Not the Grail, just a regular one - Bablokos!!! - page 76

 
Lastrer:

No, you've got it wrong. Read it more carefully. The probability of reaching 4 tails before a series of three eagles falls. Those are

orororororororororororororororororororororororororororororororororororororororororor and the unlucky outcomes of ooo or rorororororororororororororororororororororororororororororor.


Yeah, that sounds really messed up... What does this have to do with the probability of DO? The probability is always the same. For any series of 4 shots, it is 0.5^4 =1/16, regardless of when the shots are made, before, after, or right now.

You can of course fantasise as much as you like and come up with some rules and secret combinations, but it's all in your head, the coin doesn't care about them, there are no combinations for it, there are only 2 outcomes: heads and tails with equal probabilities.

 
prikolnyjkent:

And I don't mean predicting the outcome of one particular shot. I mean the visible properties of a SERIES of throws.

For example, isn't it possible to take advantage of the fact that none of the 1000 series of 1000 throws has a maximum deviation from the x-axis greater than 120? (https://ru.wikipedia.org/wiki/Задача_о_разорении_игрока)

Because obviously you can. (And I would say you have to...).

Well, you take 10 000 series and obtain the deviation of e.g. 500. What next? Why do you have to take exactly 1,000? And where do you start this series from?

If in one of 10 000 series you get a deviation of 500, then why this series can't be in your sample of 1000 series?

 

Yeah....

See how it turns out I do not know how to count, and when it came to ask a specialist, it turned out that all was lost. Don't talk about what and what not, you need to count, not ..... And in fact the problem is directly related to the current minimax this time having a non-zero MO. All the more, not a word, besides these words are somehow meaningless.

 
Meat:

So take 10,000 series and get a deviation of e.g. 500. Then what? Why do you have to take 1,000? And from where to start this series?

If in one of 10 000 series you will receive a deviation of 500, then why this series can not be in your sample of 1000 series?

It can, but the probability is much lower. It's like a gopher that no one sees, but it's there. In this case it would be a black swan, which happens but not all the time. So you have to calculate the capital to survive it. Let's say we double up more often than we lose. We may lose on swans from time to time, but it is still a very profitable strategy.

Von wrote about falling out in one of the casinos 40 times black in a row, and Che it will be only a black swan for a strategy that can survive say 15 in a row. But it is not fatal because earnings will be much higher than the loss before the next such swan. You don't have to bet everything, that's all.

 
Meat:

So take 10,000 series and get a deviation of e.g. 500. Then what? Why do you have to take 1,000? And from where to start this series?

If one of 10 000 series will have a deviation of 500, then why this series can not be in your sample of 1000 series?

It can... Of course it can. That's why I asked the question earlier - "Is the dependence of the probability of a graph reaching some value on that value itself linear?" In other words - is the graph twice as likely to reach a value twice as distant? After all, if the dependence is NOT linear, then there "give money"...
 
prikolnyjkent:
It can... Of course it can. That's why I asked the question earlier - "Is the dependence of the probability of a graph reaching a certain value on that value itself linear?". In other words - is the graph twice as likely to reach a value twice as distant? After all, if the dependence is NOT linear, then there "give money"...


What difference does it make to you whether it is linear or non-linear? After all, you are interested in further developments after this difference is reached, right? That is, at this point you want to open a trade. Well, the further movement will be equally likely: up as well as down. Correspondingly, you can make both profit and loss with probability 0.5, and so for all subsequent trades. This probability has nothing to do with your last series, because it is a separate event, everything started from scratch.

The coin has no memory, and it does not know at what point you started your observations, now or 1000 flips ago. Therefore, it has no way of knowing from which moment it needs to measure your deviation of 120 to please you.)

 

I'm sick and tired of being part of this madhouse... People don't want to think with their heads, let them do it themselves. Customers like that are a big hit with the casinos. They lose money once, decide it's just "ammo wasn't the right calibre", start fixing the system, replace eagles with tails, etc. Then they come to play again, then again and again...

 
Meat:


So what difference does it make to you, linear or non-linear? After all, you are interested in further developments after reaching this difference, right? That is, at this point you want to open a deal. Well, the future movement will be equal: up as well as down. Correspondingly, you can make both profit and loss with probability 0.5, and so for all subsequent trades. This probability has nothing to do with your last series, because it is a separate event, everything started from scratch.

The coin has no memory, and it does not know at what point you started your observations, now or 1000 flips ago. So it has no way of knowing from when exactly it needs to measure your deviation of 120 to please you.)


Nah... I want to open a trade at the "beginning of coordinates". And I want to set stop-loss and take-profit at DIFFERENT Distances.

So it turns out that if the relationship is NOT linear, then my stop, which is twice as far away as my profit, will not be 2 times as large as my TP. This means that by "playing with" the distance to the orders, I can choose a PROFIT MODE.
 

Tell me, what are you doing on a programming forum if you can't even program the simplest algorithm and check your inventions?

 
prikolnyjkent:

Nah... I want to open the trade at the "beginning of the coordinates". And I want to set stop-loss and take-profit at different distances.

So it turns out that if the relationship is not linear, my stop that is twice as far away from profit will not affect 2 times as much as my TP. And so, by "playing" with the distances to orders, I can pick up a PROFIT MODE.
There are several EAs in the kodobase that allow manual trading in the tester. Try it out, spend a couple of evenings, a lot of questions will disappear.