If backtesting is nothing, then why is it there in many software? Also, backtesting is like testing of any thing you make. You create something then test it with your own data to see if it works. Then,you check it out with real world data to see if it works correctly. Do we not use the same way in proving out theorems at school? I think the same applies to software (here trading ones).
Backtesting software is like guitars. Tons of guitars are sold but how many owners ever perform for audiences and make a living at at? You don't even have to pay for a backtester nowadays, Ninjatrader's is free, you'd think that tons of people would have automated systems up and running and the economy would be booming...
Backtesting is not nothing. It is also not everything. It is just something you should use only if you understand how to do it properly and its limitations.
Thanks for the replies, maxpi and intradaybill. You both have your points. But, I feel backtesting is worth while and just because we have not been able to use it properly and well, does not mean it is useless. Also, I think it is the way of application that makes it worthy, just like we all know that internet is a place to earn a lot, but how many of us have managed to earn? I think it is the same with backtesting, how many of us have managed to make it work to our advantage?
I think part of the conclusion can be found in this paper from another thread. The author of the paper, a well-known expert in this field, claims that two widely used backtesters have significant defects and produce wrong results because of basic errors in their algorithms Backtesting Flaws in two Popular Programs
So true. The influence though is inversely proportional to the timeframe of the trades. For example, trend following is not affected that much and the differences should be insignificant, like a bit of slippage, etc.