Hard to Gem: 22.8 Million PSA Slabs Say the Decade Decides, Texture Helps, and the PSA 10 Premium Runs the Other Way

Ask why a card is hard to gem and you get a story about the card: dark borders, silver holofoil, textured surfaces that show every scratch. The registry tells a plainer story. The Gem Rate Curve showed that the era a card was printed in predicts its gem rate better than anything else. This piece cuts the same registry the other ways a collector cuts it, by rarity, by language, by illustrator and by whether the card's number sits past the set total, and then asks the question the gem rate actually exists to answer: what is a hard gem worth? Our daily PSA population snapshots cover 24,175 registry lines; after removing provider collisions and the lines that have never recorded a 10, 17,730 catalog cards carry 22,812,572 PSA slabs and 11,289,732 tens.

Everything else in this piece has to be read through that table. Japanese cards, a third of all slabs, gem at 73.3% pooled and 60.5% at the median line; only 17.8% of Japanese slabs are 9s. English cards gem at 36.4% and 38.0% of their slabs are 9s. Japanese Registries Are Twice as Deep found the same shape from the price side. Two explanations are usually offered, better centring off the Japanese presses and selection, since the Japanese cards that reach PSA are overwhelmingly modern; the registry cannot separate them, but the era effect below does much of the work either way. The 926 lines with no rarity recorded in our catalog, which gem at 74.8% on 3.5 million slabs, are almost entirely Japanese promos, including the 329,337-slab promo Pikachu.

Read down the table and the folk theory fails in two places. First, the rarities that are supposed to be hard, the textured and full-art ones, sit in the top half: Special Illustration Rares gem at 41.1% pooled and 41.8% at the median line, Illustration Rares at 33.7% and 26.6%. Within Scarlet & Violet alone the gap is the same way round, 36.3% for the textured SIR against 28.3% for the flat IR, on 1.3 million and 1.2 million slabs. Second, the plain rarities are the hard ones: Rare gems at 16.1%, Holo Rare at 13.8%, the Diamond & Pearl LV.X at 3.3%. That is the era speaking through the rarity label. 'Rare' is a WotC, EX and Diamond & Pearl word; 'Special Illustration Rare' is a 2023 word. The common line, 52.7% pooled but 27.2% at the median, is a reminder that a pooled rate is a weighted average of the lines people actually grade, and the commons people grade are the MEP promo starters and the Japanese ones.

Hold the era fixed and rarity still moves the number, but by less than the era does. Inside Sword & Shield, promos gem at 31.0% and VMAX holos at 67.9%. Inside Scarlet & Violet, Illustration Rares gem at 28.3%, Special Illustration Rares at 36.3%, Ultra Rares at 42.7%. Inside WotC, everything sits between 4.1% and 19.0%, and the commons gem more often than the rares; one reading is that a cheaper card is judged less severely, which the registry does not say. The one outlier against the era story is Sword & Shield Rare, a 21.2x premium on a thin group of 24 lines. The Mega Evolution era, a year old, is already the easiest English era on the registry: 57.2% pooled across 602 lines, with its Special Illustration Rares at 62.0%. The Gem Rate Drift argued PSA is getting stricter card by card. The era table says something compatible and larger: the cards being made now are easier to gem than the cards made in 2000 by a factor of six or more, and no amount of drift closes that gap.

This is the table that turns a curiosity into a pricing rule. Across 2,609 English cards with a real PSA 10 sold median, the correlation between a card's gem rate and the log of its PSA 10 premium over raw is -0.60; across the 27 era-and-rarity groups above it is -0.80. Cards that gem under 10% of the time sell as PSA 10s at a median 28.1 times their raw price; cards that gem three times in four sell at 5.2 times. The premium tracks the registry. A PSA 10 of a card that gems at 60% is a common object, and the premium collapses toward the grading fee plus a little; a PSA 10 of a WotC holo is a survivor of a process that fails 95 times in 100, and the premium is the price of those 95 failures. The Fee Is a Fixed Cost and Grading ROI work the submission math from the other side; this table is why the answer is almost never 'grade the easy card'. The expected value of submitting is the premium times the gem rate, and the two move against each other.

The secret slot is where the money in a modern set lives (The Set Inside the Set) and it is also the easy gem: 54.0% pooled and 44.6% at the median line, against 30.5% for in-set numbers, with a PSA 10 premium of 7.8x against 14.7x. Most of that is the era again, since secret slots barely existed before 2019 and are now the whole top of every checklist; this table does not control for era, so read it as a description, not a second effect.

