system analyst

[pepsi]

G'day,

have a couple of methods that i would like to test regarding past results. does anyone know of any software out there that could actually do the work. have not used the methods personally but would like to see how they hold up and if they actually turn anything over.

cheers,

[ryesteve]

It depends on what comprises the system. There's software out there with db capabilities, but the only way you'd be able to test these methods is if the software already uses all the variables you need. I'm going to guess that the chances of that aren't great. More than likely, you'd need someone with a db to do some custom programming and research it for you.

[pepsi]

it does depend on the perameters of the particular system. ok the reasoning behind this, i have a handful of paper and pencil methods that i have never used simply because of the time factor.(and if someones wondering...yes i will share this info with anyone who is interested) now if these methods could be past tested and if the wheels stay on any of them for a ROI no matter how minimal. these methods could be placed on cd rom format and selections would be easy peesy and these could be added to the coup...or am i just dreamin'

thanks for the reply ryesteve.

cheers mate

[ryesteve]

No, you're not dreaming at all. My belief is that ALL of these mass-marketed pencil and paper systems should be put through a database wringer to show exactly how they perform. If I remember correctly, Nunamaker (or somebody else) actually did such a study some years ago, and the results were quite ugly.

[jhilden]

I’ve dabbled a bit in running paper and pencil tests on my database. Programs like HTR and HSH are good for testing some of the systems but to really dig deep, you need to export much of the data and run tests through the database. The bad part is if you are starting out and do not have a database, you need to purchase it or wait a good year after subscribing to accumulate enough data to run the tests on.

As mentioned above, some of theses systems perform miserably when following system instructions as is. I only found a few that performed decent but did better when I dug into the data and found out were I could adjust the parameters of the system. By reviewing the details of the test, I also was able to uncover the tracks the system performs better.

One of the systems that I had success with was the paper and pencil Pro Simulcast System by Bob Pandolfo. I was able to query my database for the horses that met the conditions of one of the plays in the system. Once, I had all these horses, I was able to see how many won out of the sample. For the Maiden special weight (MSW) to Maiden claiming (MCL) system, the results were good with a ROI that was just below the break even mark. After carefully reviewing all the data, I was able to uncover the types of horses that bring the hit rate and roi down. Most notably, the MSW to MCL is a drop you are looking for, but using the system, you would select all horses meeting this drop if they met certain criteria – like beaten lengths at designated points of call. But by excluding horses that already had a history of a MSW to MCl drop further back in the PP’s (and obviously failed), the hit rate and ROI made this system a great play.

After testing systems in a database, you have to be aware of performing too much back fitting. To perform in the future, the system cannot be modified dramatically. In the example above, the flaw I found with the horses jumping back and forth between the MSW and MCl ranks applied to many horses. In a sample of more then 5000 starters, the exclusions knocked out about 500.

[ryesteve]

The bad part is if you are starting out and do not have a database, you need to purchase it or wait a good year after subscribing to accumulate enough data to run the tests on.
... or you can just partner with someone who already has the data. There are many such people out there.

[Light]

Programs like HTR and HSH are good for testing some of the systems

In what way can HTR test a paper and pencil system?

[jhilden]

How can HTR test a paper and pencil system? Well, it depends on the type of system that you want to test. The robot feature can be used to test simple systems but you would have to use the export feature of HTR and use a database program like MS Access if you want to test systems that contain more in depth factors. So, yes, you could say that just using HTR alone, there is not much in way of testing paper and pencil systems.

The MSW to MCl drop I mentioned above uses beaten lengths at the 2nd call and finish as one of the main selection criteria. HTR lets you export the running line information for the last ten PP’s for each horse. With this information I can test the system using the exact criteria set by the method. The improvement with using HTR data is that I can add in the class ratings or other proprietary information to the query and seek a more in depth analysis to improve the original system; if it can be improved.

Once I establish a good system or systems, I can export the races for the whole day to Access and have the program spot my plays for the day. Doesn’t mean I will blindly play all the spot plays, but at least I have some plays to look at more deeply to start the days handicapping.

