What do/would you do with a complete database?

[thorobasePA]

I'm investigating database technologies. But I want to hear about what horseplayers would want from a racing database.

Let's assume you have all the data. Every chart, every track, every year since 1991. What would you use the database for? Which of these matter most to you?

* Printing up your own PPs for the next days racing?
* Creating your own custom charts?
* Creating your own speed ratings?
* Experimenting with algorithms?
* Statistics (Trainer, Horse, Jockey form data, combinations etc.)?
* Odds analysis/line projections?
* Other (please describe)?

What would not be important to you?

Would it have to be very fast i.e. instant results from any query you can think of? Or just pre-scheduled reports? Highly automated or would you like to manually play around with the data?

How about if you could share and exchange queries and algorithms with other horseplayers?

The more you can describe, the more it helps me. Thanks!

[thorobasePA]

No-one? Alright, I'll scrap this idea.

[sjk]

I am surprised that no one replied to the question. When I first joined PA there was lots of discussion about databases and how to build and use them and many people seemed to be going to the trouble of parsing charts.

It seems that all except the few that have persisted and done it on their own have lost interest.

[Ghostrider5]

Their are several good database programs out right now. Such as paceplus, jcapper, htr, alldatani(still under development) and other hdw software vendors. If you plan on creating database software. Your'e going to need propriety factors nobody else is using. Also the ability to use bris or multicap files. Our some other kind of cheap datafile and result combo package. Top notch customer service to answer questions. Currently i am using jcapper as my database program of choice.

ghostrider5

[Warren Henry]

I'm investigating database technologies. But I want to hear about what horseplayers would want from a racing database.

Let's assume you have all the data. Every chart, every track, every year since 1991. What would you use the database for? Which of these matter most to you?

* Printing up your own PPs for the next days racing?
* Creating your own custom charts?
* Creating your own speed ratings?
* Experimenting with algorithms?
* Statistics (Trainer, Horse, Jockey form data, combinations etc.)?
* Odds analysis/line projections?
* Other (please describe)?

What would not be important to you?

Would it have to be very fast i.e. instant results from any query you can think of? Or just pre-scheduled reports? Highly automated or would you like to manually play around with the data?

How about if you could share and exchange queries and algorithms with other horseplayers?

The more you can describe, the more it helps me. Thanks!
I would want ALL historical data for each horse (basically all PP data and all results data) for all his activities going back as far as possible.

I would use such a database to do my own research into relevant patterns. I would develop my own factors (speed, class, etc) in ways that would be different from what everyone else was doing. It would need to be fast enough that I could enter a structured query and get back the results while I still remembered why I asked the question. Instant is not necessary.

I would want it to be capable of automatically updating itself with the daily data feeds from whatever data provider was the original data source.

What did you have in mind.

[ranchwest]

Their are several good database programs out right now. Such as paceplus, jcapper, htr, alldatani(still under development) and other hdw software vendors. If you plan on creating database software. Your'e going to need propriety factors nobody else is using. Also the ability to use bris or multicap files. Our some other kind of cheap datafile and result combo package. Top notch customer service to answer questions. Currently i am using jcapper as my database program of choice.

ghostrider5

Isn't AllDataNI a spreadsheet program?

[thorobasePA]

I would want ALL historical data for each horse (basically all PP data and all results data) for all his activities going back as far as possible.

I would use such a database to do my own research into relevant patterns. I would develop my own factors (speed, class, etc) in ways that would be different from what everyone else was doing. It would need to be fast enough that I could enter a structured query and get back the results while I still remembered why I asked the question. Instant is not necessary.

I would want it to be capable of automatically updating itself with the daily data feeds from whatever data provider was the original data source.

What did you have in mind.

Thanks Warren and Ghostrider5. I must watch some of those JCapper videos to see what they are doing.

I don't have any specific plan in mind but wanted to know what horseplayers look for in database handicapping. I was particularly interested in the idea of sharing (queries etc.) but it seems for many they are mostly interested in tackling things on their own. If I was going to go down the database route, I probably would try to do something very different rather than replicate what others have done.

[raybo]

Isn't AllDataNI a spreadsheet program?

Yes, it's an Excel spreadsheet, at this time, but the future of AllData-NI will include an Excel database and batch processing. It's a ways down the road, but, that's the plan for it's future.

[aratrace]

Interesting thread.....I'll add my 2 cents.

For me a Racing DB is a Query Tool to produce effective outputs!

