All mathmatical systems will beat any crap, roulette or other similar game

How, it's simple,

The truth is if you play by the fixed rules of the game as set out you must win with a system. Why is that not possible, because casinos change the rules. Ie there is no rule in craps that sets a house limit. That is a house rule not part of the game of craps. If it were a fixed rule all house limits would be equal. They are not, so they can not be fixed rules of the game.

To be correct in saying no system will work you must use the added on non-fixed modifyed rules. Take away those rules and the "system" can't loose.

That would be true, except for one thing - you are implying that "your bankroll is limited" is a "rule".

If you had an unlimited bankroll and unlimited time and the ability to bet whatever amount you wanted, then even a Martingale system will work eventually, just as if you put six monkeys in front of six typewriters, eventually they will write Hamlet.

Of course, "eventually" is a very, very long time - even if you had one billion monkeys typing at 100 characters per second, and limited them to the 26 letters (all lowercase) and the space bar, it's highly unlikely that, even one million years later, any of them typed, "to be or not to be".

Quote: Whiskeyjack

How, it's simple,

The truth is if you play by the fixed rules of the game as set out you must win with a system. Why is that not possible, because casinos change the rules. Ie there is no rule in craps that sets a house limit. That is a house rule not part of the game of craps. If it were a fixed rule all house limits would be equal. They are not, so they can not be fixed rules of the game.

To be correct in saying no system will work you must use the added on non-fixed modifyed rules. Take away those rules and the "system" can't loose.

Okay, if the house limits are always $10000 maximum, $5 minimum, across the board.... how does your mathematical system beat the table game?

Remove the maximums. Now give me a system that allows me to take a $1000 bank roll and double it (guranteed) playing roulette.

Still doesn't exist.

It's the rules of the game (that you are -EV on every spin) that mean no system will help you, not the externalities.

(I'd like to avoid a debate about infinite bank rolls given infinite time. Just to say, I'm not convinced a infinite bankroll'd Martingale does actually work).

Quote: Whiskeyjack

The truth is if you play by the fixed rules of the game as set out you must win with a system.

True: The system can not lose, if you play by the fixed rules of the game as set out.

The system here is the casino.

I've said it before, I'll say it again. The Martingale, or other progressive systems, DON'T work in theory, even with no limits.

It is possible that you will lose indefinitely, until the end of time. There is no law in probability that requires your luck to ever change. You could lose all the money in the world, bet again, and still lose.

Systems don't work. And the house limit is not the reason.

sure, but that is like saying that you could win over and over for ever?! there is no law that requires your winning luck to change either? the question is about what are the chances..... i have a question on this,

what is the probability that from any hand in baccarat the subsequent hands will show the player to have won 80 more than the bank?

what is the formula for this if i wanted to change the 80 to 100 or 200?

lets say x is the point at which we are beginning to record the results, and from x, we see that (regardless of the number of hands played or the history of results) the player has won 80 more times than the bank...... PLEASE HELP with a formula

thank you so much. this can beat the casinos if the chances are low enough.

The formula is (p/(1-p))^i, where p is the probability of winning and i is the number of hands more that you want to see. Derived from (one-sided infinite) gambler's ruin.

Now if you can figure out how to beat the casinos with this information, I will be impressed.

To The OP:

Quote: dwheatley

The formula is (p/(1-p))^i, where p is the probability of winning and i is the number of hands more that you want to see. Derived from (one-sided infinite) gambler's ruin.

Now if you can figure out how to beat the casinos with this information, I will be impressed.

THANK YOU so much for doing that just 2 questions, is (1-p) meant to be the probability of losing? because it's not equal to 1-p so can i just put the probability of losing in there? also can you please explain how and why you get that formula? thank you so much and you will be the first to know if it works!! :)

You can ignore ties, so p is the probability of winning, ignoring ties, while 1-p is the probability of losing, ignoring ties. I assume p < 1-p.

The formula comes from the gambler's ruin formula which gives you the probability of getting to N before 0 from any amount i in a losing game. When N goes to infinity, it becomes your problem backwards: what is the probability of getting to 0 from any amount i in a winning game?

also what is the resulting number? for example if heads and tails is .5/.5^50 equals 1, what is the 1?

The formula isn't meant for the H/T case, you should have p < 0.5.

But it still makes sense: the 1 means the as the number of flips goes to infinity, there is a 100% chance you will have 50 more heads than tails at some point.

Quote: dwheatley

You can ignore ties, so p is the probability of winning, ignoring ties, while 1-p is the probability of losing, ignoring ties. I assume p < 1-p.

The formula comes from the gambler's ruin formula which gives you the probability of getting to N before 0 from any amount i in a losing game. When N goes to infinity, it becomes your problem backwards: what is the probability of getting to 0 from any amount i in a winning game?

this is meant for baccarat which i believe has the best odds of any game for a win (bank bet)

so the numbers would be p=0.458597.........player wins are 0.446247 and then tie is 0.095156

could you please show how that fits it with the formula?

Quote: dwheatley

The formula isn't meant for the H/T case, you should have p < 0.5.

