Compare long and short trades fairly

Assign direction at position level

Classify the trade by the position opened. A sell that closes a long position is not a short trade. Treat partial exits and position reversals consistently, or you will compare execution types instead of trading decisions.

Split the eight sample trades

In our synthetic NQ sequence, trades 1, 4 and 7 are short trades. Their net results of USD 145, −155 and 95 total USD 85. The five long trades total USD 25. Together they still produce USD 110 after USD 40 in costs.

The short group averages USD 28.33 per trade and the long group USD 5. Three and five observations are insufficient to establish a stable preferred trading direction.

Compare within similar conditions

Then split by setup, session and unit of risk. Were most longs opened during a falling session? Were short positions larger? If a subgroup contains just one trade, document the case without presenting it as robust statistics.

Define the next observation

Write down the question before the next sequence and include every new trade. Do not remove difficult days after the fact. Check whether the difference repeats and which process characteristics accompany it.