One Cantrip, Three Cards: How a $1.18 Marvel Card Reached cEDH's Top Ten
Cast a one-mana cantrip, draw three cards. That's what a $0.98 aura from 1998 does once it's sitting on Crystal, Inhuman Princess — and it's why a $1.18 Marvel card went from zero tracked cEDH entries in May to the seventh most-played commander in the format. We tested whether the curve survives the obvious objections, and we say plainly where our numbers disagree with the other tracker's.
Cast Brainstorm. Draw three cards.
Not the three Brainstorm puts in your hand — three extra ones, on top. That's what a single Curiosity does when it's sitting on Crystal, Inhuman Princess in a four-player pod, and it is the entire reason a card you can buy for $1.18 is now the seventh most-played commander in our cEDH corpus. In May she was in zero of the 9,683 cEDH tournament entries we tracked. She became legal on June 26. Over the last 30 days she's in 163 of 6,887 — 2.4%, ahead of Vivi Ornitier, ahead of Magda, Brazen Outlaw, one rank behind Thrasios + Tymna.
The interesting part isn't the curve. It's that the engine driving it is a 1998 uncommon that four generations of Commander players have already dismissed as a cute Niv-Mizzet toy.
The card, exactly
Three lines, and every one of them matters:
Flying Whenever you cast a noncreature spell, Crystal deals X damage to each opponent, where X is the number of colors that spell is. : Add , , , or .
Start with the last line, because it's the one people misread. Crystal costs — a Gruul creature by her mana cost. Her colour identity is WURG, all four, because that mana ability puts and symbols in her rules text. She's a three-mana commander who unlocks a four-colour deck and taps for any of those four colours the turn she needs them. Cheap, colour-fixing, and she's your commander, so she's always there.
The middle line is the payoff, and it's deliberately modest on its own. A mono-coloured spell is one colour, so X is 1: one damage to each opponent. Three damage spread across the table isn't a clock, it's a rounding error.
Except damage isn't what you're buying.
The interaction
Three auras, three different sets, one identical line of text:
Whenever enchanted creature deals damage to an opponent, you may draw a card.
Note what it says — and what it doesn't. Not "deals combat damage." Not "deals damage to a player." An opponent, any damage, any amount. Wizards has ruled both halves of that explicitly. From the 2011-09-22 ruling: "Any damage dealt by the enchanted creature to an opponent will cause Curiosity to trigger, not just combat damage." And from 2023-09-01: "You draw one card each time the enchanted creature deals damage to an opponent, no matter how much damage it deals."
Each time. One damage or ten, it's one card — and Crystal deals her damage to each opponent. In a four-player pod that's three opponents taking damage from one trigger, which is three separate instances of "dealt damage to an opponent," which is three draws.
So the loop, step by step:
- You cast a noncreature spell — any one, a cantrip is fine.
- Crystal's ability triggers and goes on the stack above it.
- It resolves. Crystal deals 1 damage to each of your three opponents.
- Curiosity sees three instances of damage-to-an-opponent and triggers three times.
- You draw three cards, then the original spell resolves and does its own job.
A one-mana cantrip that replaces itself now nets you three cards and puts your opponents on a three-turn clock they can't ignore. Cast two more and you've drawn nine. That's not a value engine, it's a different game.
And it scales the wrong way for everyone else: X is the number of colours the spell is, so a gold two-colour spell deals 2 to each opponent. The damage doubles. The draws don't — the ruling is one card per instance regardless of amount. Which tells you exactly how the deck is built: as many cheap spells as possible, colour irrelevant, damage a side effect.
The catch nobody mentions
X is the number of colours the spell is. A colourless spell is zero colours. X = 0. No damage — and an ability that triggers on damage doesn't trigger when nothing is dealt.
Which means Sol Ring, sitting in 99.42% of these lists, draws you exactly nothing. Same for Lotus Petal and Chrome Mox at ~98.9% and 98.4%. The fastest, most-played acceleration in the format is invisible to the engine. That's the real cost of the deck, and it's the thing that separates a Crystal list from a generic storm pile: your rituals and rocks buy you mana, and only your coloured spells buy you cards.
The corpus
Here's the adoption, in disjoint calendar months — not overlapping look-back windows, so the movement between them is real:
| Month | cEDH entries tracked | Crystal entries | Share | |---|---|---|---| | May 2026 | 9,683 | 0 | 0.0% | | June 2026 | 8,543 | 39 | 0.5% | | July 2026 | 7,950 | 184 | 2.3% | | Aug 1–25 2026 | 6,570 | 158 | 2.4% |
June is four days of legality. July is the first full month, and she lands at 2.3% immediately — no ramp. August is holding.