But it still makes sense: the 1 means the as the number of flips goes to infinity, there is a 100% chance you will have 50 more heads than tails at some point.

does that mean that at some point you will a discrepancy of 1000000000 more heads than tails at some point??

also even though the probability of win is less than .5 you must have a loss to get to ruin, which means the 1-p is not accurate because the loss probability is only .446247??

the probability of not losing is greater than .5, so is this still the formula to use??

thank you so much :)

Quote: dwheatley

I've said it before, I'll say it again. The Martingale, or other progressive systems, DON'T work in theory, even with no limits.

It is possible that you will lose indefinitely, until the end of time. There is no law in probability that requires your luck to ever change. You could lose all the money in the world, bet again, and still lose.

Systems don't work. And the house limit is not the reason.

I'm not sure why so many people on here say this frequently. It isn't true. Do it in excel. I did it and posted the spreadsheet on another thread. If you have no limits, martingale works, and gains over time.

Quote: slackyhacky

I'm not sure why so many people on here say this frequently. It isn't true. Do it in excel. I did it and posted the spreadsheet on another thread. If you have no limits, martingale works, and gains over time.

No, it doesn't. It's really pretty easy: if you're equally likely to win and lose, you'll make $0 over infinite time by martingaling (or using any other cash management system). If you're more likely to lose than win, you'll lose infinite money over infinite time by martingaling (or using any other cash management system). If you're more likely to win than lose, there are better cash management systems to use to maximize your gains.

Quote: rdw4potus

No, it doesn't. It's really pretty easy: if you're equally likely to win and lose, you'll make $0 over infinite time by martingaling (or using any other cash management system). If you're more likely to lose than win, you'll lose infinite money over infinite time by martingaling (or using any other cash management system). If you're more likely to win than lose, there are better cash management systems to use to maximize your gains.

Show me.

Make an excel spreadsheet. Let's not say he said, she said. Show me. Not with formulas - they are like statistics (which from mark twain we know are lies).

show me.

Quote: jms9smith

this is meant for baccarat which i believe has the best odds of any game for a win (bank bet)

so the numbers would be p=0.458597.........player wins are 0.446247 and then tie is 0.095156

could you please show how that fits it with the formula?

you ignore ties, and normalize the banker bet to p=0.506824381, and player p=0.493175619.

The banker is expected to win more often, so eventually any number of banker wins more than player is expected. The problem is the commission.
You can still use the formula on player p=0.493175619.

Quote: jms9smith

does that mean that at some point you will a discrepancy of 1000000000 more heads than tails at some point??

That's the idea: at some point, any ridiculously large discrepancy is possible, and expected, in a H/T game of infinite length.

Quote: slackyhacky

Show me.

Make an excel spreadsheet. Let's not say he said, she said. Show me. Not with formulas - they are like statistics (which from mark twain we know are lies).

show me.

No. Also, Mark Twain was an author, not a statistician. It isn't even his quote. He was just grumpy about not being able to understand numbers.

Quote: slackyhacky

Show me.

Make an excel spreadsheet. Let's not say he said, she said. Show me. Not with formulas - they are like statistics (which from mark twain we know are lies).

show me.

You show me how to make a spreadsheet that includes infinite time and money without using formulas, and I'll be happy to comply.

Quote: rdw4potus

You show me how to make a spreadsheet that includes infinite time and money without using formulas, and I'll be happy to comply.

Come on. That's radiculous.

I got sick of waiting. here is mine.

https://docs.google.com/open?id=0B3eruTU5Xu_jOTg0YjgxMGMtMTk5Yi00NzllLWEzZGUtMGZiNzE0OTE2YTNm

download the file to your desktop so you can look at the formulas and run it as many times as you want.

Hit F9 to refresh. Hit it as many times as you want. the roulette martingale makes money EVERY time. It doesn't take infinite anything. I also have a lay system as well (you need to multiple x3 plus the vig to cover the loss). This doesn't work as well, but OFTEN makes money, but looses a bit sometimes.

Quote: slackyhacky

Show me.

Make an excel spreadsheet. Let's not say he said, she said. Show me. Not with formulas - they are like statistics (which from mark twain we know are lies).

show me.

First off, there's no point in using a spreadsheet without formulas, so let's skip the spreadsheet. Want to see how the Martingale can't change the house edge? Let's start with two assumptions.

One is that you agree that a player who makes flat bets doesn't expect to beat the game. Let's use the red bet on roulette. Do you accept that a player flat-betting red on an American roulette table has a house edge of 5.26%?

Two is that the results of a player at a first roulette table are totally independent of a player at a second roulette table. Whether the second player wins or loses has nothing to do with whether the first player wins or loses.

