How to find the variables that matter?

Back to topics list To post a new topic, please log in or register
avatar
2017
gooly 2015.07.12 21:34
 

Hi,

does anybody know (a link) how can I statistically find the variables that influences the result of the trades?

Imagine that I created an csv-sheet (for Excel) with one column of the trade-results and lots of other columns with values of the variables at the moment the trade was opened.

Now I would like to use the statistical functions in Excel to find those variables that influences the result (which will become part of the optim. process) and the others which I now know that I can skip them.

What kind of stochastic functions can I use - anybody with an idea or link?

Thanks in advance,

Gooly

avatar
250
ffoorr 2015.07.17 09:52 #
 

From what I understand, there is nothing like that. Price are not predictible, statistics are done afterward, it's money management.

If the trend is up, almost all buy orders will be winner, sell order loser.


if such statistics exist, maybe they would come from indicators, otherwise it could come from the reaction of the market on support and resistance and analysis of volume. 

avatar
13004
WHRoeder 2015.07.17 13:42 #
 

Create a spread sheet with every variable (include pair/TF) and profit/loss.

Then sort by a variable and for every value add average profit, std deviation, and ap/sd.

The value with the largest ap/sd is the most consistent.

avatar
2017
gooly 2015.07.17 14:28 #
 
WHRoeder:

Create a spread sheet with every variable (include pair/TF) and profit/loss.

Then sort by a variable and for every value add average profit, std deviation, and ap/sd.

The value with the largest ap/sd is the most consistent.

Well actually I calc. for every variable the average and deviation and calc. the profit if the var.-value is bigger than.... and some "Find-Maximum, ...

I have hoped that there is a known stochastic calculation that would it make easier.

But I have asked this question in 3 different math./stochastic forums - with only one reaction: I don't understand your question, what do you want?

It seems there is no method for this kind of question.. what a pity :(

avatar
313
JD4 2015.07.17 20:32 #
 
gooly:

Well actually I calc. for every variable the average and deviation and calc. the profit if the var.-value is bigger than.... and some "Find-Maximum, ...

I have hoped that there is a known stochastic calculation that would it make easier.

But I have asked this question in 3 different math./stochastic forums - with only one reaction: I don't understand your question, what do you want?

It seems there is no method for this kind of question.. what a pity :(

The answer to that question, I think, is like the constant search for that "holy grail" trading system that everybody and their brother is looking for.  While it is possible that a system like that actually exists, finding it is not so easy to do.  To me, a part of the problem is that one main thing that influences the market the most is not always the same.  Because the market is traded by people (even the programs were designed by people and that influences it a little), there is nothing that can give 100% certainty.
Back to topics list  

To add comments, please log in or register