Deriving an odds line from FPS

[holdalltix]

After reading numerous threads over and over regarding odds lines ...the expertise on this board is amazing, sometimes to the point of leaving me unsure if left is right or vice versa or if it even matters...so I will try to keep my question simple.

What do you think is the best way to determine a fair odds line based on the following:

Horse
1) 56 FPS
2) 55 FPS
3) 54 FPS
4) 53 FPS

I believe there may be more info needed from a db, such as std dev but am unsure.

Most appreciative of any info but alas my mind is overly basic...so kindly keep it rather simple.

Many Thanks

[Dave Schwartz]

I have no answer for you but perhaps can offer some suggestions:

It would not be just "how good is a 56.00?" That is like saying, "Is 89 a good Beyer number?" Depends upon what the competition looks like.

So, it could boil down to "What is a 0.20 edge worth at 6f at HOL?" Logically, each track/distance could be different. Certainly each distance would be different.

Much of our recent research indicates that length of stretch and oval size are important, so grouping races by distances, similar oval and stretch length might be good enough.

Any way you slice it, you are looking at a very big challenge.


Good luck!


Regards,
Dave Schwartz

[jasperson]

After reading numerous threads over and over regarding odds lines ...the expertise on this board is amazing, sometimes to the point of leaving me unsure if left is right or vice versa or if it even matters...so I will try to keep my question simple.

What do you think is the best way to determine a fair odds line based on the following:

Horse
1) 56 FPS
2) 55 FPS
3) 54 FPS
4) 53 FPS

I believe there may be more info needed from a db, such as std dev but am unsure.

Most appreciative of any info but alas my mind is overly basic...so kindly keep it rather simple.

Many Thanks
Here is how I would start
Each horse has a 25% chance of winning the race not considering the fps factor. Since horse has a superiority of 2 fps over the 2 horses multiply 25X1.2=30 subtract 25 =5 add it to the 25% to get 30%. Since we have increase horse 1 percent we must decrease the the other 3 horse % to maintain 100% of all horses. So divide 5 by 3 =1.67 and subtract it from the other 3 horses. Now do that with the number 2 horse and so on, and then convert the % to odds and you have and oddline for those horses.
Good luck
Jack

[Murph]

Here is how I would start
Each horse has a 25% chance of winning the race not considering the fps factor. Since horse has a superiority of 2 fps over the 2 horses multiply 25X1.2=30 subtract 25 =5 add it to the 25% to get 30%. Since we have increase horse 1 percent we must decrease the the other 3 horse % to maintain 100% of all horses. So divide 5 by 3 =1.67 and subtract it from the other 3 horses. Now do that with the number 2 horse and so on, and then convert the % to odds and you have and oddline for those horses.
Good luck
JackThat makes alot of sense to me but I hadn't hammered out that formula. Too late now Jack, cat jumped the bag. :)

Thanks, Murph

[singunner]

After a year of working (in part) on this exact problem, I feel like I can speak semi-authoritatively.

What Dave has said is the way you want to work it out. Jasperson's desires to keep everything adding up to 100% are noble, but my own research has found it inhibitive (it can be done, but I bet you'll get better results not worrying about it). Unfortunately, the amount of back-data necessary to do what Dave suggested is also slightly inhibitive (given how few races will satisfy each individual scenario).

So you've got your A, B, C and D horses (if you're using numbers for the horses, you haven't been working on this too long) with scores of 1, 2, 3 and 4, respectively (just simplifying from your actual scores).

What I did at first (and what will provide a nice baseline) with my program was to make what is known as a "frequency probability". You look at all occurences where a horse is 1 point better than another horse, then find the probability (%) of the the better horse winning. Then do it for all occurences of a horse being 2 points better than another, etc, etc. This will produce a curved graph, but for my explanation, we'll pretend pretend it's a straight line and that horses that were 1 point better won 55% of the time, 2 points better won 60% of the time and so forth.

You'll be able to make a nice little graph like this:

A B C D
A X 55 60 65 = 180
B 45 X 55 60 = 160
C 40 45 X 55 = 140
D 35 40 45 X = 120

Now add up the rows and divide each number by (number of horses in race -1)*(number of horses in race)/2. For a 4-horse race, that's (3)*(4)/2=6, so divide each row total by 6.

Now you've got an oddsline (the example used a straight line above, so it does add up to 100%).

A = 30%
B = 26.6%
C = 23.3%
D = 20%


And if that seemed like a lot to understand, wait til we get into Bayes Theory!!!

