intradaybill raises a really good question and one I've been wondering about myself. I suspect that many people are using JMA, AMA, etc. to smooth price for basic technical strategies like MA crossovers. It seems unlikely that something so facile could produce consistent actual results in an entirely systematic strategy. So, kut2k2, I'd be curious to either see test results or read a very general one or two sentence description of what you or anyone else is doing with a price filter in a profitable systematic strategy. I'm not saying it's impossible to trade TA + filters but I'd be really surprised if anyone had something that's held up for more than a few weeks b/f failing.
Results posted to the thread. BTW notice how the clown who keeps challenging my math knowledge can't even calculate how many winning trades there are in the baseline system (jalsck and I agree that there are 32 winning trades (33% of 97 total)).
Or you could just accept that noise is noise, and ignore it. You don't need any moving average for that.
Moving averages are laughable for the most part. In the case of an SMA if the trailing bar's value changes a lot it affects the current value same as if the current bar's value changed a lot.... so people use EMA's to reduce that effect but nonetheless, the trailing value still affects the current value... I'd say that the longer term averages, say 100 bars and up might have more use than the shorter ones that try to emphasize the more recent bar values.... there are setups from moving averages of short and/or longer length that have a high success rate but the trades are so infrequent and the gains are sometimes not that great... to me the whole world based on moving averages is not worth visiting very often...
Explain how you "just ignore noise" in an automated trading system. What's clear to me is that some of us have given this subject a lot more thought than others of us. If you don't know how to deal with the noise, just admit it and stop pretending like the noise is no big deal, because frankly it is.
kut2k2, Could you post the results again? I didn't see them in this thread. So, my understanding is that you're trading some type of AMA crossover strategy. Is that correct? Is it a combo of TA-type indicators or just AMA? Is it the same strategy for all instruments? What is the hypothesis? Expected drawdown? Risk level? What do the 97 trades represent? In-sample/OOS/actual? How long have you been trading the strategy live? Are the live results similar to the backtest? I'm just trying to get a general idea of what you're doing and how rigorous your development process is. Thx, Lou
Results are in the thread that was linked to. You know, the one that you quoted? Yeah, that one. The results are self-explanatory.
You are wasting your time. He is a crank. He confuses the probability space with the number of trials. (He thinks that if you have N outcomes, you do N trials to get the expected value). He confuses random variables with their values. All that is documented well. I show his post to a few statistics grad students and they were laughing on the floor. At least the other "adaptive systems" guy was honest enough to admit that his swarm system could not produce a better performance than a simple moving average: http://www.elitetrader.com/vb/showthread.php?s=&threadid=206177&perpage=6&pagenumber=13 Adaptive or non-adaptive, if you are smoothing prices you lose information and you pay for it. It boils down to this simple fact. In the markets, you are penalized for missing information. The crank has never traded and does not know that.
I was just trying to be helpful. Nevermind. You've been losing money for 5 years and have 2000 posts on ET. You must enjoy it.
So you have nothing constructive to say about the test results. Yeah, I didn't think so. You must hallucinate that you're psychic also, to think you know my P&L statement. We are talking about indicators in this thread, not trading strategies or bank accounts. If you want to talk strategy, go to the appropriate forum. You insinuated my AMA was crap, now you want to walk away sniping after being shown wrong. You must take lessons from imbecilebill. BTW if you're capable, explain how to use wavelets for real-time filtering. I've seen the claims also, but they are all remarkably void on details.