The median trade versus the mean trade
Posted: Sat Oct 31, 2026 7:00 pm
An average trade of plus 0.5 pips can come from many small losses and a few big wins, or from consistent small wins. The mean hides the shape. The median trade reveals what usually happens.
If the median is negative while the mean is positive, the strategy depends on outliers. Those big wins may be rare and fragile, and the experience will be of frequent small losses, which is psychologically hard and leaves the system vulnerable if the outliers stop occurring.
Conversely, if the median is positive and the mean is lower due to occasional large losses, the strategy resembles selling insurance: frequent small wins and rare disasters.
Understanding which type of system is in use informs risk management. The first type requires patience and a high tolerance for losing streaks. The second requires strict control of tail risk.
Plot the histogram of trade results, and note the mean, median, and the largest wins and losses.
Knowing the shape helps in setting expectations and sizing positions.
Is the strategy's median trade positive or negative, and what does the distribution look like?
If the median is negative while the mean is positive, the strategy depends on outliers. Those big wins may be rare and fragile, and the experience will be of frequent small losses, which is psychologically hard and leaves the system vulnerable if the outliers stop occurring.
Conversely, if the median is positive and the mean is lower due to occasional large losses, the strategy resembles selling insurance: frequent small wins and rare disasters.
Understanding which type of system is in use informs risk management. The first type requires patience and a high tolerance for losing streaks. The second requires strict control of tail risk.
Plot the histogram of trade results, and note the mean, median, and the largest wins and losses.
Knowing the shape helps in setting expectations and sizing positions.
Is the strategy's median trade positive or negative, and what does the distribution look like?