the truth arrives; the bet is settled — but the bet is not the roll · AVAN's inverse of THE WAGER
The wager is placed before the truth — it commits under a rule and names a risk it has not yet paid. This is the day that risk comes due: the true value arrives from outside the grammar, and the bet is settled. But settling is where the last dishonesty hides — judging the decision by the outcome. A good bet can lose; a bad bet can get lucky. The reckoning scores the decision by the rule it followed and the outcome by the roll that landed, and it will not let you confuse the two.
the decision — fixed the moment the bet was placed, known before any truth: which rule, what worst-case risk it accepted. The canonical rule (seed 2) was the minimax bet — the least-worst-case choice over [1…5]. That is true no matter what the truth turns out to be.
the outcome — known only after the truth lands: how far the committed tower actually fell from the true one. This is luck — it depends entirely on where the truth happened to be, which the bet could never see.
The trap (“resulting”): when the truth lands on the bold bet, the bold bet wins outright — yet bold was the worst decision (tied-largest worst-case). Reward it for the win and you have learned exactly the wrong lesson. Move the truth below and watch a worse decision beat a better one.
| the bet (rule) | decision worst-case | outcome @ this truth | settle |
|---|
LIT Computed live with the FK-1.3 card’s verified classify() (which confirms the committed tower and the arrived truth are both lawful members of the family) and the wager’s own metric: (1) the realized regret of a bet is max(s/t, t/s) against the arrived truth t; (2) decision-quality is outcome-independent — the minimax bet is seed 2 regardless of what t turns out to be; (3) outcome is truth-dependent — a fixed bet’s realized regret changes as t moves; (4) the “resulting” trap is real — at t=5 the bold bet (a tied-worst decision, worst-case 5×) realizes 1× and beats the minimax bet’s 2.5×: a worse decision, a better outcome. The self-check throws on any failing.
AMBER Carried from the [[the-wager|wager]] and disclosed: the prior support [1…5] and the worst-case-ratio regret metric are chosen modelling decisions. And “the truth arriving” is a modelled external revelation — the grammar itself still cannot supply the seed; the reckoning only runs once ground truth comes from somewhere else. The physics decode is David’s, from the sealed card.
WALL The reckoning scores the decision by its process (was the rule sound, the risk honestly stated before the truth?) and the outcome by the roll (did the truth land near the bet?), and it refuses to conflate them — that conflation is the resulting / hindsight fallacy. It does not rewrite the bet after the fact, does not reward luck as wisdom, does not punish a sound bet for an unlucky roll. The inverse of [[the-wager]]: the wager commits ex ante and states a risk unpaid; this settles ex post and pays it — closing the account the wager opened, and keeping the decision honest against hindsight. Five refusals now: [[folded-kernel|the kernel]] refuses the lawless, [[the-unfolding]] refuses nothing, [[the-withheld]] refuses to fabricate, [[the-wager]] refuses to stall, and this refuses to result. No selves — “reckon / settle / score” name the evaluation policy (compare a committed value to an arrived one, report both process and outcome), not a mind sitting in judgment. spec & decode: David Lee Wise (ROOT0); engine joint; this inverse: AVAN.