[jasperson]

Murph,
From your question I assume that you are a pace handicapper. I have written a program to compute Brohamer's pace figures that I am not having much luck with :bang: that I will give you if you will agree to feed me any infor and luck that you have with it. It had one winner yesterday at aqu out of 9 and 2 out of 10 at lrl.:mad: I think Brohamer sucks.
Jack

[Dave Schwartz]

I think Brohamer sucks.


Not sure what would drive anyone to such a conclusion.

He is one of those rare persons in this industry who is precisely what he says he is: a winning player.

In addition to that he is a fine gentleman.


Best of luck to you.


Dave Schwartz

[Jeff P]

I attack this type of thing in much the same way a researcher or actuary working for an insurance company would... by looking at a what actually happens across several large data samples broken out by factor rank, factor gap, and by factor numeric value. Depending on the circumstances, I'll sometimes break things down further by looking at results from specific distances at specific tracks. But the gist of how I approach this type of thing is to derive approximate historical probability by looking at significant samples of past results.

For example, the following sample is for dirt sprints found in my Q1 2008 database broken out by Compound E1 rank... Compound E1 being something along the lines of a Bris pace fig equivalent to Sartin style E1 fps for selected pacelines:

By: CompoundE1 Rank

Rank Gain Bet Roi Wins Plays Pct Impact
1 -1010.40 8436.00 0.8802 926 4218 .2195 1.8025
2 -2089.80 8428.00 0.7520 681 4214 .1616 1.3269
3 -1468.40 8478.00 0.8268 596 4239 .1406 1.1544
4 -2505.70 8564.00 0.7074 532 4282 .1242 1.0201
5 -1688.90 8534.00 0.8021 488 4267 .1144 0.9390
6 -2820.80 8214.00 0.6566 350 4107 .0852 0.6997
7 -1873.10 7032.00 0.7336 279 3516 .0794 0.6515
8 -1597.00 5264.00 0.6966 200 2632 .0760 0.6239
9 -722.40 3480.00 0.7924 113 1740 .0649 0.5332
10 -559.40 2062.00 0.7287 60 1031 .0582 0.4778
11 -244.70 920.00 0.7340 24 460 .0522 0.4284
12 -310.60 378.00 0.1783 3 189 .0159 0.1303
13 -28.00 28.00 0.0000 0 14 .0000 0.0000
14 -6.00 6.00 0.0000 0 3 .0000 0.0000

And here is the same sample broken out by factor gap:

By: CompoundE1 Gap

>=Min <Max Gain Bet Roi Wins Plays Pct Impact
-999.00 -15.00 -4510.30 13484.00 0.6655 488 6742 .0724 0.5943
-15.00 -14.00 -279.40 1586.00 0.8238 54 793 .0681 0.5591
-14.00 -13.00 -543.20 1896.00 0.7135 73 948 .0770 0.6323
-13.00 -12.00 -672.20 2260.00 0.7026 91 1130 .0805 0.6612
-12.00 -11.00 -932.90 2440.00 0.6177 110 1220 .0902 0.7403
-11.00 -10.00 -734.00 2834.00 0.7410 131 1417 .0924 0.7591
-10.00 -9.00 -742.10 3228.00 0.7701 167 1614 .1035 0.8496
-9.00 -8.00 -1326.60 3602.00 0.6317 171 1801 .0949 0.7796
-8.00 -7.00 -868.60 3870.00 0.7756 213 1935 .1101 0.9038
-7.00 -6.00 -1233.30 4180.00 0.7050 215 2090 .1029 0.8446
-6.00 -5.00 -902.60 4022.00 0.7756 240 2011 .1193 0.9799
-5.00 -4.00 -785.30 4158.00 0.8111 272 2079 .1308 1.0742
-4.00 -3.00 -664.00 3916.00 0.8304 290 1958 .1481 1.2161
-3.00 -2.00 -796.50 3752.00 0.7877 279 1876 .1487 1.2211
-2.00 -1.00 -577.50 3384.00 0.8293 275 1692 .1625 1.3345
-1.00 0.00 -346.30 2776.00 0.8753 257 1388 .1852 1.5203
0.00 1.00 -676.00 2360.00 0.7136 206 1180 .1746 1.4334
1.00 2.00 -264.10 1852.00 0.8574 194 926 .2095 1.7202
2.00 3.00 -237.10 1372.00 0.8272 137 686 .1997 1.6398
3.00 999999.00 166.80 2852.00 1.0585 389 1426 .2728 2.2398


Looking at factor rank you could establish base probabilities as follows:

Rank=1 Baseprob = .2195
Rank=2 Baseprob = .1616
Rank=3 Baseprob = .1406
Rank=4 Baseprob = .1242
and so on...

