A system with its own past performance

[Sy Abend]

Does anyone know of a book or system for sale that comes with an extensive workout of results? I’m referring to something like a year’s worth of results with a bet-by-bet listing of Date, Race, Track, Horse’s Name, Win, Place and Show. I believe this is the best way to graphically see things at a glance like losing streaks and bunches of winners. It’s also an easy way to spot if there are any way-out longshots that skew the results.

Any suggestions would be appreciated.

[boomman]

Sy: You might check out www.racehorserunner.com (http://www.racehorserunner.com/) I know Pascale well (he runs site) and he might be able to provide something along those lines for a fee...

Hope this helps you-Boom

[ryesteve]

Does anyone know of a book or system for sale that comes with an extensive workout of results? I’m referring to something like a year’s worth of results with a bet-by-bet listing of Date, Race, Track, Horse’s Name, Win, Place and Show.
A warning: if you do find a system that comes with that kind of documentation, there's an excellent chance that the system's rules were backfit to the sample of races that they're using to demonstrate its performance.

[formula_2002]

A warning: if you do find a system that comes with that kind of documentation, there's an excellent chance that the system's rules were backfit to the sample of races that they're using to demonstrate its performance.
well said

Just have them post their picks right here on PA (the "selections " forum)
It's better than them offering you a good review by the Phillips report.

Joe M

[njcurveball]

If you were to get something like this, how many plays would you want to see?

Three (3) horses a day would be over 1,000 for the year. In a typical longshot system with a 10% win, you would see 100 winners and 900 losers.

If you think you are crafty enough to whittle through all that data and make it better, or even trust it for the following year, I suggest you get some computer downloads and start setting up your own database.

As was said already, many of the great systems were simply put together by pattern recognition of winners that already crossed the finish line.

Perhaps the greatest of all time was DOSAGE. What good is it knowing it has shown a flat bet profit on the last 30 Derbys, when it has lost the last ten (just as an example).

If you think a static system from a 30 year old book still holds up today, you maybe right for a week or even a lucky month at some tracks. Back then it was "intelligent" to make a grid and check off a box if the horse ran in the last 14 days, finished within 3 lengths, had a top 10 jock, etc.

We are in 2007 and the edge is with people who have ratings that are not available to others. It is just accepted knowledge that Beyer figures were winning 20 years ago and they are losing today.

Jim

[formula_2002]

If you were to get something like this, how many plays would you want to see?

Three (3) horses a day would be over 1,000 for the year. In a typical longshot system with a 10% win, you would see 100 winners and 900 losers.

If you think you are crafty enough to whittle through all that data and make it better, or even trust it for the following year, I suggest you get some computer downloads and start setting up your own database.

As was said already, many of the great systems were simply put together by pattern recognition of winners that already crossed the finish line.

Perhaps the greatest of all time was DOSAGE. What good is it knowing it has shown a flat bet profit on the last 30 Derbys, when it has lost the last ten (just as an example).

If you think a static system from a 30 year old book still holds up today, you maybe right for a week or even a lucky month at some tracks. Back then it was "intelligent" to make a grid and check off a box if the horse ran in the last 14 days, finished within 3 lengths, had a top 10 jock, etc.

We are in 2007 and the edge is with people who have ratings that are not available to others. It is just accepted knowledge that Beyer figures were winning 20 years ago and they are losing today.

Jim

If you're refering to me, I'll say this. In any set of 80 factors, handicappers, systems etc, you are bound to find a few profitable in 10, 100, 1000 races.
They start to weed themselves out as the number of plays increase.
Just how many plays are need to weed them down to "ZERO" value, I'm not certain, but I would say that 1,000,000 times the size of my 300,000 horse data base should do the trick. but then again I not too sure about that. ;)

The point is, nothing last forever and there is good reason for it.

[Overlay]

If you think a static system from a 30 year old book still holds up today, you maybe right for a week or even a lucky month at some tracks. Back then it was "intelligent" to make a grid and check off a box if the horse ran in the last 14 days, finished within 3 lengths, had a top 10 jock, etc.

We are in 2007 and the edge is with people who have ratings that are not available to others. It is just accepted knowledge that Beyer figures were winning 20 years ago and they are losing today.

You're absolutely correct about the weakness of checklist systems that deal in fixed, yes-or-no criteria like the ones you cite. But even ratings not available to others (today) will go through the same cycle of obsolescence and required replacement as word gets out about them, if they are designed or used solely to "pick the winner" through a process of elimination or ranking, without providing insight into the actual winning chances of the horse in question, or of its competitors.

[Kelso]

[QUOTE=formula_2002]
the size of my 300,000 horse data base should do the trick
QUOTE]


Do you have race data on 300,000 horses, or do you mean data on 300,000 horse races? If 300,000 horses, are standardbreds, Arabians and/or quarter horses included? How long has it taken you to assemble so large a database?

