Blaiserboy wrote:I wonder if you could email that strategy to Popov, his email is in his profile. I am sure he will want to see the discrepancy real time so he can analyze the differences.
Hopefully my post above has everything needed, the strategy's xml file is attached.
We can not really know until someone actually runs the thing real time. I am really suspicious of Metaquotes both 4 and 5, part of the reason being that they do not provide much in the way of worthwhile statistics.
I am not so sure that a MT4 backtest will be reliable regarding slippage as compared to a real account trading.
Oh I agree totally, I'm deeply suspicious of anything Metaquotes comes up with. I don't think its any accident that "opening price" backtests in MT4 are deliberately incorrect, or that the "every tick" backtests take FOREVER to run. I mean, the one I posted took 16 hours to complete on my very powerful PC, which is obviously a real hindrance to using the platform for any kind of real backtest work, forgetting the absence of OOS testing and Monte Carlo analysis.
What I'd expect though is that MT4's backtests would tend to sugar-coat results to make them look better than they really are, as its in Metaquotes' (and their brokers) interests for you to stop researching & start trading. When therefore MT4 shows a significantly worse result than an independant product like FSB, I need to question what's going on. At the very least I'd like to see a reasonable degree of correlation between the two.
Btw in my MT4 backtests I deliberately set the spread quite high (100 points) to mimic the effect of slippage.
Edit: Quick note on Steve Hopwood's site, which I've frequented on & off over the years. There's some wonderful people & brilliant minds there but there seems to be an entrenched belief that backtesting is useless, that only forward testing is of any value. I can understand not wanting to backtest with MT4, indeed it'd be impossible with many of the multi timeframe systems in development there, but to me a rigorous backtest is essential. I simply won't waste time forward-testing a system if its backtest shows it blowing out accounts, and by refusing to test I'm afraid many of those guys are groping forward in the dark like blind men in a coal mine hoping to find a diamond.
Sorry, that's rather off-topic for this thread, I'll end my little mini-rant there
Edit #2: What I'd really love one day is the ability to export to cAlgo EAs, so I can dump MT4 once & for all.