Ask how much one observation contributes
A hypothetical record contains nine trades that each lose one unit and one trade that gains twenty. Its total is eleven units, despite nine losing observations. That arithmetic is correct, but the total alone conceals how concentrated the result is. This is an invented teaching example, not a report of a gold strategy or evidence about future returns. Win Rate Leaves Half the Question Open considers frequency and magnitude. Here the narrower question is whether one exceptional observation carries much of the recorded outcome.
Describe the concentration before changing the dataset
The mean of those ten fictional observations is 1.1 units per trade; the median, the middle of the ordered observations, is minus one. Neither description cancels the other. They answer different questions about the same record. Report the underlying sequence and calculation convention so that a reader can see why they differ. Do not remove the large gain simply because it looks unusual, and do not remove a large loss to improve the average. An observation may require checking for a recording error, but a surprising value is not itself proof of an error.
Keep sensitivity analysis separate from reported performance
You could show the original total beside a clearly labelled calculation excluding the exceptional observation. That is a sensitivity exercise, not an alternative history in which the trade never happened. Explain why the comparison was made and retain the complete result. Compare Like With Like adds the need for a consistent period and basis. The Observations That Never Reach the Journal explores how selective records can change the question before any calculation starts. A clean-looking summary cannot repair an incomplete record.
Use the example to improve the research question
Ask what circumstances produced the exceptional outcome, whether the rules allowed it in advance and whether its size was recorded under the same convention as other trades. Do not infer that such an event will repeat or that an average is a dependable income. CyPredict and The Gold Trader’s Code provide connected educational reading. The useful discipline is to show the full record, multiple descriptions and the limits of each. That gives another reader a way to assess the concentration rather than inherit the confidence of a single total.

The Gold Trader’s Code
Continue the study of rules and professional trading decisions.
Discover the book ↗Read the book overview →Sources & further reading
Sources consulted 10 November 2026. Editorial frameworks and illustrative examples are identified within the text.



