including Giles' PC style in the mix is an improvement IMHO.

DJofSD
10-02-2014, 03:39 PM
It's been a long time since I looked at the original book for QSP. Heck, I'm not even sure where I have put that book. But, as I recall, the original formulation was at the 2nd call and had a cut off for so many lengths behind the leader.

For the BRIS running style, one thing I never got around to doing was seeing if the letter assigned to a horse was static, i.e. once it was given, was it always the same, or, did it change from race to race, or, from year to year?

GameTheory
10-02-2014, 03:42 PM
Yes, that was a real example. According to Brisnet, they don't consider finish positions or finish B/Ls, only 1st and 2nd call positions and B/Ls.

The method I use needs to know how a horse "normally" runs, not only how it runs when it has a "good" race ("good" normally means 1st, 2nd, or 3rd and B/Ls of 2 or less, and I never assume any horse is going to run his "good" race). When horses have only 1 or 2 of those races (or none), and they were in the distant past, they do me no good.
If they really don't consider the finish (or filter by it in any way), then I would say the logic of making that an "E" would be something like, "When it DID contend for the lead (bl <= 1.0), it did it at the first call, therefore E. It did not contend for the lead in any other race and those races were not considered." If it had not had that most recent race, probably would have gotten an NA (unknown). It is very likely the rules are different for horses with only a few races (and/or possibly maidens generally) and so you'll see apparent inconsistencies or what seems like guessing (because it is) compared to those horses with more of a record.

Still, I don't see how they can discount the finish totally, or at least use it as a filter. It would be possible to reverse engineer if you had enough data examples...

raybo
10-02-2014, 04:49 PM
I'm surprised to hear that BRIS running styles might differ from Quirin. Including Giles' PC style in the mix is an improvement IMHO.

If I remember, Quirin only used E, P, and S styles, while Brisnet uses E, E/P, P,and S, so obviously they are not the same. Brisnet also uses both the 1st and 2nd calls for their running styles, while I think Quirin only uses one or the other but not both, for his running style calculations.

Both Brisnet and Quirin use only the 1st call for their early speed point calculations.

raybo
10-02-2014, 04:59 PM
If they really don't consider the finish (or filter by it in any way), then I would say the logic of making that an "E" would be something like, "When it DID contend for the lead (bl <= 1.0), it did it at the first call, therefore E. It did not contend for the lead in any other race and those races were not considered." If it had not had that most recent race, probably would have gotten an NA (unknown). It is very likely the rules are different for horses with only a few races (and/or possibly maidens generally) and so you'll see apparent inconsistencies or what seems like guessing (because it is) compared to those horses with more of a record.

Still, I don't see how they can discount the finish totally, or at least use it as a filter. It would be possible to reverse engineer if you had enough data examples...

Yeah, I thought of that, that they use a slightly different method for lightly raced horses. And, I agree that the only way an "E" could have been assigned to that horse was because it was a length behind at the 1st call in that one race. I can live with that, as E horses seem to be the easiest style to assign and the anomalies are much fewer than the other styles. For example, many of the E/P horses could qualify as an E horse (many of them having races in which they were 1st at both the 1st and 2nd call, but for some reason Brisnet assigns them an "E/P" instead. This leads me to assume they are using all of a horse's races, or at least the last 10 or so, and looking for the most frequently occurring styles for those races from which to assign their current running style.

Robert Goren
10-02-2014, 05:36 PM
Yeah, I thought of that, that they use a slightly different method for lightly raced horses. And, I agree that the only way an "E" could have been assigned to that horse was because it was a length behind at the 1st call in that one race. I can live with that, as E horses seem to be the easiest style to assign and the anomalies are much fewer than the other styles. For example, many of the E/P horses could qualify as an E horse (many of them having races in which they were 1st at both the 1st and 2nd call, but for some reason Brisnet assigns them an "E/P" instead. This leads me to assume they are using all of a horse's races, or at least the last 10 or so, and looking for the most frequently occurring styles for those races from which to assign their current running style.I believe I read some place that BRIS never changes a running style after they assign one. They assign the rating after the first race in the horse shows something. A horse that runs dead last all the way around the track in its first start does not get one. A horse in its first start that pops out on top will get an "E" even if it fades to last. The "E/P" looks like it goes to horses that are 2nd or 3rd in the that early race and fade. The "P" goes to a horse who runs like an "E/P" early but closes. The "S" goes to a horse that close in that rated race. Like I said before, I don't think BRIS changes the rating ever(at least I have seen it if they do), the horse do and sometimes BRIS is wrong, but they are right a surprising amount of the time.
Giles's method is by far the best method I have seen for determining the way the horse wants to run. The problem is that a horse is not always fast enough to run the way it wants to. Most of the time jockey has no say in what way the horse will run and most horses want the lead, but some will settle for second or third if they have to, but only if they have to. Some won't and will battle tooth and nail for the lead if they are not quite a bit slower than the other horse. That something most handicappers have trouble figuring out.

raybo
10-02-2014, 06:24 PM
I believe I read some place that BRIS never changes a running style after they assign one. They assign the rating after the first race in the horse shows something. A horse that runs dead last all the way around the track in its first start does not get one. A horse in its first start that pops out on top will get an "E" even if it fades to last. The "E/P" looks like it goes to horses that are 2nd or 3rd in the that early race and fade. The "P" goes to a horse who runs like an "E/P" early but closes. The "S" goes to a horse that close in that rated race. Like I said before, I don't think BRIS changes the rating ever(at least I have seen it if they do), the horse do and sometimes BRIS is wrong, but they are right a surprising amount of the time.
Giles's method is by far the best method I have seen for determining the way the horse wants to run. The problem is that a horse is not always fast enough to run the way it wants to. Most of the time jockey has no say in what way the horse will run and most horses want the lead, but some will settle for second or third if they have to, but only if they have to. Some won't and will battle tooth and nail for the lead if they are not quite a bit slower than the other horse. That something most handicappers have trouble figuring out.

So, are you saying that, without having a database of every horse's races, one will never duplicate their Brisnet running style assignments? That would severely hamper my efforts, to say the least.

Maybe if I started with it's most distant race in the last 10, and worked forward. I have been doing it the opposite way, for obvious reasons. Many horses' styles DO change over time.

By the way, if every horse wants the lead, then you'd have a heck of a scramble for it in every race, which is not the case. Almost any horse alive can run 22 or less for a quarter mile. The jockey does have some say in how they run, he does have hands and feet after all. That being said, some horses want to lead, others want to run with the leader, others want to be buried nice and safe in the middle of the pack, and some just don't care for the pack at all.

