I started trading this strategy live in July with the lowest minimum size I can possibly trade. Now I’m about to put on more size and just summarized the numbers so far.
First, here is the day to day profits:
July, August and October all showed decent profits. September was just a grind.
And here is the accumulated returns:
Now, 10 000 USD for almost four months of trading is nothing to write home about. However, I can easily trade 5 times the size I’m doing now. And I emphasize the importance of having several strategies. This is just one of 6 others I’m currently trading. I don’t want to reveal this strategy as I don’t want a crowded marketplace on the paths where I go. The strategy takes on positions during the first two hours of the trading day and exit is on MOC. The entry is based on parameters/performance going 1-3 days back.
Both long and short orders have outperformed SPY with 0.08% during the holding period. Now I’m looking to hedge with SPY those days where I’m too long or short.
There are no stops except time stop at MOC. Using my stats I can probably take profit earlier when I reach about 7-10 cents per share in profits. I have had a lot of days when my P/L have dropped significantly from the peak. So far the best day is 10 cents per share traded.
As expected the strategy has performed worse than the backtesting. Backtesting always involves a little bit of curve-fitting and better fills you’ll never get in real life.
The number of fills varies from day to day, but average is 35 fills per day. On average I have about 200 000 USD worth of stocks, both long and short. I expect to trade 1 mill USD quite soon on average. If I deposited 200 000 in a retail account (I trade prop), then I would have at least 800 000 in buying power, so 200 000 should be more than sufficient to trade this strategy. This means that the strategy has generated 5% return in four months with hardly any drawdown at all! I think that is pretty good and it shows the value of daytrading and the law of big numbers. Just a small edge can turn into acceptable numbers if done multiple times.