You can make your probability assessment more accurate by bringing in factor gap. For example, when the top ranked horse has an advantage of 3 or more points over the second ranked horse (a gap of 3) the historical probability runs at about .2728 instead of .2195.

I'll actually develop a factor gap impact value for each identifiable rank and gap segment... for example, the rank=1 gap>=3 segment would have a gap impact value of 1.2428 (.2728/.2195) which I would then multiply by the rank=1 base prob (.2195 x 1.2428 = .2728 prob)...

and prob for the rank=1 gap>=2 segment would be .2195 x .9078 = .1993...

Like I say I'll do this (or rather have the program do this) for each rank and gap segment for every factor that I'm using in my model... I actually do this for hundreds of factors.

The end result is that I get a very accurate idea about horse win probabilities.

And no, before anyone asks, betting every horse that steps into a starting gate at odds higher than a number from an odds line (a theoretical overlay) is not the road to profits. There ARE ways to use odds lines to help drive profits. It might be counter intuitive, but IMHO strict adherence to all theoretical overlays isn't the way to go.

And I couldn't agree more with Dave. Brohammer's work, if you take the time to understand it - is damn good.


-jp

.

[bigmack]

Murph,
From your question I assume that you are a pace handicapper. I have written a program to compute Brohamer's pace figures that I am not having much luck with :bang: that I will give you if you will agree to feed me any infor and luck that you have with it. It had one winner yesterday at aqu out of 9 and 2 out of 10 at lrl.:mad: I think Brohamer sucks.
Jack
Nice, Jack. Mind if we make you poster boy for the "mighty bothersome person" award?

Where is it written that even the best of pace figures "picks" winners? Put your thinking cap on, man. Then take your hat off & hand it to me so I can hand it back to you & say, "Here's your hat, what's your hurry"?

[jasperson]

I attack this type of thing in much the same way a researcher or actuary working for an insurance company would... by looking at a what actually happens across several large data samples broken out by factor rank, factor gap, and by factor numeric value. Depending on the circumstances, I'll sometimes break things down further by looking at results from specific distances at specific tracks. But the gist of how I approach this type of thing is to derive approximate historical probability by looking at significant samples of past results.

For example, the following sample is for dirt sprints found in my Q1 2008 database broken out by Compound E1 rank... Compound E1 being something along the lines of a Bris pace fig equivalent to Sartin style E1 fps for selected pacelines:

By: CompoundE1 Rank

Rank Gain Bet Roi Wins Plays Pct Impact
1 -1010.40 8436.00 0.8802 926 4218 .2195 1.8025
2 -2089.80 8428.00 0.7520 681 4214 .1616 1.3269
3 -1468.40 8478.00 0.8268 596 4239 .1406 1.1544
4 -2505.70 8564.00 0.7074 532 4282 .1242 1.0201
5 -1688.90 8534.00 0.8021 488 4267 .1144 0.9390
6 -2820.80 8214.00 0.6566 350 4107 .0852 0.6997
7 -1873.10 7032.00 0.7336 279 3516 .0794 0.6515
8 -1597.00 5264.00 0.6966 200 2632 .0760 0.6239
9 -722.40 3480.00 0.7924 113 1740 .0649 0.5332
10 -559.40 2062.00 0.7287 60 1031 .0582 0.4778
11 -244.70 920.00 0.7340 24 460 .0522 0.4284
12 -310.60 378.00 0.1783 3 189 .0159 0.1303
13 -28.00 28.00 0.0000 0 14 .0000 0.0000
14 -6.00 6.00 0.0000 0 3 .0000 0.0000

And here is the same sample broken out by factor gap:

By: CompoundE1 Gap

>=Min <Max Gain Bet Roi Wins Plays Pct Impact
-999.00 -15.00 -4510.30 13484.00 0.6655 488 6742 .0724 0.5943
-15.00 -14.00 -279.40 1586.00 0.8238 54 793 .0681 0.5591
-14.00 -13.00 -543.20 1896.00 0.7135 73 948 .0770 0.6323
-13.00 -12.00 -672.20 2260.00 0.7026 91 1130 .0805 0.6612
-12.00 -11.00 -932.90 2440.00 0.6177 110 1220 .0902 0.7403
-11.00 -10.00 -734.00 2834.00 0.7410 131 1417 .0924 0.7591
-10.00 -9.00 -742.10 3228.00 0.7701 167 1614 .1035 0.8496
-9.00 -8.00 -1326.60 3602.00 0.6317 171 1801 .0949 0.7796
-8.00 -7.00 -868.60 3870.00 0.7756 213 1935 .1101 0.9038
-7.00 -6.00 -1233.30 4180.00 0.7050 215 2090 .1029 0.8446
-6.00 -5.00 -902.60 4022.00 0.7756 240 2011 .1193 0.9799
-5.00 -4.00 -785.30 4158.00 0.8111 272 2079 .1308 1.0742
-4.00 -3.00 -664.00 3916.00 0.8304 290 1958 .1481 1.2161
-3.00 -2.00 -796.50 3752.00 0.7877 279 1876 .1487 1.2211
-2.00 -1.00 -577.50 3384.00 0.8293 275 1692 .1625 1.3345
-1.00 0.00 -346.30 2776.00 0.8753 257 1388 .1852 1.5203
0.00 1.00 -676.00 2360.00 0.7136 206 1180 .1746 1.4334
1.00 2.00 -264.10 1852.00 0.8574 194 926 .2095 1.7202
2.00 3.00 -237.10 1372.00 0.8272 137 686 .1997 1.6398
3.00 999999.00 166.80 2852.00 1.0585 389 1426 .2728 2.2398


Looking at factor rank you could establish base probabilities as follows:

Rank=1 Baseprob = .2195
Rank=2 Baseprob = .1616
Rank=3 Baseprob = .1406
Rank=4 Baseprob = .1242
and so on...

You can make your probability assessment more accurate by bringing in factor gap. For example, when the top ranked horse has an advantage of 3 or more points over the second ranked horse (a gap of 3) the historical probability runs at about .2728 instead of .2195.

I'll actually develop a factor gap impact value for each identifiable rank and gap segment... for example, the rank=1 gap>=3 segment would have a gap impact value of 1.2428 (.2728/.2195) which I would then multiply by the rank=1 base prob (.2195 x 1.2428 = .2728 prob)...

and prob for the rank=1 gap>=2 segment would be .2195 x .9078 = .1993...

Like I say I'll do this (or rather have the program do this) for each rank and gap segment for every factor that I'm using in my model... I actually do this for hundreds of factors.

The end result is that I get a very accurate idea about horse win probabilities.

And no, before anyone asks, betting every horse that steps into a starting gate at odds higher than a number from an odds line (a theoretical overlay) is not the road to profits. There ARE ways to use odds lines to help drive profits. It might be counter intuitive, but IMHO strict adherence to all theoretical overlays isn't the way to go.

And I couldn't agree more with Dave. Brohammer's work, if you take the time to understand it - is damn good.


-jp

.
I glad that all this makes sense to you and I would like to say that I would like to learn anything new to make my handicapping better. Maybe I am missing something in his book and would like to be enlightened in what I am a doing wrong. I select races that I thinks demonstrates the best horses abililty at this distance and class to the best of my abibility from his most recent races. I them print out the prediction of his best ap horses in order along with the best speed rating of each horse's selected race and the best speed rating horse wins more often than the brohamer top rated horse hands downs. Granted I select the race to evaluate as the best speed rating for that horse at the distance and class with no trouble lines. If I am doing something wrong I would like to know what it is. I am a prostitute in handicapping I will switch to any thing that makes money, so don't like any one handicapping system over an other,and if I am doing something wrong I would lilke to correct it. My data just shows the speed rating over shadow brohamer's ap rating. If I am wrong somebody point it out to me
best
Jack

[bigmack]

If I am wrong somebody point it out to me
You insulted one of the greats of the game in this public forum. Have the decency to apologize and accept that it had nothing to do with him. Until then, give us a reason to help you out.

[jasperson]

Nice, Jack. Mind if we make you poster boy for the "mighty bothersome person" award?

Where is it written that even the best of pace figures "picks" winners? Put your thinking cap on, man. Then take your hat off & hand it to me so I can hand it back to you & say, "Here's your hat, what's your hurry"?
I am an engineer(retired) and I bought Brohamer's book and programmed my computer to hack his numbers. I was neither a believer or disbeliever, but hoping he had something. Alas as Daniel said to Nebuchadnezzar about the handwriting on the wall "it was weighted in the balances and found wanting". I didn't want it turn out that way but it just did. If I am wrong tell me what I am doing wrong and I will correct it. Provide me some data that you have the proves me wrong along what I am doing and I will be glad to change my approach to the problem.
Best Regards
Jack

[aaron]

Just my opinion,but I think a lot of people spend much to much time trying to create a line. While you must know what is the minimum price you will accept,there are many factors that can make any line incorrect,so if you are a winning player who has a good opinion you will pick enough overlays to compensiate for any mistakes in your line.
I'm not a big believer of going 4 deep in a race in search of an overlay. If you have an opinion on a race,it will jump out at you. I believe if nothing has changed after scratches to alter your opinion,then the only option is how to bet that opinion.

