Every lot, not the interesting ones.

Two of these are worth reading in the opposite direction to how they sound. The worn-over-issued ratio looks far larger in raw form — one sale reads at 10× — but that gap is player tier, not provenance: starters get the worn items and bench players get the issued ones. Anyone quoting much above 2× is measuring who wore it.

The name effect is the cleanest of these, because the product is held constant. Thirty NBA head coaches, the same custom Air Max 90, the same charity, the same sale, all sold within a week of each other. Median $1,024. Steve Kerr's pair made $21,760 — twenty-one times the median and 5.7 times the next coach. Nothing about the object differs. The only variable is whose feet were in it.

One refinement the partial capture couldn't see: the series is thirty-three lots, not thirty. Three are broadcast-crew pairs rather than coaches, and one of them — the NBC crew — outsold every coach except Kerr. Counted that way Kerr is 5.3 times the next lot rather than 5.7 times the next coach. Both are true of different populations, which is why the population gets named.

The photo-match number is the one I'd trust most, because it's the only finding here that has been tested three times against independent populations. Heritage basketball runs 6.78× in 2024, 3.62× in 2025 and 5.61× in 2026 — separate years, separate consignments, same direction every time. I published the 5.61× first, from 2026 alone. Three years later it turns out to sit inside the range rather than being the range, which is roughly the best outcome a single-year number can hope for.

And the garment hierarchy is relative, not absolute. Shorts sat at the bottom of one Finals sale and took the top four slots in an All-Star weekly, because the jerseys had been held back for the premier sale. What a category is worth depends on what else is in the room.

Every finding above about estimates comes from ordinary consigned sales, where the house is guessing at what someone will pay for an object it was handed. One sale in the dataset isn't that. It was single-owner, direct from the athlete, sold once, for a foundation — and it inverts the result completely.

61 of its 81 lots closed at or above the low estimate. Median 1.50×. In the two consigned sales, 63.5% closed below. Same house, same premium, same season, opposite direction.

The top lot is the cleanest illustration of what's being paid for. A warm-up-worn pair carrying a $3,000–5,000 estimate closed at $76,800 — twenty times the low. Its only real distinguishing feature was that it was worn on the first night of the run everybody was following.

And the premium doesn't travel. A game-worn, dual-signed, inscribed pair from that sale closed at $70,400 in April. The same shoe resold at a different house ten weeks later for $48,800. Down 31%, with nothing about the object changed. What changed is that you could no longer say you bought it from him.

Two access sales in the dataset now, and they're consistent with each other. That's an observation with a sample of two, which is why it's described here rather than stated as a law.

Two lots in the same sale, described word for word identically — same player, same product, same series. One closed at $512. The other at $2,560.

Five times apart, for objects nobody could tell apart. That's not a pricing insight, it's the noise floor, and no forecasting method beats it. Some part of every miss on this site is simply which minute a lot happened to close in. It gets stated in every report for the same reason the sample sizes do.