How much of the backtest actually survived?
A backtest is the best case a set of assumptions can produce. This puts that report next to what the strategy did afterwards and measures the gap — including the two ways this comparison can quietly flatter itself.
Diagnosis
Read the first two entries before the percentages: overlapping periods and the same file loaded twice both make the comparison look perfect.
Every metric, side by side
Ratios are forward divided by backtest, so 100% means the figure came through unchanged.
| Metric | Backtest | Forward | Retained |
|---|
A ratio is left blank when it would mislead — most often because the backtest side is zero or already negative, where a positive ratio would read as "held up" while both sides lose money.
Both equity curves, normalised
The periods and the trade counts differ, so both curves start at 100 and the axis runs from the start to the end of each report.
What the two reports actually are
Before comparing results, check that the two files describe the same strategy over periods that do not overlap.
| Field | Backtest | Forward |
|---|
How this is calculated
Both files go through the same reader as the Backtest Report Analyzer: statistics are recalculated from the trade list rather than taken from the report's own summary. That is what lets an MT4 report and an MT5 report sit in one table, and it is why a live account statement can be compared against a tester run.
Degradation is normal; collapse is not. Costs, latency and a broker's actual fills take something off every strategy, so a profit factor that keeps most of its value is the expected outcome. The number that matters is how far below that the forward result sits, and whether the trade count moved — a strategy that stopped trading did not degrade, it stopped meeting its own conditions.
Slippage cannot be measured here. That would need the same trades on both sides, and a backtest and a forward run are different trades in different weeks — subtracting their average prices measures where the market was, not how it filled. What is measurable is whether the edge per trade shrank while the hit rate stayed put, which is the shape execution cost takes.
Nothing is uploaded. Every number on this page is computed in your browser, and the link you copy carries only the values you typed.
Related tools
A forward result that holds is evidence the assumptions were survivable, not proof the strategy is safe. It is one path through one market with one broker.