In developing trading strategies and specially with bots, you always need a way to measure the performance of a system. There are heaps of different numbers out there and even more ppl arguing which number to pursue. 🧵
Many numbers are useless on their own. Like Winrate: Have a 90% winrate but avg. looser loses 10x your average winner? you will feel like a winner while the big losses get you broke. 10% WR with avg win of 10x your avg loss? you feel like a looser but lambo soon!
Same with win/loss-ratio: means nothing without the winrate. For me, the importance in those numbers is that they need to fit your mindset. You need constant gratification? shoot for high WR. You don't care about looses but hate cutting winners? pump that win/loss ratio.
You need a system that fits you, cause you need to execute it. So shoot for performance numbers that fit. For me, the most important number is yearlyProfit/maxDD. Cause thats what makes all the diff systems comparable. I have a level of max pain that i allow for every system...
so i scale the position size up to that level. more position size more profit, but also more DD. So i set posSize to have an expected maxDD of 10%. Now my number tells me what to expect per year. its kinda scaling-invariant with fixed level of max pain.
Second number i look at is the underwater-days: so the maximum number of days between equity ATHs. Or the max time of pain. Cause being under water sucks. Being under water for a long period of time makes you question your strategy and do stupid things. so better reduce it.
All the other numbers are mainly of informative nature for me. I look at time, i monitor them during trading, but i don't care if high or low. I just look for deviations from the backtest.
and i hate pain.