If you accept these two basic facts as true, then it is easy to see that the Martingale doesn't "work". That's because the Martingale is equivalent to a bunch of different players at different tables, all of them flat betting different amounts, and then adding up the total results. Here's how:

At table 1, Andy bets $1 on red. He keeps betting $1 until he loses. When he loses, he shouts to Bill that it's his turn.
At table 2, Bill waits for Andy to yell at him. When Bill hears Andy, Bill bets $2 on red. If he wins, he shouts back to Andy to start again. If he loses, he shouts to Chuck.
At table 3, Chuck waits for Bill. When Chuck hears Bill, Chuck bets $4 on red. If he wins, he shouts back to Andy. If he loses, he shouts to Dave.
At table 4, Dave waits for Chuck. When Dave hears Chuck, Dave bets $8 on red. If he wins, he shouts back to Andy. If he loses, he shouts to Edgar.
At table 5, Edgar waits for Dave. When Edgar hears Dave, Edgar bets $16 on red. If he wins, he shouts back to Andy. If he loses, he shouts to Frank.
This pattern continues for another 5 or 10 tables, depending on the maximum house limit. Maybe the rest of the players are named Gerald, Hank, Izzy, Joe, Kurt, Leif, Mark, Nick, Oswald, and Paul -- they all wait until the player at the table before them lost, and then they make a flat bet that is double the last player's wager. Except for Andy, they never bet twice in a row, and Andy only bets multiple times in a row when he goes on a winning streak.

All of these players are flat-betting. They never change their wagers, and they just sit around most of the time waiting for their designated time to bet. And if you were just walking by the roulette table where, say, Frank was playing, you'd never say "wow, Frank just waits around and bets $32 on red every once in a while. That's a winning system!" In fact, you would expect Frank to lose his fair share of those $32 bets and overall have results that approach the house edge. You would never think that Frank was somehow a guaranteed long-term winner.

But here's the point: if Andy isn't a guaranteed long-term winner, and Bill isn't a guaranteed long-term winner, and Chuck isn't a guaranteed long-term winner, and so forth for the rest of them, then the combination of all of them can't be a guaranteed long-term winner either. After all, you're just adding up their results, and if all of them are losers individually, they're all losers together.

Now consider that instead of using Bill, Chuck, Dave, etc., to make bigger bets, Andy just runs from one table to the next, making the flat bet of the appropriate size on the table. If he loses, he goes to table 2 and bets $2. If he loses there, he goes to table 3 and bets $4. If he wins, he goes back to table 1 and starts over at $1. Now it's all Andy's money, but it's exactly the same results as if Bill, Chuck, Dave, etc. had been betting some of it. Andy still isn't a guaranteed long-term winner.

And finally, consider that instead of Andy running from one table to another, he gets lazy and just sits at the same table making the bets he was going to make at all the other tables. He starts at $1, then flat bets until he loses and then bets $2. If he loses he bets $4; if he wins he starts over at $1. That's the Martingale system. Still think it's a guaranteed long-term winner?

Quote: slackyhacky

Come on. That's radiculous.

I got sick of waiting. here is mine.

https://docs.google.com/open?id=0B3eruTU5Xu_jOTg0YjgxMGMtMTk5Yi00NzllLWEzZGUtMGZiNzE0OTE2YTNm

download the file to your desktop so you can look at the formulas and run it as many times as you want.

Hit F9 to refresh. Hit it as many times as you want. the roulette martingale makes money EVERY time. It doesn't take infinite anything. I also have a lay system as well (you need to multiple x3 plus the vig to cover the loss). This doesn't work as well, but OFTEN makes money, but looses a bit sometimes.

I noticed you went as high as 8 losses in a row, with the top bet of $512 winning. A person doing this will inevitably continue catching those losses at 1024, 2048, 4096 etc., etc. I wouldn't want to be that guy whose bankroll becomes busted.

Quote: MathExtremist

If you accept these two basic facts as true, then it is easy to see that the Martingale doesn't "work". That's because the Martingale is equivalent to a bunch of different players at different tables, all of them flat betting different amounts, and then adding up the total results. Here's how:

All of these players are flat-betting.... then...... he gets lazy and just sits at the same table making the bets he was going to make at all the other tables. He starts at $1, then flat bets until he loses and then bets $2. If he loses he bets $4; if he wins he starts over at $1. That's the Martingale system. Still think it's a guaranteed long-term winner?

Owned. ME sinks the 30 foot putt for eagle!

Quote: slackyhacky

Come on. That's radiculous.

I got sick of waiting. here is mine.

https://docs.google.com/open?id=0B3eruTU5Xu_jOTg0YjgxMGMtMTk5Yi00NzllLWEzZGUtMGZiNzE0OTE2YTNm

download the file to your desktop so you can look at the formulas and run it as many times as you want.

Hit F9 to refresh. Hit it as many times as you want. the roulette martingale makes money EVERY time. It doesn't take infinite anything. I also have a lay system as well (you need to multiple x3 plus the vig to cover the loss). This doesn't work as well, but OFTEN makes money, but looses a bit sometimes.

OK. Even though I didn't bother to examine the formulas, I tried it.

In the 100 times I hit F9, there were some net profits, and some net losses. Never more than +/- 150,000.