I have some systems that require the use of a speed chart that is included with the system. I was successful in testing these systems by creating rules that converted the fraction times to the authors speed chart and then was able to build more rules that followed the systems steps. These types of systems crashed and burned with no survivors.

I bought many systems as long as they were cheap (Like on Ebay), for the sole reason of running them through my database just to see how they performed and could be improved.

I will add that oddly, I am not much of a systems player; I like to stick to my comprehensive method of handicapping. I guess since my education is in engineering, I like to test and see if I can improve current methods. I enjoy the research / handicapping part of horseplaying a lot more then I do with betting at the track – but I have to support my HDW subscription:)

[Light]

Excuse my lack of knowledge of HTR,but in order for one to test a paper and pencil system w/HTR,sounds like one would be required to purchase the expensive data that HTR runs on?

If so, it would make it impractical as a diagnostic tool to test paper and pencil systems whose authors would not want to purchase expensive data to test every unproven paper and pencil system they come up with.

[ryesteve]

Excuse my lack of knowledge of HTR,but in order for one to test a paper and pencil system w/HTR,sounds like one would be required to purchase the expensive data that HTR runs on?
Isn't it a given that no matter what software you use for testing, you're going to need data?????

[Light]

Yes

[pandy]

Thanks for the informative note about my Pro Simulcast system. We ran a computer test on the maiden claiming angle and the drop to non winners of 2 claimers and the system showed a loss of less than 1% with no tweaking at all, so it is a very good system. Your point about eliminating horses that already raced and lost for a tag is interesting and makes sense.

Pandy

[ryesteve]

Yes
Then why single out HTR as being "impractical" because one would need data to go along with it?

[Light]

Too expensive for the not interested in learning or using HTR. Access or Excel can do much more analysis than HTR without the cost or restriction of being tied to one source for data. This is obvious. It also comes accross like a plug when you recommend HTR or HSH for system analysis over the obvious.

[ryesteve]

Too expensive for the not interested in learning or using HTR. Access or Excel can do much more analysis than HTR without the cost or restriction of being tied to one source for data. This is obvious. It also comes accross like a plug when you recommend HTR or HSH for system analysis over the obvious.
I still don't know what you're getting at. It's not like Access is free. And I don't see anyone calling you a Microsoft shill for suggesting it.

Do I think HTR is the best tool for what this thread started out asking? No. Do I even think it's usable for this purpose? I have no idea, since I've never used it. But deciding it's unusable because it needs data, and then calling it too expensive while recommending Access, are totally bizarre arguments.

[Light]

Still dont get what I'm getting at? I wont buy a program like HSH for $600 without a trial. I wont buy a program like HTR that requires $100+ a month (even though KM came in 2nd at the NHC) without a trial.Would you buy a car without having it checked out or without a warranty? So why does someone promote these programs that are not primarily created for the task of analyzing self made systems when it is clear that Excel and Access were and can do more for less. No brainer.

[Hosshead]

...I wont buy a program like HTR that requires $100+ a month (even though KM came in 2nd at the NHC) without a trial.Would you buy a car without having it checked out or without a warranty? So why does someone promote these programs that are not primarily created for the task of analyzing self made systems when it is clear that Excel and Access were and can do more for less. No brainer.By the way, HTR is free, it has a free demo that comes with weeks of some free racefiles.
It has an export function for Access and Excel. This also allows you to export some proprietary info that you won't find in other data.
So if you're complaining about the Difference in data costs between HDW and say Procaps (about $60/mo. diff.), I can understand that.
Especially if you already have years of racefiles/results you want to use.
But if you found a way to make money with new (HTR) info. (exported to Access) that you can't find in other data, it would be worth it, no?

[ryesteve]

Still dont get what I'm getting at? I wont buy a program like HSH for $600 without a trial. I wont buy a program like HTR that requires $100+ a month (even though KM came in 2nd at the NHC) without a trial.Would you buy a car without having it checked out or without a warranty? So why does someone promote these programs that are not primarily created for the task of analyzing self made systems when it is clear that Excel and Access were and can do more for less. No brainer.
You're still not approaching the original question from the right direction. Anyone who's interested in testing a pencil and paper system is NOT going buy ANY of those programs, wait two years while they collect data, and then see what the answer is. The obvious course of action is to partner with someone who already HAS the software and the data. The question would never be, "Should I buy HTR/HSH to test a system"; it's, "Would a person who ALREADY owns HTR/HSH be able to test a system". In other words, your issues aren't applicable.

