A free tool i just build to find algorithmic strategies

Thanks for the detailed post. Going to look it over more closely later.

You said you used 100% of the data.

Can it train (and validate, if it does that) on, say, 50% of the data, then backtest on the out-of-sample other 50%?

Backtesting on OOS data will give a more accurate clue as to real world performance.

 

@userque sure you can train on 50% of the data. You will see the profit factor on OOS data, which probably wont be as good but you will see.

TBH i would try to find something on 50% of the data and store the patterns but then i would also try it again with 100% and search for a strategy similar to the one with 50% just to have the best latest values.

gl

 

Ok, that's good that it can do that. I'm intrigued by your work. Nice job! I hope to explore it more this weekend.

 

thx if people would use it i would feel some accomplishment and i would also continue my work on it
you can also try http://trading-level.com/
(to find good places for TP and SL)

 

I don't know the right answer, of course, but some of the rule generation attempts I've tried are
relating indicators among multiple instruments,
another relating indicators among multiple instruments,
relating astronomical data for cyclic predictions, and
relating price waves.

One thing I've noticed is finding one set of rules that work well across multiple instruments is harder than finding rules that work on a single instrument. And I've generally tried creating rules specific to a single instrument. This probably led to more overfitting.

 

hm interesting

again i remember Larry Williams on this i think he also used "astronomical data" but in the end it was a month cycle and he noticed in some days of the month the stock index tends to go up (maybe it was because of people getting their salaries or something). But i guess it wont be that easy these days.

 

So i made an improvement (among other things that arent obvious) you can upload csv data from yahoo finance.

I am yet to try some individual stock data, any suggestions ? i will share my findings, perhaps apple can be first

 

I tried it with a file downloaded from yahoo finance which resulted in empty graphs:

 

Hi, use http://algostrategies.io/ plz
...
http://trading-level.com/ is an older tool i build before but it only accepts Meta trader csv file structure.

but thx for trying (it may take some time till you get some results gl.)

 

So i made couple more improvements on http://algostrategies.io/

First i noticed that the whole app can become unresponsive once there is a lot of data uploaded.
Sadly this fix makes old uploaded data useless so you need to upload the data again.
Also playing around i noticed it can crash when chrome decides that there is too much data uploaded.

Another change is about the stocks/data. So i did try apple and noticed quickly that the chart sucks when the data changes overtime rapidly (so from 10$ to 400$). Basically the equity on the chart goes sideways and only last fraction goes suddenly up. My fix is to have 2 charts, one is basic, second/new is adjusted so it takes percentage changes rather then changes in absolute numbers.
Lastly i also fixed the fitness function so that the profit factor is counted based on percentage changes.

sooo i think its better.

I do my research on apple again tomorrow