The basic problem with most database' is that they use the same data supplied by Equibase and before them both DRF/Equibase and before that DRF.

Sub-providers of this data (BRIS, DRF, TrackMaster, and others) create and add certain proprietary information, such as Speed Ratings, Variants, Fractional and Pace Ability Numbers as well as their take on Breeding info. This proprietary information is quite similar across the broad base of these add-on data suppliers. This information has inherent flaws in its creation and when run through DB Queries generates exponentially flawed results and outputs. This speaks directly to the age-old axiom.....GIGO....garbage in - garbage out.

You cannot create accurate True-Track Times (Track Records), create Length-based Variants, compare pace or fractional splits across the broad base of North American Tracks if you cannot identify & quantify by: Race Class, Surface, Surface Condition, Turn Radius, Banking, Run-up to timer, number of turns, distance to first turn, distance to finish from final turn and Class based Deceleration Tables. Without this ability you cannot distinguish yourself from the masses, whether mechanical or computerized calculations.

It took me 30 years to create and validate the above adjustments, another 5 years to generate tables and a computer program that interact with any BRIS/Equibase "purchased datafile". The last 2 years have been for testing, debugging and tweeking the program.

After I determine the desired outputs and queries to generate said output I'll consider building or acquiring an extension to my existing and growing DB.

Regards,
EJD

[InControlX]

I think the major advantage in maintaining your own database and your own examinating software is that you can test all sorts of weighting ideas and performance measuring methods that are "out of the box" from mainstream methods. The bad part is that this is a sink hole for your time.

I gave up on elaborate multi-factor corrections to "official" published split times years ago, concluding that while relative pace at a given competition level is very important, the absolute times are masked in random factors and are of little value. Of course, and here's the disclaimer... I'm not writing this from a yacht!

ICX

[CBedo]

For me, the database is nothing more than a repository of information to do further research. As much as I've done, I'm still in the early stages for sure. I wish I could use it to come up with "the" magic formula, but I'm not expecting that to happen.

The database portion is pretty easy; it's coming up with the analytic framework to use it that's the fun part.

[thorobasePA]

Thanks to all for replies.

They point to the belief that if you have all the raw data, there exists analysis techniques to discover significant wagering edges. Is this belief justified?

What I'm interested in is that the answers continue the theme of racing research being a purely individual pursuit, largely because of the fear of any advantage discovered being disseminated to the market and the edge being quickly eroded. Does this continue make sense?

The sheer volume contained within multiple years of racing results, the monopoly on the source, plus the complexity of both the structure of that data and investigating it, surely mean that the few who spend great expense (in terms of purchasing the data itself plus the time invested) are getting very little back, no?

If we are talking years (if not decades, as described in one response) as the time-frame to generate meaningful results, shouldn't there a more collaborative approach?

[sjk]

Thanks to all for replies.

They point to the belief that if you have all the raw data, there exists analysis techniques to discover significant wagering edges. Is this belief justified?

What I'm interested in is that the answers continue the theme of racing research being a purely individual pursuit, largely because of the fear of any advantage discovered being disseminated to the market and the edge being quickly eroded. Does this continue make sense?

The sheer volume contained within multiple years of racing results, the monopoly on the source, plus the complexity of both the structure of that data and investigating it, surely mean that the few who spend great expense (in terms of purchasing the data itself plus the time invested) are getting very little back, no?

If we are talking years (if not decades, as described in one response) as the time-frame to generate meaningful results, shouldn't there a more collaborative approach?


I have done quite well with the individual approach. Don't knock it if you haven't tried it.

[thorobasePA]

I have done quite well with the individual approach. Don't knock it if you haven't tried it.

That's a fair comment. Re-reading what I wrote I think the following:

The sheer volume contained within multiple years of racing results, the monopoly on the source, plus the complexity of both the structure of that data and investigating it, surely mean that the few who spend great expense (in terms of purchasing the data itself plus the time invested) are getting very little back, no?

should be:

The sheer volume contained within multiple years of racing results, the monopoly on the source, plus the complexity of both the structure of that data and investigating it, I would assume to mean that few who spend great expense (in terms of purchasing the data itself plus the time invested) are getting something significant back, no?

Yes, I'm making a flat out guess here but this is what my gut is telling me after creating a multi-year racing database and this was dealing with a free data source. And the effort and time involved in working with it the database I felt may have been too much.

sjk, you said yourself:

When I first joined PA there was lots of discussion about databases and how to build and use them and many people seemed to be going to the trouble of parsing charts.