One methodological note, because it's the reason we trust this curve at all: our cEDH corpus is known to carry date leakage for brand-new cards, where lists dated before a card's release somehow contain it. Crystal shows exactly zero entries in May. The pre-release control passes cleanly, so the timeline above is the card's real timeline.
Where that puts her over the trailing 30 days (2026-07-27 → 2026-08-26, n = 6,887 entries):
| # | Commander | Share | Entries | Top-cut conversion (n) | |---|---|---|---|---| | 1 | Kraum + Tymna ("Blue Farm") | 7.3% | 500 | 33.70% (359) | | 2 | Kinnan, Bonder Prodigy | 6.8% | 466 | 32.00% (350) | | 3 | Rograkh + Thrasios | 5.3% | 366 | 31.73% (271) | | 4 | Rograkh + Silas Renn | 4.6% | 317 | 25.47% (212) | | 5 | Sisay, Weatherlight Captain | 4.1% | 285 | 29.13% (206) | | 6 | Thrasios + Tymna | 3.3% | 230 | 30.27% (185) | | 7 | Crystal, Inhuman Princess | 2.4% | 163 | 33.86% (127) | | 8 | Vivi Ornitier | 2.2% | 150 | 24.11% (112) | | 9 | Ral, Monsoon Mage // Ral, Leyline Prodigy | 2.2% | 150 | 29.92% (127) | | 10 | Nick Fury, Agent of S.H.I.E.L.D. | 2.1% | 144 | 34.29% (105) |
Share here is share of tracked tournament entries — every list we scrape from an event, not just the ones that made top cut, and not a field share of all cEDH played anywhere. Conversion is the share of a commander's entries that reached their event's top cut, computed only where full standings exist; that's what the n in parentheses counts. Note that her 33.86% sits on 127 entries, which is a small enough base that it belongs in a sentence like this one rather than in a headline. The 2.4% is the number with weight behind it.
Is any of it real?
That's the question worth asking about any two-month-old commander, and there are two obvious ways this kind of curve turns out to be a mirage. We tested both.
"She only shows up at tiny local events." A $1 commander should be over-represented at the smallest, most casual-adjacent tables. Over the window 2026-06-27 → 2026-08-22 (16,448 entries, 1,201 events, 368 of them Crystal):
| Event size | Crystal entries | Corpus entries | Crystal share | |---|---|---|---| | 64+ players | 59 | 2,356 | 2.50% | | 32–63 players | 135 | 5,084 | 2.66% | | 16–31 players | 126 | 4,270 | 2.95% | | Under 16 players | 48 | 4,738 | 1.01% |
Flat across the board, and lowest at the small events. Whatever is driving this, it isn't kitchen-table overflow.
"It's five grinders registering the same list every weekend." Those 368 entries come from 225 distinct players across 247 distinct events. The ten most active pilots account for 57 entries — 15.5% of the total — and 167 of the 225 players appear exactly once. It's broad.
One honest caveat on the absolute counts: six of the 368 rows (1.6%) look like the same human entry ingested twice, where one event got listed under two names — "commanderliga.de – Mythic Intuition" and "Mythic Intuition … (Gelsenkirchen, Germany)" are the same tournament. Every row carries a distinct Topdeck permalink, so this is upstream double-listing rather than our deduplication failing, and it inflates numerator and denominator together, which leaves the shares alone. Read the entry counts as ±2%.
Where we disagree with the other tracker
edhtop16, which runs on Topdeck.gg data, lists Crystal at 102 entries and a 0.33% meta share. We have 368 entries and 2.2% over their own stated window. That's not a small difference and we're not going to smooth it over.
It isn't a date-range effect — we ran our corpus over their exact window to check. It isn't a top-cut-versus-entries effect either. What we can say is that our set looks like the broader one: we cover 1,201 events in that window, and every single one of the nine finishes they cite reproduces in our data, from Kenny Gerry's 7th of 50 on June 27 to the August 22 win. Their share figure appears to be computed against a larger denominator than the 102 entries they report next to it, but without their exact event-inclusion filters we can't reconcile the two from the outside. Our number is ours, our counting basis is stated above, and you should weigh the two accordingly.
The list that won
On August 22, Yannik H won commanderliga.de's Mythic Intuition — a National & European Qualifier, 58 players — with Crystal.
