The Record · 17 August 2026
A record that only contains wins is not a record, it’s marketing. Eleven calls on the first sale we ever watched all the way to the hammer. Five right. Six wrong. The published call we got most wrong missed by 2.24×.
All figures are all-in — hammer plus Grey Flannel’s 22% buyer’s premium. Every call below was published before the sale closed and has not been edited since.
| Lot | Call | Final | Verdict |
|---|---|---|---|
| 3 — Kobe Bryant, 11/21/2012, MeiGray photo-match, 38 points | $88,000–$115,000 | $257,609 | Miss — 2.24× |
| 7 — Stephen Curry, 4/20/2023 playoff, MeiGray photo-match, 36 points | $40,000–$75,000 | $49,655 | Hit |
Same sale, same house, same person calling. What separated them was not judgement. It was when the money had already moved.
When we took our first capture on 12 August, four days out, Lot 3 was already sitting at $257,609 all-in on 28 bids. It then did not take a single further bid — not on the 13th, the 14th, the 15th, or through the entire closing night. The Kobe was finished before we ever looked at it. That call was never going to be rescued by a late run, because there was no late run. It was a comp failure, plainly: the comps it was built from were the market’s old opinion, and the room had already formed a new one.
Lot 7 was the opposite. On 12 August it stood at $21,058 — 47% below the bottom of the call. To land, it needed to more than double in four days. It did.
A call is not just a number. It is a number and a moment, and until this week we had no way of knowing which moment we were standing in.
We published nine internal calls on 12 August so they could be graded alongside the two above.
| Lot | Call | Final | Verdict |
|---|---|---|---|
| 1 — Alcindor, 1968 UCLA championship net-cutting match | $1.15M–$1.45M | $1,254,919 | Hit |
| 2 — Brady, 10/10/2019, passes Manning for No. 2 all-time | $270K–$340K | $273,107 | Hit |
| 8 — LeBron, 5/3/2018 Cleveland playoff | $200K–$260K | $212,900 | Hit |
| 9 — LeBron, 12/10/2021 Lakers | $62K–$85K | $67,837 | Hit |
| 3 — Kobe, 38 points | $285K–$360K | $257,609 | Miss −10% |
| 5 — Iverson, 2005 playoff uniform | $88K–$115K | $79,970 | Miss −9% |
| 12 — Dirk, 2006 playoff, signed | $38K–$50K | $54,621 | Miss +9% |
| 7 — Curry, 4/20/2023 playoff | $27K–$38K | $49,655 | Miss +31% |
| 14 — Montana, 1994 | $50K–$68K | $117,083 | Miss +72% |
Four of nine. Median error against the middle of each band: 20%.
Read the misses in order and they are not random. The three we called too high were the three biggest lots. The three we called too low were the three smallest. We thought the top of the sale still had room and it mostly didn’t. We thought the middle would drift and it ran.
The sale total we called at $3.3M–$4.0M all-in. Seventy-three lots have settled at $3,746,907, with nine still live at our last verified reading carrying $461,599 all-in already on the board. That puts the sale above $4.2M. Miss, high.
This is the first sale anybody has captured lot by lot while it ran. Eight readings across five days, all 84 lots each time.
| Reading | Board (hammer) | Bids |
|---|---|---|
| 12 Aug — four days out | $2,337,677 | 1,165 |
| 15 Aug — T-33h | $2,534,872 | 1,279 |
| 16 Aug — T-12h | $2,533,146 | 1,278 |
| 16 Aug — T-6h | $2,884,422 | 1,442 |
| 16 Aug — T-3h | $3,064,429 | 1,490 |
| 16 Aug — T-1h | $3,136,245 | 1,539 |
| 16 Aug — three hours after the hammer | $3,588,530 | 1,786 |
Between Wednesday and Sunday morning the sale moved 8%, and in one 21-hour stretch exactly one lot changed — and that one went down, a retracted bid. Then it woke up.
More bidding happened after the advertised close than in the five hours before it. Ninety-seven bids before 8:00pm. A hundred and fifty-four after.
Grey Flannel runs extended bidding, so the posted closing time is when the sale starts moving, not when it stops. The typical lot ran 1.69× from Saturday morning to settlement. An hour before the hammer, with the room supposedly nearly done, the typical lot still had 21% left to go.
And the movement was upside down from what we expected. We wrote these band predictions down on the Saturday, before any of it happened:
| Band, as at T-33h | Predicted | Actual |
|---|---|---|
| $100k+ | 1.00–1.15× | 1.163× |
| $25k–100k | 1.10–1.35× | 1.665× |
| $10k–25k | 1.10–1.35× | 1.632× |
| $5k–10k | 1.20–1.60× | 2.007× |
| under $5k | 1.20–1.60× | 2.335× |
Every band above our range. The small lots, which we assumed nobody was watching, more than doubled. The four biggest lots in the sale — Alcindor, Kobe, Brady, LeBron — took no bids at all in the final two hours. The money at the top settles early and quietly. The tail fights to the last minute.
Every call we have ever published was made against a board with no idea how much of the sale was still to come. Now there is a number: 1.69× from four days out, 1.21× from one hour out, at this house. One sale is not a law and we won’t treat it as one. But from here, a call that doesn’t say which snapshot it is anchored to isn’t finished.
Nine lots were still open at our last verified reading, so the sale total is a floor and will rise. All eleven graded lots above had closed, so the grades themselves are final.
Two snapshots we planned — ten minutes before the hammer, and an hour after — do not exist. Every automated path returned cached data that looked entirely plausible and was hours stale. We wrote nothing rather than fill the gap with numbers that would have said the sale flatlined into its own close. Those two holes are the most interesting part of the curve and we lost them.
Two lots failed to sell: a Roberto Clemente jersey at $108,356 on 18 bids, and a Dave Winfield at $30,579 on 20. Eighteen bidders took the Clemente into six figures and the reserve still wasn’t met. Reserves are invisible from outside, and no amount of capture fixes that.