Except for the three times that I lost the entire 10 million!

hi ME,

very good post.
In my opinion your fact number two (which is true, of course) and the fact that result A at time tA does not influence result B at time tB at the same table is something people (at least those believing in martingale) do not grasp at the same thing.
Most believers of the martingale system believe that the last result doesn't influence the next result. They hear losing twice in a row happens only 25% of the time. So after they lost one time they think that now they are more lucky to win. They do not understand that the first 50/50 chance has already happend and the next is still 50/50. To make it more visible. Here are the possible results of 3 50/50 events:

LLL = 1/8
LLW = 1/8
LWL = 1/8
LWW = 1/8
WLL = 1/8
WLW = 1/8
WWL = 1/8
WWW = 1/8

Now because LLL only has a chance of 1/8 they think if they lost twice in a row and losing 3 times in a row is just so unlikely (1/8) they are due to win. What they don't realize is that their are already at point LL (1/4) and still at a 50/50 chance to either hit LLW or LLL.

For all of you who don't want to download this document, it's a craps simulator where the player is laying the 10 and tripling his bet after a loss.

So, if you bet $21, you win $10. If you lose, you bet $63 to win $21. This is over 40,000 rolls, which represents about 400 hours on a craps table or 8 hours a week for a year.

So, after pressing F9 a few times, here's the minimum and maximums in bankroll (starting with a 10,000,000 bankroll)

1. 9,078,106
2. 9,727,625
3. 9,923,911
4. 9,702,425
5. 9,717,499
6. 9,975,735
7. 9,938,614
8. 9,908,321
9. 9,392,203
10. 623,821 (loss of 9.7 million)

In 100 trials, I lost the kitty (all of it) FOUR times.

Quote: slackyhacky

Come on. That's radiculous.

I got sick of waiting. here is mine.

https://docs.google.com/open?id=0B3eruTU5Xu_jOTg0YjgxMGMtMTk5Yi00NzllLWEzZGUtMGZiNzE0OTE2YTNm

download the file to your desktop so you can look at the formulas and run it as many times as you want.

Hit F9 to refresh. Hit it as many times as you want. the roulette martingale makes money EVERY time. It doesn't take infinite anything. I also have a lay system as well (you need to multiple x3 plus the vig to cover the loss). This doesn't work as well, but OFTEN makes money, but looses a bit sometimes.

Nice spreadsheet. But it's not a martingale system, it uses a fixed bankroll (not infinite), and it's not making money.

Quote: rdw4potus

Nice spreadsheet. But it's not a martingale system, it uses a fixed bankroll (not infinite), and it's not making money.

uhhh......

where to start...

please screen shot the roulette side showing it not making money. It does everytime for me.

martingale doubles your losses. My spread sheet clearly does this. Please define your definition of martingale.

I don't care about infinite. What does that have to do with it? Show me a spreadsheet that uses a system that doubles your losses (whatever you want to call it) that doesn't work. I have shown you mine, it works everytime I have run it (probably over 3 million now) with a gain.

Let's not talk. let's show. How about that? post the image. post the screen shot.

Quote: boymimbo

For all of you who don't want to download this document, it's a craps simulator where the player is laying the 10 and tripling his bet after a loss.

So, if you bet $21, you win $10. If you lose, you bet $63 to win $21. This is over 40,000 rolls, which represents about 400 hours on a craps table or 8 hours a week for a year.

So, after pressing F9 a few times, here's the minimum and maximums in bankroll (starting with a 10,000,000 bankroll)

1. 9,078,106
2. 9,727,625
3. 9,923,911
4. 9,702,425
5. 9,717,499
6. 9,975,735
7. 9,938,614
8. 9,908,321
9. 9,392,203
10. 623,821 (loss of 9.7 million)

In 100 trials, I lost the kitty (all of it) FOUR times.

I have both roulette and craps on there (as I thought I mentioned.)

Quote: gameterror

To make it more visible. Here are the possible results of 3 50/50 events:

LLL = 1/8
LLW = 1/8
LWL = 1/8
LWW = 1/8
WLL = 1/8
WLW = 1/8
WWL = 1/8
WWW = 1/8

No because LLL only has a chance of 1/8 they think if they lost twice in a row and losing 3 times in a row is just so unlikely (1/8) they are due to win. What they don't realize is that their are already at point LL (1/4) and still at a 50/50 chance to either hit LLW or LLL.

And gameterror holes out from the sand for birdie!!! OH MY GOODNESS!

Quote: slackyhacky

Quote: boymimbo

For all of you who don't want to download this document, it's a craps simulator where the player is laying the 10 and tripling his bet after a loss.

So, if you bet $21, you win $10. If you lose, you bet $63 to win $21. This is over 40,000 rolls, which represents about 400 hours on a craps table or 8 hours a week for a year.

So, after pressing F9 a few times, here's the minimum and maximums in bankroll (starting with a 10,000,000 bankroll)

1. 9,078,106
2. 9,727,625
3. 9,923,911
4. 9,702,425
5. 9,717,499
6. 9,975,735
7. 9,938,614
8. 9,908,321
9. 9,392,203
10. 623,821 (loss of 9.7 million)

In 100 trials, I lost the kitty (all of it) FOUR times.

I have both roulette and craps on there (as I thought I mentioned.)

You've definitely mentioned that. the spreadsheet opens to the craps sim, which is on the right side of the page. I just now noticed that the roulette is there if I scroll left. after looking at it for 5 minutes, it appears that the roulette is coded slightly incorrectly. You have it looking for a random number between 1 and 38, and paying even money on 20 or over. Actual roulette basically looks for a random number between 1 and 38, pays even money on 18 or under, and take's the player's money on 19 or over. 37 and 38 would map to 0 and 00. Your "roulette" sim has no house edge, but the house edge in American roulette is 5.26%.

