A ranking is not a return leaderboard
A 50% gain over six months and a 50% gain over six years are not the same pace. Period normalisation exposes that difference, but it also magnifies short bursts. A short annualised number is not evidence that the pace could persist for a full year.
Match the account boundary
Before placing two records side by side, ask whether each is a personal cash account, a fund, a competition account or a model portfolio. Confirm the start and end dates, cash-flow treatment, fees and account currency. Do not merge a competition division with a personal audited series just because the trader is the same person.
Read the ranking inputs
The site's composite methodology considers more than pace: independent documentation, risk and consistency affect comparability. Use this calculation as a diagnostic, then inspect each profile's evidence tier and the published scoring method. Missing drawdown or account scope should remain missing, not estimated.
Worked comparison
An illustrative 50% total gain over 18 months annualises to about 31%. The same 50% over six months annualises much higher, but the second figure is based on a far shorter observation window. The calculator is useful for spotting a mismatch in periods; it cannot supply the missing risk history. Put maximum drawdown, starting capital, leverage and the number of independent observations beside the period-adjusted figure. If any field is absent, mark it unknown rather than filling it with an industry average.
Comparable or not comparable
Keep a row for each account with the original start and end statements, cash flows, fees, account currency and whether the record is personal, managed or competitive. Do not compare a backtest with live capital on the same evidence tier. If two traders differ in leverage or drawdown disclosure, retain those differences visibly even when the headline returns match. A ranking can be useful for discovery while still declining to assert that every pair of records is economically equivalent.
Stress-test the ordering
A ranking should show whether a modest change in evidence quality, return period or drawdown treatment reverses adjacent positions. Where the order is sensitive, describe the records as a close comparison rather than claiming a precise superiority. Preserve the underlying inputs so a reader can reconstruct the judgement. A transparent tie or an unranked record is more informative than a decimal score that hides missing data.