The strategy tester was handed 23 years of gold. The report header of this run reads M1 from 2003.05.05 to 2026.08.17, which is the deepest window our data world offers. Quantum Athena X opened its first position on 2018-01-17 and its last on 2026-08-11. Everything before that first date, 14 years and 8 months of tick data, produced no trade at all. That is the first thing anyone should know about the equity curve on the sales page, because the curve does not start where the test starts.
One deterministic tester run, dissected deal by deal together with the tester journal. 3,976 trades from 7,952 deals, pairing confidence exact, journal coverage 1.0, chronology violations 0, duplicate deal ids 0, randomizer prints 0. The reconstructed final balance matches the report to the cent. Our catalog protocol runs a fixed 0.01 base lot on a 100,000 USD frame at 1:5000, which are measurement settings that keep the tester from censoring the trade sequence rather than a recommendation. On top of the tester result we retrofit commission at 3.50 USD per lot per side, so our net of +4,649.80 USD sits below the header figure of +4,967.88 USD.
A backtest that ignores 14 of its 23 years is unusual enough to be worth a cross check, so we ran one. Three other audits in our catalog come from the same vendor and were offered the same window. Quantum Queen and Quantum Queen X both place their first trade on 2018-01-17, the same day as Athena. Quantum Titan starts on 2020-01-02. For a control we took an EA from a different vendor on the same symbol and the same M1 window, Gold House, and it trades from 2003-05-06. So the data world does deliver the early years and other software does use them. Whatever holds these products back is in the products.
What follows from that is simple and it is not an accusation. The tested history of Athena is 3,128 days long, not 23 years, and those days cover one long bull market in gold. Any claim that this strategy was proven across two decades of gold regimes is not supported by the run we measured. Our Quantum Queen X review looks at the same vendor from a different angle.
Our structure forensics rate this red. The EA added 1,574 entries into an open losing basket against only 44 that followed the move, it held up to 12 positions at once, and 43.8% of all entries stacked onto something already open. There were 11 moments with long and short open at the same time. Position size does not escalate, the median adverse add is 1.0 times the basket average, so this is averaging without a martingale multiplier.
The result is the smooth curve that averaging always produces. The deepest end of day drawdown across the whole run is 0.30 USD on a six figure balance, and our reconstructed intraday equity never drops more than 171.66 USD, on 2026-01-29. Losses do not show up in the balance line because losing baskets stay open until the market comes back. The tester journal also records 354 failed entry attempts, which is a caution in our data quality dimension. In live trading those attempts are the moments a broker says no.
The EA labels its own trades, so we can split them. Five setups appear, four of them long only and one short only. Their results add up exactly to the run total, 1,455.34 plus 937.00 plus 1,238.49 plus 737.57 plus 281.40 equals 4,649.80 USD, and their trade counts add up to 3,976. Every one of the five wins between 74% and 78% of its trades. The short setup contributes 244 trades and 281.40 USD, which is 6% of the profit.
Over the traded span our engine measures the instrument itself rising from 1,338.65 to 4,410.38 in the entry price series, a move of 229.46%. Holding 0.01 lots of gold from the first trade to the last would have earned 3,071.73 USD. Athena earned 4,649.80 USD. The machine is ahead by 1,578.07 USD after 3,976 trades and 810.33 USD of commission and swap.
That comparison is approximated from entry prices, because a tester report carries no independent price series, and buying and holding gold means sitting through every gold correction of those years while Athena kept its balance line flat. The point is not that holding is better. The point is that 93.9% of the profit came from the long side with 93.9% long trades, in a market that tripled. Our regime dimension flags that as long beta suspicion, and the honest reading is that a large part of this record is the direction of gold rather than the timing of the entries.
Divide the net by the trades and each one is worth 1.17 USD. That is also the exact cost shock at which the entire run breaks even. Our sensitivity table charges a flat extra cost per trade and watches the result. At 0.50 USD more per trade the run keeps 2,661.80 USD and the profit factor falls from 4.33 to 2.275. At 1.00 USD more it keeps 673.80 USD, the profit factor is 1.216 and the win rate collapses from 76% to 38.0%. At 2.00 USD the run loses 3,302.20 USD.
Costs already take 14.8% of gross profit in our world, and the larger half of that is financing rather than commission. Swap costs 492.25 USD against 318.08 USD of commission, spread over 738 overnight trades. A gold spread wider than our measurement world is the first thing to check here, because half a dollar more per trade already removes 43% of the net.
We replay the observed deal sequence as a monthly withdrawal to start balance. Across 104 months the run pays out in 104 of them. No dry month in eight and a half years. The median paid month is 23.05 USD, the best is 276.98 USD, and the total withdrawn is 4,648.65 USD. On the 100,000 USD measurement frame that is 0.52% per year.
Read those two facts together rather than apart. A perfect record of 104 positive months is what an averaging system looks like when nothing forces it to realise a loss, and the annual yield on the measurement frame is half a percent. The vendor recommends a 1,000 USD deposit instead, and on that basis none of our 7,952 simulated start points dies, because at 0.01 lots the baskets stay small. Both statements describe the same trade sequence at different sizing, and the sizing is where the risk lives.
Quantum Athena X was published on 2026-07-23 and updated twice since, last on 2026-07-31. Of the tested history, 3,109 days and 3,918 of the 3,976 trades lie before the release. The only slice that is guaranteed out of sample, after the last vendor update, is 11 days long with 47 trades and +77.73 USD. That window is far too short to confirm or refute anything, and we say so rather than quoting its profit factor of 11.46 as if it meant something.
The deal level census, the setup split and the cost shock table need full deal reconstruction, which is what we sell. Any catalog audit is $19, and EAs with several audited presets or versions come as one bundle. Your own tester report? The browser check is free.