Point the DeckCortex engine at it and the shape comes out immediately. Four complete two-card combos, and none of them are the aura:
Sewer-veillance Cam pairs with both Retraction Helix and Banishing Knack — the engine reports infinite artifact enter- and leave-the-battlefield triggers plus infinite storm count. Jeskai Ascendancy pairs with the same two instants for infinite untaps and looting. It's worth noticing that Ascendancy reads "Whenever you cast a noncreature spell" — the exact trigger condition Crystal has. The deck isn't running one engine card in the command zone; it's running the same clause twice.
The bracket-fit model reads the list as cEDH — 10 Game Changers (Rhystic Study, Force of Will, Underworld Breach, Mystical Tutor, Enlightened Tutor and six more), four two-card combos, four pieces of fast mana, five tutors. That verdict is a deterministic threshold model over published constants, not a guess, but the combo count is one of six inputs feeding it — a signal, not the bracket itself. And these four combos are what's in this list, not what's in all 163.
The aura package, meanwhile, is close to mandatory across the archetype: Curiosity in 100% of the 163 lists, Keen Sense and Ophidian Eye in 99.04% each. Nobody's picking one. They're playing all three, because they're the same card with three different names and you want to draw one early.
Not a new idea, a cheaper one
The pattern is ancient. Niv-Mizzet, Parun plus Curiosity is one of the oldest two-card kills in the format, and Spellbook still lists it — our engine flagged both `Curiosity + Niv-Mizzet, Parun` and `Niv-Mizzet, Parun + Ophidian Eye` as one piece short in that winning list. Vivi Ornitier does the same job in Izzet and sits at rank 8 in the table above, one slot below Crystal.
So what did Crystal actually change? Three things. Niv-Mizzet costs — she costs three. Vivi is two colours — she's four, with white for Silence (99.04% of lists) and Enlightened Tutor. And she taps for mana, which the other two don't.
What she gives up is real, though. She's a 2/3, which dies to everything. Her trigger only fires on noncreature spells. And the colourless-spell blind spot above means her draw engine idles through exactly the acceleration that makes turn-two kills possible. She is not strictly better than the decks she's borrowing from — she's cheaper, faster to deploy, and greedier in colours.
What it costs
The commander is $1.18. The engine is not:
| Card | USD | Presence in Crystal lists | |---|---|---| | Crystal, Inhuman Princess | $1.18 | 100% | | Curiosity | $0.98 | 100% | | Keen Sense | $3.26 | 99.04% | | Ophidian Eye | $14.46 | 99.04% | | Underworld Breach | $10.91 | 98.36% | | Mystic Remora | $13.04 | 98.86% | | Silence | $9.27 | 99.04% |
Ophidian Eye is the odd one out and the only structural squeeze in the package — a Time Spiral common from 2006 with exactly two printings to its name, which is how the third copy of an effect you can otherwise buy for a dollar ends up at fifteen. Everything above still sits inside a list carrying Wooded Foothills at $18.43 and a full free-spell suite, so nobody is building this on a budget. The cheap commander lowers the barrier to trying the deck, not to finishing it. We looked for a price spike timed to the adoption curve and couldn't evidence one.
What to watch
Two months in, our corpus has her winning 15 events outright, the largest that 58-player qualifier, and holding a 2.4% share across two consecutive months. That's a real deck, not a spike — but "holding" is doing work in that sentence, and the honest read is that she's arrived rather than that she's still climbing.
The number to watch is the conversion rate as its sample grows. At 33.86% off 127 entries she's converting at roughly Blue Farm's rate, which would be a genuine story if it holds at three or four times the base. If it drifts toward the mid-twenties as more pilots pick her up — the pattern Vivi shows at 24.11% — then what we're looking at is a very good cheap deck rather than a top-tier one.
And if you want to know where the next one of these comes from, watch the auras, not the commander. Curiosity was printed in Exodus in June 1998 and nothing about it changed this summer. What changed is that somebody printed a three-mana body that deals damage to three players at once and taps for four colours. The other half of the combo has been sitting in bulk boxes for twenty-eight years.
Data: DeckCortex cEDH tournament corpus, 112,803 canonical entries all-time; 30-day window 2026-07-27 → 2026-08-26 (n = 6,887 entries), explicit window 2026-06-27 → 2026-08-22 (n = 16,448 entries, 1,201 events). Share = share of tracked tournament entries, not field share; conversion = top-cut entries ÷ entries with known standings, n shown. DeckCortex tracks no match win rates. August 2026 is a partial month (1–25). Card text and colour identity from Scryfall, rulings from Gatherer via Scryfall, all pulled 2026-08-26. Prices are Scryfall paper market aggregates, not a live TCGplayer feed. The two-card combo count is one of six inputs to the bracket-fit model — a signal, not the final 1–5 bracket. External figures from edhtop16 are labelled as theirs where cited.