Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
5 posts • Page 1 of 1
Hello fellow Blox people, I recently read something from David Druz that I found pretty interesting. The theory was that the higher the expectation of a system, the harder it is to trade from a trader mindset or "comfort" perspective, so that the way a trader earns the gain is from withstanding a "difficult" equity curve. I have a short term system that has historical win rate of 22% but recently I have noticed very long losing strings which terminate with 1 or 2 very very large profit trades, sufficient enough to wipe out the previous losses and end with a profit. In light of what I said above, are these just cycles of profit patterns that oscillate with every liquid trading instrument?? I mean do systems normally go from having 20 losses and 1 big winner and then later in time may have 6 losses 2 small winners, 5 losses and 1 medium size winner etc. thus providing a "smoother ride" to profit. Any insights to this thorny question would be a big help to me. Thanks, Zoopy12
There is no way to say 'what systems normally do' -- as every system is designed differently for different purposes. However you can be sure that the markets by design will make it as difficult as possible to profit -- if it were easy everyone would do it. So certainly no pain, no gain, and 'most systems' should expect extended periods of draw down which if the system is robust should eventually spring back into profit -- usually violently to catch people off guard. The typical moment for a reversal is when everyone has given up. Typical psychology.do systems normally
I would agree that the higher the expectation, typically the more volatile the equity curve. However by tracking the standard deviation of the equity curve dynamically one could adjust the system parameters and reduce some of that volatility. Trading Blox is well suited to develop any strategy based on any design.
Hello Tim, Thank you very much for providing some keen insights to my questions that help me accept the nature of the game. Your response did bring up a follow up question. When you mention "tracking the standard deviation of the equity curve dynamically to adjust parameters" I am unclear as to the idea behind it, how would it help?? Is it by showing that once the system has "gone bad" by a certain degree, some component of it needs to be changed?? Thanks again for your generous sharing of key knowledge. I need to take the dive into learning how to code and thus be able to test my systems. My method relies on price bar interpretation of intraday data. I know exactly what my formation looks like and the objective rule set, but I don't how to translate that into code. I'll dig into your software info, perhaps it is much easier to do than I think it is. Thanks, AL
Dynamic adjustments of parameters are not a condemnation of the system, but rather part of the system design. There are many aspects of a system that could be adjusted, based on ongoing performance. It's an area to explore, and are able to help with consulting services for scripting and system design if that is helpful or desired.
Hello Tim, Thanks for clarifying the idea. When I posted the original question "expectancy versus tradability", it was because I was seeing long loss strings in a Single symbol, specifically the YM. Given my own experiences and what many others have said, since I am trying to catch trends, the addition of different symbols using the same method is something that I think could help smooth things out, the two caveats being each new symbol requires more risk allocation and sometimes ( a lot lately, it seems) the action gets choppy across many assets that you would think would be uncorrelated. That being said, I paper trade a portfolio of 6 different SP 500 plain vanilla stocks using the same trading logic and the results are certainly pretty smooth and I can see the edge showing itself by the math, over many trades. As volatility contracts, of course the overall gain suffers some, but the + edge over (608 trades feb 2018 to present) remains. I am open to exploring whatever services you may offer in regards to system refinement. As a very small trader I am somewhat budget constrained, but again I can learn what is available if not for now then for the future. Thanks, Zoopy 12