It seems that all except the few that have persisted and done it on their own have lost interest.

Perhaps you could detail what you had to do (from simply building the infrastructure to doing the data analysis) to succeed and if a collaborative solution would have been of greater benefit?

[sjk]

I don't know that I could recommend what I did to others since it was something of a leap of faith that all of the effort would pay off in the end.

Happily for me although it took a couple thousands of hours of work to develop a database program that was effective I have been on autopilot for over 10 years and the effort was well worth it.

I was using free data for a good many years but once things had proven out the cost of data from HDW for my personal use is relatively insignificant.

As to the alternative of collaboration, I looked on it as a personal challenge to do it myself. In the end the effect on the odds of others playing the same bets could be significant if those others made large bets.

[JWBurnie]

Owner stats. Many may think I'm crazy, and see no value. But I honestly think these numbers would show some amazing %'s. I think you would find that the majority of trainer’s win-place-show %'s would be heavily favored toward <25% of their owners. As we all know, some owners have better stock, some are more aggressive, some make better claims, etc...



Owner w/ Trainer (s)

Owner w/ Claimer / allowance / stake

Owner at track(s)

Class Up/Down

Owner by Surface

Bred / Sale / Claim

........



Thoughts?

[windoor]

I would very much like to see a program that can import “selected data” from a bris file to build a database and have the capability of using Boolean Operators to test the profitability of different handicapping methods.

Everything I have looked at seems to insist pace and speed numbers are the only things that are important.

I have been trying on my own and failing miserably thus far.

Regards,

Windoor

[completebill]

To face reality in horse racing today, the key piece(s) of missing information is the VET being used by the trainer. My son is a former trainer, is still working in the industry, and is in constant touch with, and maintains extremely close relationships with, all the major trainers on the East Coast, and many others across the country.
He tells me that it is no coinciodence that the most succesful trainers in any given area (on any "circuit") are all using the same vets!!

[takeout]

To face reality in horse racing today, the key piece(s) of missing information is the VET being used by the trainer.That’s something I’ve been wondering about ever since I saw two trainers seesaw in the standings some years ago. I don’t know if it’s true or not but someone did say that the vet that was working for trainer “a” switched to trainer “b”.

[DeltaLover]

Just extracting the past performance data from a db is a very primitive use of it. Some of the things I am using my db is as follows:

1) From a UI I am able to query the db for row times, my speed figures, par times, trainer / jockey and sire stats etc

2) Using the PP and the charts in the end of the day I create a database of primitive handicapping factors calculating ROI and IV.

3) Run my genetic algorithms against the data base results optimizing my winner picker script that is created automatically in my DSL and storing the 'best' script in the db as an expression tree..

4) During the betting I keep in the database bankroll related data so using upcoming race predictions the platform can make Kelly based suggestion for the size of my next bet.

[BIG49010]

Owner stats. Many may think I'm crazy, and see no value. But I honestly think these numbers would show some amazing %'s. I think you would find that the majority of trainer’s win-place-show %'s would be heavily favored toward <25% of their owners. As we all know, some owners have better stock, some are more aggressive, some make better claims, etc...



Owner w/ Trainer (s)

Owner w/ Claimer / allowance / stake

Owner at track(s)

Class Up/Down

Owner by Surface

Bred / Sale / Claim

........



Thoughts?

This is a tough one, many owners have multiple partners, and the field that bris in particular provides cuts off many names and shortens them. Very tough to deal with other than a large owner like a Mike Gill. My 2 cents

[BIG HIT]

Just try see what angles work what methods work and if combination of angles\method are any good.Nothing ground shakeing for sure,
Trainer pattern's so many thing's variables i find very interesting also love it if read any kind of pp bris drf tsn and pdf

[acorn54]

i doubt that you could really do anything with the data publicly available as far as turning it into a profit if that is your goal. as far as answering questions about certain handicapping factors, yes a database could be of some merit.

