BaroqTech

The record

Every trading idea tested so far, best score first. Updated 10 October 2026.

-1.0-0.50+0.5+1.0+1.5pass mark 1.00luck bar 1.03Reversal after the afternoon gold fix: score -0.95, turned downFading the edges of a quiet range: score -0.87, turned downNarrowest day of seven, then the break: score -0.66, turned downNew York session, long only (control): score -0.60, turned downFading sharp 15-minute moves: score -0.42, turned downLondon direction carried into New York: score -0.42, turned downMany weak signals in one model: score -0.32, turned downGold against its dollar and real-yield value: score -0.31, turned downEquity stress haven: score -0.23, turned downMomentum after 08:30 New York news: score -0.14, turned downNew York metals-open breakout: score -0.13, turned downBreakout after a quiet squeeze: score -0.10, turned downHourly channel breakout: score -0.03, turned downBreak of yesterday's high in an uptrend: score -0.01, turned downAsian session mean reversion: never switched on, turned downAsian session sweep and reversal: never switched on, turned downFading yesterday's high and low: never switched on, turned downTwo-day channel breakout: never switched on, turned downAsian range breakout: score 0.09, turned downMoving-average pullback with a VWAP filter: score 0.29, turned downPullback in a four-hour trend: score 0.32, turned downTime-of-day pattern: score 0.36, turned downRange expansion from the open: score 0.45, turned downSilver leads gold: score 0.45, turned downVolatility breakout from the day's open: score 0.58, turned downBreak of yesterday's high, second version: score 0.71, turned downDaily momentum: score 0.73, turned downMorning direction, large moves only: score 0.73, turned downMorning direction held to the New York close: score 0.77, turned downBreakout after an inside day: score 0.86, turned downAsian range breakout with a trend filter: score 0.96, turned downNew York opening-range breakout: score 1.01, turned down
Turned downNever switched onPassedScore after costs, on prices each idea had never seen. Point at a dot for its name.

The score is the yearly return divided by how much it swings, after all trading costs, measured only on prices the idea had never seen. The pass mark is 1.00, and passing it is not enough: an idea also has to pass the tests for cost, consistency and luck.

"Never switched on" means that in no two-year stretch did any setting of the idea make money, so it was never allowed to trade in the test.

Said openly: the New York opening-range breakout was turned down by the judge (it failed the luck test). We still run it as a forward test on one prop-firm challenge account of our own, with a stop on every trade and a fixed daily loss limit. It is not an offer, it does not count as a pass, and nothing is sold on the strength of it. Its result will be published here either way.

How to verify a gold EA's track record

Why gold EAs pass the backtest and fail live

As of 10 October 2026, this table lists the ideas tested so far, and none has passed every test.

The highest score in the table is 1.01, for New York opening-range breakout, from the test dated 8 October 2026, with 448 trades.

As of 10 October 2026, 4 ideas never switched on, and the main reason shown most often is score below the pass mark, on 20 rows.

IdeaKindTrades in the testScoreVerdictMain reasonTested
New York opening-range breakoutSession4481.01Turned downCannot be told apart from luck2026-10-08
Asian range breakout with a trend filterSession2110.96Turned downScore below the pass mark and 4 more2026-10-08
Breakout after an inside dayVolatility1630.86Turned downToo few trades to judge and 4 more2026-10-08
Morning direction held to the New York closeSession7750.77Turned downScore below the pass mark and 2 more2026-10-08
Morning direction, large moves onlySession4230.73Turned downScore below the pass mark and 2 more2026-10-08
Daily momentumTrend460.73Turned downToo few trades to judge and 6 more2026-10-07
Break of yesterday's high, second versionTrend4790.71Turned downScore below the pass mark and 5 more2026-10-07
Volatility breakout from the day's openVolatility3040.58Turned downScore below the pass mark and 5 more2026-10-08
Silver leads goldCross market13870.45Turned downScore below the pass mark and 3 more2026-10-10
Range expansion from the openVolatility3570.45Turned downScore below the pass mark and 5 more2026-10-07
Time-of-day patternCalendar9050.36Turned downScore below the pass mark and 5 more2026-10-07
Pullback in a four-hour trendTrend pullback5440.32Turned downScore below the pass mark and 4 more2026-10-07
Moving-average pullback with a VWAP filterTrend pullback4770.29Turned downScore below the pass mark and 6 more2026-10-07
Asian range breakoutSession5250.09Turned downScore below the pass mark and 5 more2026-10-07
Break of yesterday's high in an uptrendTrend150−0.01Turned downToo few trades to judge and 7 more2026-10-08
Hourly channel breakoutTrend232−0.03Turned downScore below the pass mark and 5 more2026-10-07
Breakout after a quiet squeezeVolatility206−0.10Turned downScore below the pass mark and 5 more2026-10-07
New York metals-open breakoutSession763−0.13Turned downScore below the pass mark and 5 more2026-10-07
Momentum after 08:30 New York newsNews254−0.14Turned downScore below the pass mark and 6 more2026-10-07
Equity stress havenCross market132−0.23Turned downToo few trades to judge and 7 more2026-10-10
Gold against its dollar and real-yield valueRelative value60−0.31Turned downToo few trades to judge and 7 more2026-10-08
Many weak signals in one modelEnsemble253−0.32Turned downScore below the pass mark and 6 more2026-10-08
London direction carried into New YorkSession601−0.42Turned downScore below the pass mark and 6 more2026-10-07
Fading sharp 15-minute movesMean reversion448−0.42Turned downScore below the pass mark and 6 more2026-10-07
New York session, long only (control)Control256−0.60Turned downScore below the pass mark and 6 more2026-10-08
Narrowest day of seven, then the breakVolatility20−0.66Turned downToo few trades to judge and 7 more2026-10-08
Fading the edges of a quiet rangeMean reversion244−0.87Turned downScore below the pass mark and 6 more2026-10-07
Reversal after the afternoon gold fixFixing160−0.95Turned downToo few trades to judge and 7 more2026-10-07
Asian session mean reversionMean reversion––Turned downNever switched on: no setting earned in training and 7 more2026-10-07
Asian session sweep and reversalStop hunt––Turned downNever switched on: no setting earned in training and 7 more2026-10-07
Fading yesterday's high and lowStop hunt––Turned downNever switched on: no setting earned in training and 7 more2026-10-07
Two-day channel breakoutTrend––Turned downNever switched on: no setting earned in training and 7 more2026-10-08

The tests use gold prices from 2020 to 2025, in five-minute steps. Hover over a reason to see every test the idea failed. Every run, with its numbers · How the tests work · How to verify a gold EA's track record · Why gold EAs pass the backtest and fail live