PDA

View Full Version : Adjusting par times


banacek
12-01-2013, 12:26 PM
I have been using Horsestreet Pars the last few years and have been satisfied. I use them as a basis and make adjustments based on the tracks I play regularly. But have a question to throw out to anyone (including Dave)

The difference between the final times of 2 distances at the same tracks seems to stay the same regardless of the class. The splits seems to vary, but not the final times. I'll use an example from the 2010 pars:

AQU M10000 6f par 113.37 6.5f par 120.06 or a difference of 6.69 seconds

AQU GRADED STAKES 6f par 108.45 6.5f par 115.14 difference of 6.69 seconds.


It seems wrong to me that this would be the case. Shouldn't the difference get somewhat smaller as the class rises..like a sliding scale?

banacek
12-02-2013, 09:45 AM
No response. I'll try to explain a different way.

Again looking at the 2010 pars...I'll use Bay Meadows, so I'm not giving anything away:

Difference in final time pars between M8000 and Graded Stakes:

at 4.5f 3.58 seconds
at Mile 1/4 3.58 seconds
at Mile 3/8 on Turf 3.58 seconds

I'm creating my own pars for the tracks I play. I have been using Horsestreet ones to adjust for shippers from tracks that I don't make pars for. Does anyone know if the Cynthia pars work the same way?

raybo
12-02-2013, 11:08 AM
On the surface of what you have posted I would tend to think that something isn't quite right, but then I don't use Dave's pars. Maybe Dave should be the one to ask about these par differentials. It seems unlikely, IMO, that the class differentials would be the same amount at different distances.

Maybe the point of the pars is not just the final times, but rather the pace that was encountered in order to run those final times. You mentioned that the splits' differentials were not the same, so that tells me that those splits are the key to the par figures.

I have often said that class could be expressed through faster fractionals while attaining the same final time. Some players wonder how a horse that has run faster final times gets beat by a horse that has run slower final times, and the answer obviously is in how those final times were accomplished.

Dave Schwartz
12-02-2013, 11:48 AM
Sorry, I missed this until now.
(BTW, make sure you read to the bottom.)

The world is broken into two camps:

1. Beyer View: The value of a length changes in proportion to the distance.
2. Quirin View: 1 point=1 length

The Beyer view is that the value of a length should diminish as the distance gets longer. This is very logical.

The idea is that if you and I have a foot race, and you are faster by (say) 10 lengths (or seconds, or whatever) at 100 yards, you must be better by 20-somethings if we double the length of the race.

The problem with this actually becomes obvious if you think of the above example. Really... consider a race between you and someone you know (perhaps from your youth - LOL). If you raced a particular distance of ground, and you were 10 lengths better, would you really be 20 lengths better if the distance was doubled?

The answer is, "probably not." Perhaps this would be true if you ran (say) 10 yards then 20 yards... a distance where you could go full speed all the way. But the reality is that in a race between horses (or people) you don't run that way. There is just naturally a "pacing" issue.

This becomes very obvious in turf racing. In some turf races it looks as though nobody really wants the lead. The leaders spend most of their time looking over their shoulders to see who is where. How much of the race is all out?

My test for this was to look at beaten lengths. It seems to me that, if in fact, the lengths-proportionate-to-distance concept was right, then the average margin between the first two finishers would increase proportionately as we go from 6f to 9f. In other words, the average beaten lengths would increase by 1.5 from 6f to 9f.

They do not.

They do increase, but only slightly.

IMHO, there SHOULD be a slight increase. However, it is no where near 1.5 times from 6f to 9f.

Between the two options, I chose the Quirin approach because it is closer to reality. It is also simpler.

BTW, I actually have a switch in my par times to extrapolate the pars outwards using the increasing theory. I did this several years ago, took the resulting pars to import races and tested the results. By comparison, the results were just terrible. I mean, really bad.

Again, just to be clear... there should be SOME change. I could certainly make those changes, but then the system would become much more difficult for everyone to use. Slightly more accurate, but tougher to apply.

One more thing... to use the human-to-human example, there comes a point where the beaten horse, uh... guy... just gives up. As a result, the losers' speed figures are always underestimated if they finish far behind.


