Strategy profiles

Long-form writeups on strategies that survived our gates, plus autopsies of notable failures. Generated by the lab, human-approved before publish.

39 published
profile2026-05-20

SMA crossover with ADX filter passes all 6 gates

Golden cross in low-ADX regimes posts 2.2% per trade, 100% win rate, and a deflated Sharpe of 11.55 across 10 coins.

profile2026-05-20

SMA 21/50 Cross With ADX Filter Graduates With 12.7 Sharpe

Golden cross entry with ADX regime filter passes all gates, but zero forward trades yet and narrow win-rate ceiling demand caution.

profile2026-05-20

SMA Golden Cross in Calm Markets Passes All Six Gates

SMA 21/50 cross with ADX-below-50 filter achieved 2.4% per trade and perfect validation win rate across ten coins.

profile2026-05-20

SMA Golden Cross with ADX Filter Passes All Six Gates

Simple moving average cross protected by ADX regime filter hits 2.5% per trade with 96.6% win rate and 0.2% fluke probability.

methodology_note2026-05-20

The bug that silently hid 1,560 of our backtests

We thought 5 new strategy families produced 0 graduates. One character in the wrong place was actually producing zero TRADES. Here's what we found, fixed, and what it means.

profile2026-05-20

SMA Cross in Sideways Markets: Perfect on Paper, Untested Live

Golden cross filtered for low-ADX environments passed all gates with 100% winrate, but zero live trades raise coverage questions.

profile2026-05-20

Williams %R + MFI mean-reversion killed by gate audit

Mean-reversion entry on oversold Williams %R and MFI failed strict gates with -1.90 Sharpe, killed after 19 forward trades.

profile2026-05-20

RSI + SMA-200 + ADX Filter: Passed, But Zero Live Trades

Oversold entry above 200-SMA in weak-trend regimes. Strong backtest, 2.76 deflated Sharpe, but no forward trades yet.

profile2026-05-20

RSI + SMA 200 + ADX Range Finder: 0.85% Per Trade, 87% Wins

Mean-reversion play in ranging markets passed 6 gates with 4.76 deflated Sharpe, but zero forward trades yet.

profile2026-05-20

RSI + SMA Long in Low-Volatility Windows

Passed all gates with 1.07% per trade and 87.5% win rate, but zero forward trades yet. High fluke risk at 47%.

profile2026-05-20

RSI(14) Oversold + EMA(200) Filter Passes All Six Gates

Mean-reversion play buying RSI dips below 30 on 15-minute bars, exits at 1% profit. 90.5% win rate, 0.673% per trade, 24.22 deflated Sharpe.

profile2026-05-20

RSI-14 Oversold + EMA-200 Filter Passes All Six Gates

Mean-reversion strategy buying RSI dips above the 200 EMA cleared gates with 1.28% per trade and 90% win rate across ten coins.

profile2026-05-20

Session filters killed every entry we tested

We crossed 4 proven entry patterns with 5 time-of-day windows across 283 candidates. Zero graduated. Here's what that tells us.

profile2026-05-20

SMA 21/50 Golden Cross With ADX Filter Graduates

Classic moving average crossover with trend filter passes all gates on 10 coins, but zero forward trades raise deployment questions.

profile2026-05-20

SMA cascade killed at audit: insufficient validation trades

Four-moving-average short strategy died at audit with only 11 validation trades, despite perfect 5-trade forward run.

profile2026-05-20

SMA Cross With ADX Regime Filter Passes All Gates

Golden cross in low-ADX regimes clears validation with 2.2% per trade and 11.55 deflated Sharpe, but zero forward trades yet.

profile2026-05-20

RSI/SMA combo with ADX filter passes 6 gates, zero live trades

Mean-revert strategy using RSI 14 below 35, SMA 200, and ADX under 30 cleared validation with 0.638% per trade but hasn't triggered in forward testing yet.

