The catalog reads like twenty six independent products. The vendor line inside the audit files says something narrower. Across the 26 default audits there are 13 distinct seller names, and six of those names carry 19 of the 26 files. The two largest carry five each. That is not a judgement about anybody, it is a count of a field we copy out of the product listing, and it changes what a shortlist of three robots can promise.
Every audit ends with a block called catalog_meta that carries five fields, the slug, the product name, the vendor, the MQL5 listing id and the tested version. The build copies them straight out of our catalog file, which records what the product page states. We do not resolve companies, we do not merge spellings and we do not guess. A seller who lists one product under a personal name and the next under a company would appear here as two sellers, and the audit would never know.
The listing id carries the other half of the arithmetic. Thirty three audit files hold a vendor line and only 26 distinct ids. The gap is the variant files. Quantum Titan holds three files under id 113770, Smart Gold Hunter four under 170050, Scalping Robot Pro two under 149463 and The Gold Reaper two under 111357. Those are presets and versions of one listing rather than separate products, so a count of products has to run over ids and not over files.
The other seven names appear once each. Nineteen of the 26 audits therefore sit behind six names, which is the reason to open this field before a shortlist gets built rather than after.
Each audit closes with six verdict lamps, and the one named prop fit is the harshest of them in this catalog. Over the 26 default audits it stands at red 15 times, at caution 5 times and at ok 6 times. Those six green lamps are the interesting part. They belong to Gold Snap and to all five files under the Puscasu name. Not one of the other eleven sellers holds one.
Before that means anything, the lamp needs its mechanism. A dimension starts at ok and is raised to the strongest flag that fired inside it, so a green lamp is the absence of findings rather than a pass. In the prop dimension five checks can fire, no death free position size, survival only at a quarter of the tested size or smaller, the per position loss limit breached at one third size, the consistency rule breached in at least one year, and open positions against a profile that bans them overnight or over the weekend. Each set of five runs once per prop profile, and every catalog entry carries the same three profiles. A green prop lamp therefore means five checks stayed quiet on three profiles each.
Now the honest subtraction. Quantum Queen and Quantum Queen X are one measurement in two files. Both report 3,357 paired trades, a tester head of 4,865.60 USD, an audited net of 4,597.04 USD and a first entry at 2018-01-17 03:00:00 over the same window, and our catalog file records the Queen run as deal identical to Queen X version 4.4 after the product was delisted. Quantum Emperor is a second subtraction of a different kind. Its audit still runs on a vendor report from another broker world, GBPUSD with an .ls suffix on H1 from 2020.01.01 to 2026.08.11, 559 trades and 7.59 USD of net, so it is not comparable to a run on our own gold history.
What survives the subtraction is still a pattern. Four distinct measurements on our own history carry a green prop lamp, and three of the four sit under one seller name. The fourth is Gold Snap.
The clearest thing the vendor line buys you is a comparison that nobody advertises. TwisterPro Scalper and ThunderGold Scalper are two listings, ids 166740 and 188424, two versions, 3.20 and 1.3, and two different recommended deposits on their product pages, 250 USD and 100 USD. Both audits run XAUUSD on M15 from 2003.05.05 to 2026.08.17, because our test setup fixes the symbol and the history for everything we test.
What the world does not fix is what the robot does inside it, and there the two land almost on top of each other. 986 trades against 984. An audited net of 329.72 USD against 314.65 USD. Six verdict lamps that read identically, red on data quality, ok on structure, ok on costs, caution on concentration, caution on regime and red on prop fit. The published one line summaries run parallel finding by finding. TwisterPro refuses to trade the first 2.7 years, breaches the consistency rule in 9 years and makes 80 percent of its profit in 41 days. ThunderGold refuses the first 3.0 years, breaches in 11 years and makes 80 percent of its profit in 42 days.
They are not the same file, and the differences say so. The first trade lands on 2006-01-06 in one and on 2006-05-04 in the other. The tester head reports a maximum balance drawdown of 105.70 USD against 26.13 USD, and a profit factor of 1.57 against 1.73. The trade comments carry different prefixes, TwisterPro in one report and ThunderGold in the other, which is the EA naming itself. Two builds then, close in behaviour and far apart in the detail. Whether they share code is a question a tester report cannot answer, and we do not answer it here.
The other family of five reads the opposite way. Under the Eriksson name the prop lamp stands at red for Range Breakout EA with Range Filters, Prop Firm Gold EA and Gold Atlas, and at caution for Market Anomalies EA and Pulse Engine. It reaches ok in none of the five. The concentration lamp behaves the same way, red in four of the five and caution in the fifth.
The same subtraction applies here as well. Market Anomalies EA and Pulse Engine are measured on vendor reports from another broker world, USDJPY and EURUSD with an .ls suffix over 2020.01.01 to 2026.08.11, and those two audits are not published for exactly that reason. Of the three measured on our own gold history, the prop lamp is red three times out of three.
Every audit in the catalog prints the seller, the listing id and the tested version at the bottom of the page, and the six lamps at the top. Both ends are free to read.