Want an interesting adjustment approach? Try this... take any lengths behind at the finish that exceeds 8 and treat it like an 8 when making the speed numbers. IOW, you have a race that goes in 102. There are two horses that finished 8 & 15 lengths behind, respectively. They both get 94 ratings.

You will be amazed at the results.

All of a sudden, those "impossible" longshots that win are not so impossible.


Currently we are selling or 2013 Par Times. The 2014 Pars will be out in March. If you purchase the 2013 Pars now, you will automatically receive the 2014 pars when they are available. (If you order print pars, we will charge you shipping twice.)

I am telling you this because right now you can purchase everything in our store for 40% off!

So, if you are going to purchase the 2014 Pars, this is the best time to do it. EVEN IF YOU ALREADY PURCHASED THE 2013 PARS!

http://www.practicalhandicapping.com/desktop/package/PublicImages/StoreImages/CyberWeekSale70.jpg (http://thehorsehandicappingauthority.com/cyber-week-2013-savings/)

raybo
12-02-2013, 12:06 PM
The Beyer view is that the value of a length should diminish as the distance gets longer. This is very logical.



Dave, what do you mean by the "value" of a length? Are you talking about the time value of a length? If you are, wouldn't it be just the opposite of what you stated. Wouldn't the time value of a length be larger in a longer race, because the horses are traveling at a slower speed, thus more time required to run a length? Or are you talking about the value of a length regarding the calculation of a speed figure, ie; 1 length = 1 point in a speed figure.

I thought the OP's question was about the par "times", not par "figures". Just a bit confused, as I've never read Beyer much nor have I ever made traditional speed figures.

Dave Schwartz
12-02-2013, 12:15 PM
No, points-to-lengths.

banacek
12-02-2013, 12:33 PM
IMHO, there SHOULD be a slight increase. However, it is no where near 1.5 times from 6f to 9f.


Thanks for the detailed response. I agree it wouldn't be 1.5 times, but it would have to be more than 1.

A stakes horse that does 1:08 for 6 furlongs and 2:00 for a mile and an eighth.

A 10000 claimer might do 1:10 for 6 furlongs, certainly not 2:02 for a mile and an eighth.

Like a 3:45 miler and 9.70 100 meter runner.. Someone who does 3:47 would be world class. Someone who runs 11.70 would only beat me :)


It just seems illogical to me, especially to have the same difference in a Mile and 3/8 race on the Turf and a 4.5 f race on the dirt- not that I'd use one to project the other.

And Dave, I'm not knocking your pars- a tremendous amount of work must go into them. I plan on still using them, but I think I might massage the numbers a bit with local knowledge. I am just doing my owns pars for a couple of tracks and the difference appears to be there.

Thanks for your response again.

thaskalos
12-02-2013, 12:50 PM
The world is broken into two camps:

1. Beyer View: The value of a length changes in proportion to the distance.
2. Quirin View: 1 point=1 length

The Beyer view is that the value of a length should diminish as the distance gets longer. This is very logical.

The idea is that if you and I have a foot race, and you are faster by (say) 10 lengths (or seconds, or whatever) at 100 yards, you must be better by 20-somethings if we double the length of the race.

The problem with this actually becomes obvious if you think of the above example. Really... consider a race between you and someone you know (perhaps from your youth - LOL). If you raced a particular distance of ground, and you were 10 lengths better, would you really be 20 lengths better if the distance was doubled?



I think you are making a small mistake when stating the "Beyer view" above, Dave; this wasn't the idea behind Beyer's assertion that the value of a length should diminish as the race gets longer.

This is what Beyer presented as proof that a value-length adjustment should be made for the various distances:

When a human athlete runs a mile and his time is one second slower than the world record for the mile...then he is considered one of the world's best milers himself. But when he runs a 100-meter dash in a time that is one second slower than that of the 100-meter world record...then he is considered as nothing special at all.

This is hard to refute...IMO.

But the original poster is making a different point...which Beyer also addressed in one of his books:

The OP is talking about the adjustment that par charts make when projecting the horses' 6-furlong speed out to 6.5 furlongs.