Thank you.

[betchatoo]

Since 2004, I have collected data on between 9,000 and 10,000 races annually. These are all thoroughbreds and each year represents approximately 70,000 horses.

[PriceAnProbability]

Does anyone know of a book or system for sale that comes with an extensive workout of results? I’m referring to something like a year’s worth of results with a bet-by-bet listing of Date, Race, Track, Horse’s Name, Win, Place and Show. I believe this is the best way to graphically see things at a glance like losing streaks and bunches of winners. It’s also an easy way to spot if there are any way-out longshots that skew the results.

Any suggestions would be appreciated.

http://www.redboard.com (http://www.redboard.com/) might be a guess. If that doesn't work, maybe backfit.com might.

[Sy Abend]

Thanks for the links, but they're both dead-ends. If you get a chance, could you please review them?

[banacek]

Thanks for the links, but they're both dead-ends. If you get a chance, could you please review them?

I love this board!

[Murph]

http://www.thorostats.com

[Kelso]

Since 2004, I have collected data on between 9,000 and 10,000 races annually <snip> each year represents approximately 70,000 horses.


Just trying to be certain ... same as with my question to Formula. Do you mean 70,000 runners in those 9000-10000 races ... with a lesser number of different horses, most of whom race several times a year? (Gotta admit, even if they're not 70,000 different animals ... that's a whole lotta data!)

Thank you.

[betchatoo]

Just trying to be certain ... same as with my question to Formula. Do you mean 70,000 runners in those 9000-10000 races ... with a lesser number of different horses, most of whom race several times a year? (Gotta admit, even if they're not 70,000 different animals ... that's a whole lotta data!)

Thank you.

You are correct. That's 70,000 total runners and includes horses that have run multiple times, they are not unique, different horses

[singunner]

Since we're on the subject, about how many unique horses run per year? If you can't do it across all US tracks, how about an average number of unique horses at a single track per year?

[betchatoo]

Since we're on the subject, about how many unique horses run per year? If you can't do it across all US tracks, how about an average number of unique horses at a single track per year?

I don't have all my info imported into a database, so I can't do that type of query.

[sjk]

For 2006 I show 72450 which should be in the ballpark.

[PriceAnProbability]

Thanks for the links, but they're both dead-ends. If you get a chance, could you please review them?

Ugh!!

"Redboarding" and "backfitting" are the names of the technique you are using.

Another is "regression analysis."

The problem with these paradigms is, well, shift happens.

[sjk]

Doesn't anyone else want to take a stab at how many unique horses ran last year?

Equibase has made database handicapping an endangered endeavor.

[Rook]

Doesn't anyone else want to take a stab at how many unique horses ran last year?

Equibase has made database handicapping an endangered endeavor.

SJK,
I've got 72508 horses that ran last year. I hope the missing 58 horses aren't going to drive you crazy like the next poster who is going to break me into a sweat by telling me that there were actually 72566 horses that ran :eek:

[gm10]

Rook, I just saw in another thread that you started a few years ago, that you wrote your own software to come up with an odds-line. Do you still do that? How are you finding it?

I'm doing pretty much the same as you, btw.

[Rook]

Rook, I just saw in another thread that you started a few years ago, that you wrote your own software to come up with an odds-line. Do you still do that? How are you finding it?

I'm doing pretty much the same as you, btw.

Hi gm,
Yes, I am still using my software. I am quite happy with it although I am still always on the lookout for ways to improve its performance.

Good luck with your program.

Rook

[skate]

say Rook;


are you including colts, because i ....

i came up with...:lol: :ThmbUp:

[banacek]

Hi gm,
Yes, I am still using my software. I am quite happy with it although I am still always on the lookout for ways to improve its performance.

Good luck with your program.

Rook

Nice to see you back Rook. I always enjoyed your posts. You were playing Woodbine pretty seriously if I remember correctly. How has the polytrack changed things for you (and your program)?

[Rook]

Nice to see you back Rook. I always enjoyed your posts. You were playing Woodbine pretty seriously if I remember correctly. How has the polytrack changed things for you (and your program)?

Thanks for the greeting Banacek. Any guy who bashes Woodbine's HPIBET is a friend of mine.:)

I absolutely love the new polytrack. Woodbine was my second most profitable track in 2006 and this was a dramatic turnaround because I was losing money on that temporary track.

And although I will likely never change my opinion of them being a bunch of incompetents and thieves, I have to give Woodbine credit for being an early adopter of Trackus. When you have 9 horses in a super ticket it sure makes it easy to follow what is going on.

Good luck in whatever tracks you are currently playing.