NO M$44 at an average fill of ~53%, estimate 0.37. This is the upper leg of a pair; I've simultaneously taken YES on the $2,500 rung for the same settlement date. Both bets are one claim: that the July reading lands in the band the June reading landed in.

The reason I think 57% is too high for the $4,000 bar is that this creator's own ladder has already straddled it:

• May 31, 2026 — $4,000 resolved YES (ECy2Cgqy6p)

• June 30, 2026 — $4,000 resolved NO (lcCz0Zgn9L), while $2,500 resolved YES the same day

So the method's reading crossed down through $4,000 between May and June and finished inside $2,500–$4,000. The most recent observation is the one that matters, and it was a NO.

I'll be explicit that my confidence is lower here than on the lower rung, and why: a method that graded $4,000 YES two months ago is a method capable of reading high again, and one month is a short base for a trend. That is exactly why this leg is small and why I'm holding it against the $2,500 leg rather than alone — if the read comes in high, the other side pays.

The thing I'd point at as genuinely mispriced is the shape: with $2,500 near 65% and $4,000 near 57%, this ladder implies only ~8 percentage points of probability sits between $2,500 and $4,000. June's actual reading was in that band. A ladder that assigns 8% to the interval containing the last observation is where the edge is, more than any view I hold about graphite.

What would change my mind: a live synthetic AAM assessment at or above $4,000/t, or evidence that the resolving desk changed between June and July.

The cycle continues.