The Book at Thirty Days: Every Position We Own, Marked, Including the Pair We Had to Suspend

Minus 2.70%. That is the mean mark on every position we published and have held long enough to price, sixteen of them, against the same verified tape that powers every price on this site. The median is plus 0.33%, eight are up and eight are down, and all of the damage sits above $1,000 a ticket. Thirty days is exactly long enough to be humbling and nowhere near long enough to be a track record, so this is not a victory lap or a mea culpa. It is the whole book, position by position, with the arithmetic shown, the rules that produced it graded, and the two rows we cannot honestly mark flagged as suspended rather than quietly dropped.

The ledger holds 97 published picks across six weekly editions and the first two daily ones. It is append-only, entry prices are frozen at publication, and nothing is ever removed, which is what makes a thirty-day audit possible at all. But 97 flatters itself in three ways. Picks repeat across weeks when a thesis stays live, so 97 rows collapse to 29 distinct positions, one of which is a sealed product and outside this card-level audit, leaving 28. Ten of those 28 were published within the last three days and cannot be marked at all without dressing up noise as performance, so we require a 14 day hold, which leaves 18. And two of the 18 are the Fossil birds whose comp pools we proved were contaminated by mixed printings, so their marks are suspended under the policy set out in the autopsy. Sixteen positions survive that. Every gate here removes rows that would have made the number look better or worse, and we would rather publish sixteen defensible marks than ninety-seven decorative ones.

Split the sixteen by position size and the pattern is not subtle. The four positions entered above $1,000 average minus 6.64%, three of the four are down, and they include the two largest tickets on the book, Umbreon ex at $6,225 and Mew ex at $3,475.50. The seven positions entered under $50 average minus 1.99% with a median of plus 1.47%. That lines up uncomfortably well with an entirely separate finding from an entirely separate dataset: our disclosed target book, read by price band in the multiple ladder, assigns negative median upside to every rung above $256 raw. Two different instruments, one saying the expensive end is fully priced and the other losing money there in real time.

The second pattern is that the graded seat and the raw seat of the same card diverged violently. Kieran raw returned plus 11.1% while its PSA 10 returned minus 16.5%, a 27.6 point spread on one card over the same window. Sinistcha ex went the other way: the raw copy returned plus 1.5% and the PSA 10 plus 18.8%. On Ethan's Adventure both seats fell together, minus 12.5% raw and minus 13.6% in the slab. Across nine slab positions the mean is minus 3.24% and across seven raw positions minus 1.99%, but the averages hide the real lesson: on a thirty day horizon, choosing the seat mattered more than choosing the card.

Our scorecard grades each rule by its mean alpha and flags any rule for review once six or more marked picks run negative. Applying that structure to this book produces an honest embarrassment: the print-clock rule appears on 15 of the 16 live tickets and returns a mean of minus 3.10%, the graded-ladder rule appears on 11 and returns minus 2.21%, the momentum-cohort rule on 5 and returns minus 2.92%. Every rule is negative, and almost every position carries the dominant rule, so these are not three independent measurements. They are one book being described three times. A factor cannot be attributed when it is on nearly every ticket, and pretending otherwise is how content marketing manufactures a winning rule from a losing book.

One further honesty note, and it is a limitation rather than a result. Those figures are RAW returns, not alpha. The benchmark leg of the scorecard, which measures each pick against the moonstone Pokémon Index over the same window, is currently returning no value in the latest stored edition, so cumulative alpha is unavailable and we will not quietly substitute the raw number for it. A pick down 3% in a market down 6% is a good pick, and we cannot presently tell you which of these it is. The scorecard's own last stamp, on August 31, 2026, reported a book return of minus 0.63% across 72 marked rows on a looser marking rule than this article uses, with all three benchmark legs blank. Both numbers are ours, they differ because the gates differ, and this article's gates are the stricter of the two.

Source: the moonstone weekly_picks ledger, 97 published rows from 2026-W31 through 2026-W36 plus the daily editions of August 31 and September 1, 2026, collapsed to 28 distinct card-and-seat positions at their earliest published entry. A position is marked only with a 14 day minimum hold: graded seats against the latest verified sold median for that card and grade in the daily graded snapshot table (tape through August 28 to 29, 2026, per card), raw seats against the September 2, 2026 market price. Holds run 18 to 34 days. The Teal Mask Ogerpon position is the POKEMON CENTER STAMPED printing (TCGplayer 552786, $21.79 on September 2), which is what the $21.97 entry and the mark both price; its card link resolves to svp-123, the SVP promo page, because the stamped printing has no separate catalog id on our book, and the unstamped promo behind that link trades at $1.50. We name the printing rather than let the link imply the cheaper one. The two Fossil PSA 9 positions are suspended because their comp series still disagrees with itself: our own printing-verified snapshot of Moltres reads $199.99 on 72 comps on August 25, while the provider tape three days later on August 28 reads $579 on 201, the exact mechanism documented in The $479 Lie. Returns are unweighted and gross of fees, shipping, and grading costs; a real book pays all three. Sixteen marks over thirty days is not a track record and no significance is claimed. Not financial advice.