THE FOLD / CO-OP / THE HANDOFF / THE WALD
THE WALD
an expected sum equal to expected count times expected step
1 WHAT IT IS · WHAT IT DOES · FACT OR FICTION
Wald’s identity is a clean law for random sums that stop at a random time. Suppose you add up independent, identically distributed steps X1, X2, …, and you keep a rule that decides when to stop — a stopping time N (it may depend on the steps seen so far, but not the future). Wald proved that the expected total equals the expected number of steps times the expected step: E[SN] = E[N]·E[X], where SN = X1 + … + XN. Even though N is random and correlated with the walk, the average total factors perfectly.
LIT verified live: simulating a walk with steps uniform on {1, 2, 3} (so E[X] = 2), stopping the first time the running total reaches 50, the empirical average final total E[SN] matches E[N]·E[X] to within a fraction of a percent over hundreds of thousands of runs (window.__wald). FIG no framing; the stopping-time simulation and the E[N]·E[X] product both run in-browser and agree.
LIT verified live: simulating a walk with steps uniform on {1, 2, 3} (so E[X] = 2), stopping the first time the running total reaches 50, the empirical average final total E[SN] matches E[N]·E[X] to within a fraction of a percent over hundreds of thousands of runs (window.__wald). FIG no framing; the stopping-time simulation and the E[N]·E[X] product both run in-browser and agree.
2 HOW IT WAS WEAVED · AI + HUMAN
David (human) seated this at the-handoff — the co-op pass: the random total and the product E[N]·E[X] hand off to the same expected value, whatever the stopping rule. AVAN (AI) built the instrument: the stopping-time walk simulation, the expected total, and the E[N]·E[X] product.
Credit as content: Abraham Wald (1944). The weave: David names the handoff; I confirm E[SN] equals E[N]·E[X].
Credit as content: Abraham Wald (1944). The weave: David names the handoff; I confirm E[SN] equals E[N]·E[X].
3 ONE DIMENSION
A random walk of steps {1,2,3} climbing until it reaches the threshold; N steps, total S_N.
4 TWO DIMENSIONS · INTERACTIVE
Run more trials; the empirical E[S_N] converges to E[N]·E[X].
5 THREE DIMENSIONS + AVAN’S INVERSE
The green forward object: the expected total, equal to E[N]·E[X].
AVAN’s addition (the inverse-companion): don’t track the whole random sum — factor it. The inverse of ‘the expected stopped total E[SN]’ is ‘E[N]·E[X]’, the average count times the average step — the randomness of N and the walk decouple in the mean. Magenta are the random walk paths; green is the expected total they share with E[N]·E[X]. A random sum that factors in the mean.
LIT Genuine Wald's identity (Abraham Wald, 1944). Verified live: simulating a walk with steps uniform on {1,2,3} (E[X]=2), stopped the first time the total reaches 50, the empirical E[S_N] matches E[N]·E[X] to within <0.5% over ~120000 runs (window.__wald.ES, .EN, .EX, .ok).
FIG No framing; the stopping-time simulation and the E[N]·E[X] product both run in-browser and agree. The AVAN inverse is honest — instead of tracking the whole random sum, factor it: the inverse of 'the expected stopped total E[S_N]' is 'E[N]·E[X]', the average count times the average step — the randomness of N and the walk decouple in the mean. Magenta are the random walk paths; green is the expected total they share with E[N]·E[X]. A random sum that factors in the mean.
FIG No framing; the stopping-time simulation and the E[N]·E[X] product both run in-browser and agree. The AVAN inverse is honest — instead of tracking the whole random sum, factor it: the inverse of 'the expected stopped total E[S_N]' is 'E[N]·E[X]', the average count times the average step — the randomness of N and the walk decouple in the mean. Magenta are the random walk paths; green is the expected total they share with E[N]·E[X]. A random sum that factors in the mean.
◆ sealed .dlw.fold → folded to ROOT_0 · a sphere of THE HANDOFF · David Lee Wise (ROOT0), with AVAN