The Same 54 Bernstein Configurations on Gold: What Survived the Move From Nasdaq

The short answer

Across 2,621 eligible gold sessions between 2016-01-04 and 2026-09-25, 8 of the 54 configurations finished with a positive simulated net result per contract, and 38 beat the same-session buy-the-open, sell-the-close reference (−$27,640 over 2,621 sessions). At an assumed $20 round-trip charge instead of $5, 1 remains positive.

Set against the NQ results for the same calendar years, the two markets agree on the sign of 24 of 54 shared configurations; 6 are positive on both, out of 34 that were positive on NQ. The rank correlation between the gold and NQ net results is 0.32. A configuration that worked on Nasdaq futures tells you little about whether it works on gold.

Data, session and conversion

Prices are Dukascopy’s spot gold (XAU/USD) 1-minute bid bars, the same vendor and resolution as the NQ study’s CFD series, which we had validated against exchange futures. Spot gold is not the COMEX contract: there are no roll gaps and no contract months, and its price runs a few dollars from the front-month future. Results are expressed per GC contract at $100 per point, with one tick of slippage on every market order and a $5 round-trip commission, exactly as for NQ.

The day session is 09:30 to 16:00 ET, the window the course and the NQ study use for index futures, so that every rule keeps its meaning. It is not COMEX’s own pit session, which opens and closes earlier. Only full sessions on the NYSE calendar with all 390 minutes present are used; the rest are excluded, not patched.

Every year is covered in full.

The configurations are the NQ study’s, unchanged in logic. 12 parameters denominated in Nasdaq points were converted under one rule, fixed before the run: rules stated in ticks keep their tick count in gold ticks; rules stated in points keep their dollar value at gold’s point value; rules stated in percent, bars or dollars are unchanged. The specific settings remain course material and are not printed.

All configurations, gold, sorted by net result

Each ID links to the configuration’s row in the NQ article and each label to the index entry there that describes the method and its source. Net figures are per contract after slippage and $5 commission. The last column is the same configuration on NQ over the same years, for the sign, not the size: an NQ point is worth $20 and a gold point $100, and gold moves far less in points.

Year by year for the leaders

The five largest net results plus the base breakout and its full-target variant, against the buy-open, sell-close reference. Picking the top five after the fact is selection; the table is here to show whether the totals come from one year or many.

Weekday rows show net result and, in brackets, the number of positions. With a few hundred positions per configuration split five ways, a weekday that stands out is not a rule.

Which configurations to test next on gold

This is a research shortlist, not a trading recommendation: the configurations that pass five gates applied to all 54 alike, with no hand-picking. Positive net result; still positive at a $20 round-trip charge; profit factor of at least 1.05; positive in at least half of the years with trades; and at least 100 positions. 1 passes.

The other 7 configurations with a positive net result, and the gate each fails. Most fall at the cost gate: they trade often enough that an extra $15 a round trip turns them negative.

The course’s 15-minute breakout family, the reason most readers ask about gold, is not on the list: 2 of its 13 variants finished positive and the best of them made +$4,491 per contract over the whole period. On this evidence a trader testing that method on gold is testing a hypothesis the data already argue against. The shortlist above is where a forward test, on a demo account or with the TradingView tester on the actual GC contract, has something to confirm or refute. A forward test that starts now is the first untouched evaluation these configurations will have had on gold.

What this study cannot conclude

• Spot gold is a proxy. The contract’s roll gaps, exchange session and slightly different price path are not simulated; the day-session choice is the study’s, not the exchange’s.
• The parameter conversion is one defensible rule among several. Converting by percent of price, or by gold’s own volatility, would give different settings and different results; none of them is the course’s own recommendation for gold.
• 54 configurations, one instrument, one cost model. The tables do not correct for multiple testing; the leaders are the best of many draws.
• Where coverage is incomplete, annual figures are not comparable across years and the totals are not comparable with the NQ totals in size.
• No configuration is recommended. Hypothetical results on one period do not establish an edge.