profile2026-05-20

RSI 14 Below 35 With SMA 200 Filter, Low ADX Only

Long-only RSI dip-buyer with trend and regime filters passed all 6 gates with 88.9% win rate, but zero forward trades yet.

profile2026-05-20

RSI+SMA+ADX Chop Filter: 0.44% Per Trade, 88.9% Win Rate

Long-only RSI mean reversion above 200-SMA, filtered to low-ADX chop. Passed 6 gates with 16.83 deflated Sharpe.

profile2026-05-20

RSI+SMA+ADX Range Filter: 0.67% Per Trade, 89% Win Rate

Long-only RSI dip buying above the 200 SMA, restricted to low-ADX ranges. Passed all six gates with 8.47 deflated Sharpe.

profile2026-05-20

RSI-14 / SMA-200 Low-ADX Mean Reversion Passes 6 Gates

Long when RSI under 35 and above 200-day MA, exit at profit target. Passed with 5.35 deflated Sharpe, but zero live trades yet.

profile2026-05-20

RSI 14 Under 35 + SMA 200 Filter, Low ADX Regime Pass

RSI dip buy above the 200-day, filtered for low ADX ranges. Passed 6 gates but zero forward trades yet.

profile2026-05-20

RSI-7 mean reversion killed by its own validation data

ADX-filtered RSI strategy soft-killed after backtest validation showed negative Sharpe and insufficient edge despite brief forward test gains.

profile2026-05-20

RSI 7 below 25 with ADX filter: rescued by regime fit

Mean-revert strategy failed validation gates but showed profitable regime behavior. Now in observation, not for general use.

profile2026-05-20

RSI-7 <25 with ADX filter: restored, still watching

Countertrend entry passed forward test but failed strict backtest gates. Restored for regime-specific observation.

profile2026-05-20

RSI-Bollinger mean reversion cleared gates with 90% win rate

Oversold RSI plus lower Bollinger band entry, filtered by low ADX, passed all six gates but traded zero times live.

profile2026-05-20

Phase K: Volume filters killed every oversold and breakout signal

We tested 224 volume-confirmed entry variants. Zero graduated. Requiring above-average volume stripped edge from known patterns.

profile2026-05-20

BB Lower + RSI Oversold in Ranging Markets

Mean reversion on 15-minute bars passed all gates with 1.4% per trade and 90% win rate, but zero forward trades yet.

profile2026-05-20

Ranging Low-Vol BB+RSI Oversold on 15-Minute Bars

Buys Bollinger lower band touches with RSI oversold, filters out strong trends. 1.7% per trade, 90% win rate, 10 coins.

profile2026-05-20

ROC-12 Mean Reversion in Low-ADX Regimes

Buy dips when 12-period rate of change drops below -2%, but only when ADX signals low trend strength. Passed 6 gates.

profile2026-05-20

ROC-12 Mean Reversion Killed by Insufficient Validation Data

Strategy buying 12-period ROC dips in low-ADX regimes soft-killed at audit with only 151 validation trades, negative Sharpe.

profile2026-05-20

ROC-12 mean reversion killed by negative validation Sharpe

Rate-of-change dip buying failed strict gates with -0.03 Sharpe and negative P&L despite 82% win rate.

profile2026-05-20

ROC-12 reversal with ADX filter killed at audit

Mean-reversion play on 12-period pullbacks failed strict gates with negative Sharpe and negligible edge despite 73% win rate.

profile2026-05-20

BB + RSI Entry with ADX Filter Passes All Gates

Bollinger Band lower touch with oversold RSI, trading only in low-ADX regimes, graduated with 27.88 deflated Sharpe.

profile2026-05-20

Bollinger RSI oversold above EMA200: 0.625% per trade, 85% wins

Mean reversion strategy buying RSI oversold dips inside lower Bollinger Band passed all 6 gates with 7.83 deflated Sharpe.

profile2026-05-20

BB-RSI Oversold Above EMA200: 85% Win Rate, 7.3 Deflated Sharpe

Bollinger Band / RSI mean reversion with regime filter passed all gates. 0.94% per trade, 85% win rate, 0.1% fluke risk.

profile2026-05-20

CCI + MFI Double Dip: Restored But Still On Watch

Oversold combo passed forward test but failed strict backtest gates. Restored to observation while we fix the regime framework.

profile2026-05-20

MFI+RSI double oversold: killed by negative validation Sharpe

Dual oscillator mean reversion looked promising early but negative Sharpe and insufficient validation trades killed it at audit.

null_result2026-05-20

Phase I: We tested 220 volatility-envelope breakouts. Zero survived.

