Seems to me that backtesting is just like the pattern or the history of the market and it is reliable in predicting the future, though there are other factors to be considered, don't rely on this thing fully.
The strategy which has one exception in the past 10 years and was proved to be false once. Can mislead, hundreds of times in the future. Well, everything will have an exception. Backtesting is useful, and you might find that your strategy is accurate. Proper management to your risk and reward, you can be a winner with it.
by definition, technical analysis is relying on the pair's history, in order to predict its future. saying you can't rely on backtesting equales to saying you can't rely on technical analysis.
this is partially true, however, if you do your homework around the fundamental area, you should be covered. (unless, of course, Ben Laden strikes again, or whatever... some things are never to be predicted - and that's where money managment kicks in)
If you have a trading strat that involves moving average crossoveres etc and you wanted to test that then its useful to see how often these crosses take place in your time frame, whether they would have made any money for your risk etc.
If there are a lot of whipsaws etc
So it has its place and is useful in that respect.
I think backtesting is necessary for us to learn how to trade...I personally learned all the things that dont work...which is most things.
All systems work some of the time, including my flip a coin sytem, none of the systems work all of the time, including my flip a coin sytem. So its up to you whether to backtest or flip a coin.
Backtesting will tell you very quickly what does "not" work. If you learn what not to do....then you will know what to do. Altough I have developed a number of systems backtesting, there are many other factors involved in trading. Amibroker has a great, and affordable, backtester along with free forex data.
For what it's worth, InterbankFX also has a back testing option built into their trading platform. Haven't used it, but for those interested in exploring various systems it's one more up for dissection.
i understand that you want to verify that your system is working like you want it to. But why backtest it if you can use a live demo account? Any clarity on this would be appreciated.
It comes down to time and quantity. Backtesting will allow you to evaluate potentially hundreds or thousands of trade signals in a relatively short time. How many live demo trades can you make in an hour? Not enough to validate your system idea I should think.
That said, not a bad idea to back test using historical data, then forward test in a demo account.