Capper Al
10-02-2014, 06:49 PM
If I remember, Quirin only used E, P, and S styles, while Brisnet uses E, E/P, P,and S, so obviously they are not the same. Brisnet also uses both the 1st and 2nd calls for their running styles, while I think Quirin only uses one or the other but not both, for his running style calculations.

Both Brisnet and Quirin use only the 1st call for their early speed point calculations.

Too much to remember. I'll find my old Quirin book and look it up when I get some time.

GameTheory
10-02-2014, 06:52 PM
I believe I read some place that BRIS never changes a running style after they assign one. They assign the rating after the first race in the horse shows something.That's tough to believe. They assign a running style after a few races on a young horse and THAT'S IT forever? (Even after 10, 20, 30, 100 more races and years have passed?) They can't possibly be that stupid, can they?

Tom
10-02-2014, 09:07 PM
Find two sets of PPs for the same horse, this year and last or longer - see if that is true.

raybo
10-02-2014, 10:35 PM
Someone with a "real" database that uses Bris data could do that easily. Unfortunately, I don't have a "real" database, and I haven't used Bris data for a couple of years. That's the problem, since I switched data providers my running style/pace pressure win elimination method doesn't work nearly as well as it did with Brisnet running styles. So, I've been trying to program them into my program, without success.

GameTheory
10-02-2014, 11:02 PM
Someone with a "real" database that uses Bris data could do that easily. Unfortunately, I don't have a "real" database, and I haven't used Bris data for a couple of years.
So let's find some horses that are still running where you have BRIS RS designations from a couple years back (to see if they change).

If you have enough old files, I could make an attempt at figuring out how they designate them...

DeltaLover
10-02-2014, 11:21 PM
Someone with a "real" database that uses Bris data could do that easily. Unfortunately, I don't have a "real" database, and I haven't used Bris data for a couple of years. That's the problem, since I switched data providers my running style/pace pressure win elimination method doesn't work nearly as well as it did with Brisnet running styles. So, I've been trying to program them into my program, without success.

I will do it tomorow and post the results

Speed Figure
10-02-2014, 11:22 PM
In the Belmont Tonalist was an EP5, In the Jim Dandy he was an EP8, In the Travers he was an EP7, and In the Jockey Club Gold Cup was an EP7.

raybo
10-02-2014, 11:22 PM
So let's find some horses that are still running where you have BRIS RS designations from a couple years back (to see if they change).

If you have enough old files, I could make an attempt at figuring out how they designate them...

How many matching cards and results files would you need?

Speed Figure
10-02-2014, 11:28 PM
Wicked Strong last 6 running styles.

Wood Memorial P2
Kentucky Derby P0
Belmont P0
Jim Dandy P1
Travers P3
Jockey Club Gold Cup EP5

GameTheory
10-02-2014, 11:29 PM
How many matching cards and results files would you need?
Don't see why I'd even need results, but I've got them going back a long time if I do...so just "lots"...

GameTheory
10-02-2014, 11:30 PM
Wicked Strong last 6 running styles.

Wood Memorial P2
Kentucky Derby P0
Belmont P0
Jim Dandy P1
Travers P3
Jockey Club Gold Cup EP5
EP -> P, maybe it is mainly about the speed points? (which would taken from last 10)

raybo
10-02-2014, 11:31 PM
In the Belmont Tonalist was an EP5, In the Jim Dandy he was an EP8, In the Travers he was an EP7, and In the Jockey Club Gold Cup was an EP7.

Yeah, like GT, I find it hard to believe that Brisnet would assign running styles and Quirin early speed points in a horse's first 2 or 3 races, lifetime, and never update them after that. I'm pretty sure they use software, exclusively, to calculate them, and they have the lifetime races for every horse that has run races in the US for years and years, so the likelihood that they would not adjust them, on an ongoing basis, seems extremely remote to me.

raybo
10-02-2014, 11:33 PM
EP -> P, maybe it is mainly about the speed points? (which would taken from last 10)

I think the early speed points are just a representation of how hard the horse fights for the lead, so to go from a P to an E/P would indeed represent that his more recent races have shown that he is fighting harder for the lead, or to be closer to the lead.

raybo
10-02-2014, 11:38 PM
Don't see why I'd even need results, but I've got them going back a long time if I do...so just "lots"...

Let me see if I can find a bunch of my old Brisnet files. I don't remember if I put them on a thumbdrive, or my external hard drive, or if I just deleted them. I have about 88 Brisnet .mcp and .drf card files on my PC that I have been using for this research, is that enough? Only have matching results files for about 15 of those.

GameTheory
10-02-2014, 11:43 PM
Let me see if I can find a bunch of my old Brisnet files. I don't remember if I put them on a thumbdrive, or my external hard drive, or if I just deleted them. I have about 88 Brisnet .mcp and .drf card files on my PC that I have been using for this research, is that enough? Only have matching results files for about 15 of those.
Not what I would consider lots. I have some sitting around, let me see what I can do with them...

raybo
10-02-2014, 11:45 PM
Not what I would consider lots. I have some sitting around, let me see what I can do with them...

Appreciate it!

raybo
10-03-2014, 12:30 AM
This is what makes "sense" to me, regarding preferred running styles:

An "E" performance = 1st at either or both the 1st call and/or 2nd call, or 1st, 2nd, or 3rd at either the 1st or 2nd call AND less than or equal to 1 BL at that call.

An "E/P" performance = 1st, 2nd, or 3rd, at either the 1st or 2nd call, AND less than or equal to 3 BLs at that call. (already we have some overlap between E and E/P assignments).

A "P" performance = 4th through 7th position at the 1st or 2nd call OR more than 3 BLs but less than or equal to 7 BLs at either of those calls.

An "S" performance = Worse than 7th position OR more than 7 BLs at the 1st and 2nd call

A "NA" performance = less than or equal to 1 race lifetime AND was not 1st, 2nd, or 3rd at the 1st or 2nd call if he had a race, or less than 3 races lifetime AND never showed anything in any of those races (ran at the back of the pack for the whole race, in all of those races).

But, I have programmed those criteria (and many others) and they don't match the Brisnet running styles nearly enough of the time. There is also the problem with coming to a final running style based on separate running styles in multiple races. I was using the last 10 races for each horse (or as many races as they actually had, less than 10) and using the running style that he was assigned the most often. If there was a tie between 2 running styles I used the slower running style, so a tie between E and E/P would give me an E/P, a tie between E/P and P would give me a P, a tie between E/P and S would give me a P, P and S would give me an S, an E and a P would give me an E/P, an E and an S would give me P, etc..

