I record every trade I make in Excel. I categorize them into short/long, which day of week, date, where the market opened, how the market performed the day before, how the stock performed the day before, average volume in the ticker, P/L in the ticker and which strategy. Of course, collecting this is very boring, but sometimes this is an invaluable tool.
This morning I discovered one pattern in one of my strategies which I don’t think is curve fitting: Certain stocks/industries perform much better than the rest in this strategy. Looking at my numbers I discovered the following P/L diagram during the last 50 days (this is a daytrading strategy. Ignore the numbers on the x-axis. P/L is on the y-axis):
- Win ratio per day is 67%
- Average win day is 693 USD
- Average losing day is 462 USD
Hence, expected profit per day is 358 USD. This is actually very good. Size per position is just 5 000 USD and on average this strategy gives 29 fills per day (29 different stocks). This strategy trades both long and short every day. Backtesting this further back I get even better results (but the the trading environment were better).
Different sectors have different patterns. Some tend to correlate more or less with market (S&P 500), some don’t go with the market at all, others hardly move at all etc. There are so many ways to trade. You have to keep looking for patterns which makes logical sense. You can’t trade oil stocks the same way as you trade utilities.
Every morning I look at my numbers looking for patterns or looking for something that I have not tested before. Usually nothing turns up, but the only way to make money is to keep looking.