Every grading debate eventually hits the same wall: is PSA stricter than it used to be, or do collectors just remember their tens fondly? Nobody could answer because nobody had the tape. We do now. Our registry tracker records every card's population AND its gem count daily, which means for any card we can compare its lifetime gem rate against the gem share of ONLY its recent submissions. Across the 99 cards that added at least 200 slabs in our window, the answer is: yes, measurably stricter, and wildly unevenly.
The strictest drifts (lifetime gem rate vs recent submissions)
| Card | Lifetime | Recent | New slabs |
|---|---|---|---|
| Charizard ex, Paldean Fates | 40.6% | 15.8% | 285 |
| Squirtle, Stellar Crown | 37.6% | 14.5% | 234 |
| Pikachu with Grey Felt Hat | 42.9% | 21.9% | 736 |
| Pikachu, 151 | 26.9% | 8.5% | 329 |
| Mew, Crown Zenith GG | 28.3% | 10.7% | 318 |
Why drift happens, and why it clusters
Three forces move a card's gem share and none of them require a conspiracy. Supply quality degrades: the cleanest copies get graded first, so late submissions draw from handled stock. Submitter mix shifts: after a card spikes, casual owners flood in with binder copies that a dealer would have pre-screened out. And the grader's own line moves: when the Van Gogh Pikachu went from gemming 43% to 22% across 736 fresh submissions of a promo that is barely two years old and was distributed IN a protective sleeve, supply-quality decay alone struggles to explain it. Whatever the mix of causes, the drift is real money: at current medians the gap between a 9 and a 10 on that card is the whole grading thesis.
The loosening tail is smaller but just as tradeable. Latias ex gems 53.7% recently against a 35.0% lifetime rate, and Pikachu ex from Prismatic Evolutions runs 40.6% vs 26.6%. Cards drifting looser are quietly getting CHEAPER to gem: the expected cost per PSA 10 falls as the hit rate rises, which changes the grade-or-hold math our ROI table publishes before the market reprices it.
Cohort: every tracked card whose PSA population grew by at least 200 slabs across a window of at least 10 days ending August 28, 2026, with a starting population of at least 500 (99 cards). Lifetime gem rate = gem count over total at the window start; recent gem share = gem-count delta over total delta across the window. Averages are unweighted across the cohort; the cohort skews toward heavily-submitted modern chases by construction. Population figures are PSA registry counts recorded daily by Moonstone. Drift attribution (supply decay vs submitter mix vs standards) is discussed qualitatively; the data measures the drift itself, not its cause. Not financial advice.