[takeout]

This is a tough one, many owners have multiple partners, and the field that bris in particular provides cuts off many names and shortens them. Very tough to deal with other than a large owner like a Mike Gill. My 2 centsThat’s something I’ve noticed between the BRIS PP’s that I get and the track program. Both have their pluses and minuses but the track program actually has better (more complete) owner stuff. Example:
BRIS - Own: Fiddlestix Inc
TP - Owner: Fiddlestix, Inc. (A. Wolf)

[traynor]

One of the most useful aspects of database handicapping might be testing what you think you know to find out if it can be applied in the real world. For example, constructing queries to test an "overall profitable" spot play on selected samples. If you only look at the overall values, you may never discover that you are generating a 200% ROI in a few very specific situations, and running in the red on most others.

[aratrace]

I would very much like to see a program that can import “selected data” from a bris file to build a database and have the capability of using Boolean Operators to test the profitability of different handicapping methods.

Everything I have looked at seems to insist pace and speed numbers are the only things that are important.

I have been trying on my own and failing miserably thus far.

Regards,

Windoor


Windsor -

We are on the same page. It took a couple of years, and a son willing to do the programming for me and the end result is a Program that does what you describe. FYI, the programing language is Pearl.

For the past 4 years.....Every morning at 1am, my program logs on to BRIS and automatically downloads every new MCP (multicaps) datafile and every new .xrd file (results). The program then backs-up said downloads to a 2nd archive drive.

The Program then processes every MCP file against the results files and creates proprietary (Length-based) Speed Ratings, Variants for every Track, Distance, Surface and Surface Condition, after which it creates Final Time Abillity Numbers for every race run.

A Variant Directory is created and stored for each and every race and track.

When handicapping and wagering, I download specific MCP files from BRIS.

The program then acts on each file by pulling specific fields or ranges of fields, and proceeds to re-calculate these fields using my formulae, algorithms and adjustments. The Program then creates new and writes new datafiles that now include my science and outputs. Some of these changes are my Speed Ratings, Length Variants, Fractional and Final Ability times, Running Styles, Race Shape, Fractional and Final Pars and Breeding ability Values for FTS, Turf, Dirt, All-Weather, including distance ability.

Some of the files that are created can be used in the Multi-Caps, Lightening Profits and BRIS PP Generator programs. The uniqueness of each demands that a separate file must be written for each.

The Program also creates 3 other output reports in Excel format. They include a Pace Generator, 15 Pattern Angle's.

Not sure how long testing will continue, but the next step is combining said datafiles, results files, Variants files, etc. into a DB that will allow for the creation of IV based user-friendly value output and selections.

I use the Program for daily and Contest play, successfully! In a more perfect world, with a functional DB, I expect it will provide a narrower scope and increased success, with much less grunt work on my part.

Regards,
EJD

[PhantomOnTour]

Owner stats. Many may think I'm crazy, and see no value. But I honestly think these numbers would show some amazing %'s. I think you would find that the majority of trainer’s win-place-show %'s would be heavily favored toward <25% of their owners. As we all know, some owners have better stock, some are more aggressive, some make better claims, etc...



Owner w/ Trainer (s)

Owner w/ Claimer / allowance / stake

Owner at track(s)

Class Up/Down

Owner by Surface

Bred / Sale / Claim

........



Thoughts?
I follow owners very closely in NY and have found some solid combos. Problem is (as another poster put it) some owners form groups or multiple groups and they can be tough to follow. For example, Graham Motion had many live runners for a guy named Donald Adam in 2006-2007 and probably before that(TD Vance,Adriano to name a few) but I havent seen him have a winner for this guy in NY in 2 plus years...did Mr. Adam retire? pass away? form an ownership group with a new name? etc...

Some owner/ trainer combos are too well known and overbet:
Shug/Phipps
Dutrow/IEAH (sometimes-this guy still gets value)
Levine/Repole Stable
Violette/Klaravich Stable
Romans/Ramsey
Hushion/Schwartz
Mott/Live Oak Stud

But some others fly a little under the radar:
Rice/Obviously NY Stable (but she will be hammered in NY this summer)
Galluscio/Majesty Stud
Hough/Robsham

Alright, that's enough.

[ranchwest]

i doubt that you could really do anything with the data publicly available as far as turning it into a profit if that is your goal. as far as answering questions about certain handicapping factors, yes a database could be of some merit.

Seriously?

[traynor]

I recently completed a project that involved data mining of a database containing every harness race run in Australia in the last five years. More than 92,000 races, close to a million entry data lines. The most interesting thing about the study is that almost everything that is "common knowledge" about harness racing doesn't hold up to scrutiny over time. From recent good races, to close up finishes, to positive driver changes, to finish in the money last--all proved to be false and misleading assumptions based on limited data.