[pepsi]

just wanna thank you blokes for this insight. down under we have a program caled GTX (ozzy racing only) and it is what it is! i do not own it but some contacts do have it and of course like anything it cant use some parameters that r to vague, but all in all i have the impression that it is a good program. different strokes for different folks...i would still like to test some of the paper and pencil methods...anyone up for it??

cheers,

[AwolAtPA]

By the way, HTR is free, it has a free demo that comes with weeks of some free racefiles.


I am a 'seasoned' horse player whose research goes back to the days when BRIS charged ten cents a minute for connect time. My data cost really went down when the NEW 2400 baud modem were sold. The data was ONLY raw chart information which I had to write an extraction program to feed my data base!! In other words, I know the value of good data and the time it takes to collect and process the raw data.

enuf history, I want to describe the opportunity with HTR data for research. The system has MANY export files which are described (FREE) on the HTR website. So, Pepsi, can read those file descriptions to confirm IF his parameters are available in the data. Any '..pencil and paper..' method can be automated with a complete description of the processing rules. The tool for the processing may be Excel, Access, Visual Basic, or whatever the programmer knows how to use. However, if the rules included something like the '..Gray horses owned by top stable..' then HTR could not be used because top stable could be defined but the horse's color is NOT one of the data values.

OK, let us say Pepsi could just use the Robot feature of HTR. The issue about testing with FREE data. Well, Ken states that a user can contact him to get a FREE weekend to try out the system. However, the HDW data is available for the last fourty-five (45) days. So, Pepsi, prepares for his free week end and downloads 45 days of data. THEN, Pepsi, finds sometimes it is working but wants more data. Well, HDW works on a calendar month. So, on the first day of any month, Pepsi pays $119 for one month, downloaded the last 45 days and at the end of that month he will have...hmmm, 45 + 45 + 30, that is ABOUT four (4) months! Now, is four months of data sufficient, to start betting, maybe not, but you will have a good clue about spending more time with this tested pencil-paper method.

So, the data cost works out to about a dollar a day for ALL the tracks that are covered by HTR/HDW. For Feb 9-10-11, I downloaded 17+19+18 tracks for a total of 54 tracks over a three day period!

Oh, but, what if, Pepsi's system is ONLY for one track, say Aqueduct. Well, TSN data is available for fifty cents a day. Those 120 days would be only sixty dollars. However, who has the data that will satisfy Pepsi's system. Ahh, only Pepsi and his development partner knows and they will know how much programming and data preparation time will be ESTIMATED to be necessary to get started.

Disclaimer: I am not a shill for Ken Massa but a satisfied customer!

also, the method I described to sign up for a month, grab two and half months of data, then stop the HTR subscription is known to be used by more than one researcher. In other words, I am not suggesting something shadey that would get you in trouble with HTR/HDW. If you doubt me, then just ask Ken. Personnally, when I sign up, I plan to stay on board for at least six months to cover my current research plan.
---------
Pepsi:

if you find HTR has your data and want to try it,
THEN, I could give you some help to get thru the blizzard of conditions during start up process. With so much power by the many choices, the WHAT do I do first can be confusing. After you have done it once, I expect you will find the downloading and processing is simple and straightforward.

so, on the HTR web site, ready the export file spec to confirm if HTR might work for you.

duane

[AwolAtPA]

---(A) I wont buy a program like HTR that requires $100+ a month== (B)=(even though KM came in 2nd at the NHC) without a trial.
---So why does someone promote these programs that are not primarily created for the task of analyzing self made systems when it is clear that-- (C)==Excel and Access were and can do more for less. No brainer.

well Light, may I be the third person to reply to the above.

A) the HTR program is FREE. The data cost $119 a month with the number of tracks covered ranging from a low of five on a Tue to as high as nineteen on a Sunday. Refer to my posting #20 for details.

