Borrowed magic.
Universes Beyond, mass reprints, denser text boxes — a data essay on what Magic: The Gathering is made of in 2026.
Magic: The Gathering shipped 295 cards in August 1993 — fifteen at a time, in two-and-a-half-million-print-run packs that cost two dollars. By 2026, Wizards of the Coast prints more cards in a single Standard expansion than existed in the entire game for its first two years. The game survived from one number to the other. The question is whether anything else did.
Pull the full archive MTGJSON has recorded — every printing back to Alpha — and run three measurements against thirty-three years of release dates. Different game, or the same one borrowed?
What follows is what the numbers said. Three Acts: how Wizards started licensing in, how the back catalog became the front shelf, and how the text box got loud.
The game went borrowing.
For thirty years Magic was made out of itself. Then, almost overnight, it wasn’t.
Both are creatures. Both shipped legal-for-tournament. One was drawn by Anson Maddocks for a game its author was still finishing. The other was painted from a New Line still and licensed by a publisher in Renton, Washington.
| Year | Universes Beyond share | Native MTG share |
|---|---|---|
| 1993 | 0% | 100% |
| 1994 | 0% | 100% |
| 1995 | 0% | 100% |
| 1996 | 0% | 100% |
| 1997 | 0% | 100% |
| 1998 | 0% | 100% |
| 1999 | 0% | 100% |
| 2000 | 0% | 100% |
| 2001 | 0% | 100% |
| 2002 | 0% | 100% |
| 2003 | 0% | 100% |
| 2004 | 0% | 100% |
| 2005 | 0% | 100% |
| 2006 | 0% | 100% |
| 2007 | 0% | 100% |
| 2008 | 0% | 100% |
| 2009 | 0% | 100% |
| 2010 | 0% | 100% |
| 2011 | 0% | 100% |
| 2012 | 0% | 100% |
| 2013 | 0% | 100% |
| 2014 | 0% | 100% |
| 2015 | 0% | 100% |
| 2016 | 0% | 100% |
| 2017 | 0% | 100% |
| 2018 | 0% | 100% |
| 2019 | 0% | 100% |
| 2020 | 0% | 100% |
| 2021 | 0% | 100% |
| 2022 | 8% | 92% |
| 2023 | 33% | 67% |
| 2024 | 16% | 84% |
| 2025 | 30% | 70% |
- The all-internal era. For its first twenty-five years, Magic drew exclusively on its own settings — Dominaria, Mirrodin, Innistrad, all of them invented inside Wizards. Every printed card was native to the game.
- 2019 — Secret Lair drops. Wizards opened a direct-to-consumer channel: limited-run drops sold straight from a website, bypassing the booster-and-distributor pipeline. The first drops were art reskins of native cards. The mechanism that would carry licensed IP existed first.
- 2020 — The Walking Dead. The first Secret Lair drop with outside IP — five tournament-legal cards reskinned with AMC’s zombies. It sold out in 24 hours. It also drew a public-letter response from competitive players objecting to non-canon cards being tournament-legal.
- 2022 — 8% baseline. Warhammer 40k Commander shipped a full product line; UB was 8.4% of every printing that year. Still small enough to read as an experiment.
- 2023 — The cliff. LotR: Tales of Middle-Earth and the Doctor Who Commander decks pushed UB to 33% of printings — a fourfold single-year jump. One in three new cards now belonged to a story Wizards didn’t own.
- 2025 — A standing rotation. Final Fantasy. Marvel’s Spider-Man. Fallout the year before, Assassin’s Creed alongside it. UB stopped being an experiment and became a recurring product line, baked into the release calendar.
- 8% → 33%Share of printings, 2022 vs 2023
- 7Franchises licensed since 2021
- 30%Share of 2025 printings, Universes Beyond
Reprints became the product.
Magic always sold old cards alongside new ones. The engine that produces the old ones has changed shape three times.
Identical card. Identical rules text. One mana for two colorless, then and now. Sol Ring didn’t change. Commander did: Wizards shipped a wave of pre-constructed decks through 2023, each one packing a fresh printing.