Quote: rdw4potus

Quote: slackyhacky

Quote: boymimbo

For all of you who don't want to download this document, it's a craps simulator where the player is laying the 10 and tripling his bet after a loss.

So, if you bet $21, you win $10. If you lose, you bet $63 to win $21. This is over 40,000 rolls, which represents about 400 hours on a craps table or 8 hours a week for a year.

So, after pressing F9 a few times, here's the minimum and maximums in bankroll (starting with a 10,000,000 bankroll)

1. 9,078,106
2. 9,727,625
3. 9,923,911
4. 9,702,425
5. 9,717,499
6. 9,975,735
7. 9,938,614
8. 9,908,321
9. 9,392,203
10. 623,821 (loss of 9.7 million)

In 100 trials, I lost the kitty (all of it) FOUR times.

I have both roulette and craps on there (as I thought I mentioned.)

You've definitely mentioned that. the spreadsheet opens to the craps sim, which is on the right side of the page. I just now noticed that the roulette is there if I scroll left. after looking at it for 5 minutes, it appears that the roulette is coded slightly incorrectly. You have it looking for a random number between 1 and 38, and paying even money on 20 or over. Actual roulette basically looks for a random number between 1 and 38, pays even money on 18 or under, and take's the player's money on 19 or over. 37 and 38 would map to 0 and 00. Your "roulette" sim has no house edge, but the house edge in American roulette is 5.26%.

There's also something wrong with column L. You have it coded as 2^(corresponding value in column k). I'm not sure what you're trying to do there, but that makes the bankroll go up by 2 for every winning 1 unit bet.

Quote: rdw4potus

You've definitely mentioned that. the spreadsheet opens to the craps sim, which is on the right side of the page. I just now noticed that the roulette is there if I scroll left. after looking at it for 5 minutes, it appears that the roulette is coded slightly incorrectly. You have it looking for a random number between 1 and 38, and paying even money on 20 or over. Actual roulette basically looks for a random number between 1 and 38, pays even money on 18 or under, and take's the player's money on 19 or over. 37 and 38 would map to 0 and 00. Your "roulette" sim has no house edge, but the house edge in American roulette is 5.26%.

BOOM. Roasted.

Quote: rdw4potus

Quote: rdw4potus

Quote: slackyhacky

Quote: boymimbo

For all of you who don't want to download this document, it's a craps simulator where the player is laying the 10 and tripling his bet after a loss.

So, if you bet $21, you win $10. If you lose, you bet $63 to win $21. This is over 40,000 rolls, which represents about 400 hours on a craps table or 8 hours a week for a year.

So, after pressing F9 a few times, here's the minimum and maximums in bankroll (starting with a 10,000,000 bankroll)

1. 9,078,106
2. 9,727,625
3. 9,923,911
4. 9,702,425
5. 9,717,499
6. 9,975,735
7. 9,938,614
8. 9,908,321
9. 9,392,203
10. 623,821 (loss of 9.7 million)

In 100 trials, I lost the kitty (all of it) FOUR times.

I have both roulette and craps on there (as I thought I mentioned.)

You've definitely mentioned that. the spreadsheet opens to the craps sim, which is on the right side of the page. I just now noticed that the roulette is there if I scroll left. after looking at it for 5 minutes, it appears that the roulette is coded slightly incorrectly. You have it looking for a random number between 1 and 38, and paying even money on 20 or over. Actual roulette basically looks for a random number between 1 and 38, pays even money on 18 or under, and take's the player's money on 19 or over. 37 and 38 would map to 0 and 00. Your "roulette" sim has no house edge, but the house edge in American roulette is 5.26%.

There's also something wrong with column L. You have it coded as 2^(corresponding value in column k). I'm not sure what you're trying to do there, but that makes the bankroll go up by 2 for every winning 1 unit bet.

Here, I lose 4 times in a row from the start. I have bets of 1, 2, 4, and 8. After those losses, my bankroll should be 999999, 999997, 999993, and 999985. The negative impact of the 4th loss doesn't seem to be reflected in the simulation.

Quote: DJTeddyBear

Quote: slackyhacky

Come on. That's radiculous.

I got sick of waiting. here is mine.

https://docs.google.com/open?id=0B3eruTU5Xu_jOTg0YjgxMGMtMTk5Yi00NzllLWEzZGUtMGZiNzE0OTE2YTNm

download the file to your desktop so you can look at the formulas and run it as many times as you want.

Hit F9 to refresh. Hit it as many times as you want. the roulette martingale makes money EVERY time. It doesn't take infinite anything. I also have a lay system as well (you need to multiple x3 plus the vig to cover the loss). This doesn't work as well, but OFTEN makes money, but looses a bit sometimes.

OK. Even though I didn't bother to examine the formulas, I tried it.

In the 100 times I hit F9, there were some net profits, and some net losses. Never more than +/- 150,000.

Except for the three times that I lost the entire 10 million!