Bollinger Band breakouts with ADX filters, trend confluence, and Donchian confluence. All failed our 6-gate stack on 3 years of 10-coin Binance data.

Research Lab Findings

Every strategy the research engine has tested. Ordered by validation P&L on held-out data. We publish the losers too, not just the winners. That's the whole point.

Test your own strategy →
35,946
Candidates tested
1607
Passed all gates
4.47%
Pass rate
5/31/2026
Last test

Research pipeline progress

A2 expanded
154
A2_expanded
A3 higher tf
63
A3_higher_tf
A single
442
Phase A: single indicator sweep
B combo
46
Phase B: winning entry combinations
D explore
27,460
21 queued
Phase D: random exploration
E exit taxonomy
3,436
Phase E: exit taxonomy (entry × exit grid)
F user stack
53
F_user_stack
G cross
1,108
537 passed
G_cross
H regime target
742
19 passed
H_regime_target
I vol breakout
220
I_vol_breakout
J session filter
270
J_session_filter
K volume confirm
224
K_volume_confirm
L multi tf
167
6 passed
L_multi_tf
M refine
1,524
1045 passed
M_refine
regime targeted
37
regime_targeted

How the research lab works

  1. Candidate strategies get queued automatically (entries across RSI, MFI, StochRSI, Williams %R, CCI, ROC, BB, Donchian).
  2. Each candidate runs against 3 years of Binance data on 10 coins at 5m timeframe with real fees.
  3. Walk-forward validation splits data 60% train / 40% test. In-sample numbers that don't survive out-of-sample are flagged.
  4. Strategies pass through one of three approval lanes: (a) standard — positive walk-forward P&L, 15+ trades, 30%+ WR, Deflated Sharpe fluke probability < 50%; (b) near-breakeven with strong fundamentals — P&L > -0.02% with 100+ trades, 60%+ WR, fluke < 50%; (c) high-conviction selective — 8+ trades with 75%+ WR and clear positive P&L. We tightened these lanes on 2026-04-30 after a gate audit; 549 prior approvals were retroactively retired as noise.
  5. Whatever survives graduates to forward testing. Whatever fails is published here with the fail reason.

See the full engine architecture →Is the regime classifier even working? →

About what you see below

Losing strategies show every detail. We publish the specific indicators, thresholds, and parameters that didn't work because the honest map of what failed is more useful than most research results. Passing strategies show the indicator family and a coarse performance bucket. The exact thresholds, timeframes, and exit params for winners stay private while we forward-test them ourselves. Edges die when they get crowded. When one graduates to real-capital deployment we publish the full post-mortem with performance data that's already baked in.

What passed (1 patterns)

Exact params redacted until deployed. Patterns grouped by indicator family.

sma_21

15m · long

500 variants
PnL: +0.5%+/t
WR: 65%+
Val trades: 30-49
Below: strategies that failed.The detailed numbers are here because losing configurations have no IP value and the honest map of what doesn't work is the most useful thing we can publish. Sort or filter to browse.
EntryExitVal P&LVal WRVal tradesSharpeFluke %ConfidenceStatus
No findings match these filters.

See a combination that interests you? Test your own variant with real fees:

Test a strategy →

Research findings are from our ongoing automated backtesting. Walk-forward validation splits data 60%/40%. “Passed” means all 5 statistical gates cleared. Numbers shown are on held-out (40%) validation data. Not financial advice.