So the real value in database analysis (especially data mining) is both finding what works, as well as finding what doesn't work. If you have a database, it is fairly easy to run a data mining application to locate correlations and predictors. One of the easiest to use is available free from Waikato University in New Zealand:

http://www.cs.waikato.ac.nz/ml/weka/

While it may sound intimidating, it shouldn't. Data mining applications are used routinely by people who are relatively novices in the database field. The advantage is that once WEKA (or another data mining application) points out positive correlations (many of which are invisible to ordinary study) you can execute queries much more productively. That is, once you have an idea what to look for, it is a lot easier to find it.

[upset]

Windsor -

We are on the same page. It took a couple of years, and a son willing to do the programming for me and the end result is a Program that does what you describe. FYI, the programing language is Pearl.

For the past 4 years.....Every morning at 1am, my program logs on to BRIS and automatically downloads every new MCP (multicaps) datafile and every new .xrd file (results). The program then backs-up said downloads to a 2nd archive drive.

The Program then processes every MCP file against the results files and creates proprietary (Length-based) Speed Ratings, Variants for every Track, Distance, Surface and Surface Condition, after which it creates Final Time Abillity Numbers for every race run.

A Variant Directory is created and stored for each and every race and track.

When handicapping and wagering, I download specific MCP files from BRIS.

The program then acts on each file by pulling specific fields or ranges of fields, and proceeds to re-calculate these fields using my formulae, algorithms and adjustments. The Program then creates new and writes new datafiles that now include my science and outputs. Some of these changes are my Speed Ratings, Length Variants, Fractional and Final Ability times, Running Styles, Race Shape, Fractional and Final Pars and Breeding ability Values for FTS, Turf, Dirt, All-Weather, including distance ability.

Some of the files that are created can be used in the Multi-Caps, Lightening Profits and BRIS PP Generator programs. The uniqueness of each demands that a separate file must be written for each.

The Program also creates 3 other output reports in Excel format. They include a Pace Generator, 15 Pattern Angle's.

Not sure how long testing will continue, but the next step is combining said datafiles, results files, Variants files, etc. into a DB that will allow for the creation of IV based user-friendly value output and selections.

I use the Program for daily and Contest play, successfully! In a more perfect world, with a functional DB, I expect it will provide a narrower scope and increased success, with much less grunt work on my part.

Regards,
EJDHow do you make your variants? Doing that many tracks you must just average the races or it would take all day

[acorn54]

Seriously?


that has been my experience. i have been using jcapper since november of 2004 and i have never come up with any combination of factors available in the datafiles offered for sale by bris or formerly tsn that show a positive roi, over any length of time.

[JustRalph]

that has been my experience. i have been using jcapper since november of 2004 and i have never come up with any combination of factors available in the datafiles offered for sale by bris or formerly tsn that show a positive roi, over any length of time.

???????

How you can say that........ I don't get it. Is this really what you meant to say?

read back what you wrote. Over any length of time ? Any factors?

[aratrace]

How do you make your variants? Doing that many tracks you must just average the races or it would take all day

Upset -

The Variant creation process is totally automatic, processed daily immediately after downloading the MCP files and the corresponding Results files.

The program acts on True Track Times Tables (every track, distance and surface) that create "Speed Ratings" and then on corresponding "Race" class tables (every track, distance and surface). All calculation and output values are created using deceleration models and are reported as 1 point = 1 length at all distances. Deceleration models include "number of turns" and "turn radius" adjustments based on each and every tracks class rating.

eg. Speed Rating + Race Class Value.

Each Variant is stored in a group (one of ten) and the average calculated for like surfaces, distances and conditions. A number of validation checks are made to ensure an accurate Variant is reported for each group. Anomalies are eliminated.

The Variants are created and stored in a file for 10 potential categories. Sprint Dirt Fast, Sprint Dirt Wet, Route Dirt Fast, Route Dirt Wet, Sprint AW, Route AW, Sprint Turf Firm, Sprint Turf Wet, Route Turf Firm, Route Turf Wet.

The process enables processing BRIS files using my stored Variants to adjust all PP's for each horse. Having an accurate "length" variant allows the program to create fractional ability values as well as final time values and the MAX projected Ability number for each horse.

The program essentially writes my fully adjusted values back to the BRIS file which I can then view in the BRIS PP Generator and a few Excel reports.