Robert Goren
10-03-2014, 07:20 AM
I want to make myself clear. The BRIS running style never changes, the number changes all the time. Once an "E", always an "E". I can not find where I read that. I am still sure I read it someplace.

TexasDolly
10-03-2014, 09:41 AM
I want to make myself clear. The BRIS running style never changes, the number changes all the time. Once an "E", always an "E". I can not find where I read that. I am still sure I read it someplace.

They do change. In the first at TDN today witt's rule is running as an "ep", last out on 17 sept. witt's ran as an "s" . Before the race the horse was designated as an "s" and the
performance that day graded the horse as an "s" . Before witt's race on the 31st of aug the horse was designated as "s" and his performance was rated as an "ep".
The determination of the current race not clear to me.
TD

PhantomOnTour
10-03-2014, 09:46 AM
I completely ignore the BRIS running styles and the speed points assigned to them.
Figuring out who will be where in the early stages is crucial, and something I like to determine myself....what about jock switches and barn changes and cutbacks/stretchouts? What if they go from Leparoux to Chris Decarlo on board?

If I can't surmise the early pace and who will be where then I don't play that race.

DJofSD
10-03-2014, 09:46 AM
There appears to be a conflicting assertions.

To both: what is the source of your assertions, BRIS ultimate PPs, raw data from BRIS from the "Software Data Files" section of the web site, or, something else?

TexasDolly
10-03-2014, 09:52 AM
There appears to be a conflicting assertions.

To both: what is the source of your assertions, BRIS ultimate PPs, raw data from BRIS from the "Software Data Files" section of the web site, or, something else?

If the question was addressed to me the source was the software
data files.
TD

DJofSD
10-03-2014, 09:54 AM
If the question was addressed to me the source was the software
data files.
TD
Yes, in part, and, THX.

DeltaLover
10-03-2014, 10:01 AM
I want to make myself clear. The BRIS running style never changes, the number changes all the time. Once an "E", always an "E". I can not find where I read that. I am still sure I read it someplace.

I think you misunderstood what you have read.

Take a look, in the following spreadsheet. it contains a few thousand horses with their assigned running styles per race. The third column contains the number of different styles per starter:

running-styles (http://themindofagambler.com/running_styles.xlsx)

raybo
10-03-2014, 12:22 PM
I think you misunderstood what you have read.

Take a look, in the following spreadsheet. it contains a few thousand horses with their assigned running styles per race. The third column contains the number of different styles per starter:

running-styles (http://themindofagambler.com/running_styles.xlsx)

DL, how are those individual race running styles listed? Most recent on the left, or on the right? That will go a long way in determining how they assign the running styles in their PPs. You can have 10 running lines listed in the PPs for a horse, having more than one style, it is important to know how today's running style is assigned, regarding those previous 10 styles for that horse.

Thanks!

GameTheory
10-03-2014, 12:23 PM
I think you misunderstood what you have read.

Take a look, in the following spreadsheet. it contains a few thousand horses with their assigned running styles per race. The third column contains the number of different styles per starter:

running-styles (http://themindofagambler.com/running_styles.xlsx)
Delta would it be possible to find out the ES pts range for each style? They seem to go hand in hand. In other words, I'm assuming all of the E's are E8 or E7 but not E2, right? And there are no S8s. Are they just taking the speed pts and putting a label on it or are there overlaps? EP6 and also E6s out there?

DJofSD
10-03-2014, 12:34 PM
DL, how are those individual race running styles listed? Most recent on the left, or on the right? That will go a long way in determining how they assign the running styles in their PPs. You can have 10 running lines listed in the PPs for a horse, having more than one style, it is important to know how today's running style is assigned, regarding those previous 10 styles for that horse.

Thanks!
Just a bit of information.

The single file format of the software data files contain the characters for the running style in field 210. This is the only field which has any such data. In other words, there is not a field in the portion of the data for each running line which has a code for the BRIS running style.

This implies the data in the spreadsheet represents instances of data from field 210 for as many different separate files as there are codes listed in the cell for each horse.

Please correct me if I am wrong.

GameTheory
10-03-2014, 12:36 PM
Just a bit of information.

The single file format of the software data files contain the characters for the running style in field 210. This is the only field which has any such data. In other words, there is not a field in the portion of the data for each running line which has a code for the BRIS running style.

This implies the data in the spreadsheet represents instances of data from field 210 for as many different separate files as there are codes listed in the cell for each horse.

Please correct me if I am wrong.
Yes I think that's right. There is no running style for each line, just one applied to the horse for "today" (the file date) based on the last 10 or whatever...

DJofSD
10-03-2014, 12:44 PM
... based on the last 10 or whatever...

It's not clear to me how those designations are assigned by BRIS and what BRIS is trying to tell us.

If they are mutable, what is the criteria?

I treat them like the Prime Power (which are very mutable) -- interesting and the intellectual property of BRIS.

DeltaLover
10-03-2014, 01:05 PM
DL, how are those individual race running styles listed? Most recent on the left, or on the right? That will go a long way in determining how they assign the running styles in their PPs. You can have 10 running lines listed in the PPs for a horse, having more than one style, it is important to know how today's running style is assigned, regarding those previous 10 styles for that horse.

Thanks!


Delta would it be possible to find out the ES pts range for each style? They seem to go hand in hand. In other words, I'm assuming all of the E's are E8 or E7 but not E2, right? And there are no S8s. Are they just taking the speed pts and putting a label on it or are there overlaps? EP6 and also E6s out there?


Sorted from older to newer.
I modified the document, adding racing date + QSP
Take a look:
http://www.themindofagambler.com/running_styles.xlsx

GameTheory
10-03-2014, 01:34 PM
It's not clear to me how those designations are assigned by BRIS and what BRIS is trying to tell us.

If they are mutable, what is the criteria?

I treat them like the Prime Power (which are very mutable) -- interesting and the intellectual property of BRIS.I'm assuming they are based on the last ten because the speed points are and that's part of it, but maybe they go back more...

DJofSD
10-03-2014, 01:40 PM
I'm assuming they are based on the last ten because the speed points are and that's part of it, but maybe they go back more...
They could be based upon a lot of different factors -- both in kind and in number.

So, what I would like to know is how do the running style for today races gets assigned for a horse that runs on both dirt and grass? By only using the grass lines?

P.S. the only things I've ever found that ever explains the running style is on page 2, here: http://www.brisnet.com/library/premexpl.pdf

DeltaLover
10-03-2014, 02:24 PM
I am not sure, I would like to know how the Bris Running Style (or any other similar metric) is created.