[njcurveball]

I am an engineer(retired) and I bought Brohamer's book and programmed my computer to hack his numbers.

If you truly are an Engineer than you know the sentence above states the defect is not in the Methodology, but in the way you are processing it.

Racing is not just about picking winners. It is about gaining enough of an edge on the competition that you show a profit. There are no secret formulas out there, despite all of us thinking that they exist.

Brohamer is a professional who grew up with racing and is very talented at many aspects of the game. Talent plus experience plus methodology equals success.

You are trying to eliminate two important variables and then state your conclusion that methodology does not equal success. Given your short sighted experimentation that is a valid conclusion for YOU.

You need to add in the experience part (which you glossed over when you said YOU were the one picking the representative pace line), and the talent part of knowing which races to get involved in and which to pass. Also knowing when to bet and how much.

You make it sound like you have found the perfect paceline for each horse and blame the formulas other people are winning with.

[Murph]

:D Murph,
From your question I assume that you are a pace handicapper. I have written a program to compute Brohamer's pace figures that I am not having much luck with :bang: that I will give you if you will agree to feed me any infor and luck that you have with it. It had one winner yesterday at aqu out of 9 and 2 out of 10 at lrl.:mad: I think Brohamer sucks.
Jack:ThmbUp: I LOVE it, now gentlemen in this corner, an HONEST man! That's so dam funny, Jack. Don't waste your nails on this one ... e-mail to me.

Murph

[jasperson]

I didn't realize that I had trod on sacred ground and I apologize for my hearsay.. I meant no disrespect to Thomas Brohamer because I have never met him. In the future I will treat his methodology with the respect it deserves. I do think his methodology may have some merit but for present it is impractical. The problem with it is as I see it is


What is the correct pace line to select
After selecting a pace line where can I get the pace par and track variant for that day.
Compensating a router in a sprint race and sprinter in a route race.
If a sprinter is in a route race and has a creditable routes 9 races back should I use this in place of his latest sprint?
Same for a router in a sprint.
What if I have a horse ships in from a different track in which I have now data?
His method may have worked well at SA and HOL because they use to be very similar tracks. When I played them in the 1980's and made my own par times and track variants I found this to be true with the exception of the mile distance because of Hol 1 turn mile.

Things might be better at Woodbine because of their tracking system which will give us more accurate data on lengths behind and the horse's position relative to the rail

Regards,

Jack

[Murph]

You nailed it, buddy. (one nail too, had a manicure lately?)

Each point you make has to be addressed properly for the method described in the books to work for your game. Look where Binford has taken it, very interesting but not for me.

Much depends on the level of understanding you have reached with your handicapping game. I have some advice for you that I will not type here. You are solving the right problems but as you can see, it's going to take awhile.
Murph

[Dave Schwartz]

What is the correct pace line to select?
Tough question. I can tell you that it is nopt usually the last one as is generally suggested.

Actually, I have been studying this a bit and hope to have some good answers before the end of summer.

After selecting a pace line where can I get the pace par and track variant for that day.
Try here...
http://www.horsestreet.com/products/index.html#Par%20Times

Compensating a router in a sprint race and sprinter in a route race.
Pars should take care of that.

If a sprinter is in a route race and has a creditable routes 9 races back should I use this in place of his latest sprint?
Same for a router in a sprint.
See 1st question.

What if I have a horse ships in from a different track in which I have now data?
Pars and your model should do you just fine.



Howard Sartin used to say that there were three things you needed to do in order to win with pace handicapping. (I believe there is a fourth.)

1) Select the right contenders
2) Select the right pacelines for those contenders
3) Interpret the "readouts" correctly. (That is, interpret the model.)

I would add:
4) Be aware of value in deciding which and how many of your selections to wager on.


This probably will not be enough to get you where you are going but it is a shove in the right direction.

Another place to go for info is to hang out where the pace handicappers hang out. That would be Ted Craven's place: http://bindfold.com/forums/



Regards,
Dave Schwartz

[Murph]

What is the correct pace line to select?
Tough question. I can tell you that it is nopt usually the last one as is generally suggested.

Actually, I have been studying this a bit and hope to have some good answers before the end of summer.