The key to it all is the mathematical accuracy of the True Track Times and Race Class Value tables. They include adjustments for Run-up, Radius, Number of Turns per distance, etc. All-in-all a huge advantage.

Hope this answers your question.

EJD

[SOUTHERN SLEW]

I purchased Jcapper last October 2009 after having tried several of the more popular public programs. I have always used a black box type program in the past so when I purchased Jcapper, the idea of Data base handicapping was entirely new to me. I can honestly say that Jcapper and DB handicapping has opened my eyes to an entirely new concept of handicapping, it gives you the ability to analyze the raw data and formulate reliable betting strategies that have proven to be successful in the past, therefore giving you a tremendous edge over your fellow handicappers who utilize traditional methods that are not adaptable to the constant changes in surface fluctuations, track trends and the myriad of changes that the handicapper is always subjected to.

So yes I would highly recommend a DB software , I have traded the financial markets for many years and have always had success using powerful and highly sophisticated DB software to decipher the best trading strategy. so Jcapper was a no brainer when I first discovered it on the PA site as it offers the same high level of sophistication and to a certain degree in my opinion a very sophisticated Artificial Intelligence platform, in that it has the ability learn from the raw data therefore adapting to changes , isolating factors that have proven to produce high ROI and ultimately producing a unique methodology, that not only fits your style of play but gets you more winning tickets at the track and payoff that you always dreamed of.

In Summary I would without hesitation highly recommend any well designed DB software and it does not have to be Jcapper, just choose one that you are comfortable with, or select one that offers the criteria you demand, but from someone that has been handicapping horses for more than 20 years, you can take my word to the bank, Jcapper is the best I have seen, I have hit more longshots with Jcapper than I have ever had with any other handicapping program, 30-1, 10-1, 5-1 etc and the consistency is mind boggling. Just remember that the data is all the same its what you do with it that matters and in Jcapper, I have discovered what I consider a unique Artificial Intelligence ability to learn from the data, so like anything else in life the more data it is subjected to, the more proficient it becomes.

[ranchwest]

that has been my experience. i have been using jcapper since november of 2004 and i have never come up with any combination of factors available in the datafiles offered for sale by bris or formerly tsn that show a positive roi, over any length of time.

Sometimes you have to think outside the box to come up with profitable plays.

What is valuable about data files is not what comes in them, but what is created from them.

[46zilzal]

In a parimutuel game, DO NOT do what the crowd is doing.

[acorn54]

???????

How you can say that........ I don't get it. Is this really what you meant to say?

read back what you wrote. Over any length of time ? Any factors?


i can say that ALL of my racing angles that show a positive roi over years of testing in the jcapper software have proprietary factors unique to jcapper.
NONE have ONLY the publically available factors in the bris data files and the formerly tsn data files. that is my experience and i stand by it.
if you could make a profit exclusively from the bris data files or formerly tsn datafiles you don't have to purchase a database software like jcapper for many hundreds of dollars you can just buy a database software like the whatif analyzer for a measly seventy dollars https://www.whatifanalyzer.com/WIAhome.html

[ranchwest]

i can say that ALL of my racing angles that show a positive roi over years of testing in the jcapper software have proprietary factors unique to jcapper.
NONE have ONLY the publically available factors in the bris data files and the formerly tsn data files. that is my experience and i stand by it.
if you could make a profit exclusively from the bris data files or formerly tsn datafiles you don't have to purchase a database software like jcapper for many hundreds of dollars you can just buy a database software like the whatif analyzer for a measly seventy dollars https://www.whatifanalyzer.com/WIAhome.html

There's no reason why a person can't create proprietary factors out of most any data files. There's nothing magic about it. It's typically a lot of hard work, but it can be done.

I know I've programmed a lot of ideas I've read in books or online. I'm sure that very few people if anyone has programmed some of these concepts. I find them very useful and I'm confident that these ideas are not in common use.

That's the key -- to use values that work and are not in common use.

Then when you find an odds-on favorite that is the unlikeliest for the win, that's sweet.

[Dave Schwartz]

-- to use values that work and are not in common use.

The biggest problem is that parts of everything are used by just about everything.

You are correct, of course. The key is a unique view point into the races.

[ranchwest]

The biggest problem is that parts of everything are used by just about everything.

You are correct, of course. The key is a unique view point into the races.

You are correct that parts of nearly everything are used by just about everything.

However, transforming data into different forms can be complex.

