Over the weekend I have been working on a few system related projects. For one I implemented an equity curve filter filter into Thor. If you’re not familiar with EC filters then please take a look at my pertinent post from two years ago. Having observed over a year of live Thor action now it seems to me that both sell offs and earning periods come in clusters. The idea here is to limit drawdown periods of course until market conditions are more favorable. It is an approach that has served us very well on CrazyIvan and I think it’ll work with Thor as well.
Talking of CrazyIvan – now that Scalpius is out of beta and appears to be running error free I am [...]
Over the weekend I decided to finally catch up with a chore I was planning for a while now and one that was long overdue: To sift through nearly 2500 CrazyIvan live entry logs and to separate the wheat from the chaff. Clearly sufficient statistical evidence was present and over the past 18 months we have gone through very distinct market phases. I was curious how some of the entry patterns responded and if perhaps some of them could be eliminated.
What I found out was rather interesting and unfortunately, as the system is proprietary, the exact details I must withhold. The most salient part however I am happy to share, which is that inside periods (on the 480 minute [...]
In early 2014 we decided to implement an equity curve (EC) filter into CrazyIvan. The reason for doing so was the realization that CrazyIvan exhibited high dependency, meaning it was cycling through distinct clusters of winning and losing periods. Here is a recent reason example that demonstrates how our EC filter works:
Granted this is a rather idealized occurrence and our EC filter may sometimes activate or deactivate a bit late, which is unavoidable. As there apparently is a high correlation between winning and losing streaks between the symbols CrazyIvan trades and winning campaigns can run for a day or more the EC filter keeps an open cache of open positions in [...]