Back to my live day trading test. After 13 days of live trades, the statistics are similar to the paper trades except profit came down from 1.2% to about 0.5%. However, that is not what bothers me. What really bothers me is when I drilled down to the details, the statistics looked too much like a random process: a 50/50 coin flip. Here are the statistics:
Live Trades:
# of trading days: 13
# of winning days: 8
# of trades: 317
% of winning trades: 49.7%
Average # of trades per day: 24
Min # of trades in a day: 6
Max # of trades in a day: 37
Average profit per day in %: 0.5% of the trade size
Paper Trades:
# of trading days: 53
# of winning days: 46
# of trades: ~1500
% of winning trades: 50%
Average # of trades per day: 28
Min # of trades in a day: 10
Max # of trades in a day: 50
Average profit per day in %: 1.2% of the trade size
I can calculate the p-value to see if the outcome is statistically significant but I already know the answer: It is not, because if I removed two big winning days out of 13, it is breakeven. Those two days are statistical anomalies.
I will complete the test in a month but as of now I don't think I have a winning strategy unless I can make some changes.