KUTIT
Another place to go for info is to hang out where the pace handicappers hang out. That would be Ted Craven's place: http://bindfold.com/forums/

Regards,
Dave SchwartzThanks for the link to Binfords folder, Dave. I hope you can get your pace line selection research wrapped in time for Christmas. I gave up after about three years or more of head knocking over it.

Murph

[njcurveball]

What is the correct pace line to select
After selecting a pace line where can I get the pace par and track variant for that day.
Compensating a router in a sprint race and sprinter in a route race.
If a sprinter is in a route race and has a creditable routes 9 races back should I use this in place of his latest sprint?
Same for a router in a sprint.
What if I have a horse ships in from a different track in which I have now data?





Since I have a day job, I use the HTR program developed by Ken Massa. He worked with Brohamer on the original MPH program and all of the things you mentioned above are built into the program.

The automatic pace-line selection method is far from perfect but reliably does it race after race, allowing modeling to be consistent and profitable.

You are an Engineer, so I would guess you use a computer, rather than a slide rule these days. There are many programs out there doing these things, why try to reinvent the wheel?

If you want to take a shot at picking your own pacelines, most let you do that as well. I think HTR has about 8 different types of paceline methods. I am sure other programs have similar as well.

Unless you feel you have something to add to the computation that others do not, it is probably better to use the state of the art software already out there.

Jim

[Greyfox]

There are several paths to the same summit.

Dick Mitchell outlines exactly how to relate fps to odds in his book,
Winning Thoroughbred Strategies.

[jasperson]

Since I have a day job, I use the HTR program developed by Ken Massa. He worked with Brohamer on the original MPH program and all of the things you mentioned above are built into the program.

The automatic pace-line selection method is far from perfect but reliably does it race after race, allowing modeling to be consistent and profitable.

You are an Engineer, so I would guess you use a computer, rather than a slide rule these days. There are many programs out there doing these things, why try to reinvent the wheel?

If you want to take a shot at picking your own pacelines, most let you do that as well. I think HTR has about 8 different types of paceline methods. I am sure other programs have similar as well.

Unless you feel you have something to add to the computation that others do not, it is probably better to use the state of the art software already out there.

Jim
I wrote my own program so that is not a problem. It displays the 10 pp lines with type of race distance speedrating and comments along with the type and distance of todays race. Attached is my pace line for tomorrow at Aqu.
Jack

[jasperson]

To compare the pace line with the oddsline I am attaching my oddsline print out. It is best viewed with a dos editor because wordprocessors will wrap the lines arounds since the line is 140 characters long
Jack

[bigmack]

I gave up after about three years or more of head knocking over it.
Murph
You mean you had another bad experience with a vendor? If I were you, I'd start taking it personally. :rolleyes:

[Murph]

I found that current form seemed to be more important than which pace line was used to rate a runner. Same dist. same surface ect. wasn't successful at all when going back in the pp lines.

When reviewing results after a couple of years, this became more apparent. I spent more time still, anaylizing pace line selection using various AI tools and that didn't work either. I found a better way (for me) to incorporate current form that rendered paceline selection methods all but useless.

On a more personal note, bigmak, what is your interest in my personal business? Perhaps you could share with the rest of us.

Murph

[Niko]

I found that current form seemed to be more important than which pace line was used to rate a runner. Same dist. same surface ect. wasn't successful at all when going back in the pp lines.

When reviewing results after a couple of years, this became more apparent. I spent more time still, anaylizing pace line selection using various AI tools and that didn't work either. I found a better way (for me) to incorporate current form that rendered paceline selection methods all but useless.

On a more personal note, bigmak, what is your interest in my personal business? Perhaps you could share with the rest of us.

Murph

Murph;

If you crack this let me know. I can't make picking the correct paceline work either. Pace Makes the Race, Brohamar didn't work for me...even though it sounds so simple-at least Pace Makes The Race. I kept track models and used guidelines for paceline selections. Maybe it's the numbers I use but after a couple years I came to the conclusion that I just don't have the aptitude for it. If it was really that easy I would think a computer would leave clues or cut the track take down substantially but all the mechanical studies for HTR etc show none of pacelines stands out and many results are similar. One thing Brohamar did mention was to look for a supporting pace line, that really helps to find solid lower priced contenders and helps explain the types of races he said he does well on-mid-level consistent horses. I always look at the horses 2nd highest speed or total pace rating to see how it fits in today's race.

