Does martingale really works?

[sed]

I really think the folks arguing here would get more out of it if they would just stop and read the wikipedia article on martingale betting strategy.

[boumalo]

I really think the folks arguing here would get more out of it if they would just stop and read the wikipedia article on martingale betting strategy.

http://en.wikipedia.org/wiki/Martingale_(betting_system)

Yes there are a lot of feelings and misconceptions in this thread

" It is therefore a good example of a Taleb distribution – the gambler usually wins a small net reward, thus appearing to have a sound strategy. However, the gambler's expected value does indeed remain zero (or less than zero) because the small probability that he will suffer a catastrophic loss exactly balances with his expected gain. (In a casino, the expected value is negative, due to the house's edge.) The likelihood of catastrophic loss may not even be very small. The bet size rises exponentially. This, combined with the fact that strings of consecutive losses actually occur more often than common intuition suggests, can bankrupt a gambler quickly."

[makebitcoin]

I always feel so sad for the "martingale believers". They are absolutely certain their method works and for the ones that keep playing it ALWAYS results in being 100% wiped out. They lose all their bankroll. And most of the time all they have ever won is a few bet units. Win small till you lose it all.

[Acidyo]

Martingale would work if you had infinite BTC.

[boumalo]

Martingale would work if you had infinite BTC.

Which is not even technically possible

[MoneyGod]

Martingale would work if you had infinite BTC.

Which is not even technically possible

Just because of this we can say its not working for any one never try and never lost your hard working money 

[BayAreaCoins]

Martingale would work if you had infinite BTC.

Martingale will always fail if there is a house edge at all.

[kuverty]

Martingale would work if you had infinite BTC.

Martingale will always fail if there is a house edge at all.

But you can take the house edge into account by not doubling your bets, but making them, say, 2.2 times higher each time you lose.

[Phildo]

Martingale would work if you had infinite BTC.

Martingale will always fail if there is a house edge at all.

But you can take the house edge into account by not doubling your bets, but making them, say, 2.2 times higher each time you lose.

You are still making a series of -ev bets, which add up to a -ev adventure. Doing 2.2x instead of 2x gives you a slighty bigger reward if it works, but makes the catastrophe of failure come faster.

[kuverty]

Martingale would work if you had infinite BTC.

Martingale will always fail if there is a house edge at all.

But you can take the house edge into account by not doubling your bets, but making them, say, 2.2 times higher each time you lose.

You are still making a series of -ev bets, which add up to a -ev adventure. Doing 2.2x instead of 2x gives you a slighty bigger reward if it works, but makes the catastrophe of failure come faster.

Does that really matter, martingale will never work in real life where you have limited funds and limited maximum bets. If you want better than negative expected value then you must have unlimited funds and no limit. If you have those, you might just as well raise your bet by a hundredfold each time you lose.

[Zombie23]

In my personal experience no. The best way I have found is to gamble in alt-coins. Preferably one that is only worth less than 100 Satoshis. Then find a site where you can bet .00000001 of that coin. Play until you have lost like 5 in a row then start martingale at 1 or 2. Sure the profit will be slow, but this gives you more protection. The only real way to make money on this would be to actually use a program to run it for you, but this is just my idea.

[kuverty]

In my personal experience no. The best way I have found is to gamble in alt-coins. Preferably one that is only worth less than 100 Satoshis. Then find a site where you can bet .00000001 of that coin. Play until you have lost like 5 in a row then start martingale at 1 or 2. Sure the profit will be slow, but this gives you more protection. The only real way to make money on this would be to actually use a program to run it for you, but this is just my idea.

It does not offer any protection. If the game works like it should, and it will or it would be bankrupt by now, the events are distinct. You lose 5 times in a row before betting big; so what? The next events do not depend on those, you still have the same chance to lose.

Let's imagine you throw a coin 10 times and get 10 heads; on the 11th, is it more probable that you get heads or that you will get tails?

[Zombie23]

In my personal experience no. The best way I have found is to gamble in alt-coins. Preferably one that is only worth less than 100 Satoshis. Then find a site where you can bet .00000001 of that coin. Play until you have lost like 5 in a row then start martingale at 1 or 2. Sure the profit will be slow, but this gives you more protection. The only real way to make money on this would be to actually use a program to run it for you, but this is just my idea.

It does not offer any protection. If the game works like it should, and it will or it would be bankrupt by now, the events are distinct. You lose 5 times in a row before betting big; so what? The next events do not depend on those, you still have the same chance to lose.

