What holding would have paid: 13 of 22 audits earn less than the price move

2026-09-03. Every figure quoted here is a field in a published audit JSON, and the listed audits are linked in place.

Lizard traded gold from 2003-05-12 to 2026-08-12 and finished 183.99 USD ahead. The same 0.01 lots, bought on the first of those days and held to the last, moved 4,070.66 USD. Both numbers are fields in the same audit file, measured on the same run, and the second one is the competitor almost no product page mentions. Twenty two of our twenty six default audits carry that comparison. Thirteen of them finished below it.

How the comparison is built

The block is deliberately narrow, because a tester report carries no independent price series. Everything it needs comes out of the trade list the rest of the audit already uses.

The span is the EA's own. First entry to last entry, not the test window. An EA that sat out the early years is not charged for them, which makes the benchmark easier, not harder.
The price comes from entry prices. The start is the mean of the first few entries, the end the mean of the last few, so a single wide spread print cannot set the level.
The contract value is calibrated from the run itself, as the median of price profit divided by price move times lots. On every gold audit the calibration lands on the same 100.0 USD per price unit per lot, which is the check that it measured the contract rather than noise.
The size is the EA's own median lot, 0.01 on every audit named here, so nobody is compared against a larger position than they actually ran.
It fails closed. Fewer than 50 usable trades, a median at or below zero, or a relative deviation above 5 percent and the field stays empty. Four default audits carry no hold number at all for that reason, Waka Waka, Pulse Engine, Quantum Emperor and Market Anomalies EA. No figure is invented to fill the gap.

The thirteen that came second

The gap is rarely small. Lizard ends 183.99 USD against 4,070.66 USD of price move, a shortfall of 3,886.67 USD on a market that rose 1,161.43 percent while it worked. Adaptive Gold Scalper, the cleanest mechanical run in the catalog, earns 381.48 USD against 4,025.88 USD. Gold Snap makes 1,922.00 USD against 3,923.31 USD, ThunderGold Scalper 314.65 USD against 3,698.74 USD, and Twister Pro Scalper 329.72 USD against 3,631.55 USD.

Two of the thirteen are worth naming for the size of the miss. Logan finished at -24,334.65 USD while its instrument rose 390.31 percent over the same span, a distance of 27,842.67 USD. Scalping Robot Pro finished at -53,234.68 USD against 4,008.85 USD, and it took 302,798 trades to get there. Activity and outcome are separate quantities, and this block is where the separation becomes a number.

The closest of the thirteen is CryonX at 3,456.01 USD against 3,765.92 USD, short by 309.91 USD after 4,127 trades.

The nine that came first

They exist, and the margins are real. Gold Trade Pro is the widest at 13,446.15 USD against 3,887.99 USD, a margin of 9,558.16 USD. Gold House follows with 11,054.32 USD against 4,066.07 USD, then Gold Atlas at 8,239.71 USD against 4,039.17 USD and Prop Firm Gold EA at 6,769.79 USD against 4,022.90 USD. Quantum Athena X clears its 3,071.73 USD by 1,578.07 USD, and Quantum Queen X clears 3,036.77 USD by 1,560.27 USD. One honest note on that count, Quantum Queen and Quantum Queen X are the same measured run published under two slugs, so the nine contain eight distinct measurements.

The two narrowest wins say the most. The Gold Reaper beats its hold by 399.59 USD across 1,742 trades. OilVector X, the only non gold audit in the comparison, beats crude oil by 1.52 USD, 35.23 against 33.71, over almost nine years. Nine years of machinery for a dollar and a half of edge is a result, and it is the kind of result that only appears once the benchmark is written next to the profit line.

Direction is not the answer either

The obvious objection is that these robots were short in a rising market. The block answers it, because it splits the same net result by side, and the two halves add back exactly. In Gold House the long trades made 6,726.59 USD and the short trades 4,327.73 USD, which is the 11,054.32 USD net to the cent. In Gold Trade Pro it is 8,004.37 USD long and 5,441.78 USD short. In Lizard it is 125.75 USD from 12,410 long trades and 58.24 USD from the remaining 9,114, near symmetry in a market that multiplied.

Where a side does dominate, it is not the side the objection predicts. Logan lost 27,293.59 USD on its long trades and made 2,958.94 USD on its short ones while gold rose. Scalping Robot Pro lost on both, 27,291.42 USD long and 25,943.26 USD short. OilVector X never sold at all, all 1,192 of its trades are long and its short result is exactly zero. The shortfall is not a direction problem, it is a cost and selection problem, and the other blocks of each audit take that apart.

What this comparison cannot say

It is not investment advice and it is not a fair fight in every direction, so the asymmetries belong in the open. The hold pays no financing and no commission. Holding a position for years would carry a nightly swap charge that this figure ignores, while the EA result is after the full commission retrofit. That asymmetry works in favour of the hold, and it is the reason the number is a floor to clear rather than a score to match.

The spans also differ, so the amounts are not comparable between audits. Quantum Titan traded from 2020-01-02 and is measured against 2,868.09 USD, while Gold House started on 2003-05-06 and faces 4,066.07 USD. Reading one number against the other tells you about the two windows, not about the two robots. And the risk profiles are not equal either. A held position carries its full drawdown uninterrupted, which is a different experience from a strategy that closes, and the drawdown blocks of each audit are where that belongs.

Run these checks yourself

Read the span before the amount. A hold figure over six years and one over twenty three years are different questions. The span is printed next to the number in every audit.
Add the two sides. Long result plus short result equals the net result in the costs block, to the cent, in every audit we publish. If a vendor report does not survive that addition, the numbers are on different levels.
Ask the vendor for the benchmark. Any backtest can produce the price move of its own instrument over its own span. A product page that shows profit without it has left out the cheapest competitor the buyer has.

Every figure above is public in the audit files linked in place, and the category pages list the verdict lamps for free. Your own tester report? The browser check is free.

The honest limits. The hold value is an approximation, and the audit says so in the field itself. It is derived from entry prices because a tester report contains no independent price series, the contract value is calibrated from the run's own paired trades rather than taken from a specification, and the whole block is skipped when that calibration is not stable. It carries no swap, no commission and no slippage, so it is generous to itself. It is not a recommendation to hold gold, and it says nothing about what any instrument will do next. What it does say is narrow and checkable. Over the exact days a given EA was in the market, at the exact size it traded, this is what the instrument moved, and this is what the EA kept.