Betting Strategies, Back Testing, and WinCraps Scripts

[heavy]

Hey, that's a big title for what is going to be a relatively short port. There has been much discussion over back testing betting strategies in WinCraps based on certain indicators that call for the player to turn bets off, press bets, or take them down. I've long said that it's impossible to codify all of the variables that exist between a craps player's ears. This weekend, while reading the biography of Titanic Thompson, I stumbled across a quote by Thompson regarding testing of poker decisions made by humans versus machines. This was in the early days of computer sims for poker. Thompson said that the computer would never beat the human because "you cannot program deviousness into a computer." That pretty succinctly sums up my opinion as well, and I think it applies very well to DI's.

[Mad Professor]

I agree with you Heavy AND I don't bank on WinCraps results because I don't know who wrote the scripts or how competent is the programmer who wrote the scripts for my computer to crunch on. I think there are too many variables in the game and way too many opportunities for programmer error and oversight.-Little Joe

Okay, I hear you both on all that; but isn't there some point at which you can trust the computer'?

We trust our drive-by-wire cars, which were programmed by humans...we trust our fly-by-wire airplanes...hell, we even trust the machine that we are reading this on right now, which was programmed by someone else.

I don't think you guys are saying that Steen's WinCraps software itself is flawed to the point of producing only untrustable garbage, or that because Maddog's BoneTracker was programmed by a human that it is somehow inherently untrustable; but that's the way it might be misread by some.

Re-read LittleJoe's comments and then tell me his blanket-statement about bet-scripts being untrustworthy isn't the case. It seems to make the argument that WinCraps betting-scripts aren't valid because they were written by someone else; is a false premise.

Each script is 'open' so that any user can comb through it bet-command by bet-command to see that it conforms to what it is advertised as. If you are suggesting some kind of nefarious 'hidden-code' that intentionally derails or understates/overstates the results; then that's a pretty serious allegation.

Again I'm sure that's NOT what is being intimated here; but a re-reading of the above post might be construed that way.

I get Heavy's point about the fact that his bet-decisions are so intricate and so complicated that no computer would ever be able to factor in all of the decision-elements that he uses when he is at the tables; because he has made that argument before.

But doesn't advantage-play dice-influencing really boil down to betting when you have an advantage, and not betting (or betting far less) when you don't have the advantage?

If we aren't talking about dice-influenced outcomes, and instead are theorizing about random outcomes; then fine, all bets are off (literally, at least for me ).

But if we are talking about de-randomized positive-expectation outcomes, we can make it all super-complicated and make our bet-decisions hinge on all sorts of other things like today's bio-rhythm charts, the cut of the cocktail waitresses dress (and the color of her hair); but is there a point where we actually say, "Okay, this is my basic A-P betting plan, and I'll modify it on the fly from there."

Or are even all of our dice-influenced bet-decisions almost entirely based on mood, feelings, urges, hunches, and quasi-superstitions?

Is there some point where we trust the math; and if so, where and when? Or have we gotten to the point where we don't trust any machine if it's more complicated than a shovel, simply because it contains a microchip that we ourselves didn't program?

Can we start with WinCraps and BoneTracker first.

Do either of you guys actually think that their coding/scripting is inherently flawed? I don't think that is what you are saying, but in reading LittleJoe's post above; that's what some might interpret is being said.

If that's not the case, are you suggesting that the guys who wrote the various bet-scripts intentionally or accidentally are offering flawed scripts? Each script is 'readable'; so I'd be interested in someone pointing to the most inherently flawed ones first.

If on the other hand, you are simply saying that those scripts don't "allow" for something like Heavy's "lucky red-head" factor; then yes I agree, both of those pieces of software fall WOEFULLY, PITIFULLY, and EMBARRASSINGLY short.


MP

[Mad Professor]

This post has been intentionally deleted and then reposted below in order to maintain thread-continuity.

MP

[Mad Professor]

Hi LittleJoe,

Thanks for clarifying your position. It was as I hoped it would be, as opposed to what it could have been misconstrued as being. I appreciate that.

However, you bring up a point about WinCraps scripting that makes it look much more complicated than it actually is.

WinCraps is structured so that it avoids the need for the script-writer to know any kind of complicated coding-language. Instead, bet-instructions are 'built' using a menu-based system; such as:

"When X happens, bet $Y, then if L happens, change this bet to $A, that bet to $B, and this other bet to $C.

If M, N, O, or P occurs; then bet $GG on this wager, but change that wager to $YY, and remove Bet-T, Bet-R, and Bet-D..."

In other words, anyone who can operate a drop-down menu, and knows how to structure a sentence in English, can write a bet-script in WinCraps. Now here's the great thing about WinCraps; if you structure it wrong; it automatically rejects that bet-instruction and won't accept it until it is right.

On the other hand, if you 'meant' to enter a $5 bet, but you fat-fingered it and entered it as $500; then no, the system doesn't extend to the mind-reading realm; but it does require that a properly-phrased wager is made before it accepts it (just like in a casino if a dealer doesn't understand what you are trying to bet).

I know it's all the fashion to denigrate and undermine all of the effort that some of these guys have put into the various bet-scripts (by saying stuff like, "It doesn't include a feature that allows for in-casino humidity-fluctuations, so it isn't a valid system for the way that I bet"); but if you think they are mathematically-faulty or bet-structured incorrectly, it would be helpful if you would point to those specific errors so that Steen can correct them.

However, simply saying that you don't trust anyone whom has written these bet-scripts seems a bit disingenuous.



MP

[Mad Professor]

Here's an article that, to my mind, really cuts to the heart of not only what WinCraps will or won't do; but it also contains a much better example of how these bet-scripts are actually written (or rather, how each bet-statement is made by way of a drop-down menu and fill-in blank bet-amount basis).

Using WinCraps to War-Game your Current D-I Skills Against Various Betting-Methods



MP

[Mad Professor]

But now I'm curious. How many BMs don't care enough to learn how to program WinCraps and how many crapsters could care less about learning how to make a computer simulated betting scheme?

That's a superb question, LittleJoe.

I would venture to say it's about equivalent to the number of skilled-shooters who aren't able to make sustainable bankroll-growing profits from their current betting-schemes.

Not that one has anything to do with the other.



MP