Probability when combining multiple percentages

[Abb1987]

Hi. I've been working on a program for automated trading of stocks and crypto currencies which combines multiple trading strategies.

I have no background in mathematics so have no real idea of how to calculate probabilities.

If:
Strategy 1 has an average win rate of 55%
Strategy 2 has an average win rate of 57%.
Strategy 3 has an average win rate of 60%

I was wondering how to calculate the probability of the trade going in the right direction when two or more of these strategies generate entry signals at the same time??

 

[pka]

Hi. I've been working on a program for automated trading of stocks and crypto currencies which combines multiple trading strategies.
I have no background in mathematics so have no real idea of how to calculate probabilities.
Strategy 1 has an average win rate of 55%
Strategy 2 has an average win rate of 57%.
Strategy 3 has an average win rate of 60%
I was wondering how to calculate the probability of the trade going in the right direction when two or more of these strategies generate entry signals at the same time??

This is a help site not a tutoring or doing site.
In order it receive help , one must post work in progress done by you along with questions you have about your work.
That way we know where to begin.

 

[Abb1987]

This is a help site not a tutoring or doing site.
In order it receive help , one must post work in progress done by you along with questions you have about your work.
That way we know where to begin.

I see.
OK this is where I'm at:

Let's take 1 and 2

(55/100) + (57/100) = 1.12
= 112% Chance of being right.
Can't be right.

So
(55/100) * (57/100) = 0.31
= 31%
I'd doubt the probability would drop in my scenario so this can't be right either.

And now I'm stuck.

Will you help me now?

 

[Abb1987]

I also highly doubt its

55 + 57
-------------
100 + 100

I tried googling to find the correct calculation to no avail.

All I asked for was the method, not a telling off.

 

[Cubist]

With this information I don't think you can draw any conclusions. What if strategy 1 looks for signs of a price fall, and strategy 2 looks for signs of a price rise. If they both trigger at the same time then will the price be likely to stay the same (leaving no room for profit)?

I recommend that you trade on paper for a long time (without money). Gather real life statistics to calculate your probabilities. Wait until strategy 1 & strategy 2 have triggered at the same time - on MANY separate occasions. Study what happened as a result.

 

[Abb1987]

With this information I don't think you can draw any conclusions. What if strategy 1 looks for signs of a price fall, and strategy 2 looks for signs of a price rise. If they both trigger at the same time then will the price be likely to stay the same (leaving no room for profit)?

I recommend that you trade on paper for a long time (without money). Gather real life statistics to calculate your probabilities. Wait until strategy 1 & strategy 2 have triggered at the same time - on MANY separate occasions. Study what happened as a result.

The strategies work both ways. I'm asking how to calculate probability if two indicators give off signals in the same direction.
I just want to know how to calculate.
I don't need to know, I'm just curious.

Haha I'm a bit beyond paper trading but that would would good advice for a new guy.

This is for a trading algorithm which has several "trigger signals."
Each signal has been backtested over several thousand trades to give a percentage win rate on its own.
I just wondered how to calculate the probability when multiple signals fire at the same time (in same direction).

 

[Cubist]

I'm don't think there's a formula for what you're asking.

The answer completely depends on:-
- the details about how each trigger works
- and how the particular commodity behaves, on average, after a multiple trigger event

It seems reasonable to assume that there would be a greater probability of profit after a multiple trigger event. However, don't bank on this. What if there's a cap on the price movement of commodity A because there's an alternative commodity B available (which has a stable price) and everyone would switch to buying B if A's price moves too far upwards. This could mean NO EXTRA chance of winning. Also, what if the triggers are basically looking for very similar indicators - if so then it wouldn't give rise to an increased chance of winning when they all trigger.

I'm pessimistic by nature, but there is also the chance that a double trigger might have a disproportionately large increase on the probability of winning. It could indicate a "run" might happen on the commodity (a mad scramble where lots of people move to buy a limited resource).

There really is no way to know the resultant probability with the information you have. I think the only way to estimate the probability is to back-test these events or try to build a reliable financial model.