In almost all the par charts we see...the difference between 6 and 6.5 furlongs is always either 6.4 or 6.6 seconds -- regardless of class. A stakes horse who runs the 6f in 1:08.6 is expected to run the 6.5f in 1:15 (a difference of 6.4 seconds)...while the 5,000 claimer who runs the 6f in 1:13.6 is expected to run the 6.5f in 1:20 (the same difference of 6.4 seconds).

What the OP poster is asking is what Beyer himself asked...and what I -- and many other players -- have been asking for years now:

How is the 5,000 claimer able to negotiate the extra half-furlong of the 6.5f race in the same exact time as the stakes horse...when their ability levels are so different?

Shouldn't it take longer for the 5,000 claimer to travel that extra half-furlong than it takes a stakes horse?

Horseplayersbet.com
12-02-2013, 01:00 PM
I think you are making a small mistake when stating the "Beyer view" above, Dave; this wasn't the idea behind Beyer's assertion that the value of a length should diminish as the race gets longer.

This is what Beyer presented as proof that a value-length adjustment should be made for the various distances:

When a human athlete runs a mile and his time is one second slower than the world record for the mile...then he is considered one of the world's best milers himself. But when he runs a 100-meter dash in a time that is one second slower than that of the 100-meter world record...then he is considered as nothing special at all.

This is hard to refute...IMO.

But the original poster is making a different point...which Beyer also addressed in one of his books:

The OP is talking about the adjustment that par charts make when projecting the horses' 6-furlong speed out to 6.5 furlongs.

In almost all the par charts we see...the difference between 6 and 6.5 furlongs is always either 6.4 or 6.6 seconds -- regardless of class. A stakes horse who runs the 6f in 1:08.6 is expected to run the 6.5f in 1:15 (a difference of 6.4 seconds)...while the 5,000 claimer who runs the 6f in 1:13.6 is expected to run the 6.5f in 1:20 (the same difference of 6.4 seconds).

What the OP poster is asking is what Beyer himself asked...and what I -- and many other players -- have been asking for years now:

How is the 5,000 claimer able to negotiate the extra half-furlong of the 6.5f race in the same exact time as the stakes horse...when their ability levels are so different?

Shouldn't it take longer for the 5,000 claimer to travel that extra half-furlong than it takes a stakes horse?
This makes sense to me.

Slower horses should run at slower velocities than faster horses the longer they run. You would expect that if a stake horse is expected to run the last half furlong in a 6 1/2f race to be faster than a maiden 10k claimer. The greater the distance, the more the differential should be.

raybo
12-02-2013, 01:11 PM
This makes sense to me.

Slower horses should run at slower velocities than faster horses the longer they run. You would expect that if a stake horse is expected to run the last half furlong in a 6 1/2f race to be faster than a maiden 10k claimer. The greater the distance, the more the differential should be.

That is my contention also. However, when talking about higher class horses, don't they normally run faster fractions, overall? If so, then maybe the final time differential, between lower and higher classes, is less than what we think it should be. Interesting topic.

Dave Schwartz
12-02-2013, 01:18 PM
Thaskalos, that is I addressed.

I just used more of an extreme, from 6f to 9f. Same principle.


Shouldn't it take longer for the 5,000 claimer to travel that extra half-furlong than it takes a stakes horse?


In principle, yes.

Quirin had a table that he suggested. It showed how to increase the adjustment away from $10k. The problem is, "Why $10k?" Why not $8k, or $25k?

If you put this into a spreadsheet you will see that the "fulcrum" point makes a big difference. Couple this together with the problem of "a $16k race doesn't always have $16k horses" and you have a nightmare.

Tom
12-02-2013, 02:25 PM
Loking at straight $10K claimers for older males, and using Beyers, I get "10k pars" from 72 - 85. Not such a great fulcrum, but probably more reality than error. A 10K horse at Beu is not a 10K horse from Aqueduct.

cj
12-02-2013, 02:35 PM
I think there are two different things being discussed here, and there is some confusion. How is the value of a "beaten length" relevant to par times, which are based on the winner? They seem like two different things to me.

