Testing a Betting System
Poll
| No votes (0%) | |||
| No votes (0%) | |||
| No votes (0%) | |||
| No votes (0%) | |||
| No votes (0%) | |||
| No votes (0%) | |||
| No votes (0%) | |||
| No votes (0%) | |||
| No votes (0%) | |||
 | 1 vote (100%) |
1 member has voted
In craps: 1 hour = 15 shooters = 150 - 200 dice rolls
Much easier and quicker to use math and/or simulation.
How long would it take to prove a system doesn't work? It's already been proven that they don't work many, many times over. People have been trying for hundreds of years and they will be trying until the end of time.Quote: lovetosing2011I always read how betting systems cannot be profitable in the long run, how long do you think it would take to prove this to be true for the best of systems?
In craps: 1 hour = 15 shooters = 150 - 200 dice rolls
There's no such thing as the "best system", it's all just a matter of taste.
You should research gambler's falicy.
When you can show the math and it adds up to a positive number.Quote: lovetosing2011I guess I'm trying to figure out how long it takes before its safe to say it's not just luck anymore.
It's possible to get lucky for a very long time.
Quote: AxelWolfWhen you can show the math and it adds up to a positive number.
It's possible to get lucky for a very long time.
So how long is a long time?
When all the casinos ban you for wining to much to often.Quote: lovetosing2011So how long is a long time?
Quote: lovetosing2011So how long is a long time?
Depends on the variance, the way the system plays out, and the degree of certainty.
If you're unsure, do the math or run a sim. If you wanna know how many rounds to test for, I'd start at at least 500,000 rounds, and run the sim several times. If the results aren't strikingly obvious, then run 10B round sim.
I have a feeling he wants to test it in a casino?Quote: RSDepends on the variance, the way the system plays out, and the degree of certainty.
If you're unsure, do the math or run a sim. If you wanna know how many rounds to test for, I'd start at at least 500,000 rounds, and run the sim several times. If the results aren't strikingly obvious, then run 10B round sim.
Let's pretend that you thought a roulette wheel or dice were slightly biased. Then how many rolls or spins would you want before you were were 100% confidante it was?
Quote: RSDepends on the variance, the way the system plays out, and the degree of certainty.
If you're unsure, do the math or run a sim. If you wanna know how many rounds to test for, I'd start at at least 500,000 rounds, and run the sim several times. If the results aren't strikingly obvious, then run 10B round sim.
The math is super complicated...I need to talk to a statistician. And I would run a simulator, but I'm having trouble with the coding. Know anything about coding with WinCraps?
Quote: AxelWolfI have a feeling he wants to test it in a casino?
Let's pretend that you thought a roulette wheel or dice were slightly biased. Then how many rolls or spins would you want before you were were 100% confidante it was?
Yep that is exactly what I was wondering.
Quote: AxelWolfI have a feeling he wants to test it in a casino?
Let's pretend that you thought a roulette wheel or dice were slightly biased. Then how many rolls or spins would you want before you were were 100% confidante it was?
Depends on the level of bias or irregularity. Say a coin has a bias so there's 49% heads and 51% tails. To be confident in this, you need many many trials. If the bias is more significant, say, 10% to 90%, you don't need as many trials. That's, of course, assuming you're flat-betting. (ie: With a 49% to 51% bias, as well as using a martingale or some martingale-type system, you would need many more trials than if you were to flat bet.)
Quote: lovetosing2011The math is super complicated...I need to talk to a statistician. And I would run a simulator, but I'm having trouble with the coding. Know anything about coding with WinCraps?
What is the basis for this, anyway? Do you think you have developed (or are in the process of developing) a system which can beat the odds? Or do you think you've found something that's "not right" -- ie: uneven/unbalanced dice?
How "super complicated" is the math? Can you give an example or explain your reasoning?
I didn't read all if your post(It's late/early been out since 7:30AM).Quote: RSDepends on the level of bias or irregularity. Say a coin has a bias so there's 49% heads and 51% tails. To be confident in this, you need many many trials. If the bias is more significant, say, 10% to 90%, you don't need as many trials. That's, of course, assuming you're flat-betting. (ie: With a 49% to 51% bias, as well as using a martingale or some martingale-type system, you would need many more trials than if you were to flat bet.)
What is the basis for this, anyway? Do you think you have developed (or are in the process of developing) a system which can beat the odds? Or do you think you've found something that's "not right" -- ie: uneven/unbalanced dice?
How "super complicated" is the math? Can you give an example or explain your reasoning?
But I knew you would say it depends on amount of bias.
But what if you don't know and that's what you are trying to find out? A system player can never tell you what percentage he has(because he doesn't really have one)
Dude is trying to prove something that's impossible to prove (a betting system) and he wants to know how many rolls it would take to convince us it actually works.
The question is...how many rolls of winning does it take to say,.OK this sh*t isn't random, something strange is going on?
How many licks does it take to get to the center of a tootsie pop?
(Oh cool, Steely Dan song came on in the casino, see below)
But, just because less may be necessary to prove a system works....more rounds would work as well. If you're unsure, try running a 10 billion round sim. That should get you going in the right direction.
Longer than your lifetime. Seriously, there are people who are ahead lifetime on keno and lottery games, both of which have enormous house edges. The games weren't rigged or broken, they just got lucky. There's no trick to it.Quote: lovetosing2011So how long is a long time?