Instead I am more interested on how this metric is reflected to the the race's outcome and I think that this is where we need to focus.

Clocker
10-03-2014, 02:57 PM
Instead I am more interested on how this metric is reflected to the the race's outcome and I think that this is where we need to focus.

I don't see how this metric for any one horse has any direct, quantifiable relationship to the outcome of the race. The metric has relevance only in relationship to the particular mix of the running styles of the rest of the field, given the probable (i.e., predicted) pace of the race. It seems to be more of a qualitative than a quantitative matter.

DeltaLover
10-03-2014, 02:59 PM
If you had to create a race level metric, indicating “how much” early speed it has how would go?

Does this kind of a metric makes sense?

How it can become useful?

How it can be validated?

Should we use derivatives (like QSP or BRL) or rely elusively to raw data?

DeltaLover
10-03-2014, 03:00 PM
I don't see how this metric for any one horse has any direct, quantifiable relationship to the outcome of the race. The metric has relevance only in relationship to the particular mix of the running styles of the rest of the field, given the probable (i.e., predicted) pace of the race. It seems to be more of a qualitative than a quantitative matter.

I do not say anything different

Jeff P
10-03-2014, 03:59 PM
I don't see how this metric for any one horse has any direct, quantifiable relationship to the outcome of the race. The metric has relevance only in relationship to the particular mix of the running styles of the rest of the field, given the probable (i.e., predicted) pace of the race. It seems to be more of a qualitative than a quantitative matter.Assume for the sake of argument that track surface or course layout has the ability to play a part in determining race outcomes.

Assume for the sake of argument that a player creates factors of his own making and uses same in his own model (logistic regression or otherwise.)

Further assume for the sake of argument that one of the factors created by the player is something along the lines of "fit for track profile" and that one of the many inputs en route to making that factor is run style of individual horse.


-jp

.

raybo
10-03-2014, 04:19 PM
Just a bit of information.

The single file format of the software data files contain the characters for the running style in field 210. This is the only field which has any such data. In other words, there is not a field in the portion of the data for each running line which has a code for the BRIS running style.

This implies the data in the spreadsheet represents instances of data from field 210 for as many different separate files as there are codes listed in the cell for each horse.

Please correct me if I am wrong.

Sorry I didn't get back to you earlier, got busy. The run style in field 210 is for today's race. It is their calculation of "current run style". That is why I asked how those run styles in your spreadsheet were listed, most recent on the far left, or most recent on the far right.

Running styles are calculated for each race the horse runs, but that run style is a combination of previous run styles assigned for previous races he/she ran. How many races do they go back for the calculation? I don't know, but it could very well be a combination of all its lifetime starts, or it could be a combination of its last 10 races, or whatever. But whatever that number of races is, it is still affected to some degree by each of its previous races, because each run style calculated, includes previous races run, if there were any (in the case of young horses).

DJofSD
10-03-2014, 04:30 PM
Ray, thanks for the reply. Minor point: the spreadsheet is not mine.

To Jeff and Ray: I have a couple of things I would like to explore regarding the BRIS running style. Nothing earth-shattering -- just curious to see what would fall out from doing some explorations. But, I would need to be able to reproduce the running style designation.

Let me try to describe one of them. I would like to study the difference in the running style "predicted" for today's race (what is contained in field 201) v. what actually was run.

raybo
10-03-2014, 04:44 PM
Delta would it be possible to find out the ES pts range for each style? They seem to go hand in hand. In other words, I'm assuming all of the E's are E8 or E7 but not E2, right? And there are no S8s. Are they just taking the speed pts and putting a label on it or are there overlaps? EP6 and also E6s out there?

GT, Brisnet states that the early speed point (ESP) is calculated at the 1st call only. Brisnet also states that their running styles are calculated using both the 1st call and the 2nd call. So, while they may seem to go hand in hand, they really are calculated at different calls.

My understanding of the early speed points, is that they DO NOT measure early speed. They measure a horse's preference for the lead, or its willingness to fight for the lead. That is why we have "E0" horses. They want the lead but either don't have enough early speed to compete for the lead, or they just don't want to fight hard for it. I think this is a big understanding by many players, they think the ESP measures early speed, when in fact it doesn't, it measures "will". So, when you see someone say a horse is an "E" horse, that only means that the horse likes to have the lead, or has to have the lead to win, but it doesn't say how much early speed it has.

What we have is a running style that expresses how a horse normally runs (or for some data providers, how it normally runs when it runs a "good" race).

Then we have the ESP which measures willingness to go for the lead, or fight for it if pressured.

And then we have early speed, however you measure it (I use adjusted fractional velocities; FR1, and FR2), which tells you how fast horses can run early, which really tells you if the horse can compete with others in the race for the lead, without expending too much energy doing it.

IMO, all 3 of those must be used in combination to accurately predict how a race will unfold early, and which horses will be advantaged or disadvantaged by that early running.

The reason I started this thread was because, in my program, a portion of it utilizes a basic database (chart) of winners' running styles and ESPs and my version of the Randy Giles Pace Pressure Gauge, for eliminations from win contention (I also use my early velocities in that method to eliminate potential early speed matchups - pace duels and potential meltdowns). The Brisnet running styles work better than the JCapper running styles (the data provider I now use) for that portion of my program.

raybo
10-03-2014, 04:56 PM
They could be based upon a lot of different factors -- both in kind and in number.

So, what I would like to know is how do the running style for today races gets assigned for a horse that runs on both dirt and grass? By only using the grass lines?

P.S. the only things I've ever found that ever explains the running style is on page 2, here: http://www.brisnet.com/library/premexpl.pdf

I can't say 100%, but running styles are positional only, what running position a horse held at the 1st and/or 2nd call, and/or how many lengths behind the leader they were at those calls. It makes no difference whether the race is on the dirt or on the grass. Yes, horses that run both dirt and turf may run them differently, but I think that is more a factor of a horse's preference for a surface rather than whether it wants to run the same way on both, and the jockey also plays a part in the way it runs on both surfaces, as well as the early pace in each race they run (which typically is slower on turf than on dirt).

That is partially why I prefer to use a running style calculation that measures how a horse most often runs, not how it runs on dirt versus turf, or how it runs only when it is successful. A horse that is not successful in a race can still affect the way the race is run, and that is what I want to know, how this race will be run.

Clocker
10-03-2014, 05:57 PM
My understanding of the early speed points, is that they DO NOT measure early speed. They measure a horse's preference for the lead, or its willingness to fight for the lead.

As designed by Quirin, Speed Points are based on a simple count. I don't know, but I have always assumed that Bris and others use the same method because it is objective.

