- Search Metals Mine
- 150 Results (1 Thread , 149 Replies )
- mikkom replied Nov 11, 2011
Oh the beauty of noncorrelation and huge list of equities. Here is the first screenshot I have posted anywhere of my new stock/etf trader.
- mikkom replied Jun 30, 2011
+25%? Also, relevant performance metric would be reward vs risk. Some of your years look pretty bad on risk basis - there are some VERY high drawdowns, 2003 for example - this tells about far too high risk ratio and will give you very nice looking ...
- mikkom replied May 15, 2011
Yes it can handle multiple positions. I don't see why this would have anything to do with using out-sample as real out-sample.. Here is one run with about 8 hours of time from point zero (I started it after I switched back to original fitness algo). ...
- mikkom replied May 14, 2011
Here are two examples, I can repeat both my changing some aspects of fitness algorithm and running the program for few hours (the strategies that are combined to form these curves are the same on both examples). The curves will not be totally the ...
- mikkom replied Mar 2, 2011
What a difference a good fitness algo does — I have been testing a new fitness algo for a few days and I have to say I'm a bit astounded.. I thought there is a bug or something but after trying different methods (mar etc) and comparing ...
- mikkom replied Dec 8, 2010
I have been pondering this a lot, what would be an ideal way to find which systems are robust and which are more random occurences with few "long tail" trades that might cause the numbers look better than they actually are. With the introduction of ...
- mikkom replied Nov 26, 2010
Just got my new server that is going to run this program from now on.. core 7 with 12 GB of ram Here is some tech and curve pr0n. Yes, the layout is almost exactly the same as before because I like my development environment as I'm used to.
- mikkom replied Nov 19, 2010
More pretty pics The stats are zero but the number of simulated trades is correct (500k+ trades for that curve) That is a combined curve of N systems (after only ~160 rounds of random systems/mutations/crossovers). pic2 after 1 mil+ trades. I ...
- mikkom replied Nov 18, 2010
Yay — One of the first visualizations of curve on the new system I'm waiting for new computer right now - on my few years old laptop the current speed is about 0.1 seconds per result when the system has stabilized (ie. (including ...
- mikkom replied Sep 24, 2010
Just for fun, here are my last 30 days. 4 pairs, 20 genetic algo trading methods.
- mikkom replied Aug 26, 2010
Two last out-sample years zoomed.
- mikkom replied Aug 26, 2010
I would rather stick with something you can decipher by yourself (I very much dislike neural networks because I can't reverse engineer what they do). In fact my work with gp has only backed my previous thoughts on how market works. It might be ...
- mikkom replied Jun 10, 2010
I don't believe in luck Some people wanted to see what genetic programming can do.. so here is 2007- to today eurusd curve (at the moment I only have poor quality eurusd data that far but I believe my own tester much more than mt4 tester..) ...
- mikkom replied May 21, 2010
I'm totally aware of this, however the trades are not as tight as they might seem to be - as I mentioned some time ago, I for example consider sl to be hit if same candle that the trade is opened with touches it and I still trade daily stuff even ...
- mikkom replied May 20, 2010
This is one of my latest runs (fx, not equities) - after some light bulb moments (as silly as it might sound, after about 10 years of trading - with less trading in the middle of the mentioned years - I think I have finally made myself clear what ...
- mikkom replied Apr 16, 2010
I do understand what you are saying, I have been experimenting with different combinations a lot (and I mean a lot, I did have a personal server running GA/GP stuff for stocks in our companys server room 24h some 10 years ago ) - that didn't lead ...
- mikkom replied Apr 16, 2010
Sorry I don't quite follow, what do you mean by multiple dimensions? Time? At this point I'm not even sure if it's the fitness function (I use very simple one now, just a simple check that few of the parameters are above treshold) or the portfolio ...
- mikkom replied Mar 22, 2010
As I said, I wanted to do some breakout studies.. Well after I did some of the studies I did a small test for one of the models that came into my mind and generated algos only with a very specific order management scheme and here is the result - the ...
- mikkom replied Mar 10, 2010
I just had to get a screenshot from this.. Note the ** part
- mikkom replied Feb 14, 2010
Here is one other indicator that might be of interest if you are interested in p&f, I did it some time ago - it shows p&f in the main chart as a channel.