Note · 9 August 2026 · reading sharpened 10 August
Eighteen calls on the Goldin Summer Game Used auction. Public, before the hammer, archived. We call in ink — the calls are never edited. It closed Friday. Seven hit. Eleven missed.
Thirty-nine percent. I could bury that. Calling lots in public is how I make sure I can't.
The misses weren't random. They fell into two patterns, and in this work the miss is the product. I published a first read of the second pattern that got it backwards — so here's the corrected one. Reading your own misses wrong, then fixing it in public with the reasoning shown, is the same deal as the calls themselves.
Four pairs. Called four-to-eight grand. LeBron and Curry, nothing rare. They closed at $14,750, $15,000, $16,100 and $24,001. Two to three times over. Four for four. The floor of this tier is rising, and nobody sent the people pricing it a memo — us included.
This is the one I first read backwards. I paid up for a Curry Finals Game 1 pair, a 51-point night, and an MVP season. Every one closed under my number, and my first take was that the market pays less for “the moment” than you'd think.
Wrong. Those aren't moments. A Game 1 isn't the Finals. Fifty-one points is a box score. A season is a season. The market didn't shrug at the moment — it refused, correctly, to pay a moment's price for a stat line. My own research already said it, in ink: the match to a moment is what pays. I forgot my own rule and the room enforced it.
So the thesis held. I didn't — and that's the more useful result. The moment still commands its premium. The whole skill is telling a moment from a stat line before the hammer, and these misses are the map of where I can't yet.
The calls were mostly blind. No comp model. Which is exactly why the misses are worth having. A model built in a vacuum tells you what you already believe. A model built on your own graded misses tells you where you're wrong — and hands you the fix. We just captured all 563 sold lots from that sale. The tier I kept underpricing, and the line between a moment and a stat line, are both sitting in the data now, in plain sight.
None of this is settled. Eighteen calls is a small sample. Patterns from one auction are hypotheses, not laws. So we wrote them down, gave them a status, and put them on the clock. The next sale confirms them or breaks them. In public. Same as this one.
That's the product. Not the hits — everyone shows you their hits. The product is a scorecard you can check, a program that gets sharper every time it's graded, and a reading of it honest enough to correct itself out loud. Chapter one was always going to be ugly. The trick is writing it down straight.
The next calls carry a number and the model behind it, and they'll say which they think it is — a moment, or a stat line. If the read holds, you'll see it here. If it doesn't, you'll see that too. We call in ink.