Speed Points (http://www.horseracinghandicapper.com/QSP.pdf)

I have heard some describe them as a relative probability that the horse will be on the lead or very close at the first call.

DJofSD
10-03-2014, 06:00 PM
That is partially why I prefer to use a running style calculation that measures how a horse most often runs, not how it runs on dirt versus turf, or how it runs only when it is successful. A horse that is not successful in a race can still affect the way the race is run, and that is what I want to know, how this race will be run.

Yes, the "dirty little secret" is the BRIS running style does appear to be a visual running style. My second little idea has to do with looking at the BRIS style and comparing to my own.

raybo
10-03-2014, 06:12 PM
Sorted from older to newer.
I modified the document, adding racing date + QSP
Take a look:
http://www.themindofagambler.com/running_styles.xlsx

Thanks DL!! This will help to understand how horses' running styles (RSs)and early speed points (QSPs) change over time.

Here is a horse that shows its race record, from 26 Nov 2008 through 19 June 2009. Note that the running styles and QSPs have been placed in front of the race date, that is its "current" RS and QSP at that time, because that race has not been run yet. This is how I envision the progressions of run styles and QSPs.

BUEA OAK

(1st time starter, so it has no assigned running style or QSP yet, current RS and QSP = "NA:0")

NA:0 - 26Nov08 race 1

(After 1st race, it ran well enough to the 1st call to warrant a 2 QSP, and probably dropped to the rear of the pack through the 2nd call, so still no distinguishable running style through the 2nd call (not enough data yet), new current RS and QSP = "NA:2")

NA:2 - 11Dec08 race 2

(After 2nd race, It ran probably as a "P" horse so when combined with previous efforts it now is assigned an "S" RS through the 2nd call, and it ran well enough to the 1st call to keep a 2 QSP, new current RS and QSP = "S:2")

S:2 - 08Jan09 race 3

(In its 3rd race, it probably ran as a "P" horse again through the 2nd call (so combined with previous RSs it gets a "P", and its 1st call performance was similar to race 2, current = "P:2")

P:2 - 22Jan09 race 4

(In its 4th race it again ran like a "P" horse through the 2nd call, but ran worse to the 1st call, current = "P:1")

P:1 - 13Feb09 race 5

(In its 5th race it again ran like a "P" so combined with previous RSs it keeps the "P", but it ran better to the 1st call so when combined with previous QSPs it gets a 2, current = "P:2")

P:2 - 08Mar09 race 6

(In its 6th race it probably ran like an "E" horse, on the lead or within a length at the 2nd call so combined it now gets an "E/P", but it ran poorly to the 1st call and it's 2 QSP is degraded to a 0, current = "E/P:0"

E/P:0 - 02Apr09 race 7

(In its 7th race, it probably ran up front again by the 2nd call and keeps its "E/P", but ran poorly again to the 1st call and keeps its 0 QSP, current = "E/P:0")

E/P:0 - 24Apr09 race 8

(Ditto race 7, except it ran much better to the 1st call and now gets a 5 QSP, current = "E/P:5")

E/P:5 - 24May09 race 9

(Race 9 obviously was a very poor performance, probably running mid pack or farther back to the 2nd call, and terrible to the 1st call - current = "P:0") (Time for a rest or a change?)

P:0 - 31May09 race 10

(No rest or change so the horse is still running as a poor "P", current = "P:0")

P:0 - 19Jun09 race 11

(Well, there still wasn't a rest before the 11th race, so barring a major change in distance, surface, track, etc., I would bet it's RS and QSP was similar, or later, in its 11th race, but that was it's last race so we don't know what its new current RS and QSP would be.)

Tom
10-03-2014, 09:47 PM
If you had to create a race level metric, indicating “how much” early speed it has how would go?

Does this kind of a metric makes sense?

How it can become useful?

How it can be validated?

Should we use derivatives (like QSP or BRL) or rely elusively to raw data?

Speed point percentage has been useful over the years.
In HTR, each race header has the QP6 for the race - how many horses have 6 or more ES points. I filter by that metric in many spot plays.

Tom
10-03-2014, 10:57 PM
Thanks DL!! This will help to understand how horses' running styles (RSs)and early speed points (QSPs) change over time.

Here is a horse that shows its race record, from 26 Nov 2008 through 19 June 2009. Note that the running styles and QSPs have been placed in front of the race date, that is its "current" RS and QSP at that time, because that race has not been run yet. This is how I envision the progressions of run styles and QSPs.

BUEA OAK

(1st time starter, so it has no assigned running style or QSP yet, current RS and QSP = "NA:0")

NA:0 - 26Nov08 race 1

(After 1st race, it ran well enough to the 1st call to warrant a 2 QSP, and probably dropped to the rear of the pack through the 2nd call, so still no distinguishable running style through the 2nd call (not enough data yet), new current RS and QSP = "NA:2")

NA:2 - 11Dec08 race 2

(After 2nd race, It ran probably as a "P" horse so when combined with previous efforts it now is assigned an "S" RS through the 2nd call, and it ran well enough to the 1st call to keep a 2 QSP, new current RS and QSP = "S:2")

S:2 - 08Jan09 race 3

(In its 3rd race, it probably ran as a "P" horse again through the 2nd call (so combined with previous RSs it gets a "P", and its 1st call performance was similar to race 2, current = "P:2")

P:2 - 22Jan09 race 4

(In its 4th race it again ran like a "P" horse through the 2nd call, but ran worse to the 1st call, current = "P:1")

P:1 - 13Feb09 race 5

(In its 5th race it again ran like a "P" so combined with previous RSs it keeps the "P", but it ran better to the 1st call so when combined with previous QSPs it gets a 2, current = "P:2")

P:2 - 08Mar09 race 6

(In its 6th race it probably ran like an "E" horse, on the lead or within a length at the 2nd call so combined it now gets an "E/P", but it ran poorly to the 1st call and it's 2 QSP is degraded to a 0, current = "E/P:0"

E/P:0 - 02Apr09 race 7

(In its 7th race, it probably ran up front again by the 2nd call and keeps its "E/P", but ran poorly again to the 1st call and keeps its 0 QSP, current = "E/P:0")

E/P:0 - 24Apr09 race 8

(Ditto race 7, except it ran much better to the 1st call and now gets a 5 QSP, current = "E/P:5")

E/P:5 - 24May09 race 9

(Race 9 obviously was a very poor performance, probably running mid pack or farther back to the 2nd call, and terrible to the 1st call - current = "P:0") (Time for a rest or a change?)

