A calendar year starts on 1 January. Your account did not: judge by rolling windows
Most EA pages carry a year-by-year table. It looks complete, and it is one of the easiest tables in the report to misread, because every row starts on 1 January. Your account will start on whatever day you buy.
Why the calendar can hide a bad year
A bad twelve months does not care about the calendar. It can start in May and end the next April, split across two rows that each look fine. A calendar table shows you one starting point per year.
A rolling window closes that gap. Take every twelve-month stretch in the test, starting from each month in turn: January to December, then February to January, then March to February, and so on to the end. A test of 78 months contains 67 of them. The calendar table shows 6, plus a half year.
A worked example: our own index robot
Index Cadence was tested from January 2020 to June 2026, 78 months. Its calendar table runs from +40% in its weakest full year to +85% in its best. The rolling count is the stronger statement: none of its 67 rolling twelve-month windows was negative. The worst single month was -7.4% and the longest run of losing months was two.
That is still one path through one history, and the listing says so in the same sentence. It is not a promise about your next twelve months. What it does show is that the result did not depend on starting in January.
Your two-minute check
1. Count the rows of the year-by-year table. That is how many starting points you have been shown.
2. Look for a rolling-window count: how many twelve-month stretches ended below zero. If the page has none, ask the seller for it.
3. Read the worst month and the longest run of losing months next to it. They tell you what the inside of a bad window feels like.
4. If the last row is a part year, read it as a part year. It is not one more full year.
The calendar is a convenience for the report. Your account runs on its own clock.
Index Cadence on the MQL5 Market: https://www.mql5.com/en/market/product/190845