Let's take pace. There's AP, SP, EP, FX, %E, %M, deceleration, etc. They all mean nothing at all -- unless you know when, how and why to employ them.

There's obviously not many factors that significantly impact odds, so using other factors offers the potential for good value.

[46zilzal]

[QUOTE=ranchwest

Let's take pace. There's AP, SP, EP, FX, %E, %M, deceleration, etc. They all mean nothing at all -- unless you know when, how and why to employ them.

[/QUOTE]
MOST of these are redundant. Sprints are all about 2nd call, routes SP and turf deceleration.

[raybo]

MOST of these are redundant. Sprints are all about 2nd call, routes SP and turf deceleration.

You might want to add to your post, the words "in my opinion", unless you're "all knowing", which I believe there exists none on the face of the Earth.

[Dave Schwartz]

There's obviously not many factors that significantly impact odds, so using other factors offers the potential for good value.


I would have to seriously disagree with that statement.

Factors such as "last race Beyer," "Quirin Early Speed Points" and "morning line odds" have a huge impact on the tote.

In fact, IMHO, there are very few factors that do not impact the tote in a noticeable fashion.


Regards,
Dave Schwartz

[raybo]

I would have to seriously disagree with that statement.

Factors such as "last race Beyer," "Quirin Early Speed Points" and "morning line odds" have a huge impact on the tote.

In fact, IMHO, there are very few factors that do not impact the tote in a noticeable fashion.


Regards,
Dave Schwartz

I agree. There are many publicly available factors that, in and of themselves, dramatically affect the final odds.

The tote reflects changes in odds throughout the betting cycle. Whether those bets were made by expert handicappers or less knowledgeable handicappers or even non-handicappers.

Expert handicappers, IMO, combine many public factors, in various ways, creating composite factors, sometimes several composites of different types, weighing those composites against the rest of the field and against the track, surface type, it's condition, the distance of the race, the probable shape of the race, which horses will benefit and which will not, etc.. Then they add another factor, the current odds and anticipated changes, prior to the final odds being calculated. This final analysis is a result of all monies wagered or anticipated, whether individual wagers were the result of single factors, or composites, or no factors at all.

All factors, whether single or composites or whether publicly available or not, have impact on the final odds. Some publicly available factors have more impact on expert handicappers' wagers than on non-expert handicappers' wagers, and some have impact on less knowledgeable handicappers and non-handicappers.

One cannot, IMO, single out any factors as being useless or unimportant. They are all used in one way or another by different types of wagerers.

And, all of the different types of wagerers create the pools and determine the final odds.

[ranchwest]

MOST of these are redundant. Sprints are all about 2nd call, routes SP and turf deceleration.

That's a significant agreement with my point. There's a lot of numbers that are used by a minority of people and many of those numbers require an understanding of how to use those numbers.

I make a wider use of the Brohamer/Sartin numbers, but I have a similar opinion in that I just scan through most of the numbers -- there's usually only a few that are useful for a particular race.

[ranchwest]

I would have to seriously disagree with that statement.

Factors such as "last race Beyer," "Quirin Early Speed Points" and "morning line odds" have a huge impact on the tote.

In fact, IMHO, there are very few factors that do not impact the tote in a noticeable fashion.


Regards,
Dave Schwartz

You are correct that Beyers, Quirin Points and Morning Lines have a huge impact on the tote.

However, if every factor had a significant impact on odds, almost nobody could legitimately pick a longshot -- it would be pure luck. I know for certain that a lot of people are picking longshots without it being luck.

[delayjf]

I have always wanted a database that would allow me to do indebt study into trainer patterns. I'm not only looking at a number cruncher but also a tool that would allow me to examine the PPs of horses in a particular catagory. Say for example the PP of the last ten horses of a specific trainer that won off a layoff.

[Tom]

You have it - HTR exports.

[delayjf]

Tom,

I approached Ken a few years ago with my idea to see if it was feasible using HTR. There were two issues that got in the way.

One, at that time, htr didn't export all of the raw PP data necessary for my idea to work. I believe you could only export the some of the race / horse data and the ratings / rankings HTR produced from the HDW data - not the raw PP data one could download from say bris. I looked at some of the export files and I think you can now download running line data, but I'm not sure if it would give me everything I'd need like the PP data from Bris.

Two, the other issue is Access file size / record limitations. It seemed like the best way to proceed was to use SQL. Since I have neither the expertise or the money, I've tabled the idea for now. A horse player can still dream though.