Show Me the Wire
12-02-2013, 03:46 PM
How is the 5,000 claimer able to negotiate the extra half-furlong of the 6.5f race in the same exact time as the stakes horse...when their ability levels are so different?

The ability differences are based on the length of stride. A stakes horse will cover the same distance and time in less strides. Both horses can cover the distance in the same time, but one can do it in 12 strides or less, while the other will do it in 19 or more strides.

Stillriledup
12-02-2013, 04:06 PM
There are probably some 2,500 dollar claimers at Los Alamitos (thoroughbreds) who could beat a Grade 1 horse like Game On Dude in a really short race, maybe 100 yards or even 2 furlongs...but once the better horse can finally catch up to the quick cheapie, he will pass him and the margin will steadily increase depending on how long the race happens to be. A really strong 20k claimer would probably be able to beat Game on Dude to the top of the stretch before GOD reeled him in and crushed him....so, the length of the race is everything when it comes to a par time, the shorter the race, the more likely it is that a run of the mill decent cheap claimer can come close to the track record.

The Track record for 4 1/2 Furlongs at Los Alamitos for Thoroughbreds is 49.2 set in 1990. Recently, a 2,500 claimer named Buds Pal won a race at Los Al on Oct 27th in 50.4

So, he was less than 2 seconds away from the track record.

I would bet at a track like Belmont, or Santa Anita, a cheap claimer can't even come within 5 seconds of the track record at 1 mile...so, there's probably some formula that you could use to compare track records at all different distances to see how close the solid hard knocking cheap claimer can come to the record for the distance.

traynor
12-02-2013, 07:16 PM
From a developer's standpoint, coding an algorithm to create par times is a LOT easier (and seems much more symmetrical and therefore (?) more reasonable to most people) than gathering winning times by class and distance, cleaning all the dirty data, and generating figures based on real world events rather than algorithms and formulas.

I think anyone who ever questioned Quirin's pars (extrapolated from the average times of a single grade, and possibly a single distance) and started making their own pars understands that quite clearly. Specifically, reality is not as neat and symmetrical as some par charts indicate.

None of which should be construed as a criticism of Dave Schwartz's pars--I have been making my own pars for many years so I am not familiar with the processes he uses to create them.

hogoffate
12-02-2013, 09:30 PM
I use Quirin figures along with Dave's pace pars. The 10K par is 100-100 on the Quirin scale and Dave's par times are 46.2 and 110.1 ( times used are only examples). Let's say 15K level par times are 46.1 and 109.4. Would this not make the Quirin par 101-102? I currently use this method but am wondering if I'm missing something. is this too detailed or should I just give the next level up a 102-102 and try to keep a symmetrical relationship among the pars if at all possible?

Thanks for any help

banacek
12-02-2013, 09:54 PM
I think there are two different things being discussed here, and there is some confusion. How is the value of a "beaten length" relevant to par times, which are based on the winner? They seem like two different things to me.

I agree, cj. I wasn't talking about calculating pars based on Quirin or anyone else. I am talking about calculating pars based on each class at each distance. Now when you are comparing between tracks, sure something must be done, but not when comparing the same track. If this is being done to make the between track calculations easier, then it is going to mess up the comparisons of differences at the same track. That's going to cause some serious errors-therefore money down the drain.

I don't need to look at a 10000 par or a 15000 par for that. Let's say we have Whoop-de-doo Downs with 2 distances (6f and 1 Mile) and these classes with associated average times over an extended period of time so there is a significant sample size:
M5000 114.0 141.4 27.4
M20000 112.6 140.0 27.4
MSW 111.2 138.2 27.0
5000 112.4 138.6 26.2
20000 110.8 136.8 26.0
STK 110.0 136.0 26.0

Other than to massage the data a bit for any obvious discrepancies, why would I change any of them? Why would I adjust them so that the differences between the 2 distances was exactly the same for each class? I don't get it.

