Does martingale really works?

[MoneyGod]

I try this method and have lost too much once I have 15 time losing streak and second time I have 17 time losing streak so I believe its depend on luck also no method works with out this

[sed]

Your expected profit is negative.

Actually that just isn't true.  If you have a buffer of 10 rolls, and you roll 100 times.  Your expect PROFIT is going to probably be around 40 base units...

If you do that your expected "profit" is -5 base units.

I think some of the folks here don't know the definition of expectation or the definition of profit

Folks, expectation is the value you get when you calculate the long term average for your situation (it factors out all local variance).

Profit is the amount you have *above* what you started out with.  Now if you can forsee the result of all of your bets in such a way that at a real casino your expectation is positive profit, then you are a prophet.

Well that isn't true.....

I could easily go walk into a casino today, lay down 1000 dollars and probably end up + 5 dollars so I would walk away at $1005 dollars.

It hardly makes me a prophet, playing Martingale at a casino and winning 1 hand the odds are HIGHLY in your favor.  But, if you keep doing that, eventually you are going to lose the $1000 dollars...

If you walk into a casino and place a bet for 1000$ that pays 1005$ for the win, then (a) that's not a very normal game is it? (b) the long term expected average isn't calculated by simulating one roll.  It's calculated by simulating an infinity of rolls, obviously you can't simulate infinity so you just find the limit.

[Don007]

I try this method and have lost too much once I have 15 time losing streak and second time I have 17 time losing streak so I believe its depend on luck also no method works with out this

Of course you need some luck. Or, you don't need bad luck.  I think it's working well, IF you have infinite capital / funds..

Soon or later you'll deal with a long "losing streak", and you'll need a lot of funds just to play another round. If that round, let's say the round on top of your funds, fails, you're pretty screwed.

I do go to the casino sometimes, but I'm not a big spender (neither I do use martingale). I often just use a fruitmachine and play roulette. I don't see myself as a Lucky person but most of the time I don't lose much money. Sometimes I get a little profit, and sometimes a small loss. However, I've had a great evening always (and would otherwise have spend those bucks on beers or something).

[PacoMartin]

Initial bet: 1 unit
Max bet: 16 units
# of bets till you quit:  5
# of bets you want to make:  40
% chance of busting:  30
Average Profit:  16 units

I agree it would be hard to calculate expected profit.  Maybe if you just used your base earnings...it would be easier?  hmm who knows..

After 40 plays you have a 46.792% of losing 5 in a row. Even if you lose 5 in a row plays 36-40, in theory you might have won every single play beforehand, and actually have 35 chips instead of the 31 you lost in the streak.

But more than likely you will have won 20/40 and so if you started with 31 chips (the minimal bankroll) you will now have 20. You can start again with a 4 loss Martingale.

Bottom line is that playing Martingale on coin toss neither helps you or hurts you. In a real casino game with a house edge, when you play Martingale you tend to break up your bankroll into smaller units than if you were flat betting. Once there is a house edge, you don't want to break up you betting into to small of a unit. It drags the game out, but you have a much lower probability of doubling your initial bankroll.

See if you have $31 to play with, if you are doing coin tosses it doesn't matter if you play Martingale, bet it all at once, or keep flat betting $1. You have a 50/50 odds of doubling your $31 before you go bankrupt. It would just be a lot less fun to bet it all at once.

But assuming a 1% casino edge if you are betting in increments of $1 you have only a 35% of doubling your $31 before going bankrupt. If you bet in units of $3.10 (10 units) then you have a 45% chance of doubling before going bankrupt. Martingale is similar to betting $1 flat bets, because after each cycle you are only ahead $1.

[cp1]

Cliffnotes:  You have a 95% chance of winning about $45 and a 5% chance of losing $1024.

[dooglus]

Actually that just isn't true.  If you have a buffer of 10 rolls, and you roll 100 times.  Your expect PROFIT is going to probably be around 40 base units...

You are confused about what "expected profit" means.

Consider a game where 9 times out of 10 you win 1 unit, but the other 1 time out of 10 you lose 19 units.

When you play the game, you win 90% of the time.  You might say that you "expect" to win.

But that's not what expectation means when talking about expected profit.  The expected profit is the average profit per visit.  If you played 1000 times, you would win 1 unit about 900 times (winning a total of 900 units) and lose 19 units about 100 times (losing a total of 1900 units).  So you would end up down about 1000 units after playing 1000 times, and so your expected profit is -1000 / 1000 = -1 unit.

You can reasonably "expect" to win any given game, but your expected profit is negative.

[Sonny]

At 1% house edge, your expected loss will always be 1% of your wagered amount, no matter you are doing 1.1x bet or 999x bet, and no matter you have constant bet size, random bet size, martingales, or whatever betting schemes.

[leex1528]

Actually that just isn't true.  If you have a buffer of 10 rolls, and you roll 100 times.  Your expect PROFIT is going to probably be around 40 base units...

You are confused about what "expected profit" means.

Consider a game where 9 times out of 10 you win 1 unit, but the other 1 time out of 10 you lose 19 units.

When you play the game, you win 90% of the time.  You might say that you "expect" to win.

But that's not what expectation means when talking about expected profit.  The expected profit is the average profit per visit.  If you played 1000 times, you would win 1 unit about 900 times (winning a total of 900 units) and lose 19 units about 100 times (losing a total of 1900 units).  So you would end up down about 1000 units after playing 1000 times, and so your expected profit is -1000 / 1000 = -1 unit.

