Every strategy family we tested
672 strategies, 755 test runs corrected together for multiple testing. 0 passed.
Findings
Every test we have published, newest first: the rule against holding, after costs, and against random-entry copies of itself.
0 of 755 tests passed correction noneallFive widely discussed trading techniques of autumn 2026, tested
We tested five techniques that came up often on Reddit and YouTube in September and early October 2026, 16 versions in all, after costs, with the rules written down before running. None passed our pre-set test. The intraday ones are hourly crypto versions, not the 1 to 15 minute futures and stock setups people post.
0 of 719 tests passed correction noneallDoes the opening range breakout work? An hourly crypto version, backtested
Did not pass (hourly crypto analogue). Lost money before costs; 0 of 52 earlier multi-market versions passed.
A year, after costs Beat 1 of 200 random-entry copies noneallDoes the ICT Asia session liquidity sweep work? An hourly crypto version, backtested
Did not pass (hourly crypto, no structure confirmation). Beat 0 of 200 random copies; 0 of 1,057 sweep rules passed in the earlier study.
A year, after costs Beat 0 of 200 random-entry copies noneallDoes the VWAP and 9 EMA strategy work? An hourly crypto version, backtested
Did not pass (hourly crypto). Trading costs on thousands of trades swamped both versions.
A year, after costs Beat 0 of 200 random-entry copies noneallDoes the 9/21 EMA crossover work? We tested it on thousands of trades
Did not pass. Costs ate the plain cross; the filtered version beat random copies but trailed the basket.
A year, after costs Beat 39 of 200 random-entry copies noneallThe 200-day line, the golden cross and Uptober: tested on 18 coins including Bitcoin
Did not pass (18-coin basket). Smaller drops, smaller returns, also on the matched window; Uptober beat most random copies but not enough.
A year, after costs Beat 49 of 200 random-entry copies noneallOpening range breakout, 81% win rate
A viral clip says 81% wins. On NQ it won 90.4% of 1,641 trades and still lost 1.15 points a trade on the test split, net of costs.
Net result per trade, test split (points; TSLA in dollars) Volume profile POC pullback
8 configurations on 8 Binance USD-M perpetuals. 0 of 8 passed. The best lost 0.087R a trade on the test split, net of costs.
Net result per trade (R), test split: best and worst of 8 Fibonacci Retracement: does it work?
Did not pass. Crypto rule: -3.8% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 6 of 200 random-entry copies noneallCandlestick Patterns: does it work?
Did not pass. Crypto rule: 15.4% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 199 of 200 random-entry copies noneallSupport and Resistance: does it work?
Did not pass. Crypto rule: -0.6% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 56 of 200 random-entry copies noneallStochastic, CCI and Williams %R: does it work?
Did not pass. US stocks rule: 0.9% a year after costs, against 18.4% for holding the same assets (corrected q 1.00).
A year, after costs Beat 84 of 200 random-entry copies noneallADX and Parabolic SAR: does it work?
Did not pass. Crypto rule: 11.2% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 196 of 200 random-entry copies noneallFunding Rate and Open Interest: does it work?
Did not pass. Crypto rule: 6.4% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 84 of 200 random-entry copies noneallBollinger Band Squeeze: does it work?
Did not pass. Crypto rule: 0.0% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 130 of 200 random-entry copies noneallPairs Trading: does it work?
Did not pass. US stocks rule: 12.0% a year after costs, against 14.2% for holding the same assets (corrected q 1.00).
A year, after costs Beat 66 of 200 random-entry copies noneall52-Week High: does it work?
Did not pass. ETFs rule: 10.0% a year after costs, against 9.4% for holding the same assets (corrected q 1.00).
A year, after costs Beat 165 of 200 random-entry copies noneallSanta Claus Rally: does it work?
Did not pass. ETFs rule: -0.1% a year after costs, against 13.2% for holding the same assets (corrected q 1.00).
A year, after costs Beat 69 of 200 random-entry copies noneallMACD + RSI Confluence: claimed vs measured
Claimed 73 to 90% wins. We measured 30.6% over 252 trades, net of costs.
Win rate, net of costs 0%100%ClaimedMeasuredStochRSI Cross: claimed vs measured
Claimed 75% wins. We measured 54.9% over 4,763 trades, net of costs.
Win rate, net of costs 0%100%ClaimedMeasured9/21 EMA Cross: claimed vs measured
Claimed 67.2% wins. Not tested alone: in a 6,428-cell indicator sweep, 0 passed the full gate.
