Why Learn Basic Programming Skills?

[headhawg]

Originally Posted by badcompany

Traynor's premise is that programming skills can help you write a handicapping app.

However, if you are a mediocre programmer, is it wise to compete against people who are actually good at programming?

I'm basically talking about the difference between doing something as a hobby and doing it for money. If the latter, you're gonna need some talent.

I think the important thing is what traynor wrote in posts 35 and 36 -- can you use data in a way that gives a handicapping edge. That's really the main thing that sets handicapping programs apart. Dave's HSH program is going to have some proprietary numbers in it, just like Jeff's Jcapper program and Ken's HTR program. If I knew those "algorithms" I could code it -- and I'm not gifted. Is the average person going to be able to create something worthwhile? Probably not, but one never knows. Ninety-five plus percent of 'cappers lose money anyway mostly because they do the same thing losing things over and over, or are not willing to do the work that it takes to succeed. Perhaps by creating their own software they might have a chance to put themselves in the 5% category.

[badcompany]

As an aside, listening to Malan's lecture, I'm having flashbacks to sitting in a classroom when the teacher is going just a bit too fast.

Gotta love being able to scrub back and listen again.

[Tom]

Why wouldn't a good database do the trick?
You can explore lots of ideas and you can make new factors in it.
What will a program do that a DB won't?

[traynor]

Originally Posted by Tom

I did a lot of programming in basic. Most of it Sartin type stuff.
I did a program for the Sharp 360 Handheld that did the Total Pace stuff, using the Beyer numbers to figure out the variant. MY friend, Achilles (here) and I made a few bucks on the Inner track simulcasting at Finger Lakes.

When Synergism II came out, I loaded it up at Sartoga, and the hand held froze up! I erased it and quickly wrote a simple program that calculated APV and %Early. That was all I had to go by, other than the morning analysis by the Hat and DRF.

One of the best days I ever had there.

Not surprising. I used the Sharp 1260 (the small one with the single line display) for several years, with little more than Phase III-type figures generated. It was great!

[traynor]

Originally Posted by badcompany

As an aside, listening to Malan's lecture, I'm having flashbacks to sitting in a classroom when the teacher is going just a bit too fast.

Gotta love being able to scrub back and listen again.

I like fast. The faster the better. Malan goes fast, but with clarity--rather than overwhelming listeners with a torrent of words that make one think he or she is "slow" when in fact the speaker is almost inarticulate, and compensates by speaking rapidly to create the impression of knowledge. Julie Zelenski (the Stanford instructor for CS106b) and Mehran Sahami (the Stanford instructor for CS106a) go at that same rate of speed. It is the slow, ponderous, pontificating type of presentation that puts me to sleep. Or causes me to dump the class and get in another, better one.

[traynor]

Originally Posted by Tom

Why wouldn't a good database do the trick?
You can explore lots of ideas and you can make new factors in it.
What will a program do that a DB won't?

If you are familiar with designing and executing fairly complex queries, you are satisfied with the results you get, and you have nothng specific that you want to learn more about, there is probably little reason for you to consider doing anything else. The same goes for those who are familiar with, and comfortable using, Excel spreadsheets.

[HUSKER55]

where, (how), do you get the past performance data into a program? Maybe I should say how do I convert a pdf file from bris, or any place else, into any program.

I called Bris, DRF and Equibase and they all said any data downloads were for their programs only and that any other data was PDF only.

The lady at equibase said to contact them (via feedback) and perhaps they could make one one for a fee, but she thought the answer was no.

I am missing something here.

Will someone throw me a rope?

[headhawg]

The easiest way is to use a data provider's comma-delimited files. You would simply have the program read the data into the program and use it from there. How to do that is dependent on the programming language, but there would be a method/procedure out there on the Interwebs to show you how to do it. Parsing a pdf file is a whole other ballgame, and if you're paying a data provider just go with the simplest file structure.

Most people will agree that HDW provides the best data, but the most popular is probably the Bris single-file format.

[HUSKER55]

thanks everyone!

[traynor]

It also depends on how many tracks you plan to play, or are playing. The whole "I can bet 20 tracks a day!" syndrome that seems to fuel many computer users is--in most cases--sheer folly. They dabble here and poke a little there, and believe their gee whiz bells-and-whistles computer apps are "handicapping the races" for them. If they were making buckets of money, I would think otherwise. What I think is that most computer users are doing as poorly as (or worse than) most pen-and-paper handicappers--with the aid of technology to make it seem they are doing something "high tech." There is nothing "high tech" about reading the output of someone else's handicapping app unless that app is generating a generous profit for you.

Playing multiple tracks is great--if you are winning. If not, you may get WAY much further ahead by concentrating on one track, and usng your own app to analyze those races. If you don't know how to "equalize" a six furlong pace line for an entry in a seven furlong (or even six-and-a-half furlong) race, that might be a good thing. If you understand how those spiffy "handicapping programs" manage (or at least superficially seem to manage) horses running at different distances at different tracks to be "comparable" in today's race, you may undersand why so many bettors lose (with or without the aid of computers).

