#2 could potentially be bad. Entering at last trade price is not realistic. Does your backtest still look good if you lag the signal by one candlestick?
I have the impression that @nooby_mcnoob is not interested in having his backtest being critiqued. But only whether his Sharpe ratio is correctly calculated.
I welcome critiques, actually. Just starting to get back into the system development since I was distracted by a somewhat large income opportunity which has now concluded. I'm sure I'll start posting to my journal again soon. I just haven't responded to tradrjoe because in order for me to lag by one candlestick, I'd have to do a lot of modifications to my backtesting code and right now, I'm still testing the idea and not stress testing my backtest.
Any measure that uses variance has the problem of time frame. A Sharpe ratio using months will be less than one using years. It's a ratio that, the longer the time period used as input, the better it looks. How many 3 sd days are there as a percentage of all days? How about weeks, months, years? When was the last 3sd day in equities? When was the last year?
All good points. I'm sure you can game each metric individually, but you probably can't game a combination of them.
If you are trading intraday and are using last trade price to enter into positions in your backtest, any mean reversion signal would look very nice on paper purely from the bid/ask bounce. But those are fictional returns because you wouldn't be able to enter into those prices in real trading.
My entries and exits are generally hours apart, and note that I am only making 1-2 trades per day. Would that still be relevant?
If your strategy is mean reverting, then it would bias you to often enter a buy at the bid and a sell at the ask within a backtest using last prices. In real life, this would not be realistic obviously. So it could potentially be the bulk of the outperformance shown in the backtest.