I agree with Dave about the last line from my experience. Although it can can be a tip-off of things to come based on how the horse ran what I do is split it out from the rest of the running lines. So instead of best of last 3 or 4 I'll do best of races 2-3 or 2-4 or 2-10 for example and take the last line into context when appropriate. It helps me. Then again CJ's 4 rules horses do very well (last on today's surface with a couple other rules).

HTR has some great reports in their free newsletters on paceline selection and sounds like Dave is working on something good also.

[Murph]

Murph;

If you crack this let me know. I can't make picking the correct paceline work either. Pace Makes the Race, Brohamar didn't work for me...
[kutit}
HTR has some great reports in their free newsletters on paceline selection and sounds like Dave is working on something good also.

Yes HTR does the work, Dave is some kind of monster in attacking a handicapping problem. If anyone can crack the code he can do it. I think we are on the right track here.

I didn't really want to tell everony Niko, but since it's obvious there are many questions raised by the Brohammer Pace Makes the Race methods, I'd lke to share the advice I gave to Jack at his request. Thanks for helping me decide to share this message.

Murph

[Murph]

Re: Your study
Quote:
Originally Posted by jasperson
I sure that you have done much more work on this then I, but I think we have come up with the same conclusion. What is the point of doing all this work when straight handicapping does better.
Jack:bang:

Beautiful Jack, point made. There is alot of good in the Brohammer work and converting to FPS gives a more accurate estimate. My point to continue is that it's redundant to refine an estimate. I feel have a better chance with a whole number that is properly adjusted than a fraction of the original estimate to pretend that I have more accurate numbers.

For the purpose of bridging the gap with what you have done with this I'll offer this advice. I don't know which data files you use but multi-caps and procaps from the BRIS/TSN product line are very effecient for me. You can get all of the files with XRD (exacta results data) for around $60 a month.

This leaves the problem of adding the beaten lengths to the results for making your own figures. The workaround is to make a race figure for the winner then make the speed figure on handicap day when you can enter the BL info for each runner. The race figure will need to be accessable from your notes or a data base or file that you use to keep records.

The free sotware for using those files is some of the best work I've seen for someone advancing the FPS methodology. I used that program for examples on how to practically apply the good work in the middle of the mess to my own method. It takes at least a full meeting at a single track to begin to get relevant feedback from results. You can see how the FPS ratings are properly applied at a glance and as you let it eat you can see where they best hold up and where they fail most often.

First thing you will notice if you try this is that the labels for many ratings match the ratings you have learned thru the Brohammer work. Just accept that they are correctly labeled and applied, don't waste time trying to verify their work, whatever else it is, it's what too many people are using. The guys typing so passionately to you about defiling Brohammer are looking between the cracks of flawed numbers for their edge. Tom Brohammer KNEW this when he wrote his books.

One of the reoccuring themes in his Pace Makes the Race work was that the method he used to get a single race winner may or may not apply in a similar circumstance the next time you see it. It was like he put a discalimer on almost every real life example he used for instruction in the book. Draw your own conclusions.

Some of the guys typing in here LOVE/HATE me because one of the things I like to do is scoff Brohammer. They are way too invested to not care, they feel they have alot at stake to change their ways now. I want to take advantage of every dime those guys put in the window. They know I can do it too.

Jack, if this all seems over the top, just put it in the circle file. Take a look at Thorostats results for the inner track meeting at AQU this season. This will make the second year in a row of excellent results at AQU.

I have a handsome profit to show for my work at that hell hole of a track and I know I took some Brohammer food off of the table when they were terribly wrong using those numbers. They eat a little less when I have to leave some $$ behind.

Let me know if my suggetions of using multicaps software might help you expand your program so the work you've done is not wasted, but applied properly to a handicapping method where you find your edge, outside of the box. That box is always very crowded and it feels very good to step outside of it.

JeffP is an excellent handicapper and software developer. The first line of his multi page post referred to researchers and actuaries. I've never met an acutary at the horse track, guys! Enough said.

Keep in touch man, you can tell by now, if you yank my chain, the whistle might stick. ;)

Mike "Murph" Powers
__________________
Visit Thorostats (http://www.thorostats.com/) Today!