B) Ken has been very modest about his win and continued to talk about his LUCK. Well, you may read my post on the HTR forum (there I am AwolAtHTR). IF the HTR program was the magic answer then ALL tournaments would be won by HTR users. Well, only three were within the top twenty for NHC. Yes, two others beside Ken did know how to take a shot at good priced horses to be competitive in the contest.

C) I am an experienced Visual Basic programmer and use Access as my data base tool. So, your referece to '..no brainer..' is short sighted. I would not able to deliver a system using Excel. However, I know processing that VB can perform that gives Excel programmer headaches. Oh, but I would grant you the superior rating for programming Excel over my ability. That is, I can load an Excel spreadsheet into Access, Excel or Quattro Pro but even a basic formula would take me time to use the HELP functions!! With my system experience, I expect I could define processing rules that ANY spreadsheet program would be a brain teaser but be a piece of cake for Visual Basic.

I think you have achived a first for the PA forum with your ..NO BRAINER..

duane

[ryesteve]

just wanna thank you blokes for this insight. down under we have a program caled GTX (ozzy racing only) and it is what it is! i do not own it but some contacts do have it and of course like anything it cant use some parameters that r to vague, but all in all i have the impression that it is a good program. different strokes for different folks...i would still like to test some of the paper and pencil methods...anyone up for it??

cheers,
Wait, are these methods you're looking to test designed for Aussie racing?? I don't think there are many people around here who are storing up that data...

[Light]

Take the question from the back end. What if the question was I would like software to pick horses and I said Excel and Access. Well that wouldn't be too practical.HTR or HSH would be a better answer.

Here the question is I want statistical analysis. So someone says HTR and HSH. This is as ridiculous as the above answer.

I know HTR is free. Its the data that bothers me. You can go and buy the entire microsoft office package for previous versions to 2007 for much less than one month of HDW data and the learning curve is probably the same as HTR. Its just a more relevant answer to the question.

Pepsi

I think you would agree with me as some of those Aussie boards I've visited have members who are really into Excel for analysis. Thats actually why I visit them.

[AwolAtPA]

---
Here the question is I want statistical analysis. So someone says HTR and HSH. This is as ridiculous as the above answer.
---

yes, I agree, Excel is the superior statistical analysis tool over Visual Basic.

now that my Pepsi-Light moment is over.

let us get back to the thread issue and try to help Pepsi confirm his pen-pencil method because there must be history created before statistical analysis or ANY research plan can be defined.

duane

[ryesteve]

Take the question from the back end. What if the question was I would like software to pick horses and I said Excel and Access. Well that wouldn't be too practical.HTR or HSH would be a better answer.

Here the question is I want statistical analysis. So someone says HTR and HSH. This is as ridiculous as the above answer.
Not if the only data one is interested in analyzing is horse data that is already compiled by those programs. No one ever asked, "What's the best program for doing statistical analysis". If they had, the answer would have been SAS and this thread wouldn't still be meandering along.

[jhilden]

Someone asked a question and I contributed an example of a how using a programs data has helped me accomplish what the person was asking.



The original question was: Is there software out there that can do the work for testing systems? I thought that Access and excel were “Obvious” for simple database testing. I mentioned HSH and HTR because out of all the programs/data sources out there that I have used, these two programs with HDW’s data have been excellent. The two programs make exporting the raw data easy and I can spend most of my time analyzing instead of performing rudimentary maintenance.

As usual in the PA forums, we have posters that just shoot down ideas without providing any alternative advice and accuse others of being a shill. LIGHT - If you think the HTR/HDW is not the way to go, please enlighten us with your idea of how to amass large quantities of data that require little or no work to export into Access. Also, the process should require the user to not have advance skill in programming.

I do not care if you like HTR or not, shoot it down all you want; but please, at least contribute something other then false concluding statements.

[Light]

LIGHT - If you think the HTR/HDW is not the way to go, please enlighten us ..

Why waste time learning HTR just to export to Access or Excel when you can go there directly.If some guy wants to take your advice and go the roundabout way to get to Excel or Access,feel free. I'm just saying,cut out the middle man(HTR) and go directly to Access or Excel. KISS.

