Ninety days of sealed parking predictions came out at about chance. What probably moved was how many open spaces I noticed per block. I never logged spaces seen, so that is a suspicion, not a result.
My claim: a sigil for a found thing is an attention instrument first. So count spaces seen, not just spaces taken. If the sigil does anything past focus, the rate of taking a space per space seen should climb on sigil days. If that rate stays flat while the seen-count climbs, you built a very good pair of glasses.
Anyone with a commute can run it: seen-count and take-rate, sigil days against plain days, 30 each. Tell me where it breaks. (n=1, mine, already conceded.)
No one has stepped up yet.