The only way to really know this is to use pip/tick data.... For backtesting to work in Forex, you would need to consider trading during those times between the news and limit the length of trade. But this would be a labor intensive model and require more of your time to execute, thus more labor. Michael B.
With a 29% modeling quality, that backtest is useless. MT's backtester is not very reliable and doesn't look inside the bar. The highest you can get is 90% with 1 min data. I wouldn't trust the results.
I have no idea with the modelling quality given by MT even if I read their manual. But I think if the model is on a day chart, and the buy/sell decision is based on previous close, it should be fine. Isn't it? It is possible that there is a dramatically up/down, but in a daily chart basis, I think it should have 80% accuracy. Although the backtest is not 100% accuracy, but we should know how much difference between the backtest and real trade. guess 20%?
Do you use back-testing? If not, how can you make sure your trading strategy is good? To find a good trading strategy without back-testing is definitely time-cosuming, and probabaly, you will not find one at all, or lose all the money one day finally. (That is what I am thinking seriously)
Backtesting is helpful to START to consider a strategy..it is not the end all. Discretionary trading is REQUIRED in my humble opinion and you will need to learn how to trade. Putting the Market in a can is not that easy tonyzhou. You will not approach Forex logically and win...sorry... You have not even discussed the "keeping your dealer" happy aspect of trading a secondary market...(heck...there is not any solid evidence of a primary market...lol...yes Banks trade huge sums with each other...)
I have a stock trading system, runs well, I back-tested that system first, it also runs well in real situation right now. I tried to add some human intelligence , but the performance is bad. So I stick on that system now without any human intervene. I guess the same thing in forex. Too much human intervene in a trading system is dangerous, I think. So I hope to get a system and leave it untouched. Like the RSI system I posted, it is only rely on single RSI, the back-testing result is positive. The only thing I am not feel comfortable is win trade 30% / loss trade 70%. I haven't put it in real to test. But even if I test it in a demo account, will it be the same in a real cash account? I guess it will be some small difference.
ES is right, you need to be less historically dependent and more proactive to have any lasting success.
Not only is ES right, Tony, he has given you two genuine pearls of wisdom: 1) "Discretionary trading is REQUIRED in my humble opinion and you will need to learn how to trade." 2) "Putting the Market in a can is not that easy tonyzhou. You will not approach Forex logically and win...sorry..." There is no "set-and-forget" system which will win you consistent profits in forex, nor can there be one - the forex market is dynamic and in constant flux, and no mathematical algorithm can solve it for you. You must study both the technical and fundamental analytical keys to understanding it, and then hone your observational skills by watching the market move until you feel confident you can recognise what it is most likely to do next. If that sounds demanding and difficult, it is! Too bad there's no easy way, no "lazy man's road to riches" - but there just isn't.
Shhhh . . . it might take someone several years to figure that out -- you wouldn't want to rob him of that experience now, would you?