The more you learn about the basics of programming, the more you will understand the shortcomings and deficiencies of the applications so many rely on to "pick their winners" for them. And the more you will be able to overcome those same shortcomings and deficiencies in your own (relatively simple to write) program(s). Not rocket science. Just getting past the "OMG! That looks so complicated!" factor is the biggest hurdle. The rest of it is pretty easy.

In an old issue of Racing Times, there was an article that presented the details of pace analysis--the nuts and bolts code needed--and the whole thing was like half a page. Writing a pace analysis app--even for a complete novice--is almost trivial. However, because it seems complicated, and because it is easy to generate various rankings and ratings based on pace, it forms the basis of many (if not most) "handicapping programs."

One thing that is missing from handicapping apps (as they currently exist) is the implementation of basic statistical processes to analyze races. I think one of the most valuable fringe benefits of learning to write your own apps (or to write VBA code for Access or Excel, if that is your preference) is to take a block of races and feed it to WEKA, or RapidMiner, or Anaconda3 and see what comes out.

All the impressive buzzwords like "machine learning" and "artificial intelligence" and "least squares fit" and "regression analysis" and all the rest take no more to use than a free download of a decent mining app, and a couple of mouse clicks. Take what you learn from the data mining apps and implement it in your own app. Rinse and repeat.

[whodoyoulike]

Originally Posted by traynor

It also depends on how many tracks you plan to play, or are playing. The whole "I can bet 20 tracks a day!" syndrome that seems to fuel many computer users is--in most cases--sheer folly. They dabble here and poke a little there, and believe their gee whiz bells-and-whistles computer apps are "handicapping the races" for them. If they were making buckets of money, I would think otherwise. What I think is that most computer users are doing as poorly as (or worse than) most pen-and-paper handicappers--with the aid of technology to make it seem they are doing something "high tech." There is nothing "high tech" about reading the output of someone else's handicapping app unless that app is generating a generous profit for you.

Playing multiple tracks is great--if you are winning. If not, you may get WAY much further ahead by concentrating on one track, and usng your own app to analyze those races. If you don't know how to "equalize" a six furlong pace line for an entry in a seven furlong (or even six-and-a-half furlong) race, that might be a good thing. If you understand how those spiffy "handicapping programs" manage (or at least superficially seem to manage) horses running at different distances at different tracks to be "comparable" in today's race, you may undersand why so many bettors lose (with or without the aid of computers).

The more you learn about the basics of programming, the more you will understand the shortcomings and deficiencies of the applications so many rely on to "pick their winners" for them. And the more you will be able to overcome those same shortcomings and deficiencies in your own (relatively simple to write) program(s). Not rocket science. Just getting past the "OMG! That looks so complicated!" factor is the biggest hurdle. The rest of it is pretty easy.

In an old issue of Racing Times, there was an article that presented the details of pace analysis--the nuts and bolts code needed--and the whole thing was like half a page. Writing a pace analysis app--even for a complete novice--is almost trivial. However, because it seems complicated, and because it is easy to generate various rankings and ratings based on pace, it forms the basis of many (if not most) "handicapping programs."

One thing that is missing from handicapping apps (as they currently exist) is the implementation of basic statistical processes to analyze races. I think one of the most valuable fringe benefits of learning to write your own apps (or to write VBA code for Access or Excel, if that is your preference) is to take a block of races and feed it to WEKA, or RapidMiner, or Anaconda3 and see what comes out...

If it becomes so easy, can you provide an example of a Tbred race before it runs?

[whodoyoulike]

Originally Posted by badcompany

I'm not familiar with handicapping programs. but I do know a bit about financial market algorithms. They're basically a coded set of rules. The advantage is in the number crunching ability. They allow you to scan more markets and timeframes than you could longhand. Of course, the trick is to come up with rules that are consistently profitable.

Are you equating programming with running a Query versus programming to help in your analyses? Because, I think there is a big difference.

[badcompany]

Originally Posted by whodoyoulike

Are you equating programming with running a Query versus programming to help in your analyses? Because, I think there is a big difference.

If I understand the question, correctly, there are programs for both fundamental and technical analysis.

The former will look for various financial ratios, P/E, EPS etc; the latter, various price signals.

[traynor]

Originally Posted by whodoyoulike

If it becomes so easy, can you provide an example of a Tbred race before it runs?

The whole point is to learn to do it yourself.

[HUSKER55]

I have been searching opinions on the net and there are a lot of different programs out there. All geared to do one thing well, some do a lot of things reasonably well and etc.

I have come to the conclusion that anyone can do what they want on their terms, [comfort level]. We all learned to walk before we ran.

It boils down to chose your weapon and make your stand on your own terms.

One huge advantage is that the search for the right program for yourself sure does organize your awareness of what you are doing, what you want to do and what you want a program to do.

Regardless of your skill level there is a program for you. Confidence in one's numbers is a huge part of wagering.....well at least for me.