Paupergenesis 2026
Surviving metagame
Archetype share among players still alive
“Alive” means X-1-1 or better after that round. Each vertical slice adds up to 100%.
Survivorship vs. funnel
Does the selection rule change the metagame?
Why look at the tournament this way?
A survivorship view follows the tournament forward, one round at a time. After each round, I keep everyone at X-1-1 or better and remove anyone who falls below it. I am using that record as a rough stand-in for still having a realistic path to prizes. Players can keep playing after they cross the line, and some improve their final standing; they are simply out of this chart from that point on. I recalculate the archetype shares after every cut.
I had been listening to a lot of Pauper podcasts while getting ready for the event, and somewhere in that stretch I started thinking about how a large tournament changes as it goes. At the start, every deck is present. By the middle rounds, successful players are increasingly paired against one another. Near the end, only a small group still has a realistic shot at prizes. From a player’s seat, those can feel like three different metagames. A deck that fills the bottom tables might disappear from the matches that decide prizes, while a smaller archetype keeps showing up around the contenders.
I also had a personal reason to think about it. I won my first three rounds at Paupergenesis, then got snuffed out. The final standings show where I ended, but they skip over the part I remember: three rounds of being undefeated, the decks around me while I was alive, and the round where that ended.
I could build this because Paupergenesis was hosted on Melee.gg, which publishes individual matches, round-by-round standings, and decklists. With that data I could reconstruct every opponent, each player’s record after every round, and the exact point where someone fell below the line.
The funnel in my first Paupergenesis report starts with the final standings and compares the field with the Top 64, Top 32, and Top 16. This chart moves forward through the event and asks what players would have seen among the people still alive after each round. The X-1-1 line is only a consistent estimate of contention; I am not trying to reproduce the organizer’s prize math.
The two views start with the same field and end with the same Top 16, so any difference has to appear in the middle. I wanted to know how large that difference became. It grew from 7.3% after Round 2 to 12.6% after Round 4, jumped to 23.1% after Round 5, and peaked at 32.3% after Round 6. Round 6 was the surprising one. There were 31 survivors, almost the same headcount as the funnel’s Top 32 checkpoint, but the archetype mixes were furthest apart. The gap fell to 15.8% after Round 7 and disappeared after Round 8.
Here is how to read that 32.3%: in a group of 31 players, roughly 10 deck slots would need a different archetype before the two mixes matched. Mono Blue Terror held three seats in the final Top 31 and none among the X-1-1 survivors. Mono Red Rally had three survivors and only one seat in the final Top 31.
I then checked whether a 32.3% gap was unusual for two groups this small. I shuffled the archetype labels among the same players 10,000 times without changing who belonged to either group. Those random versions averaged a 29.9% difference, and 44% of them produced a gap at least as large as the real one. None of the rounds separated clearly from chance. For now, the chart describes the tournament I played. It is not evidence of a hidden metagame. I would like to repeat it across more large events and see whether the three phases keep appearing.
Composition distance by round
The share of deck slots that would need a different archetype for the two groups to match. The dark tick shows the average after randomly shuffling archetypes among players.
Chapter two · Genesis vs. Geddon
Was Paupergeddon more settled?
TL;DR: No, but the data is interesting.
When I pulled up the Paupergeddon Summer 2026 lists, they looked more samey than the Paupergenesis lists. Mono Red Madness was what made me notice. At Genesis I played two Sazacap’s Brew and two Thermo-Alchemist; in the much larger Geddon field, both choices had nearly disappeared. Maybe Geddon players really were converging on stock lists. I wanted to check whether that was true or whether a 132-deck Madness field was fooling my eye.
To put a number on it, I compared every pair of main decks within each shared archetype using Bray-Curtis distance. Since these are 60-card decks, the result is easy to describe: roughly how many card slots differ between two randomly chosen lists. I kept the 15 archetypes with at least five decks at both events, counted snow-covered basics as their ordinary counterparts, and combined the same functionally identical Pauper cards used by my deck explorer.
The effect was small. Paupergeddon lists were closer together in nine of the 15 archetypes, with a typical change of only 0.2 card slots. White Weenie tightened up much more, dropping from an average distance of 13.4 slots at Genesis to 7.3 at Geddon. Dimir Terror fell from 10.1 to 5.7. Jeskai Ephemerate moved the other way, rising from 6.5 to 10.1 slots, as did Spy Combo from 5.4 to 9.0.
Madness itself barely moved: 3.2 slots apart at Genesis and 2.9 at Geddon. Still, the cards that caught my eye really had faded. Sazacap’s Brew appeared in four of 26 Genesis lists (15.4%) and only three of 132 Geddon lists (2.3%). Thermo-Alchemist appeared twice at Genesis and zero times at Geddon. Geddon also produced a new branch of the deck, with Melded Moxite in 27 lists after appearing in none at Genesis. At Geddon, 76% of the lists were within two slots of the event’s central list, up from 69% at Genesis. Most players stuck close to the center, with Moxite as the largest organized departure.
The UMAP below arranges the decklists in two dimensions, with more similar lists generally appearing closer together. It uses the explorer’s count-plus-IDF weighting, which gives unusual choices enough weight to pull a variant away from the shared core. Point spacing is only a visual guide; the card-slot numbers above come directly from the full decklists. Pick an archetype to see where the events overlap, where they split, and which cards moved.
Average main-deck distance
Approximate card slots separating two lists from the same event and archetype.
Decklist neighborhoods
UMAP · main deck · counts + IDF
Cards that moved
Share of decks playing at least one copy.
Do stock lists survive longer?
TL;DR: No.
Once I had a stockness score, I wondered whether staying close to the central list actually helped. I measured each of the 170 Paupergenesis decks against the center of its archetype, then ranked it only against decks with the same label. Conventional Madness lists were compared with unusual Madness lists, for example, so the result would not turn into a comparison between strong and weak archetypes. I reused the X-1-1 survival score from chapter one.
I found almost no relationship. The correlation between stockness and the last round alive was -0.02, on a scale from -1 to 1. When I shuffled survival results among players in the same archetype 10,000 times, a relationship at least that large appeared 84% of the time. The most stock group stayed alive for 3.2 rounds on average, compared with 3.0 for the least stock group. After Round 4, 36.5% of the most stock group was alive versus 32.6% of the least stock group. I could not get the curves to separate consistently.
One tournament is not enough to rule out a small effect. The score also treats every unusual choice the same, whether it was a smart metagame read or a bad pet card. Pilot skill, matchup order, and familiarity with the deck are mixed into the result too. At Paupergenesis, though, the conventional lists did not survive longer.
Survival by distance from stock
Share of each within-archetype stockness group still alive after every round.