P:0 - 31May09 race 10

(No rest or change so the horse is still running as a poor "P", current = "P:0")

P:0 - 19Jun09 race 11

(Well, there still wasn't a rest before the 11th race, so barring a major change in distance, surface, track, etc., I would bet it's RS and QSP was similar, or later, in its 11th race, but that was it's last race so we don't know what its new current RS and QSP would be.)

Here is Buea Oak's racing history from my DB to go with this. The rF1 and rF2 is the horse's velocity ranking at the 1st and 2nd call for that race. HTR gives him a running style of sustained or rear every race, but the actual pace line data might be helpful.

raybo
10-03-2014, 11:32 PM
Here is Buea Oak's racing history from my DB to go with this. The rF1 and rF2 is the horse's velocity ranking at the 1st and 2nd call for that race. HTR gives him a running style of sustained or rear every race, but the actual pace line data might be helpful.

Thanks Tom! Unfortunately, it looks like HTR is using the 2 furlong and 4 furlong points for their calculations regardless of race distance, instead of 1st call and 2nd call. While that may be a better indicator of early speed, it is not what Brisnet or Quirin do/did. That would cause differences in the running style assignments between the 2 methods. It also could cause differences in the QSP rating between the 2 methods, but don't know exactly which point HTR is using for that calculation.

At some point, during the set of races I posted, the horse had to have run very well to the 1st and/or 2nd call in order to go from an "S" horse to an "E/P horse, and from a 0 QSP to a 5 QSP. Since your chart only goes to 4 furlongs, comparisons would be tough to make.

Tom
10-03-2014, 11:41 PM
No, those fields are the actual beaten lengths from the chart for that race, after it was run.

I put them is to show how the RS, QSP did predicting. You can look back at the last three lines before each race to see how many BL the horse was going in to calculate the speed points.

raybo
10-03-2014, 11:54 PM
If you had to create a race level metric, indicating “how much” early speed it has how would go?

Does this kind of a metric makes sense?

How it can become useful?

How it can be validated?

Should we use derivatives (like QSP or BRL) or rely elusively to raw data?

DL, Randy Giles created his Pace Pressure Gauge which measures how much early pressure is available in a race. While it does not really measure "speed" or "velocity", it does measure how much "willingness" to compete for the lead exists in the race. That pressure is calculated by summing the QSPs that are 5 or more, for all E, E/P and P horses in the race. That measurement can range from 0 to 30+. The higher the number the mors pace pressure is available the race. His other factor in the Gauge is the number of E horses in the race. From the combination of those 2 factors one can determine "early" or "late" regarding what kind of horse is advantaged, regarding the win.

I use a modified version of that method, and create a chart of all races and all pressure gauge calculations, as well as the winners' running styles and QSPs for all those races, for each race I process at a particular track (track specific). Then when processing a new race at that track, the Gauge for that race is calculated, and then "looked up" in the chart, returning the percentages that the winners of all races having that Gauge, ran as an E, E/P, P, and S and what the range of QSPs of the winners having those running styles earned. I then use the results of that "lookup" to determine horses, in today's race, that fall outside those boundaries and are disadvantaged in this race. Those horses are eliminated from win contention.

raybo
10-04-2014, 12:01 AM
No, those fields are the actual beaten lengths from the chart for that race, after it was run.

I put them is to show how the RS, QSP did predicting. You can look back at the last three lines before each race to see how many BL the horse was going in to calculate the speed points.

So, the "BL-2F" and "BL-4F" columns are actually BLs at the 1st and 2nd calls, which would be 4f and 6f for most route races?

Oh, I think see, you added the 2f and 4f BLs just for the QSP comparisons (2f for sprint races and 4f for route races)?

Tom
10-04-2014, 09:48 AM
Correct.

DeltaLover
10-04-2014, 09:57 AM
DL, Randy Giles created his Pace Pressure Gauge which measures how much early pressure is available in a race. While it does not really measure "speed" or "velocity", it does measure how much "willingness" to compete for the lead exists in the race. That pressure is calculated by summing the QSPs that are 5 or more, for all E, E/P and P horses in the race. That measurement can range from 0 to 30+. The higher the number the mors pace pressure is available the race. His other factor in the Gauge is the number of E horses in the race. From the combination of those 2 factors one can determine "early" or "late" regarding what kind of horse is advantaged, regarding the win.


Interesting... I am working on something similar, although following a different approach..

I think the main challenge is to create a stochastic metric expressing the likelyhood of a specific horse to change running style. As an example think of a race with a lot or 'early' speed, something like EEE for example. We all know, that this kind of information, is immediatly noted by the trainers and the jockeys, who most likely will try to change the typical behaviour of their horse.. I really believe that this kind of a scenario, presents one of the most profitable situations in horse betting and this is exactly what I am trying to express with my current development.

DJofSD
10-04-2014, 10:00 AM
Good luck to all those trainers that think they can change a horses running style.

DeltaLover
10-04-2014, 10:10 AM
Good luck to all those trainers that think they can change a horses running style.

It happens all the time DJ...

DJofSD
10-04-2014, 10:17 AM
It happens all the time DJ...
Sure, they can try but taking a confirmed need the lead sprinter and making it something else, no way.

Tom
10-04-2014, 11:22 AM
I have a list of those who did just that.
I call it my record of bets!

raybo
10-04-2014, 11:56 AM
I don't think it's anything new, jockeys and trainers knowing that their E horse is disadvantaged if it runs as an E in a race with one or more equal or better E horses. But, I also think that sometimes we look at the E running style assignment and automatically assume that the only way a horse can run well is if it runs like an E horse. That is one reason that I don't particularly care for a provider's using only horses' "good" races for calculating running styles.

For that reason, I have a button in my program that allows me to switch from my data provider's running styles to my own calculations, immediately recalculating all formulas that use running styles as a factor. Many times it is an eye opener, and causes me to look closer at horses who seemed to be disadvantaged with the raw style but are no longer disadvantaged with my own style assignment, or vice versa. The raw style uses only "good" races", mine uses the last 10 races, good or not so good.

How many times have we thought that a pace meltdown was in the offing, and that the race would be won by a closer? Only to see that the meltdown did not happen and an early horse won the race, and maybe not even the best early horse. Happens all the time. That's one reason that I bet multiple horses to win, because the projected early running doesn't always happen. And, when it doesn't, the prices on the winners are usually better. And, that's another reason that I want to know how Brisnet calculates their running styles, because they work better with what I am doing, enough so, to turn a negative ROI at some tracks into a positive one, or a positive ROI into a higher positive ROI.

