Faster MAs always cross over slower ones when trend direction changes. The idea of using the JMA (or any ama for that matter) as a proxy for price is to reduce whipsaws. I agree, the "JMA code" posted above is garbage.
I'm trying to help them, mate, but some folks are mighty attached to their old ways and won't listen.
You've been proven wrong twice (the "superiority" of the HMA; the value of patenting an algorithm) so I don't know exactly what "help" you think you're providing.
There will always be people at these levels of TA. They are taking their chances (61.8% losers). If in addition their decisions have random components, they fail 100% of the time. Copy this list, print it and frame it. Level 1: discretionary random Level 2: TA - mechanical Level 3: price action - mechanical Level 4: discretionary non-random Level 5: master of trading - (inc. heuristic high speed autotrading) Level 1: 100 % losers Level 2: 61.8 % losers level 3: 38.2 % losers Level 4: 23.6% losers Level 5: 0% losers Statistical error: +/- 2.5%
1. For starters, Wilmott.com mainly consists of quant wannabes 2. Most members of a real quant site (like NP) find discussing Jurik hilarious. 3. I still think Mark looks like someone from "The Thunderbirds"
I've used a trous wavelets, redundant Haar, HMA, looked at JMA. It didn't seem any different than a redundant Haar wavelet. Honestly, after a lot of effort I couldn't find a reason to smooth anything. Price has the most information so why would I strip out the information? In that regard, these filters can have some value if you want to detrend w/minimal phase effect so you can get a better look at the residuals. But, smoothing by itself. For what? So I can have silky smooth canned technical indicators? Quant funds aren't afraid of that. It's just not what they do.
OK, putting aside the issue of Mark Jurik, how do you decide what's a real quant site and what's a wannabe quant site?
I think Wilmott and nuclearphynance.com are probably as close as they come. Wilmott does appear to mostly be students trying to get into finance with some market professionals and total noobs thrown in. Most currently employed quants are not going to take the time to go posting on message boards. And even those two sites primarily deal with pricing derivatives so why you'd think they care about some moving average which looks like is meant for TA is beyond me. And if you think it's really that great then try writing a strategy and backtest it that can generate 3+ sharpe (and of course trade some size and reasonable tcost assumptions) and I'll help you get a job trading it (or find a headhunter) where you can be like these quants you're basically laughing at making a 6 figure salary (+ benefits) and potential 7 figure bonus.
No offence but 0% losing trades is impossible however strong the edge is... unless a trader extends stop-loss forever in vain hope of winning every time, which is madness. To the OP... people reply to threads they find relevant. Most posters on both Wilmott and Nuclearphynance specialise in pricing financial derivatives and not in systematic trading. Naturally, "improved" moving averages are of absolutely zero value to someone who values Bermudan options.
wait i thought you posted Statistical error:+/- 2.55789. Round to 2.6%,, oh damn now your whole argument is wrong.