ROTFL... very nice Edit: I'd add Bwolinsky and TraderZones to that list. Bwol - huge ego with nothing to back it up. Sense of reality completely gone... TZ - bully and OCD disorder with spammers/pretty much anyone. I kinda think he's got a real disorder (the large fonts make me think he talks *special* in real life).
I am not very good at all this back testing/system testing stuff and so I am not your ideal candidate. However I see you are up for a maximum of 15 losing bars and so I wonder on the 80/20 rule how many losing bars you must absorb on 80% of the occasions. As I say, this is not my forte, but if it were I would try to reduce the impact of my losing trades on my bottom line. Perhaps this is not possible. How many instruments are you live on at any given moment. Do you hold over night and over the weekends. sosueme
Good ideas come from everywhere, regardless of ones background. I'm not sure I understand your suggestion - my max. loser count is 15, so how does that figure into an 80/20 rule? Please explain a bit further... Losing trades are kept small, hence the high run count of losers, I can easily change that, but, I like the feel of trading with small losers in mind. I've got other strats that are the opposite, but, lets just focus on the dynamics of this particular strat for now. This strat trades about 1200 products (1k++ Stocks, 20 Futures). At any given time I may have up to 200 correlated trades working for this strat alone. Its intraday only, no overnights.
You are very generous towards my lack of skills in this area. 80/20 My question is how many consec losing trades do you have on 80% of the occasions. We know the max is 15, but what is the max for the main distribution minus the tails. I see you work up to 200 correlated trades at any given moment on one strategy ... obviously you are highly organised and very busy. What is the advantage in employing correlated instruments. I would have thought that a low correlation would have balanced your cashflow. When your strategy issues a signal, how many of the 1200 instruments go live. You have obviously committed a great deal of resource to this trading programme. sosueme
67 trades per year only on average on a 5 minute chart on the ES Mike?. I can see how you want to increase frequency by using other instruments. Not sure what to suggest at this point. Not judging anything either one way or the other. Do you have an equity chart pic for the 12 years?. Is the slope consistent?.
I would be interested to read your application of ArcSine to the coin tossing game, I can see how useful it could be sosueme
There is no advantage in employing correlated models/instruments other than when you're right - you're right big time (vice versa for losses). This is one execution model out of about 30, so while this one may take up to 200 correlated trades all losing/winning together, the other models will have just as many trades trying to exploit a different dynamic. The less correlation among independent models the better. I figure 80% of the products I trade display a similar price characterisitc, hence the need for limits on model trading activity. Some days, all all models just rock and I'm having a very good day, other days they cancel each other out (P/L wise) and I'm flat or moderately positive. Occasionally nothing works and I'm losing a good chunk. Overall, the combined equity curve from everything is very smooth. No one model can hurt the overall performance. It really depends on the market dynamic...
See the previous attachments. This model was originally constructed for a portfolio, hence its individual behavior will be different across CL, TF, YM, AD, equities etc. When looking at the portfolio results I have a ton of trades and a smoother curve, but when looking at individual products the curves are rougher as the model is not "fit" to individual products, but rather for all products with non-product-specific conditions. The 5min chart is just what I had up, its not time frame dependent.
In any not Normally Distributed stochastic process (i.e. price fluctuations in the markets) the number of zero crossings is less then normal and deviation is wider than normal. The difference = your profit. It is that simple!