Tickfloor

9/21 EMA Crossover: claimed vs measured

Last updated

In-sample, not a promise. Every figure we measured here comes from a backtest on past data.

Is the 67.2% win rate claimed for 9/21 EMA Crossover real?

We did not reproduce the 67.2% win rate claimed online for 9/21 EMA Crossover as one pooled figure, but we did test the cross directly on hourly crypto and daily US stocks. No version passed our pre-set test after costs, and the full test is on our 9/21 EMA crossover page. That is a historical backtest, not advice.

67.2% is the win rate claimed for 9/21 EMA Crossover by investorguidefinder. We did not measure a pooled win rate for this rule. We tested the cross directly instead, and no version on its full test page passed our pre-set test after costs. Context from the sweep it belongs to: In a broader 6,428-cell indicator sweep, 24 passed BH correction; none passed the full reliability gate.

Our direct test of the 9/21 EMA Crossover

This is the filtered version on hourly bars: long only, when the 9 EMA is above the 21 EMA and the close is above the 200-period EMA.

The other versions we tested, and the method in full, are on our 9/21 EMA crossover test.

Method. The 9-period EMA crossing above the 21-period EMA, long only, the trend-following entry as taught in the source making the claim, run inside a larger sweep of indicator rules across assets, timeframes and confluence combinations.

Data. 40 assets across 6 timeframes, scored with Benjamini-Hochberg false-discovery correction across all 6,428 cells in that sweep, not judged cell by cell.

Receipt: parameters, data, window and result file
Run id
not recorded
Frozen parameters
The 9-period EMA crossing above the 21-period EMA, long only, the trend-following entry as taught in the source making the claim, run inside a larger sweep of indicator rules across assets, timeframes and confluence combinations.
Data source
not recorded
Instruments and coverage
40 assets across 6 timeframes, scored with Benjamini-Hochberg false-discovery correction across all 6,428 cells in that sweep, not judged cell by cell.
Date window
2016/2017 through the 1 March 2025 discovery cutoff
Sample
not recorded
Cost rule
not recorded
Fill rule
not recorded
Source script
not recorded
Benchmark
the direct test: an equal-weight basket of the same assets, rebalanced daily, paying no costs; none for the sweep
Full result file
the direct test's files are linked from its full test page; not recorded for the sweep

This is one line in a bigger check. The research desk has run 1,346 strategies, most against an equal-weight benchmark of the same assets rebalanced daily that pays no costs (the strategies pay theirs). After correcting for how many we tried (the luck check, Benjamini‑Hochberg, 1,778 tests), 0 passed. That does not prove no strategy works, and it says nothing about versions we did not test.

Tested a rule of your own? Check whether your backtest is just luck. It is free and needs no account.

What win rate does 9/21 EMA Crossover claim online?

67.2% win rate, per investorguidefinder.

What did Tickfloor measure for 9/21 EMA Crossover?

We tested the cross directly. The filtered hourly crypto version (long only, when the 9 EMA is above the 21 EMA and the close is above the 200-period EMA) returned 62.9 percentage points a year less than the benchmark after costs, and no version on the full test page passed our pre-set test. In a broader 6,428-cell indicator sweep, 24 passed BH correction; none passed the full reliability gate. That sweep is context, not a test of this rule alone.

Is this a claim about Tickfloor's own accuracy?

No. This measures a publicly claimed strategy rule, not Tickfloor's agreement score. Tickfloor makes no win-rate or accuracy claim about itself.

See the full method and every result or the strategy family index.

Other claims we've tested

Terms explained

Next steps

General information only. It doesn't consider your objectives, finances or needs. Tickfloor holds no financial services licence and never places trades.