Let's imagine you throw a coin 10 times and get 10 heads; on the 11th, is it more probable that you get heads or that you will get tails?

It is just there to give you more protection. With the .00000001 bet you aren't really losing any value, and if the game truly is fair the distribution should bring it to the point where you would win. It is all about getting more layers of protection. Because lets be honest, you aren't going to lose 1000 games in a row, the odds are so small I'm confident it wouldn't happen because the statistics will balance it out, so you are right, you still have the same odds (an independent event) where if you bet >30 times than a normal distribution curve should apply. (This is all coming from a highschool stat class, but this is pretty basic stat)

[kuverty]

In my personal experience no. The best way I have found is to gamble in alt-coins. Preferably one that is only worth less than 100 Satoshis. Then find a site where you can bet .00000001 of that coin. Play until you have lost like 5 in a row then start martingale at 1 or 2. Sure the profit will be slow, but this gives you more protection. The only real way to make money on this would be to actually use a program to run it for you, but this is just my idea.

It does not offer any protection. If the game works like it should, and it will or it would be bankrupt by now, the events are distinct. You lose 5 times in a row before betting big; so what? The next events do not depend on those, you still have the same chance to lose.

Let's imagine you throw a coin 10 times and get 10 heads; on the 11th, is it more probable that you get heads or that you will get tails?

It is just there to give you more protection. With the .00000001 bet you aren't really losing any value, and if the game truly is fair the distribution should bring it to the point where you would win. It is all about getting more layers of protection. Because lets be honest, you aren't going to lose 1000 games in a row, the odds are so small I'm confident it wouldn't happen because the statistics will balance it out, so you are right, you still have the same odds (an independent event) where if you bet >30 times than a normal distribution curve should apply. (This is all coming from a highschool stat class, but this is pretty basic stat)

The statistics will not balance it out. Before you are to lose 1000 times in a row, you would be in a situation where you have lost 999 times in a row. Why did that happen then? Losing 1000 times in a row is just 50 % less probable, exactly the same probability again that the usual one coin flip goes heads or tails.

Your theory is flawed, to illustrate you this in the easiest way I can - trust me this is for the best of your understanding too - can you answer the following question: Imagine I was with you and I threw a fair coin 100 thousand times. Incredibly, every time it was tails. Would it be better to bet on heads the next time?

A yes/no will do.

[kuverty]

In my personal experience no. The best way I have found is to gamble in alt-coins. Preferably one that is only worth less than 100 Satoshis. Then find a site where you can bet .00000001 of that coin. Play until you have lost like 5 in a row then start martingale at 1 or 2. Sure the profit will be slow, but this gives you more protection. The only real way to make money on this would be to actually use a program to run it for you, but this is just my idea.

It does not offer any protection. If the game works like it should, and it will or it would be bankrupt by now, the events are distinct. You lose 5 times in a row before betting big; so what? The next events do not depend on those, you still have the same chance to lose.

Let's imagine you throw a coin 10 times and get 10 heads; on the 11th, is it more probable that you get heads or that you will get tails?

It is just there to give you more protection. With the .00000001 bet you aren't really losing any value, and if the game truly is fair the distribution should bring it to the point where you would win. It is all about getting more layers of protection. Because lets be honest, you aren't going to lose 1000 games in a row, the odds are so small I'm confident it wouldn't happen because the statistics will balance it out, so you are right, you still have the same odds (an independent event) where if you bet >30 times than a normal distribution curve should apply. (This is all coming from a highschool stat class, but this is pretty basic stat)

While I am waiting you to answer - I'd like you to, I have to point out that you seem to be thinking of "statistics" as some kind of a mystical power that will "balance things out". Statistics have no will. Statistics will behave in a certain manner though - you will see that 1000 losses is not very probable. But it is a 50-50 chance, that streak of 1000 losses, be it "0000000...." is just as probable as "10110011100011..." or any other sequence of wins/losses. Of course those cases where you both lose AND win some of the bets are more numerous, there are almost 2^1000 that kind of sequences, and only one sequence where you lose all bets. None of this means, that after a streak of  999 losses, the final, 1000th one will be more probably a win. Both are of equal probability. I would like to hear you simply answer yes or no to the question I posed, disregarding additional arguments, if I may ask you to do that.

[dooglus]

While I am waiting you to answer - I'd like you to, [...]

Everything you say is true, but I have little hope that you will ever convince him of that.

I see people using the "bet small until you see N the same in a row, then bet high that the streak will end" technique a lot on Just-Dice.  Usually they wait for the streak and then start Martingale betting.

What's funny is that it often works very well.

The reason for this is that Martingale *does* work very well until you hit a streak that's too long, and if you play enough times you will hit that streak.

But when you wait for a streak before you even start, it has the side effect of really reducing the number of real bets you can make per hour, which really increases the time until you reach a streak that busts you.

So the described strategy "works" in a sense, in that Martingale is a horrible strategy, and waiting for a certain condition to occur before starting each Martingale progression means you play less.

People see this, and think that the magic of waiting for a streak in the dust bets is what it working, whereas smoking a cigarette or reciting the Lord's Prayer between each bet would "work" just as well, to the extent that both would slow you down.

[Justin00]

It works up until the point you loose it all.
You kinda need to think if thier was sum magical trick everyone would be rich.....

[boumalo]

While I am waiting you to answer - I'd like you to, [...]

Everything you say is true, but I have little hope that you will ever convince him of that.

I see people using the "bet small until you see N the same in a row, then bet high that the streak will end" technique a lot on Just-Dice.  Usually they wait for the streak and then start Martingale betting.

What's funny is that it often works very well.

The reason for this is that Martingale *does* work very well until you hit a streak that's too long, and if you play enough times you will hit that streak.

But when you wait for a streak before you even start, it has the side effect of really reducing the number of real bets you can make per hour, which really increases the time until you reach a streak that busts you.

So the described strategy "works" in a sense, in that Martingale is a horrible strategy, and waiting for a certain condition to occur before starting each Martingale progression means you play less.

People see this, and think that the magic of waiting for a streak in the dust bets is what it working, whereas smoking a cigarette or reciting the Lord's Prayer between each bet would "work" just as well, to the extent that both would slow you down.

Yes; it is amazing to see how many players think they can change the outcome by the way they bet or when they bet

If you are going to use 100BTC to play a martingale 1BTC 2BTC ect. and you decide you will stop after 50BTC have been won you have better odds betting 100BTC @ 1.5 once

I was thinking about that yesterday looking at a player bet strange amounts in a strange order then say he played badly on the chat after he lost; you said it doesn't make sense to allow autobeting but it does because probably more than 50% of the players don't think in terms of probability but in other terms

[kuverty]

While I am waiting you to answer - I'd like you to, [...]

Everything you say is true, but I have little hope that you will ever convince him of that.

I've explained it to people, however dumb, if they do not completely believe in magic.

Now if there is someone who actually holds the belief that after a 10000 fair coin tosses that have all been tails, it is more probable that the next one is heads... I'd like to hear if there are any. I promise you won't be bashed for it or anything, these things are not intuitive to too many people.

[kuverty]

While I am waiting you to answer - I'd like you to, [...]

Everything you say is true, but I have little hope that you will ever convince him of that.

I see people using the "bet small until you see N the same in a row, then bet high that the streak will end" technique a lot on Just-Dice.  Usually they wait for the streak and then start Martingale betting.

What's funny is that it often works very well.

The reason for this is that Martingale *does* work very well until you hit a streak that's too long, and if you play enough times you will hit that streak.

But when you wait for a streak before you even start, it has the side effect of really reducing the number of real bets you can make per hour, which really increases the time until you reach a streak that busts you.

So the described strategy "works" in a sense, in that Martingale is a horrible strategy, and waiting for a certain condition to occur before starting each Martingale progression means you play less.

People see this, and think that the magic of waiting for a streak in the dust bets is what it working, whereas smoking a cigarette or reciting the Lord's Prayer between each bet would "work" just as well, to the extent that both would slow you down.

Yes; it is amazing to see how many players think they can change the outcome by the way they bet or when they bet

If you are going to use 100BTC to play a martingale 1BTC 2BTC ect. and you decide you will stop after 50BTC have been won you have better odds betting 100BTC @ 1.5 once

I was thinking about that yesterday looking at a player bet strange amounts in a strange order then say he played badly on the chat after he lost; you said it doesn't make sense to allow autobeting but it does because probably more than 50% of the players don't think in terms of probability but in other terms

Yeah, the first time I saw these autobetting options on a dice site, I thought the operators were dumb or something... but apparently it makes sense.