Why Learn Basic Programming Skills?

[traynor]

Originally Posted by DeltaLover

Since you are redesigning / reimplementing, I suggest a test driven development making a 100% testing coverage a goal. Although your initial speed might be a notch slower you must resist the temptation of adding "naked" components (meaning without related tests) since this approach will result to much higher level quality of code that will be easier to maintain and evolve as time goes by. Also, you might find some interesting ideas in Literate programming which might become helpful in some phases of your migration. Also check your PM for a link to one of articles that I have written about test driven approach and a video that I have put together.

I have been using TDD for quite awhile. Unless one has written subs that wander on for hundreds and hundreds of lines--that do what one intended but break at some time in the future and require a LOT of thought and work to fix--one cannot truly appreciate the need for TDD, small modules, and clear comments. (And always, always updating the comments to be current with the code.) There are few things more dismaying than trying to decipher clear, concise comments describing the intent and purpose of a block of code that (now) does something quite different from what it did when the comments were written.

[traynor]

Something you may want to consider, especially if you use Microsoft technologies.

https://blogs.technet.microsoft.com/...T&MC=MobileDev

There is little point in re-inventing the wheel, if one's intent is to use the wheel to facilitate going from point A to point B. Lots of very useful stuff (fairly) readily available to average users. No PhD in Predictive Analytics required. As usual, Microsoft does all the heavy lifting behind the scenes and makes it all look incredibly simple.

If you have ever considered "deep learning" about horse races, this might well be the most useful tool to consider in that endeavor.

[JJMartin]

Originally Posted by traynor

Something you may want to consider, especially if you use Microsoft technologies.

https://blogs.technet.microsoft.com/...T&MC=MobileDev

There is little point in re-inventing the wheel, if one's intent is to use the wheel to facilitate going from point A to point B. Lots of very useful stuff (fairly) readily available to average users. No PhD in Predictive Analytics required. As usual, Microsoft does all the heavy lifting behind the scenes and makes it all look incredibly simple.

If you have ever considered "deep learning" about horse races, this might well be the most useful tool to consider in that endeavor.

I have tried applying neural networks to large pp samples several times now with disapointing but not unexpected results. I think the deep learning technology applied to horse racing would probably yield impressive results based on visual information as has been demonstrated in other areas already. Perhaps if it was fed several thousand still images of real race horses or race replays with some consistency in distance,surface and race view angles, it could be a gold mine find. It would definitely see things that human perception is oblivious to.

[traynor]

Originally Posted by JJMartin

I have tried applying neural networks to large pp samples several times now with disapointing but not unexpected results. I think the deep learning technology applied to horse racing would probably yield impressive results based on visual information as has been demonstrated in other areas already. Perhaps if it was fed several thousand still images of real race horses or race replays with some consistency in distance,surface and race view angles, it could be a gold mine find. It would definitely see things that human perception is oblivious to.

You might find it useful to search a smaller set, purely to locate possible trends, then to search for those trends in a larger set. Too big a set initially blurs what could become very interesting (and useful). In essence, the reverse process of using a large data set to "prove" that whatever was discovered in the small set "won't work in the real world."

Also, you may find it more useful to ignore ROI (initially) and go straight for the win frequency as a target category. That lets you do some useful research without being misled by (a relatively small number of) outliers.

[Actor]

Originally Posted by highnote

Here is basic language code for Dick Mitchell's "Positive Expectation Handicapping" software. I believe this was in one of his books.

You might need to adjust some of the lines to make it work. It's a command line program.

This is the code I used to begin to learn how to write code for horse racing.

10 B$ = Chr$(7)
   Dim CL(20), CW(20), ES(20), FR(20), LS(20), RS(20), SP(20), XX(20),       Z1(20), Z2(20), Z3(20), Z5(20), Z6(20), Z7(20)
   Dim CM(20), CN(20), ML(20), P(20), S(20), K(20), L(20), MS(20), CF(20)

40 For I = 1 To 25: Print: Next I
50 INPUT "ENTER TODAY'S DATE "; A$
   GoSub 3400
   If FU$ = "" Then DA$ = A$: GoTo 90

What does the variable FU$ stand for?

[DeltaLover]

Originally Posted by Actor

What does the variable FU$ stand for?

Check what is happening in line 3400 and after..

By the way, this "code" is the perfect example why to avoid basic and you should not try to learn anything from it as it represents an extremely poor approach.

[Actor]

Originally Posted by DeltaLover

Check what is happening in line 3400 and after..

Obviously a status message since it gets assigned values like 'ERROR', 'FOUND', 'F' or just ''. Why FU$ instead of ST$ or ER$ ?