| Year | Reprints share | New cards share | Sets shipped |
|---|---|---|---|
| 1993 | 63% | 37% | 4 |
| 1994 | 62% | 38% | 7 |
| 1995 | 73% | 27% | 8 |
| 1996 | 33% | 67% | 5 |
| 1997 | 43% | 57% | 7 |
| 1998 | 22% | 78% | 5 |
| 1999 | 48% | 52% | 7 |
| 2000 | 21% | 79% | 5 |
| 2001 | 57% | 43% | 6 |
| 2002 | 5% | 95% | 3 |
| 2003 | 55% | 45% | 5 |
| 2004 | 6% | 94% | 3 |
| 2005 | 69% | 31% | 5 |
| 2006 | 31% | 69% | 7 |
| 2007 | 56% | 44% | 6 |
| 2008 | 32% | 68% | 7 |
| 2009 | 54% | 46% | 11 |
| 2010 | 54% | 46% | 12 |
| 2011 | 65% | 35% | 12 |
| 2012 | 42% | 58% | 9 |
| 2013 | 55% | 45% | 9 |
| 2014 | 63% | 37% | 16 |
| 2015 | 58% | 42% | 12 |
| 2016 | 55% | 45% | 13 |
| 2017 | 63% | 37% | 16 |
| 2018 | 64% | 36% | 17 |
| 2019 | 66% | 34% | 11 |
| 2020 | 84% | 16% | 18 |
| 2021 | 67% | 33% | 17 |
| 2022 | 76% | 24% | 20 |
| 2023 | 76% | 24% | 18 |
| 2024 | 75% | 25% | 21 |
| 2025 | 65% | 35% | 17 |
- 1993 — Reprint sets as the product. Alpha shipped, sold out, was reprinted as Beta, then reprinted again as Unlimited. By the end of the first calendar year roughly two of every three Magic cards in print were already reprints — of cards under a year old.
- 2000 — The expansion-only dip. Wizards stopped reprinting old sets and started shipping new ones every quarter. With nothing to reprint into, the reprint share collapsed to 21%. It would never settle that low again.
- 2013 — Modern Masters. A dedicated reprint vehicle. Modern Masters made the back catalog its own product line — premium boosters, the same Standard-illegal Modern staples, drafted by adults at convention halls. Reprints became something Wizards sold on purpose.
- 2020 — The Masters bulge. Four Masters sets in a single year. Reprint share hit 84% — the all-time peak. The pandemic made boosters more profitable than expansions; Wizards responded.
- 2025 — A standing rotation. Commander decks now ship alongside every expansion, and each one is roughly 90% reprints. The Masters cadence has slowed, but Commander has replaced it. Reprints are 65% of every year and climbing again.
- 63% → 65%Reprint share, 1993 vs 2025
- 84%Peak reprint share (2020)
- 4 → 17Sets shipped per year, 1993 vs 2025
The text box got loud.
The text box was flat for twenty-five years. Then in six it nearly doubled, while power per mana barely moved. The new volume is text.
Two creatures. Two and a half mana, either way. Both legal at a sanctioned table. Grizzly Bears has no rules text. Phlage has roughly fifty words of it: lifelink, trample, an enters-from-graveyard trigger, a colored escape cost. Power and toughness inched up. The rest is text.
| Year | Mean words | Power per mana |
|---|---|---|
| 1993 | 14.92 | 1.517 |
| 1994 | 16.40 | 1.418 |
| 1995 | 17.23 | 1.378 |
| 1996 | 18.01 | 1.476 |
| 1997 | 14.31 | 1.469 |
| 1998 | 13.79 | 1.514 |
| 1999 | 11.85 | 1.386 |
| 2000 | 16.12 | 1.482 |
| 2001 | 13.87 | 1.397 |
| 2002 | 18.59 | 1.380 |
| 2003 | 13.41 | 1.428 |
| 2004 | 16.78 | 1.429 |
| 2005 | 12.41 | 1.475 |
| 2006 | 17.93 | 1.467 |
| 2007 | 15.34 | 1.500 |
| 2008 | 17.74 | 1.492 |
| 2009 | 15.31 | 1.430 |
| 2010 | 15.80 | 1.581 |
| 2011 | 14.93 | 1.542 |
| 2012 | 15.34 | 1.509 |
| 2013 | 16.49 | 1.508 |
| 2014 | 16.88 | 1.554 |
| 2015 | 16.58 | 1.535 |
| 2016 | 18.85 | 1.638 |
| 2017 | 17.55 | 1.659 |
| 2018 | 17.04 | 1.609 |
| 2019 | 20.28 | 1.640 |
| 2020 | 19.32 | 1.602 |
| 2021 | 24.23 | 1.693 |
| 2022 | 24.65 | 1.637 |
| 2023 | 26.08 | 1.683 |
| 2024 | 24.26 | 1.694 |
| 2025 | 25.63 | 1.856 |
| Year | Keyword debuts | Example debuts (first three) |
|---|---|---|