If you lost 10 mil, you weren't looking in the right place. The roulette sim starts with 1mil, unless you put in your own starting value.

Quote: rdw4potus

Quote: slackyhacky

Quote: boymimbo

For all of you who don't want to download this document, it's a craps simulator where the player is laying the 10 and tripling his bet after a loss.

So, if you bet $21, you win $10. If you lose, you bet $63 to win $21. This is over 40,000 rolls, which represents about 400 hours on a craps table or 8 hours a week for a year.

So, after pressing F9 a few times, here's the minimum and maximums in bankroll (starting with a 10,000,000 bankroll)

1. 9,078,106
2. 9,727,625
3. 9,923,911
4. 9,702,425
5. 9,717,499
6. 9,975,735
7. 9,938,614
8. 9,908,321
9. 9,392,203
10. 623,821 (loss of 9.7 million)

In 100 trials, I lost the kitty (all of it) FOUR times.

I have both roulette and craps on there (as I thought I mentioned.)

You've definitely mentioned that. the spreadsheet opens to the craps sim, which is on the right side of the page. I just now noticed that the roulette is there if I scroll left. after looking at it for 5 minutes, it appears that the roulette is coded slightly incorrectly. You have it looking for a random number between 1 and 38, and paying even money on 20 or over. Actual roulette basically looks for a random number between 1 and 38, pays even money on 18 or under, and take's the player's money on 19 or over. 37 and 38 would map to 0 and 00. Your "roulette" sim has no house edge, but the house edge in American roulette is 5.26%.

No, it has the HE in. 20 ways to loose, 18 ways to win. It doesn't matter which is what - they all have the same chance of occurring. Any number 1-18 wins (on my sim), any number 19-38 (20 numbers) looses.

Quote: AcesAndEights

BOOM. Roasted.

Are you sure? Really sure?

Quote: rdw4potus

Quote: rdw4potus

Quote: slackyhacky

Quote: boymimbo

For all of you who don't want to download this document, it's a craps simulator where the player is laying the 10 and tripling his bet after a loss.

So, if you bet $21, you win $10. If you lose, you bet $63 to win $21. This is over 40,000 rolls, which represents about 400 hours on a craps table or 8 hours a week for a year.

So, after pressing F9 a few times, here's the minimum and maximums in bankroll (starting with a 10,000,000 bankroll)

1. 9,078,106
2. 9,727,625
3. 9,923,911
4. 9,702,425
5. 9,717,499
6. 9,975,735
7. 9,938,614
8. 9,908,321
9. 9,392,203
10. 623,821 (loss of 9.7 million)

In 100 trials, I lost the kitty (all of it) FOUR times.

I have both roulette and craps on there (as I thought I mentioned.)

You've definitely mentioned that. the spreadsheet opens to the craps sim, which is on the right side of the page. I just now noticed that the roulette is there if I scroll left. after looking at it for 5 minutes, it appears that the roulette is coded slightly incorrectly. You have it looking for a random number between 1 and 38, and paying even money on 20 or over. Actual roulette basically looks for a random number between 1 and 38, pays even money on 18 or under, and take's the player's money on 19 or over. 37 and 38 would map to 0 and 00. Your "roulette" sim has no house edge, but the house edge in American roulette is 5.26%.

There's also something wrong with column L. You have it coded as 2^(corresponding value in column k). I'm not sure what you're trying to do there, but that makes the bankroll go up by 2 for every winning 1 unit bet.

Nothing wrong with column L for a loss. It adds 2 for a win...hmmmm, didn't notice that before. Column K counts the number of losses in a row. Thus, each lose the money is doubled. L does this. If you loose 4 times, your bet will be (1,2,4,16). or 2^(# times lost).

Quote: rdw4potus

Quote: rdw4potus

Quote: rdw4potus

Quote: slackyhacky

Quote: boymimbo

For all of you who don't want to download this document, it's a craps simulator where the player is laying the 10 and tripling his bet after a loss.

So, if you bet $21, you win $10. If you lose, you bet $63 to win $21. This is over 40,000 rolls, which represents about 400 hours on a craps table or 8 hours a week for a year.

So, after pressing F9 a few times, here's the minimum and maximums in bankroll (starting with a 10,000,000 bankroll)

1. 9,078,106
2. 9,727,625
3. 9,923,911
4. 9,702,425
5. 9,717,499
6. 9,975,735
7. 9,938,614
8. 9,908,321
9. 9,392,203
10. 623,821 (loss of 9.7 million)

In 100 trials, I lost the kitty (all of it) FOUR times.

I have both roulette and craps on there (as I thought I mentioned.)

You've definitely mentioned that. the spreadsheet opens to the craps sim, which is on the right side of the page. I just now noticed that the roulette is there if I scroll left. after looking at it for 5 minutes, it appears that the roulette is coded slightly incorrectly. You have it looking for a random number between 1 and 38, and paying even money on 20 or over. Actual roulette basically looks for a random number between 1 and 38, pays even money on 18 or under, and take's the player's money on 19 or over. 37 and 38 would map to 0 and 00. Your "roulette" sim has no house edge, but the house edge in American roulette is 5.26%.

