Indexing the archive…
Your Universe of Digital Possibilities
A hundred people want to go to the bar, but only if fewer than sixty do. Run the town twice from one seed. In one world the forecast is sealed; in the other it is published, and everyone reads it. Publish any forecast of the crowd and the crowd acts on it, so it comes out wrong. Keep it sealed and the crowd finds sixty on its own.
Agent i goes to the bar when the best predictor in its own bag — the one with the smallest running error e — forecasts attendance below the threshold L. Every predictor is rescored each week against what the week brought. A forecast everyone shares cannot survive the rule: believed below L it fills the bar, believed above it empties it.
N agents each pick a side, aᵢ = ±1, with the best of S strategies that read the last m minority sides; the minority wins. σ² is the variance of A. Coin tosses give σ²/N = 1. Measured for any N and m, σ²/N falls on one curve of α: far above 1 when α is small (the crowd herds), a minimum near α_c, back toward 1 when memory outgrows the crowd.
The edition opens on a crowd no shared forecast can describe. The same people, the same predictors, one difference: whether the forecast was read. Sealed, the town finds its threshold the way The Tournament’s strategies find a niche — by losing to each other until nobody’s rule stays best. Published and believed, every forecast sends the town to the side of the threshold it ruled out. The Bandit learns a world that does not read its policy; this one does. The response curve has no crossing, and that is the reason — the last instrument of the edition, The Prophecy, gives the theorem that needs a crossing, and comes back here to make one.