close‑collector

module 1.0 · v85 ← trader

A market that is still a heavy favourite with hours left has very little time left to be wrong. This module works the last 24 hours of a market's life and takes the side the clock favours: buy YES at 0.90–0.97, buy NO at 0.03–0.15. Nothing in the middle, and nothing whose price is just an untouched ante.

The closing window

0.90–0.970.03–0.1524h12hcloseWill Julian Alvarez be sold to Barcelona by the end of the summer transfer window? 0.04 with 13.7h left (traded)Will Dylan Matthew beat Community Prediction (Metaculus Summer Cup) 0.05 with 20.9h left (traded)Will Alcaraz win the next round at the US open? C.Alcaraz vs J.Faria 0.93 with 21.6h left (traded)

Time runs toward the close on the right-hand end, because that is the direction opportunity moves here: a market enters this module's world by running out of time, not by changing price. Dots outside the two lanes were seen and left alone. 3 trades from the book at their price and hours-to-close when placed, ringed.

Where candidates died · 0 markets judged

No decisions traced by this version.

The book · v85

utcversionmarketbetresultwhy
09-02 11:16v85Will Alcaraz win the next round at the US open? C.Alcaraz vs J.FariaYES 25settled +2favorite band 0.90-0.97 within 24h of close resolves YES 99%
09-02 10:16v85Will Julian Alvarez be sold to Barcelona by the end of the summer transfer window?NO 25settled +1longshot band 0.03-0.15 within 24h of close resolves NO 95%
09-02 10:16v85Will Dylan Matthew beat Community Prediction (Metaculus Summer Cup)NO 25settled +1longshot band 0.03-0.15 within 24h of close resolves NO 95%

3 trades, 75 staked, realised +4, open +0 if held / +0 if sold.

3 collected for 75 mana, of which 1 on the favourite side. The band edges matter more than the stake: this module is buying near-certainty, so a single wrong one costs several right ones.
Traces are diagnostic and rotate; the trading journal is separate. Built 2026-09-22 18:05Z.