| 1993 | 14 | Banding (Benalish Hero); Defender (Wall of Swords); Desertwalk (Desert Nomads) |
| 1994 | 6 | Islandwalk (Goblin Flotilla); Legendary landwalk (Livonya Silone); Mill (Millstone) |
| 1995 | 3 | Cumulative upkeep (Cold Snap); Menace (Wind Spirit); Shroud (Autumn Willow) |
| 1996 | 4 | Flanking (Femeref Knight); Flash (King Cheetah); Phasing (Merfolk Raiders) |
| 1997 | 4 | Buyback (Anoint); Fear (Commander Greven il-Vec); Fight (Triangle of War) |
| 1998 | 2 | Cycling (Drifting Meadow); Echo (Herald of Serra) |
| 1999 | 2 | Hexproof (Taoist Hermit); Horsemanship (Guan Yu, Sainted Warrior) |
| 2000 | 3 | Domain (Strength of Unity); Fading (Defender en-Vec); Kicker (Ardent Soldier) |
| 2001 | 3 | Flashback (Ancestral Tribute); Surveil (Think Tank); Threshold (Divine Sacrament) |
| 2002 | 2 | Madness (Frantic Purification); Morph (Crude Rampart) |
| 2003 | 16 | Affinity (Assert Authority); Amplify (Aven Warhawk); Double strike (Ridgetop Raptor) |
| 2004 | 6 | Bushido (Bushi Tenderfoot // Kenzo the Hardhearted); Indestructible (Konda, Lord of Eiganjo); Modular (Arcbound Wanderer) |
| 2005 | 13 | Bloodthirst (Ghor-Clan Savage); Channel (Shinen of Fury's Fire); Convoke (Scatter the Seeds) |
| 2006 | 7 | Graft (Cytoplast Manipulator); Haunt (Absolver Thrull); Nonbasic landwalk (Dryad Sophisticate) |
| 2007 | 19 | Assemble (Steamflogger Boss); Aura Swap (Arcanum Wings); Champion (Changeling Hero) |
| 2008 | 11 | Conspire (Mine Excavation); Devour (Tar Fiend); Exalted (Akrasan Squire) |
| 2009 | 3 | Basic landcycling (Gleam of Resistance); Cascade (Ardent Plea); Intimidate (Halo Hunter) |
| 2010 | 8 | Annihilator (Artisan of Kozilek); Infect (Blackcleave Goblin); Level Up (Caravan Escort) |
| 2011 | 5 | Battle Cry (Accorder Paladin); Join forces (Alliance of Arms); Living weapon (Bonehoard) |
| 2012 | 9 | Detain (Azorius Arrester); Fateful hour (Break of Day); Miracle (Banishing Stroke) |
| 2013 | 10 | Battalion (Boros Mastiff); Bestow (Celestial Archon); Bloodrush (Pyrewild Shaman) |
| 2014 | 13 | Constellation (Harvestguard Alseids); Dethrone (Marchesa's Emissary); Ferocious (Force Away) |
| 2015 | 14 | Awaken (Encircling Fissure); Bolster (Dromoka, the Eternal); Converge (Unified Front) |
| 2016 | 17 | Cohort (Munda's Vanguard); Council's dilemma (Lieutenants of the Guard); Crew (Aradara Express) |
| 2017 | 11 | Afflict (Eternal of Harsh Truths); Aftermath (Dusk // Dawn); Embalm (Angel of Sanctions) |
| 2018 | 9 | Ascend (Pride of Conquerors); Assist (Bring Down); Commander ninjutsu (Yuriko, the Tiger's Shadow) |
| 2019 | 39 | ... Catch (Captain America, First Avenger); Adamant (Ardenvale Paladin); Adapt (Zegana, Utopian Speaker) |
| 2020 | 19 | Alliance (Rumor Gatherer); Backup (Gloomfang Mauler); Buy Information (Black Market Connections) |
| 2021 | 3 | Conjure (Teyo, Aegis Adept); Intensity (Static Discharge); Seek (Faceless Agent) |
| 2022 | 201 | A Thousand Souls Die Every Day (The Golden Throne); Aberrant Tinkering (Grell Philosopher); Advanced Species (The Red Terror) |
| 2023 | 61 | Affirmative (K-9, Mark I); Allons-y! (The Tenth Doctor); Bad Wolf (Rose Tyler) |
| 2024 | 48 | Allies (Cleopatra, Exiled Pharaoh); Alluring Eyes (Red Death, Shipwrecker); Betrayal (Cleopatra, Exiled Pharaoh) |
| 2025 | 85 | 10,000 Needles (Jumbo Cactuar); A Test of Your Reflexes! (Magitek Scythe); Ability (Battle Menu) |
- 1993 — Baseline. Mean words per card: 14.9. Grizzly Bears costs one and a green, is a 2/2, and has no rules text at all. Llanowar Elves taps for green. The card is mostly art and a name.