There's also something wrong with column L. You have it coded as 2^(corresponding value in column k). I'm not sure what you're trying to do there, but that makes the bankroll go up by 2 for every winning 1 unit bet.

Here, I lose 4 times in a row from the start. I have bets of 1, 2, 4, and 8. After those losses, my bankroll should be 999999, 999997, 999993, and 999985. The negative impact of the 4th loss doesn't seem to be reflected in the simulation.

If you add a zero in the formula in K (instead of 1) it fixes it. "=IF(G3+G4>1, 1+K3,0)" instead of "=IF(G3+G4>1, 1+K3,1). Although it makes me win half, as much it still never looses.

Quote: MathExtremist

First off, there's no point in using a spreadsheet without formulas, so let's skip the spreadsheet. Want to see how the Martingale can't change the house edge? Let's start with two assumptions.

One is that you agree that a player who makes flat bets doesn't expect to beat the game. Let's use the red bet on roulette. Do you accept that a player flat-betting red on an American roulette table has a house edge of 5.26%?

Two is that the results of a player at a first roulette table are totally independent of a player at a second roulette table. Whether the second player wins or loses has nothing to do with whether the first player wins or loses.

If you accept these two basic facts as true, then it is easy to see that the Martingale doesn't "work". That's because the Martingale is equivalent to a bunch of different players at different tables, all of them flat betting different amounts, and then adding up the total results. Here's how:

At table 1, Andy bets $1 on red. He keeps betting $1 until he loses. When he loses, he shouts to Bill that it's his turn.
At table 2, Bill waits for Andy to yell at him. When Bill hears Andy, Bill bets $2 on red. If he wins, he shouts back to Andy to start again. If he loses, he shouts to Chuck.
At table 3, Chuck waits for Bill. When Chuck hears Bill, Chuck bets $4 on red. If he wins, he shouts back to Andy. If he loses, he shouts to Dave.
At table 4, Dave waits for Chuck. When Dave hears Chuck, Dave bets $8 on red. If he wins, he shouts back to Andy. If he loses, he shouts to Edgar.
At table 5, Edgar waits for Dave. When Edgar hears Dave, Edgar bets $16 on red. If he wins, he shouts back to Andy. If he loses, he shouts to Frank.
This pattern continues for another 5 or 10 tables, depending on the maximum house limit. Maybe the rest of the players are named Gerald, Hank, Izzy, Joe, Kurt, Leif, Mark, Nick, Oswald, and Paul -- they all wait until the player at the table before them lost, and then they make a flat bet that is double the last player's wager. Except for Andy, they never bet twice in a row, and Andy only bets multiple times in a row when he goes on a winning streak.

All of these players are flat-betting. They never change their wagers, and they just sit around most of the time waiting for their designated time to bet. And if you were just walking by the roulette table where, say, Frank was playing, you'd never say "wow, Frank just waits around and bets $32 on red every once in a while. That's a winning system!" In fact, you would expect Frank to lose his fair share of those $32 bets and overall have results that approach the house edge. You would never think that Frank was somehow a guaranteed long-term winner.

But here's the point: if Andy isn't a guaranteed long-term winner, and Bill isn't a guaranteed long-term winner, and Chuck isn't a guaranteed long-term winner, and so forth for the rest of them, then the combination of all of them can't be a guaranteed long-term winner either. After all, you're just adding up their results, and if all of them are losers individually, they're all losers together.

Now consider that instead of using Bill, Chuck, Dave, etc., to make bigger bets, Andy just runs from one table to the next, making the flat bet of the appropriate size on the table. If he loses, he goes to table 2 and bets $2. If he loses there, he goes to table 3 and bets $4. If he wins, he goes back to table 1 and starts over at $1. Now it's all Andy's money, but it's exactly the same results as if Bill, Chuck, Dave, etc. had been betting some of it. Andy still isn't a guaranteed long-term winner.

And finally, consider that instead of Andy running from one table to another, he gets lazy and just sits at the same table making the bets he was going to make at all the other tables. He starts at $1, then flat bets until he loses and then bets $2. If he loses he bets $4; if he wins he starts over at $1. That's the Martingale system. Still think it's a guaranteed long-term winner?

Mathextemist

Rather than talk me into a logical corner, show me. Dazzle me with some extreme math. Actually, the math is easy so never mind about the extreme math, but show me nonetheless.

Starting with 1 million dollars, I have run the sim tons of times and never seen it go to zero - it always seems to end up around $17000 profit after 40,000 rolls.

Why is this all-powerful system not being employed in every casino in America? If it is so clear to you that your system is so dominant, why don't you put it into practice?

Quote: slackyhacky

Mathextemist

Rather than talk me into a logical corner, show me. Dazzle me with some extreme math. Actually, the math is easy so never mind about the extreme math, but show me nonetheless.

http://www.bjmath.com/bjmath/progress/unfair.htm

The info in the above link is too dazzling for most (and I freely admit to being a member of that dazzled group), but feel free to deconstruct and destroy the math behind it and claim victory for progression systems. Personally, I'm backing the math.

(I only posted the link as I didn't wish to breach any copyright or forum rules but if it's ok to post the entire content then feel free to do so).

