Probability. - page 2
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
I agree with you and I certainly won't argue with Vicky. Probability estimation comes after the event has happened.
Whether it is close to one or not is not important.
My task, purely for myself, is to determine the probability as a percentage of the outcome of the event. Specifically whether the price of the observed asset will rise or fall. I am not satisfied with an estimation of 0 or 1. I can be 60% or 30% sure, but the final result will of course still be 0 or 1.
Bookmakers know a thing or two about it)))
Well there are probability models - Bayesian, Markov chains. There is an article on the former - the Bayesian classifier. Another thing is how to apply it all :) To determine the probability of an event it is necessary to have a statistically significant sample, and based on it watch which consequence is most likely to occur in this or that situation.
But you should go to the branch of machine learning, there almost everything breathes with probabilities, so it makes little sense to even create a separate topic
My task, purely for myself, is to determine the probability as a percentage of the outcome of an event. More precisely, whether the rate of the observed asset will rise or fall.
The probability increases or decreases 50/50.
I am not satisfied with an estimate of 0 or 1. I can be 60% or 30% sure, but the final result will of course still be 0 or 1.
Bookmakers know a thing or two about it)))
If you know that the price will go in any direction with a probability of more than 50% - then you will already be winning.
Sorry. The first posts are gone.
There was a server failure. That's all right. I'll update the first notebook entry.
It is clear that few people open on rise or fall of an asset if the probability is low. There may be prediction errors in another thread. But few calculate the probability of the outcome as a percentage.
With rare exceptions I see traders who perform their analysis using percentages. They are very few and may be among the most successful ones.
Having tested the Expert Advisor in the Strategy Tester, I saw that 51% of deals close on the plus side - you know, if you use this system, the probability of a deal triggering is 51%.
I ran the Expert Advisor in the tester and saw that 51% of trades close on the upside - you know that if you work with this system, there's a 51% chance of a trade being closed.
Once again, who hasn't read earlier: "I wrote a bot with 20% of profitable and 80% of losing trades, and it ended up with a nice profit, and in the entire trading area.
So there is no need to look for 49/51 probabilities, even with a 90% probability of winning, the account can easily be drained
Well there are probabilistic models - Bayesian, Markov chains. There is an article on the former - the Bayesian classifier. Another thing is how to apply it all :) To determine the probability of an event it is necessary to have a statistically significant sample, and based on it watch what consequence is most likely to occur in this or that situation.
But this is a branch of machine learning, almost everything there breathes with probabilities, so there's not even much sense in creating a separate topic.
I visit the machine learning branch from time to time.
2. I think there is a point in a narrower context.
the probability will go up or down 50/50.
If you know that there is more than a 50% chance that the price will go either way, you will be winning.
Not really. At the moment it's 60 and in 20 minutes it will be 30.
The important thing is to determine the probability within a certain range over time. That's where my interest is focused right now.
Once again, for those who have not read earlier: "I wrote a bot with 20% profit and 80% loss, and it ended up in good profits, and over the whole trading area".
So don't look for 49/51 probabilities, even with a 90% probability of winning, the account can easily be drained.
I agree, I was showing you an example too.
Bookmakers know a thing or two about it)))
Like: you think the price will go up? Bet on it at odds of 2.1.
You think the price will go down? Bet on it at odds of 1.5.
That only happens in sports. Because there are strong teams and weak teams. And the odds are always higher on weak teams.
In forex there are no strong and weak currencies. In forex central banks keep their countries' currencies stable.
Well there are weak countries currencies... but it's all equalised by the swap...
in forex the payout is always at odds of 2. minus the spread.
if you think there is a 60% chance of the price going up and a 40% chance of it going down, you simply bet that the price will go up and 60% of the time you get a payout with odds of 2.
Although I've seen it on binary options: the odds of the price going up are 60%. The odds if the price goes down are 70%.
and every minute the coefficient changes.
they calculate it somehow)))
i think they come from bookmakers and work in the usual way. they know the odds for one event and the odds for the other.
they still pay out profit on odds that aren't 100%. so they still stay in profit.
i've come up with a ts on working with binary options odds. where they put the higher odds, you open in that direction.
I don't think they know how to make correct predictions.)
You ran the advisor in the tester and saw that 51% of trades close on the upside - you know that if you use this system, there is a 51% probability of a trade being closed.
Purely statistically it can work.
But if you think about it, each deal should have a separate analysis, and this is a new condition and a different forecast.
Once again, for those who have not read earlier: "I wrote a bot with 20% profit and 80% loss, and it ended up in good profits, and over the whole trading area".
So don't look for 49/51 probabilities, even with a 90% probability of winning, the account can easily be lost.
I thought it wasn't necessary to specify: "provided the take profit is equal to the stop loss!"