0 of 6,428 indicator rules passed the full gate noneallThe 200-day rule on SPY: 7.2% a year vs 12.5% holding
Over 2010-10-19 to 2026-09-11 the rule returned 7.2% a year against 12.5% for buying and holding, after 10bps/side per switch.
A year, after costs Golden Cross / Death Cross: does it work?
Did not pass. ETFs rule: 11.8% a year after costs, against 13.2% for holding the same assets (corrected q 1.00).
A year, after costs Beat 166 of 200 random-entry copies noneallRSI(2) Mean Reversion: does it work?
Did not pass. US stocks rule: 1.4% a year after costs, against 18.0% for holding the same assets (corrected q 1.00).
A year, after costs Beat 57 of 200 random-entry copies noneallMACD Crossover: does it work?
Did not pass. ETFs rule: 6.6% a year after costs, against 9.6% for holding the same assets (corrected q 1.00).
A year, after costs Beat 180 of 200 random-entry copies noneallBollinger Band Bounce: does it work?
Did not pass. ETFs rule: 1.4% a year after costs, against 7.0% for holding the same assets (corrected q 1.00).
A year, after costs Beat 198 of 200 random-entry copies noneallDonchian / Turtle Breakout: does it work?
Did not pass. Crypto rule: 65.1% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 91 of 200 random-entry copies noneallVWAP Bounce: does it work?
Did not pass. Crypto rule: -49.8% a year after costs, against 84.7% for holding the same assets (corrected q 1.00).
A year, after costs Beat 7 of 200 random-entry copies noneallTurn-of-the-Month Effect: does it work?
Did not pass. ETFs rule: 2.8% a year after costs, against 13.2% for holding the same assets (corrected q 1.00).
A year, after costs Beat 166 of 200 random-entry copies noneallSell in May: does it work?
Did not pass. ETFs rule: 6.6% a year after costs, against 13.2% for holding the same assets (corrected q 1.00).
A year, after costs Beat 125 of 200 random-entry copies noneall12-1 Month Momentum: does it work?
Did not pass. US stocks rule: 26.8% a year after costs, against 18.0% for holding the same assets (corrected q 1.00).
A year, after costs Beat 199 of 200 random-entry copies noneallDual Momentum (GEM): does it work?
Did not pass. ETFs rule: 4.2% a year after costs, against 7.0% for holding the same assets (corrected q 1.00).
A year, after costs Beat 184 of 200 random-entry copies noneallFear & Greed Index: does it work?
Did not pass. Crypto rule: 8.3% a year after costs, against 48.1% for holding the same assets (corrected q 1.00).
A year, after costs Beat 118 of 200 random-entry copies noneallCrypto Weekend Effect: does it work?
Did not pass. Crypto rule: 5.9% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 126 of 200 random-entry copies noneallBuy the Dip: does it work?
Did not pass. US stocks rule: 13.1% a year after costs, against 14.2% for holding the same assets (corrected q 1.00).
A year, after costs Beat 192 of 200 random-entry copies noneallThe January Effect: does it work?
Did not pass. US stocks rule: 1.1% a year after costs, against 14.2% for holding the same assets (corrected q 1.00).
A year, after costs Beat 200 of 200 random-entry copies noneallLow-Beta Crypto: does it work?
Ahead of the benchmark before correction, not after. Crypto rule: 101.7% a year after costs, against 62.8% for holding the same assets (corrected q 0.57).
A year, after costs Beat 199 of 200 random-entry copies noneallCrypto Trend Following: does it work?
Did not pass. Crypto rule: 88.8% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 200 of 200 random-entry copies noneallFear & Greed Timing: does it work?
Ahead of the benchmark before correction, not after. Crypto rule: 103.3% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 200 of 200 random-entry copies noneallVIX Term Structure Timing: does it work?
Did not pass. Crypto rule: 83.0% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 200 of 200 random-entry copies noneallRSI Trading Strategy: does it work?
Did not pass. Crypto rule: 44.3% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 200 of 200 random-entry copies noneallMoving Average Crossover: does it work?
Did not pass. Crypto rule: 86.0% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 200 of 200 random-entry copies noneallIchimoku Cloud: does it work?
Did not pass. Crypto rule: 20.9% a year after costs, against 62.8% for holding the same assets (corrected q 1.00).
A year, after costs Beat 200 of 200 random-entry copies noneall
Method and limits
Historical backtests, not live trading, and none is a recommendation. See the testing standard and the evidence page.
General information only. It doesn't consider your objectives, finances or needs. Tickfloor holds no financial services licence and never places trades.