[ryesteve]

Why waste time learning HTR just to export to Access or Excel when you can go there directly.If some guy wants to take your advice and go the roundabout way to get to Excel or Access,feel free. I'm just saying,cut out the middle man(HTR) and go directly to Access or Excel. KISS.
If all I want to do is analyze data that HTR or HSH is already designed to analyze, going to a generic db or statistical program is the exact opposite of KISSing

[pepsi]

Light, yes excel is a good spread source down here, a few people rate the use of excel for backfitting very highly...

the paper and pencil methods r for north american racing only.
like anything what works for the next man doesnt mean it will work for the previous man. i just want to test these methods before wasting money and having some programmer put them on some sort of file.

thanks,

[banacek]

Still dont get what I'm getting at? I wont buy a program like HSH for $600 without a trial. I wont buy a program like HTR that requires $100+ a month (even though KM came in 2nd at the NHC) without a trial.Would you buy a car without having it checked out or without a warranty? So why does someone promote these programs that are not primarily created for the task of analyzing self made systems when it is clear that Excel and Access were and can do more for less. No brainer.


I'm not sure it is a no brainer, but I don't disagree with some of what you say.

I was planning on trying either Synergysm, HTR, or HSH earlier this year. I do my current handicapping with home grown software on Excel and Access using Bris files. I do think that some of the stuff these programs can do would be preferable to me programming yet again something else on Excel or Access, but I have years of Bris files and results to work with.

I did ask on another thread if it would be possible to pay to more than 45 days of back data and was told it was not. Now there are tracks I was interested in that haven't run for 3 months, how do I start evaluating them without something to work with? 45 days won't do it for me. So I am sticking with my Excel this year and have done a lot of the work on it in the last month or so.

But I do think that these programs quite possibly offer something very valuable. They may be able to save me time - and that I will pay for. I'm no expert programmer, but I make my way through the job eventually. But if I can't access prior data, then I will be going into the new software blind in many cases. And I'm not putting my money at risk in that circumstance.

[njcurveball]

I was planning on trying either Synergysm, HTR, or HSH earlier this year. I do my current handicapping with home grown software on Excel and Access using Bris files.


How much are you spending a month on BRIS files?

And just for reference, the data in HTR and HSH is going to be the same as BRIS in many of the fields. Dates, Tracks, Races, Horses, Jockeys, Trainers, etc. will all be an exact match.

It would take a lil bit of maintenance, but you could check out the file descriptions first and try to create tables which can use your existing BRIS data and just leave the other HTR fields blank for the older data (you weren't using them anyway).


I think it would be dependant on how much you are currently making using the BRIS files and your own software. If you are doing really well then it would be better to just keep doing what you are doing.

Unlike other options, the HTR stuff is pretty much free to try out and see a file layout to see if it would have all of the things you are using from BRIS.

Jim

[jhilden]

I combined about three years of Brisnet data with my HTR data into one big database in Access. The process was a bit laborious, but the database export with the HTR files made a great starting point. For the Bris data, I opened the files in excel then cut and pasted the applicable fields into my main database.

I am not affiliated with HSH/HTR/HDW in anyway except that I am a very satisfied customer. I guess I tout there products because of all the errors I’ve come across in dissecting Bris/DRF data in my tables.

[banacek]

Thanks for the suggestion, but if I do this can I simulate a data file that HTR or HSH will read?

[njcurveball]

Can I simulate a data file that HTR or HSH will read?



Is your question will it read a BRIS file? At first glance I would say No, but that would be something you would nead to ask the People who developed the software.

My point was to set up a table with all of the HTR (or HSH) data items, modify your program to read that and then populate the previous months with your BRIS data to facilitate you going forward until you get enough data in the correct format.

This would only be in your own Access database. And of course, some things won't fit so your table may have two versions of the same field. For example BRIS_PED_RATING, HTR_PED_RATING. I am not saying this is easy or even preferable if you are doing well with what you have now.

Your other option would be to continue buying the BRIS files, in addition to the other stuff. I see some people do pay for more than one program. Although unless your bottom line can support it, I would not recommend it.

I cannot speak for the current HSH board, but the HTR board has lots of nice people who will help you with "research".


