Where do pre-release sneakers come from? 298,876 pairs sold before launch day

The Air Jordan 11 "Gamma Blue" released on December 13, 2025. By then, GOAT and StockX had already recorded 18,487 sales of it: 13,454 on GOAT and 5,033 on StockX, the first one 58 days before release day.

Jordan 11 Retro Gamma Blue (2025)Photo: StockXJordan 4 Retro Black Cat (2025)Photo: StockXJordan 3 Retro OG Rare AirPhoto: StockX

So where do pairs come from before a shoe comes out? The honest answer is that nobody can tell from sales records, and no one has measured the mix. What we can do is lay out the channels that are documented, and show what our record of 298,876 pre-release sales says about when they happen, what they cost, and which shoes they cluster on.

Early sales are routine, and they pile up near launch

Across the 2,590 releases we could follow (sneakers released January 2025 to February 2026), GOAT and StockX recorded 298,876 sales before release day. The typical release had 8. A few big Jordans had thousands.

Most of it happens late. 28% of pre-release sales came in the final three days, and 12% on the last day alone, when the marketplaces recorded 36,517 sales a day across these releases. Another 28% came 15 to 60 days out.

The "Rare Air" 3 sold almost three pre-release pairs for every pair sold in its first week.

The rules say the seller already has the pair

Both marketplaces let a shoe sell before its release date, but neither lets a seller wait for release to ship. StockX gives sellers two business days to ship a pre-release sale and says sellers need the item before they list it, with penalties for late shipping. GOAT cancels and charges sellers who miss its three-day window.

Taken at face value, that means tens of thousands of pairs of the biggest Jordans existed outside the official launch weeks before release. Our record cannot check whether every one of those sales actually shipped on time.

The documented channels

Reporting and public records point to several sources. None comes with a count.

• Legitimate early drops. Some pairs release earlier somewhere else: a collaborator's own site (a Travis Scott Jordan went up on his site five days before SNKRS), a partner store (Union's own Air Jordan 4 drop), or another country (atmos in Japan). StockX has even sold a shoe a week before launch itself.
• Friends and family, and samples. Friends-and-family pairs have hit resale before public release, and a former Nike manager pleaded guilty in 2014 to selling samples.
• Retailer mistakes. Stores occasionally ship an order months early.
• Theft. This has the longest paper trail: 12 people were indicted in 2026 over a cargo-theft scheme at Nike's Memphis operations, unreleased pairs were taken from trains in 2025, and StockX paused trading on some Jordans after a 2022 burglary.
• Fakes. Counterfeits of a Jordan 12 surfaced months before its release.
• Backdooring. Store staff selling pairs out the back is real enough that Nike has written rules against it, but public cases are rare. We looked at whether it pays in Is backdooring worth it?

Theft is the best documented, but that may only mean it is the channel that ends up in court.

Early pairs cost more, and the gap closes as launch nears

For the 810 releases with at least 10 sales before release and 10 in week one, the median pre-release price was 11% above the same shoe's week-one price (95% interval 9% to 13%: the range the true figure most likely sits in). Pre-release pairs sold higher in 630 of the 810 (78%).

The earlier the sale, the bigger the gap: 21% above week one when bought 15 to 60 days out (468 releases), 12% at 4 to 14 days (772), and 5% in the last three days (747). Against retail, pre-release pairs sold a median 24% above the sticker price, week-one pairs 11%.

That gap is not new. A 2014 study of eBay published by StockX found pre-release pairs selling at about a 12% premium.

The two marketplaces differ a little. StockX's pre-release pairs ran 12% above its week-one price (700 releases), GOAT's 7% (376 releases).

General releases sell early more than collabs do

Collaborations sold fewer pairs early, relative to how many they sold in week one: a median of 0.15 pre-release sales per week-one sale for 115 collabs, against 0.45 for 1,406 other releases (q < 0.001).

One explanation fits: a general release ships to many more stores and warehouses, so there are more ways for a pair to get out before launch, while collabs are held tightly by fewer hands. That is a reading of the pattern, not something the data proves.

What the data cannot tell you

A sale record has a price, a time, a size and a marketplace. It has no history of the pair. So none of this says how many early pairs were stolen, backdoored, sent to friends and family, or bought legitimately from an early drop. It shows that the early market is large, that it is concentrated on the biggest general-release Jordans, and that buyers pay up to get a pair before everyone else.

If you are checking what a new release sells for, the price checker shows recent medians, and Anatomy of a sneaker drop follows one release from its first pre-release sale to month six.

How we measured this

Releases: sneakers with a StockX release date and a retail price of $60 or more, released January 1, 2025 to February 28, 2026, adult sizes only, no slides, sandals or cleats. Sales: completed StockX sales plus GOAT sales of the same shoe, matched by style code, from 60 days before release day to 12 weeks after; the last window ends May 23, 2026, before StockX's May 22 to July 3, 2026 source gap affects any release in the set. A release counts only when its recorded sales reach back before release day (5,637 candidates, 2,590 kept), because our history is thinner for older listings; this keeps releases that had pre-release sales, so the share of sales made before release (12% here) describes these releases, not every shoe. Price comparisons use the same release before and after launch, releases with 10+ sales in each window, medians across releases with 95% bootstrap intervals; the share of releases with a higher pre-release price uses a Wilson interval and an exact sign test. The collab comparison uses a Mann-Whitney test; q-values are Benjamini-Hochberg across the two tests. Marketplace rules and the documented cases come from the linked sources, checked October 17, 2026. Past sales, not a forecast.