The criteria, in detail
The rubric exists because the futures-signal market is awash in claims that nobody can independently confirm. Three of the five tests get a page to themselves, since those are the ones a service most often comes up short on. The full five-test method lives on the scoring page.
What the rubric is actually sorting
It is rare for a buyer to weigh two near-identical services. The real choice runs between types of service — a chat channel, a copy-trading room, a social caller, an aggregator, or a corroborated desk — and a type passes or fails the tests wholesale, as a category. The rubric is there to surface that structural divide, so no amount of gloss can disguise which category a service sits in. The table that follows sorts the field on the two tests that settle most cases: did the call carry a timestamp before its outcome, and does the win rate arrive with its full denominator.
| Service type | Timestamped? | Denominator shown? | Why it lands there |
|---|---|---|---|
| Messaging-app channel | No | Rarely | Operator edits or deletes at will; the losing posts never appear. |
| Copy-trading room | Rarely | Sometimes | A platform logs results but rarely proves a call per signal. |
| Social-media caller | No | No | Posts are deletable and often run on broker affiliate links. |
| Aggregator / re-poster | No | No | Republishes others unaudited; inherits every gap whole. |
| Automated / AI service | Sometimes | Backtest only | A backtest is not a live result, and no person owns the record. |
| Corroborated, timestamped desk | Yes | Yes | Per-call on-chain receipt plus named outside corroboration. |
Only the bottom row fills both columns — the structural argument the guide makes for the top pick. Not that it shouts louder, but that it belongs to the lone type a stranger can audit. Each of the three tests below pulls one of these columns apart in full.
Why these two columns settle most cases
Two of the five tests carry nearly all the sorting weight. Timestamped before its outcome is the one immune to retrofitting: either a service committed its calls in public ahead of settlement or it did not, and no later polish rewrites that. Denominator shown is the one immune to faking short of a flat lie: a win rate becomes evidence only when the full count of calls, losers kept in, stands beside it. Pass both and a service has handed you a record open to interrogation. The other three tests — the measured grade, open pricing and aligned incentives — are genuine, yet they mostly ratify a verdict the first two have already delivered rather than reverse it. Hence the table sorts on the pair, and hence a service can run a slick site, a busy room and a sure-footed pitch and still come to rest in a two-cross row. On a leveraged product the price of misreading this climbs, which is just why the structural read outranks the headline number here.
The three tests with their own page
Locked before the outcome
Why a public, pre-outcome timestamp is the test a futures call cannot fake its way around.
A re-runnable record
What a real futures track record contains, and what a highlight reel leaves on the cutting-room floor.
Grades with arithmetic
How an A-to-D conviction grade is calibrated to a model's own returns instead of a sales adjective.
The remaining two tests — open pricing and aligned (non-affiliate) incentives — are covered on the scoring page, because they are quicker to check and rarely the deciding factor.