slug, like they said you can use NinjaTrader but you will have to make assumptions about what's wrong with the backtests. That in itself is a huge study, but if you work with daily data it might be less severe of an effect backtest vs real world. If you're trying to test tick data, you will get bogus results because you cannot assume market structure (ie that your order would really get filled in reality) 100% of the time, not even 50% of the time. Just be very cautious with your results and plan on a decrease in AvgTrade anywhere from 20% to 100% in real life.
It is if you only care about minimal features and don't care about common pitfalls. Even backtesting has a lot of gotchas that really requires a lot of time to figure out, and knowledge about how to determine fills, slippage etc. If you don't care about liquidity, slippage and commission, then sure, a trainee can do it, but as soon as your model hits the real world it will crash and burn - not because the system is bad, but because you tested it in fairy-land. I speak from experience working on backtesters and execution for the past 2 years, programming. If you use amibroker or any of the backtesting platforms you need to check if they're not lying to you - but how? Etc etc. You get what you pay for, always.
Quote from luisHK:
Snide apart I do wonder sometimes what programmers with good understanding in those fields make - I actually suspected hedge funds and big investment banks to pay them better than 100usd/hr
It's fairly easy out here to get much cheaper programmers but I wouldn't trust them in trading system delelopment. Yet backtesting through amibroker or a similar software might be easy enough, even for an IT trainee to manage.
Is it ?
It is if you only care about minimal features and don't care about common pitfalls. Even backtesting has a lot of gotchas that really requires a lot of time to figure out, and knowledge about how to determine fills, slippage etc. If you don't care about liquidity, slippage and commission, then sure, a trainee can do it, but as soon as your model hits the real world it will crash and burn - not because the system is bad, but because you tested it in fairy-land. I speak from experience working on backtesters and execution for the past 2 years, programming. If you use amibroker or any of the backtesting platforms you need to check if they're not lying to you - but how? Etc etc. You get what you pay for, always.