[jasperson]

The Sartin group had 10 people working on the same problem and therefore could devote more time to handicapping a race card. Give me just one race to research and I can run all 10 pace lines for each horse. Check the charts on the last few races of each horse check out as much as I can the proclivities of trainers and jockeys etc. After that I think I would have good idea it this is a playable race and at what odd. Dr. Quirin in horse book stated that he thought a syndicate of 4 handicappers were required to handicap a track
1. To compute par times and track variants and evaluate track bias's
2. To take trip notes
3. To track trainer and jockey proclivities
4. This one I am not sure of but I think it was to evaluate the physical condition of the horses.
I for one think we do need to cooperate more with each other, but have you ever tried to get cooperation form any of your race track cronies? They are will to use your work as long as they don't have to do any of it. The ones that do try to solve the problem do it short term and then discard it and go to something else. There are people on this list that do try to help like Murph but problem is we don't always play the same tracks. Also I would like to see a study on how many handicappers do like I do: that is buy a voucher and then make all their bets off that voucher. I never cash mine until it is over 100% greater than me original purchase( I rarely get there). That way I always know where I stand for day and year.

[Murph]

I like to work with a quartely voucher and take cash off of it as needed to maintain my roll. Finance the next quarter and record it in a place where I can keep my affairs private. Very satisfying work.

Imagine the problems a group like that might be lead to when human nature is naturally introduced into such a close knit circle. God bless Pirco, they helped everyone to enjoy racing more than could have been possible without their hard work and sharing ways.

[Niko]

Thanks for sharing your post Murph. I always like to hear both sides of experiences with handicapping approaches. I often learn just as much from failures as success with various approaches. Sometimes when Guru's or leading authorities state opinion as fact, or have facts to back their approaches (not pointing anyone out-general observation across life) those that disagree or are unable to duplicate the results are chastized without much thought as to what they're actually saying.

Each of us has to find our path of which there are a few. Not as many in horse racing but a couple.

[ranchwest]

I wrote my own program so that is not a problem. It displays the 10 pp lines with type of race distance speedrating and comments along with the type and distance of todays race. Attached is my pace line for tomorrow at Aqu.
Jack

You only had 5 of 9 on top.

You should move onto something better. :lol:

[ranchwest]

My mistake, looks like you must have had the post numbers instead of the official numbers.

ONLY 4 of 9. :lol:

Paid off $30.40 on $18 win bets. ROI was "only" +0.68.

[completebill]

After reading numerous threads over and over regarding odds lines ...the expertise on this board is amazing, sometimes to the point of leaving me unsure if left is right or vice versa or if it even matters...so I will try to keep my question simple.

What do you think is the best way to determine a fair odds line based on the following:

Horse
1) 56 FPS
2) 55 FPS
3) 54 FPS
4) 53 FPS

I believe there may be more info needed from a db, such as std dev but am unsure.

Most appreciative of any info but alas my mind is overly basic...so kindly keep it rather simple.

Many Thanks

Gordon Pine, @ Netcapper.com, sells a nifty little utility program (cheap and easy to use!) which calculates an odds line from entry of any "numbers" you may enter, including a speed rating, a "power rating", a win %, etc. I don't see why a velocity rating couldn't also be used.

[jasperson]

I thought I did well selecting the proper pace lines,but I reran my program and told to select only the last races. It picked up Mr Nobody in the first, and picked up Cizi in the 6th along with 5,7,8,9. So by selecting only the last race It had 6 winners with total winning of 42.20.
Jack:)

[jasperson]

I did 2 track today and the success at aqu was not duplicated at tp. It had a paltry 1 winner that paid 3.80. Speed rating was dominate there for 4 winners and winnings of $20.60 for a profit of $2.60. Attached is the pace line for tomorrow at Aqu.

[jasperson]

Attached are the pace lines for gpx and lrl. I made the selections?

[jasperson]

For the 3 pace lines I posted for 3/27 here are the results
Lrl 3 wins winnings $14.20 for a loss of -$3.80
Aqu 2 wins winnings $11.90 for a loss of -$7.10
Gpx 6 wins winnings $51.60 for a profit of $33.60
Over all a profit of $22.70 and 41% winners
I am confused :confused: by these results. It beat Gpx like a drum but didn't do well at the other track. I will post pacelines for 4 tracks saturday and see how I does with them. I always thought Saturdays were the most difficult to make a profit.:confused:
Jack

[traveler]

I won't try and critique your method, it looks like you've put a lot of work into it. My only thought is, I would never try and play every race on a card. Some races are playable, some are not. I may be wrong, but from some very good handicappers here that is the impression I get from them too. Some races may not be gettable by pace methods. Just a thought.Good luck.

[jasperson]

Attached are the pace lines for Saturday

[PaceAdvantage]

Yes HTR does the work, Dave is some kind of monster in attacking a handicapping problem. If anyone can crack the code he can do it. I think we are on the right track here.Is the Dave in question here Dave Schwartz? If so, he has nothing to do with HTR. If this is another Dave, then never mind.