Quote: slackyhacky

Mathextemist

Rather than talk me into a logical corner, show me. Dazzle me with some extreme math. Actually, the math is easy so never mind about the extreme math, but show me nonetheless.

Starting with 1 million dollars, I have run the sim tons of times and never seen it go to zero - it always seems to end up around $17000 profit after 40,000 rolls.

I'll copy a short explaination from my blog..hope that's ok.

As you can see in the above tables you will bust once out of 32 times. When you do you will lose a total of 31 units. The other 31 times you will win one unit each. This results in an expected value (EV) of:
(31/32 * 1) + (1/32 * -31) = 0

You can continue this table into infinite and the result will still be the same. As you can see this was done for a fair game with an EV of 0 for each single bet. The betting system doesn't influence the EV. So for your negative expectation in the casino...it won't help.

source: post about martingale

ps: hope that is ok with the forum rules and not considered spam...otherwise i'll delete it.

Quote: slackyhacky

Quote: DJTeddyBear

In the 100 times I hit F9, there were some net profits, and some net losses. Never more than +/- 150,000.

Except for the three times that I lost the entire 10 million!

If you lost 10 mil, you weren't looking in the right place. The roulette sim starts with 1mil, unless you put in your own starting value.

The version I downloaded starts with 10 mil. And after running it a few more times, I'll adjust my statement slightly: I've never lost or won more than +/- 250,000 except for the times I've lost the entire 10 mil.

Mind you, I may have had higher wins or losses mid-stream. I'm only looking at the net profit, in cell X40141.

2 Points:

(1) Your craps simulator showed losses of 10,000,000 4 times out of the 100 times I played. I don't have $10,000,000 laying around, don't have a casino that has limits between $10 and $10,000,000, and certainly wouldn't risk a 4% chance of losing 10,000,000 to gain 20 - 40,000.

(2) Your roulette simulation allows ruin. That is, it allows a bet size up and over the $1,000,000 seed. And you're right. After 40,000 trials, you are almost always making a profit. You would need to calculate the odds that you would at your about 15th or greater loss in a row (betting 32768) on your 40,000th trial. The odds of that are (20/38)^n where n> 15 or about 7,191 to 1. So, every 7,000 trials or so, you would show a loss after trial 40,000. But in the 100 simulations that I ran, you bet more than 1,000,000 (your entire bankroll) about 8 times. Given the odds are that you will be betting over 1,000,000 on any given trial is about 178,000:1, this makes sense.

Martingales don't work, in real life (as the casino has limits and your bankroll is finite), or in theory either.

Quote: slackyhacky

If you add a zero in the formula in K (instead of 1) it fixes it. "=IF(G3+G4>1, 1+K3,0)" instead of "=IF(G3+G4>1, 1+K3,1). Although it makes me win half, as much it still never looses.

I altered the formula. It didnt' fix the issue with the final loss. Here's another screenshot from this morning after I altered that formula. It still looks like the spreadsheet is not subtracting the 32 units for the last loss before the big win. The series starts at 1,000,000 and ends at 1,000,002. So, it's ending in the right place (except that it should be 1,000,001 for a net 1 unit win) but it's shorting the sequence by one loss in the internal accounting. This will make a big difference for situations when the string of losses is at the end of the session, or for when the string of losses tests the bankroll limitation.

Also, beneath the highlighted section, there's a 1 unit loss in row 23. That's followed by a 1 unit bet in row 24. The bet in row 24 needs to be 2 units to properly recoup the 1 unit loss on the previous decision.

The 32 is the bet size, not the amount lost/won. His martingale formula works fine, I think.

Row 15 Win +1
Row 16 Lose - 1
Row 17 Lose - 1 Bet Size 2 on next bet
Row 18 Lose -2 Bet Size 4 on next bet
Row 19 Lose 8 Bet Size 8 on next bet...
Row 20 Lose 8 Bet size 16 on next bet...
Row 21 Lose 16 Bet size 32 on next bet
Row 22 Win 32 Bet Size 1 on next bet.
Row 34 Win 1 Bet size 1 on next bet...

He doesn't double his bet after the 1st loss.

Quote: boymimbo

His martingale formula works fine, I think.

Point of clarification:

The formula works.

Martingale doesn't - and that's proved by the formula, and the frequency of complete ruin.

Quote: boymimbo

The 32 is the bet size, not the amount lost/won. His martingale formula works fine, I think.

Row 15 Win +1
Row 16 Lose - 1
Row 17 Lose - 1 Bet Size 2 on next bet
Row 18 Lose -2 Bet Size 4 on next bet
Row 19 Lose 8 Bet Size 8 on next bet...
Row 20 Lose 8 Bet size 16 on next bet...
Row 21 Lose 16 Bet size 32 on next bet
Row 22 Win 32 Bet Size 1 on next bet.
Row 34 Win 1 Bet size 1 on next bet...

He doesn't double his bet after the 1st loss.

I see what you're saying, but it still doesn't look quite right to me. Looking at row 4, there's a win, but no corresponding BR increase. I think it's not so much that there's no doubling after the first loss as it is that the rows are mis-aligned and the formula is mis-calculating as a result.