Five of the fifteen hardest are Neo Genesis, whose holos are the registry's proof that a 2000 print run can be nearly ungemmable: Lugia has produced 15 tens in 12,258 attempts. Three are Fossil, the set The Registry Is Closed used for the same point. Team Aqua's Kyogre EX and Team Magma's Groudon EX from Double Crisis are the modern exceptions, 2015 cards that gem at 0.19% and 0.32%, and the Kyogre's PSA 10 sold median of $3,218.75 against a $530.59 raw is the premium table in one row: a 6.1x multiple on a card that fails 99.8% of the time. Base Set Charizard, at 105,288 slabs and a 0.5% gem rate, is not in the table only because it is the largest registry on the list and its 0.5% clears the cut.

The easy end is Sword & Shield VMAX secret rares and the MEP promo starters, and the premiums are what the band table predicts: Charizard VMAX from the Shining Fates Shiny Vault, 36,375 tens in 48,658 slabs, sells as a PSA 10 at $241.64 against $135.43 raw, a 1.8x multiple; Charmander at 84.1% carries 3.2x. The Japanese list is easier still and we print it separately because none of it has an English sold tape: Eevee V from the CoroCoro Start Deck 100 at 96.4% (5,334 of 5,532), Mega Darkrai ex SAR at 95.8%, Mega Greninja ex SAR at 93.8% on 17,722 slabs, the Game Boy Dragonite promo at 93.1% on 35,500, and Lillie's Clefairy ex SAR at 92.5% on 24,240.

Painter or Pokemon found the artist explains 14% of a card's price rank. On the registry the artist explains nearly nothing once the era is known. The eight hardest illustrators are the WotC and EX-era hands, Mitsuhiro Arita among them at 27.3% on 1.16 million slabs, and Ken Sugimori at 10.3% on 746,653. The eight easiest are the Sword & Shield and Scarlet & Violet illustration-rare specialists, and their ranking is partly the language table again: every one of the eight carries Japanese lines, which gem at 87% to 91%, while the same artists' English lines gem at 40% to 60% (kantaro 44.0%, PLANETA Yamashita 40.1%, Jiro Sasumo 41.8% on English lines alone). The one studio with enough work across eras to test is 5ban Graphics, 1,037 lines and 1.76 million slabs, at 47.6%; weighting each of its lines by the pooled rate of its era predicts 40.7%, so the studio runs about seven points above its eras, the one artist effect this registry is large enough to see. For everyone else the era sets the number and the artist moves it by less than the sample can resolve.

Universe: the 24,175 PSA population lines in our daily snapshots (keyed by TCGplayer product id) as of 2026-09-13, using each line's most recent snapshot; a PSA 10 count of zero on that snapshot falls back to the latest snapshot carrying a count. Removed: 950 lines at 1,000 slabs or more whose totals agree within 1% and PSA 10 counts within 1.5% or one slab with a larger line (provider collisions; the larger is kept), 1,173 lines not in our card catalog, one known provider remap (Base Set 2 Charizard), and 4,321 lines with no PSA 10 recorded on any snapshot (see the callout). Final: 17,730 catalog cards, 22,812,572 slabs, 11,289,732 PSA 10s. Rarity is the catalog rarity string, case-folded; 'Rare Holo' is merged with 'Holo Rare' and 'Secret' with 'Secret Rare'; 'no rarity recorded' is the catalog's empty field. Era follows the set prefix (WotC through Mega Evolution) with Japanese prints as their own bucket. Secret slot means the card's number exceeds the set's printed total. Pooled gem rate is total PSA 10s over total slabs; median line gem rate is the median of per-card rates over lines with 100 or more slabs. PSA 9 share is PSA 9 count over total on lines carrying a PSA 9 count. Raw market is the median of our recorded daily market price over the eleven daily buckets from 2026-09-03 to 2026-09-13; PSA 10 sold median is the latest graded snapshot with three or more verified eBay sales, English lines only; premium is the ratio, over lines priced $5 and up. Correlations use the natural log of the premium; the group correlation covers the 27 English era-and-rarity groups with 20 or more premium lines. Card tables require 2,000 or more slabs and $20 or more raw. Illustrator groups require 15 lines of 100 or more slabs and 20,000 slabs in total. Medians on even samples average the two middle values. Not financial advice.