Avg return changes significantly when amount per trade changes

Hi Guys,

Does anyone know why the average return per trade figure in the backtest report changes dramatically when the ONLY change to the strategy is the dollar amount invested per trade? Yes the number of trades decreases as you allocate more per trade, but if you still have thousands of trades in each test and average return per trade varies up to 50% (with the only change being amount invested per trade), what would cause this?

For example, when I run a backtest that starts with $100k and I invest $1k per trade, I get 16k trades and an average trade % return of 0.8%. When I invest $50k per trade, I get 1k trades (which makes sense) yet my average trade % is 0.45%. When I invest $25k per trade, I get 3k trades (which makes sense) yet my average trade % is 0.25%.

How can the avg trade return be so drastically different when the number of trades in each backtest is still high/significant? I get that the pool of trades is different when I adjust the amount invested per trade but that's not enough to explain the vastly different figures. If the difference in avg trade were smaller then that would make more sense.

Any insight on this would be much appreciated.

Thank you,

Check you're code it seems that you have dynamic position sizing that hits the upper limit of you capital constrains plus take a look at commissions - perhaps the reason of the diminishing profits.