About
I collect sneakers. I got curious about what the game-worn stuff actually sells for. Then I got annoyed that nobody keeps score.
An auction house publishes an estimate. The lot sells. Everyone moves on. Nobody ever circles back to ask whether the estimate was any good.
That sounds like a small thing. It isn't. Across every lot in the dataset where a house published a number to be measured against — 219 of them — 63% closed under the low estimate. One Sotheby's soccer sale ran at 74%, median miss 56%.
I once published a sharper version of that claim, about vintage jerseys specifically, and had to withdraw it. It was built on a truncated set. The retraction is here, because a record that only holds the wins isn't a record.
These aren't malicious numbers. They're just unchecked. There's no feedback loop, so nothing corrects. Nobody has ever graded an auction house, which means the estimates have never had to be right.
So I started grading them. And me.
I pick lots, research the comps, and publish a price range before the sale closes. Publicly, timestamped, with the reasoning attached. When the hammer drops I post what happened — including the part where I was wrong, which so far is most of it.
Two classes of call, and I keep them apart because mixing them would flatter me:
Blind — made before a single bid exists. Sometimes weeks before the sale even opens, off provenance and comparable sales alone. Hard.
Live-bid — made after bidding has started, when there's a standing bid on screen. Easier. Graded separately for exactly that reason.
Nothing gets edited after a sale closes. If I revise a call before the hammer, the original stays on the page next to the revision. The untampered record is the whole asset — I'd rather be publicly wrong than quietly corrected.
Some of it is obvious once you see it. Some of it surprised me.
The market pays for the moment and discounts the name. A championship-clinching Alcindor jersey — signed, photo-matched to three games — sold below its low estimate. A Super Bowl-winning touchdown ball tripled its. The famous name isn't the thing. The moment is.
A photo match is worth more than almost anything else — but only to a moment worth matching. The same pair of Jordan's Finals shoes, split by half: the matched half sold for $2.2 million, the unmatched half for $102,000. Then two other Jordan pairs told me the opposite, and the reason was which game they were matched to.
Worn beats issued, about seven to four. Same team, same night, same jersey — one played in, one not. Across 58 role-player lots the ratio is 1.75. I first published this as a clean 2.00 off a 26-lot subset; the full population brought it down. The correction is here.
Ryan. I'm in California, I've got a shelf full of Travis Scotts I actually wear and three pairs of Currys I never will, and I've been buying at auction long enough to have lost a few I still think about.
I'm not an appraiser and I don't work for a house. I'm a collector who kept a spreadsheet and then couldn't stop.
The videos are on YouTube. Same job, more talking.
Everything gets graded. Every sale, win or thin, whether it makes me look good or not. Balls and strikes.
If you want the mechanics — how prices are converted, what counts as a hit, what a no-skill baseline scores — that's over here.