DJofSD
10-04-2014, 11:59 AM
I don't think it's anything new, jockeys and trainers knowing that their E horse is disadvantaged if it runs as an E in a race with one or more equal or better E horses. But, I also think that sometimes we look at the E running style assignment and automatically assume that the only way a horse can run well is if it runs like an E horse. That is one reason that I don't particularly care for a provider's using only horses' "good" races for calculating running styles.

For that reason, I have a button in my program that allows me to switch from my data provider's running styles to my own calculations, immediately recalculating all formulas that use running styles as a factor. Many times it is an eye opener, and causes me to look closer at horses who seemed to be disadvantaged with the raw style but are no longer disadvantaged with my own style assignment, or vice versa. The raw style uses only "good" races", mine uses the last 10 races, good or not so good.

How many times have we thought that a pace meltdown was in the offing, and that the race would be won by a closer? Only to see that the meltdown did not happen and an early horse won the race, and maybe not even the best early horse. Happens all the time. That's one reason that I bet multiple horses to win, because the projected early running doesn't always happen. And, when it doesn't, the prices on the winners are usually better. And, that's another reason that I want to know how Brisnet calculates their running styles, because they work better with what I am doing, enough so, to turn a negative ROI at some tracks into a positive one, or a positive ROI into a higher positive ROI.
:ThmbUp: :ThmbUp:

raybo
10-04-2014, 12:16 PM
Here's an example of the changing of perspective regarding pace pressure with 2 different running style calculations in today's Shadwell Turf Mile:

Raw running styles:
1 E 5
2 P 0
3 S 2
4 E 7
5 S 0
6 P 0
7 P 6
8 S 0


My running styles:

1 E 5
2 S 0
3 S 2
4 E 7
5 S 0
6 E/P 0
7 E 6
8 S 0

Now someone post Brisnet's running styles for the Shadwell. :jump:

Clocker
10-04-2014, 12:38 PM
Now someone post Brisnet's running styles for the Shadwell. :jump:

PPs here: Shadwell (http://www.brisnet.com/php/bw_pdf_viewer.php?track=KEE&race=9&param1=6210&param2=276&param3=1359450)

raybo
10-04-2014, 12:55 PM
PPs here: Shadwell (http://www.brisnet.com/php/bw_pdf_viewer.php?track=KEE&race=9&param1=6210&param2=276&param3=1359450)

Thanks!

Brisnet running styles and QSPs for the Shadwell:

1 E 6
2 E/P 3
3 P 3
4 E 7
5 E/P 4
6 P 3
7 E 6
8 P 2


Quite a difference!!

Clocker
10-04-2014, 01:16 PM
Thanks!

Brisnet running styles and QSPs for the Shadwell:

1 E 6
2 E/P 3
3 P 3
4 E 7
5 E/P 4
6 P 3
7 E 6
8 P 2


Quite a difference!!

The first thing that caught my eye was Bris calling the :2: an EP.

I rarely if ever disagree with Bris on their "E" and "P" classifications. The biggest problem is with their "S" horses that I think should be "P", and less often with "EP"s that should be "P".

cj
10-04-2014, 03:03 PM
Here's an example of the changing of perspective regarding pace pressure with 2 different running style calculations in today's Shadwell Turf Mile:

Raw running styles:
1 E 5
2 P 0
3 S 2
4 E 7
5 S 0
6 P 0
7 P 6
8 S 0


My running styles:

1 E 5
2 S 0
3 S 2
4 E 7
5 S 0
6 E/P 0
7 E 6
8 S 0

Now someone post Brisnet's running styles for the Shadwell. :jump:

These are TimeformUS styles and early speed ratings:

1 speed 108
2 tracker 63
3 midpack 82
4 leader 118
5 midpack 78
6 leader 88
7 midpack 100
8 midpack 73

Predicting a fast pace for this one.

Speed Figure
10-04-2014, 04:06 PM
This is a new early speed rating going into the next update on my program. This is for the 1st call only.


01. Sayaad (E 6) 105
02. Optimizer (E/P 3) 80
03. Grand Arch (P 3) 77
04. Silver Max (E 7) 113
05. Kaigun (E/P 4) 67
06. Wise Dan (P 3) 77
07. Two Notch Road (E 6) 93
08. Seek Again (P 2) 70

DJofSD
10-04-2014, 04:07 PM
Just looking at the PPS, no computer analysis, if I were to try to beat Wise Dan, I would go with Grand Arch.

GL to those wagering on this contest.

Tom
10-04-2014, 06:26 PM
Hope you played an exacta - nice price!

OverlayHunter
10-05-2014, 01:38 AM
Raybo, thanks for starting this very interesting and educational thread. And thanks as well to all the contributors.

Question: Raybo, what accounts for the significant differences in your Shadwell QSP's vs. BRIS's?

raybo
10-05-2014, 11:09 AM
Raybo, thanks for starting this very interesting and educational thread. And thanks as well to all the contributors.

Question: Raybo, what accounts for the significant differences in your Shadwell QSP's vs. BRIS's?

Not really sure, I tried to program Quirin's original calculations into Excel, not sure if I accomplished that exactly or not. From what I've read Brisnet's QSPs are supposed to use the same formula also. Maybe they got it right and I didn't or vice versa.

I had no idea the thread would go this long, or include as many differing opinions, I admit the purpose of the thread was rather selfish, but I still haven't accomplished anything along those lines- LOL.

eqitec
10-05-2014, 11:42 PM
I find it necessary to validate the running styles that come down with the BRIS .drf raw data files and change them quite often.

My method for validation or conversion requires the selection of appropriate pace lines. A recent 3 pace lines from races at same distance, and on the same surface & track conditions as today’s race is the ideal. Data from the pace lines selected is then depicted graphically.

I use the BRIS pace pars as a common benchmark against which all horses are measured. The info graphics show the horse’s average variance from the BRIS pace pars using only the pace lines I selected. The top section of the info graphic depicts that average variance from the BRIS par at each pace call (e.g., 2F pace, 4F pace, & late pace.)

The bottom half depicts running position gained or lost from race start to finish and # lengths gained or lost start to finish. The three graphics give me an instant profile of the horse’s running style assuming the pace lines I selected were representative.

The 1st example graphics below is a model for the typical running style of a presser. Nail It improved against the BRIS pace pars at the 3 points of call (top graph) and gained both positions and lengths vs. the race leaders throughout the race in a seemingly steady manner.

Lastly, I then use another graphic to compare the output of the 1st graphic between all horses in the race. See 2nd example. This gives me a clear picture of how well-matched the early speeds may be and where that speed is on the race track based on their post positions. I do this for each pace call. When the early speed seems evenly matched, I go quickly to the graphics for the late pace in search of the best closer with the highest variant vs. the BRIS late pace par.