Seabiscuit@AR
12-02-2013, 11:29 PM
banacek

you have found a problem with these pars and you are correct in what you are thinking. The only solution is to make your own pars if you want to have accurate pars

traynor
12-03-2013, 01:44 AM
banacek

you have found a problem with these pars and you are correct in what you are thinking. The only solution is to make your own pars if you want to have accurate pars

And the rewards can be generous for those willing to do the work.

Robert Goren
12-03-2013, 07:50 AM
And the rewards can be generous for those willing to do the work.You are probably not gaining that much. The one big gain is being able to spot a change in the hierarchy of tracks and races classifications more quickly than those who don't make their own pars.

traynor
12-03-2013, 11:56 AM
You are probably not gaining that much. The one big gain is being able to spot a change in the hierarchy of tracks and races classifications more quickly than those who don't make their own pars.

I think it is possible to gain a substantial advantage. I also think that one of the major deficiencies of (most) handicapping software applications is that they are comparing apples and oranges in the form of race times and the internal interpretations (by the software) of those times.

Without going into mind-numbing details of a boring topic, one of my best years ever was one in which I discovered "class levels" that seemed discrete, but were not--and the key to all that was an intimate undertanding of pars. That is, specific "class moves" were not moves at all, but no more than shuffling around in the same "grade" that seemed to be "moves." Because the majority (or so it seemed) of the bettors were wagering as if those "class moves" really existed, horses that should have been underlays became overlays, and false favorites were a cinch to toss out.

banacek
12-03-2013, 12:07 PM
I think it is possible to gain a substantial advantage.

I agree..that's why I do it. But I only do 5 tracks, which is enough work. Any other shippers I have to make estimates using the pars. I also estimate the daily variant for horses from other tracks using them. That's one place I am having the difficulty with the numbers. In the past I have tried to adapt the Horsestreet ones to deal with it, but I'm not so comfortable with that either, as I'm almost winging it. Does anyone know if the Cynthia ones work similarly? I've never used them.

Show Me the Wire
12-03-2013, 12:16 PM
Without going into mind-numbing details of a boring topic, one of my best years ever was one in which I discovered "class levels" that seemed discrete, but were not--and the key to all that was an intimate undertanding of pars.

Pars are not necessary to discover the above. One only needs to understand the condition structure used at the specific track and the type of horse that wins that condition.

traynor
12-03-2013, 07:50 PM
Pars are not necessary to discover the above. One only needs to understand the condition structure used at the specific track and the type of horse that wins that condition.

It could also be stated that betting on races is not necessary. I am dismayed at the "one only needs to understand" type of comment. I think the more one understands--including therelationships betweens pars, times, class structures, and race conditions, the more likely one is to get ahead of the crowd with their simplisitic notions of what they "need to know" to win.

Show Me the Wire
12-03-2013, 08:09 PM
It could also be stated that betting on races is not necessary. I am dismayed at the "one only needs to understand" type of comment. I think the more one understands--including therelationships betweens pars, times, class structures, and race conditions, the more likely one is to get ahead of the crowd with their simplisitic notions of what they "need to know" to win.

What you need to know is why the horse is in the race and that is in the conditions. You need to know the condition structure at the track and if that horse belongs in that condition in relation to the other horses.

Pars are not necessary as you imply, to discover the class moves you claimed seemed discreet, but were not. If you want to use pars that is you perogative, but it is not the sole way, to discover these class moves.

traynor
12-03-2013, 08:28 PM
What you need to know is why the horse is in the race and that is in the conditions. You need to know the condition structure at the track and if that horse belongs in that condition in relation to the other horses.

Pars are not necessary as you imply, to discover the class moves you claimed seemed discreet, but were not. If you want to use pars that is you perogative, but it is not the sole way, to discover these class moves.

