Courtyard pack report — September 2026

The month's real story is one Courtyard did not announce: the instant buyback stopped being a tiered rule and became a flat one. Cheap cards lost 10.3 percentage points of their cash value. Expensive cards lost nothing. We found it by measuring 33,980 recorded offers, not by reading a policy page.

Covering 1–14 September 2026. All figures measured from pulls indexed off Courtyard's public feed. September is a partial month, so month-over-month counts are not like-for-like.

The buyback became flat, and nobody said so

Through July, Courtyard's instant offer rewarded small cards. A card worth under $25 drew an average of 95.7% of its fair market value, while a card over $100 drew 84.6%. By September every band had converged on the same flat rate.

The convergence was fast. 44.1% of July 2026 offers, then 62.8% of August 2026 offers, then 96.5% of September 2026 offers sat on exactly 84.6% — a number that constant across nineteen thousand observations is a pricing rule rather than a market outcome.

Why it matters more than the headline suggests: the catalogue median pull is $26.50, which sits inside the band that absorbed the entire cut. A buyer chasing grails barely noticed. A buyer ripping regularly took all of it. On a $1,000 run of pulls that is roughly $60 of value that quietly stopped arriving, with no change to the packs, the prices or the odds.

Full working in the buyback guide.

Zero offers are rising

A smaller signal worth logging so the series can track it: the share of pulls Courtyard declines to bid on at all has gone from 1.5% in July to 3.8% in September. Still small, but it has doubled in two months, and it means the instant exit is a very good default rather than a guarantee.

Eight new packs appeared

September was an active month for launches, led by a Football run tied to the season and a new Ancient Coins line.

Most launched without a published price, so no cash-out ratio is computable for them yet — we report a verdict only where there is a price for a pull to beat. The one to watch is Baseball Mythic Pack at $5,000: five indexed pulls so far, two of which were the Ohtani PSA 10s below. Five pulls is nowhere near enough to judge a pack, and we will not pretend otherwise.

The month's biggest pulls

Six cards above $4,500 were indexed in the first fortnight, and Pokémon took four of them.

The $9,792.60 Reshiram & Charizard GX is the second-largest pull we have ever indexed, behind a $17,500 Michael Jordan in August. It also illustrates the problem with reading a single hit as a signal: its pack, Pokémon VIP Cosmic, has produced exactly one indexed pull, so its entire published expected value rests on that one card. We show no verdict for it, and the “insufficient data” label is doing real work.

Where the catalogue stands

Of the 46 packs with a known price and thirty or more indexed pulls, 4 sit at or above break-even. The average returns 88.5 cents of cash per dollar; the best measured returns +26.9% and the worst -39.7%.

That spread — more than sixty points between the best and worst pack on the same platform — has been the stable finding across every month we have measured. The identity of the four good packs changes; the fact that there are only about four does not.

Where the money went, by category

Pokémon remains the centre of gravity by a wide margin — 41 packs and 24,526 indexed pulls, about two-thirds of everything we see. Football is a distant second at 3,602, and September's three new Football packs suggest that gap will narrow over the season.

Two entries need a caveat rather than a headline. D.C. Comic tops the table at 1.048x, which would make it the best category on the platform — from a single pack and 94 pulls. Marvel Comic sits last at 0.683x from four packs and 37 pulls. Neither sample supports a category-level claim, which is why both are dimmed above. The two you can lean on are Pokémon and Football, where the counts run to thousands.

What we could not see this month

A report that only lists findings implies it found everything. These are the gaps, and they are structural rather than temporary.

We do not know why the buyback changed. We can show that it did, when, and who absorbed it. Whether it was a margin decision, a response to arbitrage on cheap cards, or a pricing-model rewrite is not visible from the outside, and we are not going to guess in public.

We cannot see inventory. Not packs remaining, not how a pool is seeded, not what is left in one. So when a pack's odds decline we can say the good cards stopped coming out; we cannot confirm they ran out. That distinction matters and nobody outside Courtyard can close it.

We see a sample, not a census. The feed is polled every five minutes, and activity between polls is gone rather than delayed. Busy periods are under-sampled by construction, so a pack's true pull count is at least what we report and possibly more.

FMV is Courtyard's estimate. Every value on this page, including the $9,793 Charizard, is their valuation rather than a sale price. Where their estimates sit above or below live market comparables, every ratio here moves with them.

What we are watching next month

Whether 84.6% holds. A rule that moved once can move again, and it is now a tracked series rather than an assumption.

Whether zero-offer pulls keep climbing. 1.5% to 3.8% in two months is a trend worth a third data point before anyone calls it one.

The September launches, once they have samples. Six of the eight new packs are unpriced or under thirty pulls, so none of them can be judged yet. By the October report most should be.

Method: every figure here is counted from pulls indexed off Courtyard's public feed, with sample sizes stated. We observe outputs — the cards that came out and what Courtyard valued them at — and cannot see inventory, packs remaining, or how any pool is seeded. Where a number is not derivable we say so rather than estimate. Full method on the methodology page.