Hub · Scoring
How the Trust Score works
Every ranking on this site comes out of one weighted grid. This page shows the whole machine: the five criteria, their weights, every sub-score for every firm, and the arithmetic that turns them into a total. Nothing here is written by hand — the numbers are recomputed from the same file that powers the comparator, so this page cannot disagree with the rankings.
Grid last revised 9 October 2026.
The five criteria
Weights sum to 100%. Each firm gets 0–100 per criterion; anything we cannot verify scores 0 rather than being quietly skipped.
Is the full rulebook published before you pay? Are drawdown and daily-loss limits stated in numbers you can compute yourself? Hidden consistency rules, minimum trading days or time limits cost points. A firm that reveals its rules after payment loses most of them.
How fast do you actually get paid, in what, and can you prove it? On-chain settlement scores highest because you can check the transaction yourself. An internal balance you have to trust scores lowest, regardless of the advertised speed.
How long has the firm operated, is it funded, does its infrastructure respond today (we verify domains by DNS and HTTP)? Since August 2026 this also covers solvency you can check yourself: a payout reserve readable on-chain scores above a balance sheet taken on trust. A firm that already went dark once starts near zero here.
Bots and copy trading allowed? News trading, weekend holding, scalping? Leverage caps and market breadth. The absence of restrictive rules is worth more than a slightly higher split.
Entry price against account size, the real profit split, refundable fees, and what a second attempt costs. An 80% split you reach beats a 92% split you never do.
The formula
Trust Score = Σ (sub-score × weight). One rule on top: a firm we have only desk-researched — public documents, no money of ours at risk — is capped, whatever the arithmetic says. Buying and trading an evaluation lifts the cap.
Worked example, live numbers — Propr.xyz
91×0.25 + 88×0.25 + 82×0.2 + 95×0.15 + 82×0.15
= 87.7 / 100
Desk-researched firms are capped at 85. “Traded” means we bought the evaluation with our own money.
Every sub-score, every firm
This table is generated from the grid file at build time. Recompute any row yourself: multiply each sub-score by its weight and add them up.
| Firm | Rules 25% | Payouts 25% | Survival 20% | Trading 15% | Cost 15% | Total |
|---|---|---|---|---|---|---|
| Propr.xyzTraded | 91 | 88 | 82 | 95 | 82 | 87.7 |
| HypernovaTraded | 82 | 96 | 80 | 80 | 88 | 85.7 |
| Vest MarketsTraded | 80 | 62 | 76 | 72 | 84 | 74.1 |
| Carrot FundingTraded | 78 | 82 | 55 | 66 | 82 | 73.2 |
| Solana FundedDocs only | 62 | 75 | 65 | 68 | 82 | 69.8 |
| BreakoutDocs only | 72 | 55 | 85 | 60 | 78 | 69.5 |
| HyperPNLTraded | 68 | 60 | 62 | 74 | 60 | 64.5 |
| GT FundedDocs only | 20 | 10 | 5 | 40 | 25 | 18.3 |
The rules the grid itself obeys
- Scores are never for sale. Sponsored slots buy visibility on clearly labelled placements — never a point on this grid. The advertise page says the same thing publicly.
- Affiliate links never touch scores. We hold affiliate deals with Propr and Hypernova, disclosed on every page that uses them. The grid was revised against a firm we earn from more than once.
- Every revision is dated and argued. The grid file carries a public changelog of what changed and why — including the 12 August 2026 revision triggered by a founder challenging us line by line. Numbers do not move quietly.
- Claims we cannot verify score zero. We do not average away uncertainty; we price it.
- Being nice to us buys nothing. Both firms at the top of this grid have publicly disputed parts of it. That is how it should work.
Who publishes their economics
An idea sharpened by Propr's own CEO: revenue minus payouts settles the profitability question — but only if revenue is published. This table tracks who lets you run that math. One framing rule: the ratio itself is neither good nor bad (a young firm selling many challenges mechanically shows a low one). The metric here is computability — can a reader verify the economics without trusting anyone?
Last partial re-read : 10 September 2026 — only the Carrot Funding row was re-read that day; every other row carries its own read date inside the cell.
| Firm | Revenue published | Payouts published | Pass rate published | On-chain reserve | Payout ratio computable |
|---|---|---|---|---|---|
| Propr.xyz | Yes$1.086M in Aug 2026 (+100% MoM, 36.0% margin) · $2.43M lifetime (26 Aug) | Yes$888.9K in Aug 2026 over 2,078 payouts · median 12 min · $1.11M lifetime (26 Aug) | Yes15.9% paid (3,515/22,157) · 13.6% free trial — read 20 September 2026 | Yes$1,025,662.96 USDC across three public wallets, read 4 October 2026 at 02:43 UTC; the payout contract itself is still funded payout by payout | Yes≈ $0.82 returned per $1 collected in Aug 2026 · ≈ $0.46 lifetime (26 Aug) |
| Hypernova | No | Yes$555.5K lifetime · avg 6.1s — read 26 August 2026 | Yes20.7% / 19.5% / 30.6% by risk tier — read 26 August 2026 | Yes$912,574.41 readable on Arbitrum — read 10 October 2026 | No |
| Carrot Funding | Yespublished on the Analytics page (read 10 Sep 2026) — amount not recorded here | Yespublished beside revenue, with a payout-hash table — 4 of 4 hashes verified by us on Arbitrum (733.22 · 384.38 · 399.24 · 799.41 USDC, 10 Sep 2026) | Yessuccess rate published on the same Analytics page (read 10 Sep 2026) | Yes~$15,016 readable in the Vault Custody Safe on HyperEVM (Felix/Morpho USDC vault), plus $98 on HyperCore — read 10 Sep 2026 | Yescomputable from the Analytics page (10 Sep 2026) |
| Solana Funded | No | No | No | No | No |
| Breakout | No | No | No | No | No |
| HyperPNL | No | No | No | Noreserve advertised as public on-chain, but no contract address published anywhere and our own $42 purchase was debited off-chain (traced 9 Sep 2026) — nothing to check | No |
| Vest Markets | No | No | No | Noaccounts are simulated (Prop Terms, 6 Oct 2026); payout = internal credit then USDC withdrawal from the exchange — no payout address, contract or reserve published (read 9 Oct 2026) | No |
Two firms in this table now publish enough for a reader to compute the payout ratio themselves: Propr, and Carrot Funding since we read its Analytics page on 10 September 2026. They are not equivalent, but the difference is smaller than we first wrote — Propr's payouts run through a verified distributor contract, while Carrot's four published hashes settle on-chain and its GitBook publishes a full address table with named roles (Vault Custody Contract, Vault Integrations, a 2-of-4 payout multisig, the payout executor EOA), which we had missed and read on 10 September 2026: the custody Safe on HyperEVM held about $15,016, with funds stated to be mid-migration from Arbitrum. Correction of ours, same day. One number on that Analytics page is not a reserve and must not be read as one: the $788,521 of "funding capital" is an internal risk metric — the capital still free before the firm stops taking on new funded traders. Any firm can join this column by publishing its revenue — that is the point.
Trust Score ≠ on-chain tiers
The on-chain tier scale is a separate, narrower instrument: it measures which of a firm's claims can be re-verified against a public chain — a readable payout reserve, payouts that settle by contract, trader state written on-chain. It does not claim a firm executes or hedges its trades on-chain, and it feeds only part of two criteria (payouts, survival). A firm can be honest and low-tier; a high tier is not an endorsement.
How we test prop firms (methodology) · Compare all firms · Sponsored slots: what they buy, what they never buy