- Search Crypto Craft
- 1,424 Results (11 Threads, 1,413 Replies)
- mikkom replied Feb 4, 2011
It's not really that hard, just distribute the risk correctly and many simple statistic-based methods work to some extent - many markets have fat tails and that can be exploited. It's too bad this article is not available on web fully but it has ...
Systematic trading
- mikkom replied Feb 2, 2011
How do you process long-tail candles like huge whipsaws that happen after nfp/other news? I see possibility of problems at those points if the price whipsaws a lot I had few genetic methods that took advantage of my older method that was a bit ...
Systematic trading
- mikkom replied Feb 1, 2011
If one uses times in their calculations, time zone should not matter as long as it's consistent.
Systematic trading
- mikkom replied Jan 28, 2011
I have added some futures contracts to my framework and I have a small problem.. I have quite a lot of 1 min data but I don't have any bid-ask spread data. So what would be a good source for bid-ask spread data for futures (and why not FX too) so my ...
Systematic trading
- mikkom replied Jan 12, 2011
This might interest some people here Algorithms Take Control of Wall Street
Systematic trading
- mikkom replied Jan 4, 2011
Yes, that's a time point (or time range) when certain type of event is happened/happening. I moved to "real time" event checking after the cache size got out of control, there is a hit to performance but not as big as I have thought. Now to ...
Systematic trading
- mikkom replied Dec 25, 2010
I have been running the ga without clearing caches for about a week now and it seems that event cache directory is > 500 GB... with only 4 pairs and not that many event types.. Looks like I have to reconsider the caching policy if I plan to test ...
Systematic trading
- mikkom replied Dec 13, 2010
Basically I would claim that you are not taking "regime changes" into equation. Most simply this kind of behaviour happens at example at breakouts from range and optimizing to certain type of strategies optimized to certain regime (like range-bound) ...
Systematic trading
- mikkom replied Dec 10, 2010
This is where a low level language like C shines, this is with 1 core and samsung spinpoint f3. Have you thought about using some other language for low level jobs like this? Some numbers for comparison -rw-r--r-- 1 mikkom mikkom 175M Dec 1 02:04 ...
Systematic trading
- mikkom replied Dec 9, 2010
Some cache data and this is with only 4 different types of events and 4 pairs
This is why I didn't bought SSD and now it looks like 1TB will be too small.. Oh well I have to get a bigger one then. cf$ du -h cache/ 4.9G cache/EURUSD_events/1 ...Systematic trading
- mikkom replied Dec 8, 2010
I added a new filter that seems to do the trick:: Top trades profit of all profit%. This seems to do the trick. I basically measure that top X trades don't account for more than Y% of the total profit. If anyone has more (maybe more robust) ideas, ...
Systematic trading
- mikkom replied Dec 8, 2010
I have done that before but not now (I know the data is pretty clean) but the data is the same as I have used in the past. Those are more likely methods that open many many small trades and ride them long when they hit a runner. When they find it ...
Systematic trading
- 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 ...
Systematic trading
- mikkom replied Nov 29, 2010
.. and now the answer would be yes
N threads for running strategies (each uses N child-threads when finding events), 1 thread for calculating combinations of strategies and one for visualization. Speed increase was quite visible (and measurable).Systematic trading
- mikkom replied Nov 26, 2010
Speed wasn't such a big issue as was IO and memory requirements. Disk IO is something that is simply outstanding with C, with java it was really, really slow.. The main reason for the java->c change was because I want to move into ...
Systematic trading
- mikkom replied Nov 26, 2010
Only 3 threads as of now (the genetic finder/combinator/visualizer) but I'm just working on multithreaded finder and event processors. By the way, the green line on chart is in/out sample line. I'm quite ruprised really that the results are this ...
Systematic trading
- 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.Systematic trading