Jim

[banacek]

Is your question will it read a BRIS file? At first glance I would say No, but that would be something you would nead to ask the People who developed the software.

My point was to set up a table with all of the HTR (or HSH) data items, modify your program to read that and then populate the previous months with your BRIS data to facilitate you going forward until you get enough data in the correct format.

This would only be in your own Access database. And of course, some things won't fit so your table may have two versions of the same field. For example BRIS_PED_RATING, HTR_PED_RATING. I am not saying this is easy or even preferable if you are doing well with what you have now.


Thanks again. But (back to Light's point) if I can't use it in the software then I don't see where the software comes in. I'm back to Access.Or am I missing something?

[njcurveball]

OK, you confused me.

Do you want to use BRIS files in HTR or HSH? I don't think you can, but ask the program people.

DO you want to use HTR data in YOUR program? You sure can, the export is simple and you gain all of the proprietary fields.

I really do not get "Lights" point as there is no time wasted learning the program unless you like the data. And there is also a price point for the data.

Why would I pay BRIS $1 per track per day and run up $200+ plus in bills to get the same data HDW sells for $119 a month?

I would use HTR JUST for the export in that case. I have an Oracle database and need to get the data from somewhere.

The fact that programs like HTR and HSH give you so much extra value is just the icing on the cake.

Now if you are paying $50 a month for a few tracks, then you have a decision. If you are paying $100 or more, it is really a no brainer unless BRIS has some proprietary data you are using to win money.

Jim

[banacek]

OK, you confused me.

Do you want to use BRIS files in HTR or HSH? I don't think you can, but ask the program people.

What I want is to test HSH or HTR on old files (more than 45 days) to test its performance before I wager any money. Those files are unavailable to buy ( and I clearly can't convert my old BRIS files to HSH files) so I have no way of testing it. No need to respond, this was gone over in another thread. I was just replying that I could understand the point that Light was making.

[formula_2002]

G'day,

have a couple of methods that i would like to test regarding past results. does anyone know of any software out there that could actually do the work. have not used the methods personally but would like to see how they hold up and if they actually turn anything over.

cheers,
If the input data is within the Bris single format file, I may be able to help you along.

http://www.bris.com/cgi-bin/static.cgi?page=drfsff
I work with this stuff all day long (except when I'm woking on the money making job!) ;)

[njcurveball]

What I want is to test HSH or HTR on old files (more than 45 days) to test its performance before I wager any money.


I see your point and you are not the first to ask HDW about purchasing previous months files.

I believe that was covered in a thread on here before. You need the correct files and the correct program to test on all of the proprietary factors.

What is your price point for BRIS files in an average month?

Jim

[PlanB]

I'm getting bored reading about HTR. It's a pay-4-service & so I think anyone curious should consult the web site (or PM its cheerleaders) and INVEST some money, maybe a modest month's try-and-see, and LEARN FOR YOURSELF. I mean, HTR cheerleaders, cease & desist.

[njcurveball]

INVEST some money, maybe a modest month's try-and-see, and LEARN FOR YOURSELF.


GREAT IDEA! And great advice coming from someone who regularly jumps from idea to another! :lol:

[raybo]

The thread starter here is interested in "testing" some handicapping factors/methods. If they are to be of any value then they probably won't be found in a commercially sold program, most of that stuff is either proprietary or available to everyone. IMO, the solution to his dilemma is to either create his own database/handicapping program using whatever applications he has available to him (Access, Excel, or whatever), or get with someone who already has a database and can run the queries he needs to investigate. Whether the DB owner would want pay for this or not would be up to them.

I am currently creating a "stock" DB that will be horse racing specific in another thread. I am doing this to address the problem this person has, namely, no DB and no DB experience. I have always thought that the handicapping process is so highly specialized between players that there needs to be a vehicle that allows modification by the user and has very few restrictions concerning what the user can get out of it. Sure there will be a learning curve initially, but that's true with all programs. But, once some basic queries are put in the DB the user could just modify them to retrieve the data he needs. The data source could be any comma-delimited type, so the user can use whatever they like and can afford, or already have on their harddrive. No money required for my part of this thing, as I am creating it for myself anyway.