You can reasonably "expect" to win any given game, but your expected profit is negative.

I think you are slightly confused about what "expected profit means.

If you have a buffer zone of 10 rolls, and you roll 100 times, your expected profit will probably be around 40 base units.  Your potential loss would be base unit ^ 10.

The expected profit is what you expect to happen if you go in, or in other words, the majority of the time you will make 40 base units. 

You are just getting confused on what expected profit is and what potential loss is. 

[leex1528]

If you walk into a casino and place a bet for 1000$ that pays 1005$ for the win, then (a) that's not a very normal game is it? (b) the long term expected average isn't calculated by simulating one roll.  It's calculated by simulating an infinity of rolls, obviously you can't simulate infinity so you just find the limit.

I didn't say anything about long term expected average.  I was just replying to someone who said you can't expect go go into a casino and make any money ever.

I can pretty much promise you I can walk into a casino once, and end up 5 dollars ahead playing martingale and if I never ever went back there, I would be up 5 dollars forever.

[cp1]

I'll bet 1 BTC that you won't make exactly 5 base units if you roll 100 times with martingale.

[leex1528]

I'll bet 1 BTC that you won't make exactly 5 base units if you roll 100 times with martingale.

I'd take that bet too, it wouldn't me 5, it would be more.

[PacoMartin]

Soon or later you'll deal with a long "losing streak", and you'll need a lot of funds just to play another round.

A problem with this pointless statement, is that the average number of "sooner or later" can be estimated mathematically. The comment "sooner or later" by itself is of no help.

For a loss of k in a row the number of plays will average out to N = (2^k - 1 )*2 or for k losses in a row N=62. If you simulate it over and over on average you will get 62 coin tosses before 'losing' 5 in a row. You are free to designate heads or tails as a loss/

For a probability of losing =q then the formula is recursive N(k+1)= (N(k)+1)/q where N(0)=0
For q=50.5% or a house edge of 1% then on average you will play 60 times before losing 5 in a row.
It should make sense that a slightly larger than 50% odds of losing, means that on average you will see the streak in fewer plays than with a 50% chance of losing. Hence coin toss is 62 plays, and 1% HE is 60 plays.

[dooglus]

You are confused about what "expected profit" means.

I think you are slightly confused about what "expected profit means.

Here's what everyone except you is talking about when they say 'expected profit':

    http://en.wikipedia.org/wiki/Expected_value

What are you talking about?

[leex1528]

You are confused about what "expected profit" means.

I think you are slightly confused about what "expected profit means.

Here's what everyone except you is talking about when they say 'expected profit':

    http://en.wikipedia.org/wiki/Expected_value

What are you talking about?

I am talking about a 1 time run where your odds are 99% chance of winning.  I have been saying that the entire time.  I have already said countless times that over the course of time, you are going to lose money.  But if you say, had a 20 bet limit, and you rolled 20 times, you have a 99.99% chance of gaining a very small amount of coins/money...etc

[cp1]

I am talking about a 1 time run where your odds are 99% chance of winning.  I have been saying that the entire time.  I have already said countless times that over the course of time, you are going to lose money.  But if you say, had a 20 bet limit, and you rolled 20 times, you have a 99.99% chance of gaining a very small amount of coins/money...etc

If you only roll 20 times then it's more like 92% chance

[leex1528]

I think its higher, it is essentially the odds of losing 20 hands in a row is it not?

[cp1]

I think its higher, it is essentially the odds of losing 20 hands in a row is it not?

No, because if you win the first 10 and lose the last 10 you're not coming out ahead.

[leex1528]

I think its higher, it is essentially the odds of losing 20 hands in a row is it not?

No, because if you win the first 10 and lose the last 10 you're not coming out ahead.

True, I guess I wasn't really thinking of a fixed roll though, I'm guessing I would roll again cause odds...but either way very good point

[ShibaWow]

which investment do you guys use as the base?

I usually take the x500 less then the amount I'm willing to gamble with

so I was gambling on primedice the other day with 0.1BTC and / 500 = 0.0002BTC

made a profit to 0.13BTC and was too afraid to continue, got to 0.09BTC investment one time and stopped soon


which is your base unit when using martingale?

[dooglus]

You are confused about what "expected profit" means.

I think you are slightly confused about what "expected profit means.

Here's what everyone except you is talking about when they say 'expected profit':

    http://en.wikipedia.org/wiki/Expected_value

What are you talking about?

I am talking about a 1 time run where your odds are 99% chance of winning.  I have been saying that the entire time.  I have already said countless times that over the course of time, you are going to lose money.

I'm asking what you mean by the term "expected profit". If you're not talking about mathematical expectation then what is it you're referring to whe you say "expected profit"?

But if you say, had a 20 bet limit, and you rolled 20 times, you have a 99.99% chance of gaining a very small amount of coins/money...etc

If you roll 20 times with a 99% chance of winning a tiny amount and a 1% chance of losing a lot, then you only end up with a profit if you win all 20 rolls. The chance of winning 20 out of 20 99% rolls is 0.99^20 = 81.79%.

So you have a 81.79% chance of winning a small amount and a ~20% chance of losing a lot.

Your mathematical expectation in such a case is negative.