Once more, I view with extreme skepticism any comment that presumes to tell me "what I need to know." That is generic, not personal.

proximity
12-03-2013, 08:32 PM
I think there are two different things being discussed here, and there is some confusion. How is the value of a "beaten length" relevant to par times, which are based on the winner? They seem like two different things to me.


a couple years back i had a lot of data for charles town in access and computed median times and standard deviations for common distances. and at least for this track, as race distances increased, the standard deviations increased. not quite proportionally, but closer to beyer than quirin. ie. 1 1/8 m times typically strayed further from the median than 4 1/2 f times.

note that if you don't have a lot of data this exercise could be flawed by having a lot of races clustering around a certain class level at one distance relative to a wider distribution of class levels at another distance, but in general is good for a track like ct with a big sample of races at only a handful of distances with a relatively wide spread (4 1/2- 1 1/8) and established older claimers competing across all the distances.

for cd or another major track you obviously wouldn't want to compare 4 1/2 f races (2 yo msw) with the longer distances where a wider range of classes are competing.

Show Me the Wire
12-03-2013, 08:51 PM
Once more, I view with extreme skepticism any comment that presumes to tell me "what I need to know." That is generic, not personal.


I view any comment about something as an absolute with skepticism, same regarding personal.

Obviously, I am not expressing myself correctly. If you want to discover these discrete class moves which are not, it can be done without pars. You need to know the condition structure and the type of horse that fits. Pars are not vital to this undertaking.

JohnGalt1
12-04-2013, 05:08 PM
I adjust my speed figures through out the year, but rarely and if I have a very good reason.

I've used the Cynthia pars for years.

The owner emailed me that they are for one year after I asked if they included January races of the current year. So the 2013 pars are for races in 2012.

To make it easier for me I convert the 10k final speed at each distance to a number from a universal par chart.

From the 2013 book I gave CD an 85 on turf routes and Keeneland a 90.

As the year went on discovered that Keeneland races were running about equal to CD.

If many CD races of 1:43 is a 81/16 drf number and the Kee races of 1:43 is 82/15 then I will make them equal and not rate the Kee race as 5 lengths slower as I did earlier. Or if I saw a horse from a Keeneland who ran a 1:41 and horse who ran at CD in 1:42 I could've missed a potential good win bet since the turf tracks are about equal.

The DRF speed figure confirmed if I should change a number.

After looking at many races I have now lowered the Keeneland rating to 85, and now the DRF speed figures are more closely matched.

banacek
12-04-2013, 05:19 PM
I've used the Cynthia pars for years.


Can you tell me if the Cynthia pars vary by distance or are they the way I explained in the original post? That is, do the pars between each distance at a particular track differ by the same amount regardless of class or do they vary?

Much appreciated.

proximity
12-04-2013, 08:10 PM
Can you tell me if the Cynthia pars vary by distance or are they the way I explained in the original post? That is, do the pars between each distance at a particular track differ by the same amount regardless of class or do they vary?

Much appreciated.
aqu (m)

dist............6f...............6 1/2 f

md 12......112 0/5.........119 3/5... (7 3/5)
stk gr.......108 3/5.........115 2/5.... (6 4/5)

hope this helps. I have to take off NOW for a poker trip to a.c., but you can pm me any questions and i'll answer when I get back (probably sunday). I don't make ratings based on Cynthia pars but do usually get them to keep a file of supplemental info like the winning move factors.....

banacek
12-04-2013, 08:31 PM
aqu (m)

dist............6f...............6 1/2 f

md 12......112 0/5.........119 3/5... (7 3/5)
stk gr.......108 3/5.........115 2/5.... (6 4/5)

hope this helps. I have to take off NOW for a poker trip to a.c., but you can pm me any questions and i'll answer when I get back (probably sunday). I don't make ratings based on Cynthia pars but do usually get them to keep a file of supplemental info like the winning move factors.....

Thanks, that's exactly what I wanted to know. Makes much more sense to me. Much appreciated!

Dave Schwartz
12-04-2013, 10:40 PM
I took a look at this today and have created something that I think will fix the problem of the pace pars. It will be in the next release.

Cratos
12-05-2013, 07:25 PM
I think you are making a small mistake when stating the "Beyer view" above, Dave; this wasn't the idea behind Beyer's assertion that the value of a length should diminish as the race gets longer.

This is what Beyer presented as proof that a value-length adjustment should be made for the various distances:

When a human athlete runs a mile and his time is one second slower than the world record for the mile...then he is considered one of the world's best milers himself. But when he runs a 100-meter dash in a time that is one second slower than that of the 100-meter world record...then he is considered as nothing special at all.