- 2003 — The flat era. Onslaught block ships Morph and Cycling — mechanics complex enough to need their own keywords. Mean words: 13.4. Lower than 1993. For the next decade, average text density barely moves.
- 2013 — Twenty years in, still under 17. Modern Masters, Theros, Bestow, Devotion, Monstrosity. Every set introduces keywords; none of them push the average. Mean words: 16.5. Power per mana: 1.51 — within a rounding error of 1993.
- 2019 — The break. Throne of Eldraine ships Adventure cards: a creature on the back, an instant on the front, both legal at once. War of the Spark gives every planeswalker a static ability. Mean words crosses 20 for the first time — and doesn’t come back.
- 2025 — 25.6 words. From 1993 to 2018, average word count climbed 14% over twenty-five years. From 2019 to 2025, it climbed another 50% in six. Power per mana is up 22% across the whole span. Most of what Magic has added in three decades, it added in the last six. As text.
- +72%Words per card, 1993 vs 2025 (14.9 → 25.6)
- +22%Power per mana, same period (creatures, MV ≤ 4)
- 16×Keyword debuts per year, 2020s vs 1990s (avg 70 vs 4)
Three lines, one game.
| Year | Value |
|---|---|
| 1993 | 0% |
| 1994 | 0% |
| 1995 | 0% |
| 1996 | 0% |
| 1997 | 0% |
| 1998 | 0% |
| 1999 | 0% |
| 2000 | 0% |
| 2001 | 0% |
| 2002 | 0% |
| 2003 | 0% |
| 2004 | 0% |
| 2005 | 0% |
| 2006 | 0% |
| 2007 | 0% |
| 2008 | 0% |
| 2009 | 0% |
| 2010 | 0% |
| 2011 | 0% |
| 2012 | 0% |
| 2013 | 0% |
| 2014 | 0% |
| 2015 | 0% |
| 2016 | 0% |
| 2017 | 0% |
| 2018 | 0% |
| 2019 | 0% |
| 2020 | 0% |
| 2021 | 0% |
| 2022 | 8% |
| 2023 | 33% |
| 2024 | 16% |
| 2025 | 30% |
| Year | Value |
|---|---|
| 1993 | 63% |
| 1994 | 62% |
| 1995 | 73% |
| 1996 | 33% |
| 1997 | 43% |
| 1998 | 22% |
| 1999 | 48% |
| 2000 | 21% |
| 2001 | 57% |
| 2002 | 5% |
| 2003 | 55% |
| 2004 | 6% |
| 2005 | 69% |
| 2006 | 31% |
| 2007 | 56% |
| 2008 | 32% |
| 2009 | 54% |
| 2010 | 54% |
| 2011 | 65% |
| 2012 | 42% |
| 2013 | 55% |
| 2014 | 63% |
| 2015 | 58% |
| 2016 | 55% |
| 2017 | 63% |
| 2018 | 64% |
| 2019 | 66% |
| 2020 | 84% |
| 2021 | 67% |
| 2022 | 76% |
| 2023 | 76% |
| 2024 | 75% |
| 2025 | 65% |
| Year | Value |
|---|---|
| 1993 | 15 |
| 1994 | 16 |
| 1995 | 17 |
| 1996 | 18 |
| 1997 | 14 |
| 1998 | 14 |
| 1999 | 12 |
| 2000 | 16 |
| 2001 | 14 |
| 2002 | 19 |
| 2003 | 13 |
| 2004 | 17 |
| 2005 | 12 |
| 2006 | 18 |
| 2007 | 15 |
| 2008 | 18 |
| 2009 | 15 |
| 2010 | 16 |
| 2011 | 15 |
| 2012 | 15 |
| 2013 | 16 |
| 2014 | 17 |
| 2015 | 17 |
| 2016 | 19 |
| 2017 | 18 |
| 2018 | 17 |
| 2019 | 20 |
| 2020 | 19 |
| 2021 | 24 |
| 2022 | 25 |
| 2023 | 26 |
| 2024 | 24 |
| 2025 | 26 |
The first cliff appeared in two years. The second peak took thirty. The third climb happened mostly in the last six. Each line tells a different story about pace, and the same one about subject. The card is still the artifact Garfield shipped in 1993: same back, same color pie, same five basic lands. The face has gone borrowed, reprinted, and text-heavy.
The game hasn’t changed. What’s on the cards has.
What did a card look like in 2025?
Move the slider. Watch the average card change.
The card on the left is a mean across that year’s sample: average mana value, average power and toughness, average words. The card on the right is real — the one printing in MTGJSON closest to that mean. Push the slider forward and the words climb, the power barely budges, and the color identity drifts toward five.