The pool that paid the most, and the question it answered
An alert said another pool paid 1807% better. Ranked on what a pool earns net, not one of them broke even.
The daemon scans every gauge it could hold and says when a better one exists. Its first alert said MSTRc pays 4.372% a day, 1807% better, and ranked the pool it already held seventh of ten.
The yield was real for about an hour. Across three scans in 90 minutes the capital MSTRc could absorb before hitting the cap went from $31,659 to $20,623 to $13,311, while the yield rose from 2.787% to 4.372%. Thinner depth pays more per dollar, until the dollar arrives. A finding is now provisional until its margin survives the worst depth seen in a week.
The history check behind that could never pass. It required the samples to span the whole seven day window, but samples are taken inside the window, so 43 of them measured 6.8184 days and it refused. The bar is now 90% coverage. Once it could answer, the pessimistic yield came out at 6.280% a day against 3.360% spot, more optimistic than the present. It can no longer beat spot.
Gross yield was also the wrong question. Holding a range costs divergence loss, which grows with the square of volatility over the width of the range. MSTRc pays 25 times what the held pool pays, and costs 14 times as much. Ranked on net, the held pool moved from seventh to second. The best ratio of emissions to divergence was 0.79. Not one pool cleared 1.00.
A one word bug undid that at once. Volatility had been re-keyed by address and was still looked up by symbol, so every lookup matched nothing, every pool came back unmeasurable, and the pick went silently back to whatever paid the most. It showed only because the gate says no measurable volatility rather than failing quietly. The alert sent at 09:43 recommending MSTRc was never retracted.
The divergence constant is now derived in code rather than asserted in a comment. It was 28.8. Nine days later it was 47.5.
Three days on, an alert said SNDKc was 1074% better. That was a percentage of a net close to zero. The gap is now stated in basis points a day, and a better pool has to win on every day of the week. SNDKc won five of seven.