This is hard to refute...IMO.

But the original poster is making a different point...which Beyer also addressed in one of his books:

The OP is talking about the adjustment that par charts make when projecting the horses' 6-furlong speed out to 6.5 furlongs.

In almost all the par charts we see...the difference between 6 and 6.5 furlongs is always either 6.4 or 6.6 seconds -- regardless of class. A stakes horse who runs the 6f in 1:08.6 is expected to run the 6.5f in 1:15 (a difference of 6.4 seconds)...while the 5,000 claimer who runs the 6f in 1:13.6 is expected to run the 6.5f in 1:20 (the same difference of 6.4 seconds).

What the OP poster is asking is what Beyer himself asked...and what I -- and many other players -- have been asking for years now:

How is the 5,000 claimer able to negotiate the extra half-furlong of the 6.5f race in the same exact time as the stakes horse...when their ability levels are so different?

Shouldn't it take longer for the 5,000 claimer to travel that extra half-furlong than it takes a stakes horse?

Try using logarithmic curves as oppose to linear curves.

JohnGalt1
12-05-2013, 08:05 PM
Can you tell me if the Cynthia pars vary by distance or are they the way I explained in the original post? That is, do the pars between each distance at a particular track differ by the same amount regardless of class or do they vary?

Much appreciated.


Using my par chart and numerical comparison rating, some tracks are equal for all sprints like Mountaineer are all 86 and Sunland are all 93.

My ratings for Santa Anita sprints are all 90 except 7f is 92 and 6 and 6 1/2f is 94.

Lone Star sprints are 5f 85, 5.5f 87 and all others are 89.

Routes are the same, some are one number and Tampa is 78 for all distances, except 9f is 75.

I use the Hambleton pace chart from Sartin's Pace Makes the Race.

Hope this helps.

classhandicapper
12-06-2013, 10:29 AM
I think there are two different things being discussed here, and there is some confusion. How is the value of a "beaten length" relevant to par times, which are based on the winner? They seem like two different things to me.

I agree.

My own research into beaten lengths at the finish is consistent with what Dave is saying. The average winning margins increase as distances lengthen, but not by as much as some people think.

But that's entirely different thing than the time relationships of various quality horses as horses stretch out in distance.

cj
12-06-2013, 01:28 PM
I agree.

My own research into beaten lengths at the finish is consistent with what Dave is saying. The average winning margins increase as distances lengthen, but not by as much as some people think.

But that's entirely different thing than the time relationships of various quality horses as horses stretch out in distance.

I was starting to wonder if anyone had a clue what I was talking about there!

banacek
12-17-2013, 02:37 PM
I took a look at this today and have created something that I think will fix the problem of the pace pars. It will be in the next release.

Hi Dave,

Can you elaborate on this a bit. Before I purchase any more pars, I need to make sure they truly meet my needs.

Thanks

Dave Schwartz
12-17-2013, 03:02 PM
The will be done on a ratio basis, using a multiplier.

In other words, as the final time increases (or decreases) the pace times will increase (or decrease) proportionately.

banacek
12-17-2013, 03:11 PM
The will be done on a ratio basis, using a multiplier.

In other words, as the final time increases (or decreases) the pace times will increase (or decrease) proportionately.

But what about the final times? Will they vary between distances based on the class? Thanks

Robert Goren
12-17-2013, 04:21 PM
Back in the day, I found that I had adjust my par times according to field size. There is probably 2/5ths of second between a 6 horse field and a 12 horse field at 6f and 3/5 at a mile 1&1/16 in cheap claimers.

Dave Schwartz
12-17-2013, 07:46 PM
Banacek,

They already do.

I think you are looking for a hand-crafted number for every track-surface-distance-class. That CANNOT exist because there will be years where a particular class NEVER ran a particular distance.

The numbers must be extrapolated.

banacek
12-17-2013, 08:02 PM
Banacek,

They already do.

