Well, this could be changing with chatgpt This artificial intelligence is very interesting to generate programming code
Exactly why the only way I would back-test a strat is to use tick-by-tick data. No other way to tell if trade entries and exits are real-world possible results. Also out-of-sample testing and forward testing is the only way to bulletproof final end results. Finally as previous mentioned, a couple of times, slippage and commissions have to be accounted for.
Find a system that works, NEVER. Took me several years to develop then eventually program scalping systems and then day trading systems, I concentrated on very low losing percentages. But still I was not happy with getting 5% and under losses. I kept working and twenty years later am very content with what I have developed. I have always looked at financial markets different than most, didn't care about the money, only percentages...knowing that if seldom lose, account equity curve is more 90-ish degrees. I found studying physics catapulted to higher level.
Your brain learns virtually nothing from a bazillion computer generated "backtest of trading ideas". Quite the opposite is true when you pour over 1000's upon 1000's of charts by hand. Plus, when you also see price moving in real-time over 1000's of sessions, all of those jumping pogo sticks called price bars reveal setups the historical charts cannot show. Letting go of a position is not a computer curve-fitted, one size fits all proposition. Laziness gets its just reward...none.
There is a definite learning experience gained from generating a strategy from a curve fitted backtest, naively being super confident and optimistic about it, only to see it crash and burn quickly in real trading. That is when you first start to realise that even strict rules based systems trading is a lot harder than it looks in hindsight.
The NT backtest engine isn't that good. Better to download replay data for every contract and run the strategy in replay. The results will be more in tune with the real market. If you're serious about developing strategies, TradeStation is the only platform that's powerful enough. Their testing engine is far superior to NT.
listen...i hate to tell you this...u maybe a long time investor...but you sound like a noob trader. one time, i did paper trade on futes by buying 6 long contract on es and and hold it overnight...next morning i was up over $30k. but i know if i did this on live trading....it would end real bad.
No, it wouldn't. If you did it with live money, you would make your $30K. It is no different than sim. The problem lies with the fact that you can do that only so many times before you get a string of losses and you lose all the money in the account. This is why Rickshaw man has never capitalized on the overnight melt-up he has touted for all these years. It doesn't work. It is why there is no overnight fund. At best over a one-year period you end up at break-even. But it takes a lot of capital to survive the DD.