Originally Posted by DeltaLover

By the way, this "code" is the perfect example why to avoid basic and you should not try to learn anything from it as it represents an extremely poor approach.

Obviously written for a primitive version of BASIC which did not provide WHILE .. WEND or IF .. THEN .. ELSE, but not so primitive that it required a line number on each line. Second to the huge number of GOTOs its biggest fault is a complete lack of REMs.

It intrigues me. It computes an odds line for each race (I think) but I have not been able to figure out the algorithm. If I can do that then I'll try rewriting it in Free Pascal using BRIS files for input instead of the totally manual input.

Thanks for your input.

[traynor]

I used it. It was a big time loser then, probably (would be) even more so now. One of my fondest memories is of Mitchell (very aggressively) "explaining" why it didn't work (at a Sartin seminar). Oddly enough, Mitchell never offered to refund the money the chumps paid for his app. A big surprise. Did anyone, anywhere, ever do anything with it that even vaguely approached living up to the hype?

Anyone know whatever happened to David Akimoto (I think the name is correct, but may be off a bit), the guy who actually developed the app? Smart kid.

[DeltaLover]

Originally Posted by Actor

Obviously a status message since it gets assigned values like 'ERROR', 'FOUND', 'F' or just ''. Why FU$ instead of ST$ or ER$ ?

Obviously written for a primitive version of BASIC which did not provide WHILE .. WEND or IF .. THEN .. ELSE, but not so primitive that it required a line number on each line. Second to the huge number of GOTOs its biggest fault is a complete lack of REMs.

It intrigues me. It computes an odds line for each race (I think) but I have not been able to figure out the algorithm. If I can do that then I'll try rewriting it in Free Pascal using BRIS files for input instead of the totally manual input.

Thanks for your input.

I understand, reverse engineering Basic code especially written with a lot of GOTO is not a fun task.

The input should be transparent to your procedure and be able to switch from BRIS to manual to whatever else just by altering a configuration setting.

Originally Posted by Actor

Obviously a status message since it gets assigned values like 'ERROR', 'FOUND', 'F' or just ''. Why FU$ instead of ST$ or ER$ ?

Try to avoid status messages as much as you can, use exceptions instead.

[traynor]

Originally Posted by JJMartin

I have tried applying neural networks to large pp samples several times now with disapointing but not unexpected results. I think the deep learning technology applied to horse racing would probably yield impressive results based on visual information as has been demonstrated in other areas already. Perhaps if it was fed several thousand still images of real race horses or race replays with some consistency in distance,surface and race view angles, it could be a gold mine find. It would definitely see things that human perception is oblivious to.

One of the funniest comments I have seen since a professor at the University of Washington described a class I was considering as "calculus for poets."

"The Python backing extends the toolkit beyond C++ development for better accessibility. C++ "was good for really serious programmers," said XD Huang, Microsoft chief speech scientist, but "our mission is simple: to democratize AI," not only for web developers, but for different disciplines, even sociologists. Python support will offer extensiveness, expressiveness, and flexibility."
http://www.infoworld.com/article/313...g-toolkit.html

If "even sociologists" can use it, it MUST be simple.

[DeltaLover]

Originally Posted by traynor

If "even sociologists" can use it, it MUST be simple.

It is true that python is simple in the sense that it allows a complete beginner to start writing simple programs very quickly but in reality it has a significant learning curve if you want to reach expert levels and write idiomatic code using concepts like meta-classes, decorators, coroutines, context managers etc.

[Actor]

Originally Posted by DeltaLover

I understand, reverse engineering Basic code especially written with a lot of GOTO is not a fun task.

That depends on your motivation. Solving a really complex puzzle can be quite stimulating.

[Actor]

Originally Posted by traynor

I used it. It was a big time loser then, probably (would be) even more so now.

I've never had anyone else's code work straight out of the box. I've always had to tinker with it, usually to no avail. But it's fun.

[traynor]

https://www.rstudio.com/products/rstudio/download/

Free.

https://www.rstudio.com/products/rstudio/download/

Free.

https://mran.microsoft.com/download/

Free.

Time to learn all this stuff--available. Providing I use the time I have been wasting trying to get Linux Mint to stop crashing a half dozen times a day. It might LOOK like a well-designed GUI, but it is clearly a LONG way from being ready for prime-time. Possibly a step up for those who primarily use consoles. A big step down for those accustomed to Microsoft GUIs.

[traynor]

Originally Posted by Actor

I've never had anyone else's code work straight out of the box. I've always had to tinker with it, usually to no avail. But it's fun.

Depends on what you paid for it. Buying software as a product, one tends to expect it to perform as claimed. Caveat emptor.