Artificial Algorithmic Systems Real Trading

System 230913_22290_B sold ENI
System230905_20948_BN bought 1490 shares of CLE at 1.338

 

System 230913_22290_B sold MLK

 

System 230913_22290_B sold ASM
System 230904_56931_B bought 78 of BCS at 25.7

 

System 230904_41122_B sold NNG
System 230913_22290_B bought 7912 shares of LES at 0.252

 

Month summary:
I'm down to 97k, so that's about 6% loss in October. Main index went up a lot but that's mainly because large moves in large caps.
So far (since end of July) I'm down about 3% and indexes are at about the same level.
I don't want to this to look like I'm trying to change my benchmark, but... had I bought every traded share at the end of July, I would be down about 4%. So in other words, my results are like trading randomly.

 

System 230913_22290_B sold SKL

 

System 230904_41122_B bought 32 shares of BCX at 61.6

 

System 230904_41122_B sold ADV at 1.105
System 230913_22290_B bought 2419 shares of VVD at 0.824
System 230913_22290_B bought 3834 shares of STS at 0.52
System 230904_41122_B bought 35 shares of PTG at 56.4

 

Either that or you have not had enough sample size to determine the expected return.

I kept questioning my day trading system because I can't disprove the null hypothesis: That the results could be produced by a random system without positive expectancy.

 

Number of trades in training period is around 100+ (depending on the system), verification period: ~200, test period: ~250 (highest 400, lowest 195). Each of periods is 5 years (actually test period is almost 6 years by now).