I too have programmed the Quirin Speed Points within my software and come up with many discrepancies vs the BRIS QSPs. When they are different, I try to see where the BRIS QSPs may have come from, but rarely succeed in that exercise.

The greatest weaknesses of my methods are 1) possible unreliability of BRIS pace pars when different distances are used in the pace lines selected, & 2) so many horses are so lightly raced these days that finding three representative pace lines is often impossible.

Clocker
10-07-2014, 09:55 AM
One issue I don't think has been discussed here is the definition of the styles, especially the "EP" and the infrequently used ""PS" styles.

Some people think of an EP horse as one half way in between an E and a P, able to run closer to the pace than a P, but not usually on the lead. Others define it as a horse that can run as a front end pace setter if the pace is slow, or as a presser if the pace is fast. This is what Randy Giles calls the "two-dimensional" horse. It is not half way between an E and a P, it is either, as circumstances demand.

My main problem with Bris is misidentification of EP horses, and I think that is because I think of an EP in terms of two-dimensional, and that Bris is calling it half way in between. I'm not sure that a computer program is capable of identifying a two-dimensional horse, but that is beyond my ability to judge.

raybo
10-07-2014, 11:05 AM
One issue I don't think has been discussed here is the definition of the styles, especially the "EP" and the infrequently used ""PS" styles.

Some people think of an EP horse as one half way in between an E and a P, able to run closer to the pace than a P, but not usually on the lead. Others define it as a horse that can run as a front end pace setter if the pace is slow, or as a presser if the pace is fast. This is what Randy Giles calls the "two-dimensional" horse. It is not half way between an E and a P, it is either, as circumstances demand.

My main problem with Bris is misidentification of EP horses, and I think that is because I think of an EP in terms of two-dimensional, and that Bris is calling it half way in between. I'm not sure that a computer program is capable of identifying a two-dimensional horse, but that is beyond my ability to judge.

I guess that depends on what you are calling E and P horses, and which races you are looking at to determine that. I have an Excel database of Brisnet data showing running style, 1c and 2c positions, 1c and 2c beaten lengths. Their E/P horses, by and large run like E horses sometimes and like P horses sometimes, and sometimes in between. The problem I see with this database is that a horse can have 9 races listed when a horse has run like a P horse and 1 race where he ran like an E horse and Brisnet calls it an E horse. And other times they will call it some other style. According to Brisnet they do not consider any other factors, only those 4 pieces of data.

The only thing I can figure out to explain this is that they are looking at more than 10 races.

eqitec
10-08-2014, 09:25 AM
Avellaneous is running in the 7th at DEL tomorrow that BRIS labeled as an S style runner.

Two of its races at 5 1/2F (12/13/13 & 3/14/14) give a hint of an "S" conclusion, but today's race is at 6F, not 5 1/2.

See the 1st attached chart. The horse has had only 8 starts. (The one missing race on the attached chart because of the filters in my software was at 5F on grass.)

At 6F, this horse shows me a typical E/P style, running close to the lead (C1 on the chart), then either finishing well to win or come close or fading.

The 2nd attachment gives me a graphic view of the average variances from the BRIS pace pars for the three pace lines I selected, which confirms an E/P style.

This is not to demean BRIS's running style labels. It is most challenging to program automatic running style conclusions for universal implementation across all the variations of racing surfaces and distances.

cj
10-08-2014, 10:58 AM
Avellaneous is running in the 7th at DEL tomorrow that BRIS labeled as an S style runner.

Two of its races at 5 1/2F (12/13/13 & 3/14/14) give a hint of an "S" conclusion, but today's race is at 6F, not 5 1/2.

See the 1st attached chart. The horse has had only 8 starts. (The one missing race on the attached chart because of the filters in my software was at 5F on grass.)

At 6F, this horse shows me a typical E/P style, running close to the lead (C1 on the chart), then either finishing well to win or come close or fading.

The 2nd attachment gives me a graphic view of the average variances from the BRIS pace pars for the three pace lines I selected, which confirms an E/P style.

This is not to demean BRIS's running style labels. It is most challenging to program automatic running style conclusions for universal implementation across all the variations of racing surfaces and distances.

I don't think this horse will be very close early, have as midpack and not enough speed to be up there even if asked.

raybo
10-08-2014, 11:54 AM
Avellaneous is running in the 7th at DEL tomorrow that BRIS labeled as an S style runner.

Two of its races at 5 1/2F (12/13/13 & 3/14/14) give a hint of an "S" conclusion, but today's race is at 6F, not 5 1/2.

See the 1st attached chart. The horse has had only 8 starts. (The one missing race on the attached chart because of the filters in my software was at 5F on grass.)

At 6F, this horse shows me a typical E/P style, running close to the lead (C1 on the chart), then either finishing well to win or come close or fading.

The 2nd attachment gives me a graphic view of the average variances from the BRIS pace pars for the three pace lines I selected, which confirms an E/P style.

This is not to demean BRIS's running style labels. It is most challenging to program automatic running style conclusions for universal implementation across all the variations of racing surfaces and distances.

That horse would be a straight P, not an E/P. To me, in the majority of the races he has run (in his last 10), if the horse was 4th or higher in position at the 1st and 2nd call, or 3 lengths or more back at those calls, then he can not be considered an E/P running style. A real E/P, IMO, should be 2nd or 3rd at both calls and/or less than 3 lengths back, in the majority of his last 10 races. A P horse, would run 4th through 7th (depending on field size) or 3 to 7 lengths back at those calls. And, an S horse would be 8th or worse (depending on field size) or 8 lengths or worse at those calls. In a short field all the positions would compress. The above would be for a field size of 10 or more. You can't be 8th in a field of 7 or less, for instance.

An E horse then, would be leading or less than or equal to 1 length back at the calls in the majority of his last 10 races.

eqitec
10-08-2014, 11:58 AM
Agreed. Avellaneous will not be able to close well enough against this field to win. The attached info graphics support your opinion.

raybo
10-08-2014, 12:58 PM
Agreed. Avellaneous will not be able to close well enough against this field to win. The attached info graphics support your opinion.

Whether or not the horse can "close" is not in question. The question is whether or not he can lead or tightly press early (an E/P style). He cannot, IMO, due to his preferred running style and his lesser early velocities. He could close late, but he will not be up front early, unless the others can't, or just decide not to, run early.

I have it next to last at the 1st call and last at the 2nd call, in FR1 and FR2 best velocities (raw running style and QSP = "P0", mine = "S0").