If you want to go through the learning exercise of programming a betting system for your chosen game, go for it. But you'll always reach the same conclusion: if you're playing a house-advantaged game, and that house edge doesn't change from one hand to the next *based on* whether you won or lost the prior hand, then changing your bets based on that criteria won't change the house edge. That should be obvious, no?
It doesn't take any. If you're working from the premise of a fair game (fair dice, unbiased wheel, etc) and you're looking at just betting patterns, you can examine a system on paper and see it doesn't work without betting a single dollar. If you're looking at a real-world set of events and you think they behave *differently than* a canonical casino game, then you're not really testing a system, you're playing that game for bias.Quote: AxelWolfThe question is...how many rolls of winning does it take to say,.OK this sh*t isn't random, something strange is going on?
1. How long have you been playing using this system?
2. Are you profitable at this point?
3. Do you have accurate records/proof of all sessions to verify your results?
That would be a starting point.
It's easy for a mathematician/statistician to say that all betting systems are worthless. To put that in a mathematical formula and prove it may be a bit more challenging.
Quote: cwwbjrThat's a good question. If you are referring to real casino play (not computer generated on line or simulated play) I would ask you 3 questions as follows;
1. How long have you been playing using this system?
2. Are you profitable at this point?
3. Do you have accurate records/proof of all sessions to verify your results?
That would be a starting point.
It's easy for a mathematician/statistician to say that all betting systems are worthless. To put that in a mathematical formula and prove it may be a bit more challenging.
A proof may look something like:
For any N < 0
For any K > 0
N*K < 0
N is house edge
K is money in action
Quote: cwwbjrThat's a good question. If you are referring to real casino play (not computer generated on line or simulated play) I would ask you 3 questions as follows;
1. How long have you been playing using this system?
2. Are you profitable at this point?
3. Do you have accurate records/proof of all sessions to verify your results?
That would be a starting point.
It's easy for a mathematician/statistician to say that all betting systems are worthless. To put that in a mathematical formula and prove it may be a bit more challenging.
You are right about a formula being challenging! lol
Quote: lovetosing2011I guess I'm trying to figure out how long it takes before its safe to say it's not just luck anymore.
I think that the general consensus here is a certain number of standard deviations of the game in question. I'm sure there is someone here who could say specifically for the game of craps. However, depending upon which bet/bets within the game you are tracking, the SD's may vary.
Quote: cwwbjrI think that the general consensus here is a certain number of standard deviations of the game in question. I'm sure there is someone here who could say specifically for the game of craps. However, depending upon which bet/bets within the game you are tracking, the SD's may vary.
Hey, I got a great idea! Why don't you go to this great website called Wizard of Odds! I'm thinking there is a chart on SD's! Different bets at craps even!! Soon, Love to sing, will be singing the blues!
And >I< completely understand that. Thinking a betting system will work is insane. His question is how many rolls would it take to prove his system works. It's hard to imagine because we know it's not possible.Quote: MathExtremistIt doesn't take any. If you're working from the premise of a fair game (fair dice, unbiased wheel, etc) and you're looking at just betting patterns, you can examine a system on paper and see it doesn't work without betting a single dollar. If you're looking at a real-world set of events and you think they behave *differently than* a canonical casino game, then you're not really testing a system, you're playing that game for bias.
NVM the math for now it's a RW test. If someone had proof over a 100k rolls (not stuff like keno, or some system where the guy makes $1 bets a million times and then bets a billion dollars and wins) At what point would you think something strange was going on?
Quote: RSA proof may look something like:
For any N < 0
For any K > 0
N*K < 0
N is house edge
K is money in action
I'm wondering what the equation would look like.
Quote: lovetosing2011I always read how betting systems cannot be profitable in the long run, how long do you think it would take to prove this to be true for the best of systems?
In craps: 1 hour = 15 shooters = 150 - 200 dice rolls
There is/was a "Betting System Challenge" that was modified for real casino play which gives the required rounds of play after which the system must be profitable to win the challenge. This may answer your question. A word of caution, beware of revealing the strategy of your system for obvious reasons.
Sarcasm?Quote: FDEAD3709You really should try another website. Many of the posters here are on the casino's payroll. If you want to know the TRUTH, you should go to John Patricks website.
If you want to know the TRUTH this is the best site.
If you want someone to praise your BS and a cheering squad join John Patrick's website.
Quote: cwwbjrI'm wondering what the equation would look like.
Equation for what?
Don't you have some real numbers to crunch?Quote: RSEquation for what?
Quote: FDEAD3709You really should try another website. Many of the posters here are on the casino's payroll. If you want to know the TRUTH, you should go to John Patricks website.
Funniest post of the week! Nominate it for post of the year!
Quote: cwwbjrA word of caution, beware of revealing the strategy of your system for obvious reasons.
You can't make this stuff up!
Indians win, Cavaliers clobber. Spurs annihilate Warriors. What a perfect night.
Quote: AxelWolfDon't you have some real numbers to crunch?
I still don't think a martingale on Caveman Keno or Ducks-In-A-Row is going to be +EV, but if you insist....
There's always Super 8 Race )-:Quote: RSI still don't think a martingale on Caveman Keno or Ducks-In-A-Row is going to be +EV, but if you insist....
-----------------------------------------------------------------------------
Times A Wastin
Quote: cwwbjrI'm wondering what the equation would look like.
First define X as your expected win from playing your betting system.
Quote: RS
For any N < 0
For any K > 0
N*K < 0
N is house edge
K is money in action
So then by definition we have
N*K = X.
There's the equation that relates expected win to house edge and money played.
And since N*K < 0, that means X < 0.
In other words, your expected win is negative.
