The proof tests, in detail
These tests exist because so much of the signal market trades on claims that a buyer has no way to check. Three of the five get their own page, since those are the ones providers trip over most. The complete five-test method lives on the method page.
What these tests are really sorting
A buyer rarely chooses between two near-identical providers. In practice you are picking between kinds of provider — a group chat, a mirror-trading room, a social caller, an aggregator, or a sealed desk — and each kind passes or misses the tests as a category. The job of these tests is to make that structural difference visible, so a glossy presentation cannot hide which category a provider belongs to. The grid below sorts the field on the two tests that decide most cases: was the call sealed before its outcome, and is the win rate shown with its full call count.
| Provider kind | Sealed? | Call count shown? | Why it lands here |
|---|---|---|---|
| Group-chat channel | No | Rarely | Operator edits or deletes freely; losers never surface. |
| Mirror-trading room | Rarely | Sometimes | Platform logs outcomes but rarely per-call proof. |
| Social-feed caller | No | No | Posts removable; income often via referral links. |
| Re-poster / aggregator | No | No | Relays others unchecked; carries every gap forward. |
| Automated / AI feed | Sometimes | Backtest only | A backtest is not a live result; nobody owns the record. |
| Sealed, reviewed desk | Yes | Yes | Per-call on-chain receipt plus a named outside review. |
The bottom row is the only category with both columns satisfied, which is the structural case the guide makes for the number-one pick — not that it is louder, but that it belongs to the one kind a stranger can re-check. The three tests below each take one column apart in full.
Why these two columns settle most of it
Among the five tests, two carry nearly all the sorting weight. Sealed before its outcome is the one you cannot bolt on after the fact: either a provider committed its calls in public before they resolved or it did not, and no amount of later tidying rewrites that. Call count shown is the one you cannot fake short of lying outright: a win rate only becomes evidence when the full number of calls, losses and all, stands beside it. Pass both and the provider has put a record in front of you that you can pull apart. The other three tests — the calculated grade, open pricing and aligned revenue — matter too, but they mostly ratify a conclusion the first two have already delivered. Hence the grid sorts on those two, and hence a provider can run a sleek site, a lively community and a self-assured pitch and still come to rest in a row carrying two crosses.
The three tests with their own page
Sealed before the result
Why a public seal at send-time is the test no provider can fake its way around.
Grades that are calculated
How an A-to-D conviction grade is measured against a model's own returns instead of a mood.
The remaining two tests — open pricing and aligned (non-referral) revenue — are covered on the method page, because they are quicker to check and rarely the deciding factor.