Pattern Reality Check / Method
METHOD
How we test a pattern
Every pattern goes through the same steps. The rules are written before the results are seen. We do not tune a pattern until it works.
The test
Eight steps, identical for all 25
1 · Data29 Binance USDT perpetuals with long history, 4-hour candles (daily for the golden cross), June 2020 to September 2026, plus real funding rates.
2 · Rules written firstEach pattern is defined with the usual textbook numbers. Detection only uses information available at the time of the signal.
3 · Realistic tradesEntry at the signal, a stop, a target, a time limit. One open trade per coin and direction. If stop and target fall in the same candle, the stop counts first.
4 · Costs0.14% round trip for fees and slippage, plus the funding actually paid or received while the trade was open.
5 · Three periodsBuild (before 19 Sep 2023), validation (to 19 Sep 2025) and held-back (after). A result that only exists in one period is not an edge.
6 · Random entries1,000 runs that enter at random times on the same coins, in the same direction, with the same stop distance, target and time limit. This shows what costs and stops alone produce.
7 · Stress testsTwo maps that vary the pattern shape and the exits, a second timeframe, costs from zero to three times, and results coin by coin. Weekly-cluster bootstrap for the 95% and 99% ranges.
8 · Verdict"Candidate" needs: 99% range above zero, positive in all three periods, better than 99% of random runs. None of the 25 patterns passed, so none is called an edge.
Limits
What this does not show
- Past data. Markets change. A pattern that failed here may work elsewhere, and the reverse.
- Mechanical rules. Real traders add judgment, context and filters. We test the rule as written in courses, applied to every case.
- Limit orders. Order blocks, fair value gaps and supply and demand zones assume a limit order fills whenever price touches the level. That is optimistic, so their results are an upper estimate.
- Simulation. No real orders were placed. Real fills, delays and psychology make results worse, not better.
- Not our strategies. We publish tests of popular patterns. We do not publish or comment on strategies we use ourselves.
Found an error or a rule you think we defined unfairly? Tell us. Education only, not investment advice.