I think you are looking for a hand-crafted number for every track-surface-distance-class. That CANNOT exist because there will be years where a particular class NEVER ran a particular distance.

The numbers must be extrapolated.
I understand that, but Stakes races are run every year as are 4000 claimers and the difference between a 6f and a mile and a sixteenth are not the same....there has to be a difference between distances for each class or I can't make reliable daily variants.

I have extrapolated them in the past, but wanted something where this already being done. Clearly for other users they are useful. but they just won't work for what I need them for. That's okay, I was just curious about the changes you were making. Thanks anyway.

Delta Cone
12-17-2013, 08:23 PM
Back in the day, I found that I had adjust my par times according to field size. There is probably 2/5ths of second between a 6 horse field and a 12 horse field at 6f and 3/5 at a mile 1&1/16 in cheap claimers.

Curious...were the races generally slower in small fields? Or faster?

Dave Schwartz
12-17-2013, 10:54 PM
Banacek,

I think you will be unhappy with our par times.

This will not change.

I do not believe that the Beyer process is correct for the reasons I have stated.

I do believe that there are superior ways to make numbers than the methods I use. However, short of making a race-by-race projection, these are very solid pars.

The projection approach lends itself very nicely to numbers produced by (say) TimeForm U.S. An annual par figure cannot possibly compete with a race-by-race projection.

However, short of that, I am quite pleased (as are many others) with the process we use.

Boulder
04-02-2014, 09:55 PM
Dave,

Just wondering if the new multiplier are working in the 2014 pars.

Gapfire
04-03-2014, 12:26 AM
I

AQU M10000 6f par 113.37 6.5f par 120.06 or a difference of 6.69 seconds

AQU GRADED STAKES 6f par 108.45 6.5f par 115.14 difference of 6.69 seconds.


It seems wrong to me that this would be the case. Shouldn't the difference get somewhat smaller as the class rises..like a sliding scale?

I agree Banacek. I would think that the graded stakes horse would run that last half furlong somewhat faster than the maiden claimer.

Sapio
04-03-2014, 08:16 AM
I think you are making a small mistake when stating the "Beyer view" above, Dave; this wasn't the idea behind Beyer's assertion that the value of a length should diminish as the race gets longer.

This is what Beyer presented as proof that a value-length adjustment should be made for the various distances:

When a human athlete runs a mile and his time is one second slower than the world record for the mile...then he is considered one of the world's best milers himself. But when he runs a 100-meter dash in a time that is one second slower than that of the 100-meter world record...then he is considered as nothing special at all.

This is hard to refute...IMO.

But the original poster is making a different point...which Beyer also addressed in one of his books:

The OP is talking about the adjustment that par charts make when projecting the horses' 6-furlong speed out to 6.5 furlongs.

In almost all the par charts we see...the difference between 6 and 6.5 furlongs is always either 6.4 or 6.6 seconds -- regardless of class. A stakes horse who runs the 6f in 1:08.6 is expected to run the 6.5f in 1:15 (a difference of 6.4 seconds)...while the 5,000 claimer who runs the 6f in 1:13.6 is expected to run the 6.5f in 1:20 (the same difference of 6.4 seconds).

What the OP poster is asking is what Beyer himself asked...and what I -- and many other players -- have been asking for years now:

How is the 5,000 claimer able to negotiate the extra half-furlong of the 6.5f race in the same exact time as the stakes horse...when their ability levels are so different?

Shouldn't it take longer for the 5,000 claimer to travel that extra half-furlong than it takes a stakes horse?

Hi thaskalos

The par charts are lacking. What is not taken into account is the variance between the classes.

Thomas Sapio

PhantomOnTour
04-03-2014, 10:51 AM
A correctly designed chart avoids the problem being discussed here.
It all comes down to the value that one places on each increment in their charts.
For example, let:
5.5f = .15 secs
6f = .16 secs
6.5f = .17 secs
7f = .18 secs

If you extrapolate using these values and make your chart, you will not run into this problem.
A horse rated a 108 will take 6.33 secs to go from 6.5f-7f
A horse rated a 100 will take 6.41 secs to go from 6.5f-7f