What are the four states?

Skilled, not separable from luck, thin sample, and too concentrated (in the API: skilled, luck, insufficient, flagged). Every wallet read on every Convexly surface lands in exactly one of them, derived by one deterministic state machine from the same inputs, so two surfaces can never disagree about the same wallet.

What does it take to earn the skilled state?

Three gates at once: at least 30 resolved positions, single-event concentration below 0.6, and a BCa bootstrap 95% interval whose lower bound sits strictly above zero. The read is retrospective and in-sample, and it is uncorrected for multiple comparisons, which is why cohort-level results apply a further correction (lesson 9).

Why does concentration outrank a great point estimate?

The states are evaluated top-down with first match winning, and the concentration check runs before the interval check. A record where one event carries 60% or more of the net result cannot be read as skill no matter how good its numbers look, because the figure cannot be separated from that one outcome. A flattering point estimate never outranks a structural weakness.

The four-state verdict

The answer first

When Convexly reads a wallet, the output is one of exactly four states: skilled, not separable from luck, thin sample, or too concentrated. One deterministic rule set computes the state from the same inputs on every surface: the analyzer, the board, the badge, the API, and the enterprise deliverable. Same wallet, same fields, same verdict, everywhere.

The intuition, with one worked example

Imagine a wallet showing a +9 point realized edge whose single biggest event carries 72% of its net result. The point estimate looks great. The verdict is too concentrated, not skilled, because the concentration check runs before the edge is even consulted. The record's headline figure cannot be separated from that one outcome, so no figure computed from it gets read as skill. This ordering is the whole point of the design: the checks that can invalidate a number always run before the number is allowed to impress you.

The actual method

The four states, quoted from the published API definitions:

Skilled (skilled)
"n_resolved >= 30, the BCa bootstrap 95% lower bound on realized edge is above zero, and single-event concentration is below 0.6, all on the resolved record. Retrospective and in-sample, uncorrected for multiple comparisons."
Not separable from luck (luck)
"Not separable from chance on the resolved record: the BCa bootstrap 95% interval includes zero, so this record does not distinguish the realized edge from chance either way."
Thin sample (insufficient)
"Too few resolved positions (n_resolved < 30) or no computable edge interval, so the resolved record cannot be read either way."
Too concentrated (flagged)
"A single event accounts for 60% or more of the net realized result, so the record is too concentrated to read a single figure as skill. More positions do not clear a concentration flag."

The gates evaluate top-down, first match wins: (0) an outcome-basis check first, so an edge scored on anything other than resolution outcomes caps the state at thin sample; (1) the concentration flag; (2) the 30-position sample floor and interval availability; (3) the interval-versus-zero read; and only then (4) skilled, when the BCa lower bound is strictly above zero. The interval used is always the BCa bootstrap interval, never a looser approximation.

Where you see this on the site

The analyzer prints the state as its headline verdict. The wallet board annotates every row with the same read, the embeddable skill badge renders it with the caveat baked in, and the API returns it as the state field with the definition attached. The full check is described end to end at /learn/skill-audit.

What this does NOT mean

Not separable from luck does not mean bad wallet; it means the interval includes zero, so the record does not settle the question either way. Too concentrated is a flag on readability, not a verdict on the wallet or a suggestion of wrongdoing. Thin sample is not an insult; it is the refusal to read 12 positions as if they were 300. And skilled is a property of the resolved past record: retrospective, in-sample, and uncorrected for multiple comparisons, which is exactly why the cohort-level work in lesson 9 exists. None of the four states is advice, and none is a forecast.

Next lesson: win rate vs edge

Convexly publishes new methodology research roughly every 6-8 weeks plus the /learn series on a rolling cadence. Get the next paper in your inbox when it ships: