If you are new here, don’t forget to check our Discord Channel. It’s free for everyone!

I also strongly advise you to accept cookies. They are essential for the website to work correctly. It’s especially important for images and decklists, which will be quite important, considering the amount of data in the article.

Paupegeddon is a wonderful phenomenon, especially in the post-Covid era of Organised Play – a small, grassroots initiative that after years of putting in the work is now gathering hundreds and hundreds of players for the main event. What’s even more astonishing, it happened in the format that was created by the community, grew for years without any support from Wizards and even now it’s not connected to the Pro circuit. Despite these challenges, the last edition that took place in Lucca saw 1076 entries, which is the all-time record for a Pauper event.
I’m fresh back home from my third Paupergeddon. At the first one in Rome last year, our team of Skura (who top 4’d the main event the same weekend), Franek Komsta (also known as Frakom, who top 4’d just last weekend in Lucca) and yours truly won the Trios event. During my second Geddon in Lecco earlier this year, I ended in top 32. And even though my third attempt wasn’t that successful – I lost my win and in on day one – I think that through these three events I’ve seen enough to recognise some patterns in terms of how we’ve analysed the data before the tournament, what type of mistakes we were doing while interpreting them and how the process can be upgraded in the future.

In this article I wanted to share with you my thoughts about how we tried to predict the expected metagame, how it turned out and why the online data was so much different compared to what we got from the large paper event. I’ll also share with you a few ideas how to not fall into that trap again and my metagame predictions for the next few months.

Pauper before Paupergeddon

Pauper has an opinion of a well-established format. After the last impactful B&R announcement a couple of months back (Basking Broodscale, Deadly Dispute, Kuldotha Rebirth banned, HIgh Tide and Prophetic Prism unbanned), the metagame had time to adapt to these changes. It turned out that – according to the data we’ve already gathered online – Pauper is in a very healthy place. There isn’t one best deck and instead, there are a few strategies that are considered top tier, almost twice as many in tier two and a sea of brews that, despite being weaker than the top choices, are capable of winning a tournament.
When our mini team (Skura, Frakom, Jędrek, Chomik, and I) was preparing for the Paupegeddon, we tried to predict the metagame based on the results from MTGO and adjust them to the changes that could happen between online and paper environments. The meta from MTGGoldfish looked like this:

What’s worth noting, there are also other websites with tournament results. MTGTop8.com for example gathers the data not only from MTGO, but from paper events as well. While it may look cool, the local paper metagames may look very different from each other, so it was risky to draw conclusions from them and apply them directly to a significantly larger event. We also didn’t know how high the level of players in these local tournaments was. On MTGO we could assume that grinders create an environment where the level of play is quite competitive.

We haven’t found any evidence that the metagame in a large paper event would change drastically from what we’ve seen online. Rakdos Madness and Jund were ebbing and flowing as the top deck of the format from week to week. In the next tier, we’ve had various red decks, Affinity, Mono U Terror and Faeries. Finally, there were some archetypes that either were underdeveloped and no one knew exactly how to build them properly – Spy and Tron come to mind – and a few archetypes that take the tournament by surprise, but overall they aren’t considered a top tier choice, for example Dredge, White Weenie or Bogles. For me personally, the biggest mystery was how popular would be the archetypes that are underrepresented online due to the amount of clicking required to execute the combo. High Tide was the biggest mystery, Altar Tron was close second. We expected them to be more represented on the Paupergeddon than online.

My interpretation of the Paupergeddon data

The top of the standings

If you haven’t seen the winning decklists from all three events held last weekend – the Main Event: Top 8, Top 16, and Top 32, the Top Pauper Player 2025 and a Paupergeddon Trial – you can find them by clicking the appropriate links. In my analysis I’ll focus on the main event because we got the biggest amount of data to talk about and I’ll just list the archetypes which did well in the smaller events. 

Trial
Decks with X-1 or better result: 2x Jund, 2x High Tide, and a single copy of Elves, Wall Spy, UR Control, Bogles, Altar Tron, and Rakdos Madness.

Top Pauper Player

1st: UW Familiars
2nd: Faeries
3-4: Red Madness and RG Ponza
5-8: Jund, Flicker Tron, Red Synth, and Wall Spy

The main event

The whole tournament was won by Caw Gates – a deck that wasn’t on anyone’s radar prior to the weekend. Pietro Carocci showed his expertise and navigated the deck like a true master – I was watching his games and was amazed how he played around opponents’ interaction and how much value he could generate out of a bunch of Brainstorms, Squadron Hawks, cats, and a bunch of reactive cards. In the finals, he defeated Francesco Fiorenzoni, playing a fairly stock version of Mono Blue Terror. Semifinalists were Franek Komsta on Red Rally and Francesco Foschi on Altar Tron – two decks that weren’t particularly popular and they both attacked the metagame from a slightly different angle. Franek went low to the ground and as wide as possible, capitalizing on the fact that no one was preparing for the go-wide, all in aggro and Francesco chose to go over the top of other decks in the format. In the top 8 we also had another Blue Terror, two copies of RG Monsters and Mardu Synth.

All archetypes from the top 32 looked like this:

6 Jund
4 Mono Blue Terror
4 UW Familiars
2 RG Monsters
2 UR Control
2 Faeries
2 Altar Tron
2 Red Madness
1 Red Rally
1 Caw Gates
1 Mardu Synth
1 Red Dredge
1 High Tide
1 RB Madness
1 Red Synth
1 UB Terror

And in the more macro archetype way:

Terror decks – 7
Nonblue midrange – 7
Red Aggro – 6
UW midrange – 5 
Combo – 3
Faeries – 2
RG Monsters – 2

The bigger picture

Day 1 metagame

This is the data we got from the organizers. According to the data from Pauperwave.com , there were 1076 players.

And Day 2 meta:

(the image from Paupergeddon.official)

While at the first glance the data looks correct and it shows the general trends, Jędrek and I were not pleased with the level of accuracy of the numbers. For example, the metagame breakdown above shows that Rakdos Madness was the worst performing deck and many would believe that it was true. But if you compare it to the day one data, you’ll quickly realize that not all the decks that were posted in the pie chart were shown in the breakdown. For some of them it’s because they didn’t have enough presence to begin with, but there’s a group of decks that performed so poorly that they got relegated to the ‘other’ category. Another issue was that the data wasn’t presented in a particularly accurate way – for example, rounding the average conversion rate to one decimal point while keeping the individual conversion rates as integers didn’t particularly inspire our confidence. That’s why we decided to analyse the data ourselves and try to draw our own conclusions. In my opinion, the results we got have changed the whole narrative of the event. I hope you’ll appreciate the effort!

*Since we didn’t get the information what archetypes were in the ‘other’ category on the metagame breakdown, we had to make a few assumptions to finish the analysis. We got the info about the sixteen archetypes presented, so everything else was packed into ‘other’, including decks that were listed as separate archetypes in day one like Affinity or Jeskai Ephemerate. We knew that they were present on day one and no more than two copies of each advanced to day two (we assume that decks that got three or more representatives in day two were separated from ‘other’ and listed separately). We calculated and separated these 17 slots based on their day one presence between all the decks labeled as “other” in the metagame breakdown. That’s why most decks from the lower parts of the table have a stable 6,22% conversion. We know it’s not ideal, but this is the best we could do. We won’t know if they performed horribly and didn’t get any players into day two or just poorly and sneaked through one or two copies, so please take it with a grain of salt. But we know for sure that they didn’t get more than two copies in day two, so if the deck had a strong representation in day one, it could be a signal that it underperformed heavily.

** Mardu Synth and Dredge ended in eighth and ninth place respectively, so we knew that at least one copy got through. Based on our analysis, they should get something around one copy in day two, so we assumed that those two players were the only pilots of their respective decks in day two. That’s why according to the table Dredge did poorly, but in reality it could have a second copy in day two, and its conversion rate would be close to the average. We still wanted to include them for the sake of thoroughness, but just like the previous category of decks, they should be taken with a grain of salt.

The analysis

Before we start, you need to know that despite Paupergeddon being a massive tournament, the sample sizes of certain less popular archetypes are quite low, so they might not be statistically significant in any real way. But we can notice some patterns or at least spark a discussion that will create a background for further developments.

TL:DR

• Familiars was by far the best deck of the weekend, with Mono Red Synth as a distant second.
• Altar Tron, Terror, WW, and High Tide all did well. 
• Elves, Faeries, RG Monsters, and Red Madness had an average performance.
• Jund, Gardens, and Spy slightly underperformed
• Affinity, Rakdos Madness, and Jeskai Ephemerate had a terrible weekend.

The winners:

• The best performing deck? UW Familiars. 40% conversion rate is no joke. Familiars showed how powerful they are. It’s even more interesting when we take into account that this deck wasn’t on anyone’s radar, at least based on the online data. Familiars is one of the archetypes that are hard to play online due to MTGO limitations. In paper though, it can show its true potential. Note that four out of five players who made into day two ended in the top 32. The deck also won the Top Pauper Player event. Given all that, even though Familiars didn’t end up in the top 8 of the main event, it’s clearly the deck of the weekend. In my opinion, these excellent results happened because the nature of the deck (control-combo deck that makes a lot of steps to progress and eventually finish the game) and the fact that’s very hard to play on MTGO pushed away newer players from it and only very experienced and dedicated players decided to give it a try.
• Mono Red Synth took the metagame by surprise. It looks like no one was ready to stop a horde of small creatures – Breath Weapon and its ilk were not as popular as they were in the Kuldotha era, and even Krark-Clan Shaman’s numbers dwindled in the last few months. It’s worth noting that in this archetype there are multiple sub archetypes and it’s not clear which one is the best. For example, Franek Komsta was playing Burning-Tree Emissary + Rally at the Hornburg with the intention of going as wide as possible, while most lists focused on Kessig Flamebreather, Lava Dart, and a bunch of burn spells. Only time will tell which one will be better moving forward.
• UR Terror (also known as UR Control) had a really good conversion and it secured two slots in the top 32. Even though it has Tolarian Terror, it’s more of a Monarch deck that relies on cheap interaction and Murmuring Mystic as a way to not lose Monarch and draw a lot of cards. The question is whether it’s better than the UW decks that took down both events this weekend?
• Altar Tron, just like UR, got a good conversion rate and two copies in the top 32. But unlike UR, it feels a bit underdeveloped for me. My gut tells me the deck still has room for improvements, so I’m pretty sure we’ll see more of it in the future.
• U Terror was just a solid choice. It got a decent conversion rate and four copies in the top 32. Good deck.

Solid performances:

• White Weenie is in a weird spot. It got a nice conversion rate, but it couldn’t get a single top 32 result. Maybe it was a good choice for the day one meta, but wasn’t that good in the winners’ metagame? 
• High Tide was the biggest mystery going into the weekend. Everybody knew that it had legs, as MTGO data showed it’s capable of winning tournaments, but many thought that it only showed a fraction of its power and the only thing that kept it in check was how tedious it is to execute the combo on MTGO. Since the combo is much more smooth in paper, many were scared that Tide would dominate and get swiftly rebanned. Luckily for us, it didn’t happen. It had a decent conversion and even one player in the top 32, but it was far from dominating. It turned out that there are enough ways of keeping it in check. In my opinion High Tide is in the perfect spot for a pure combo deck – it’s good enough to compete as a solitaire combo, but not too overpowered to break Pauper in half. That’s why I’m a big advocate of keeping it in the format.
• There’s a bunch of decks that either got okay conversion rates or were able to secure spots in the top 32. Elves, Faeries, Red Madness, and RG Monsters are all decent roleplayers that will likely continue to show up. 

Below expectations:

• Jund had a truly puzzling weekend. It got six players into the top 32, so it should be considered a good performer. But with over 15% of the metagame share and an underwhelming 11,52% of conversion rate with the biggest sample in the room, we should question if Jund is as good as everyone thinks it is. In my opinion there’s one thing that keeps Jund in check. It suffers from the metagame being a bit too wide for it – it theoretically has tools to stop nearly every strategy, but it doesn’t have enough sideboard slots to cover them all at the same time, which is a classic midrange dilemma.
  We should also take into account that midrange decks that want to grind their opposition to dust can end up with more draws than the average Pauper deck. And since in most tournaments, drawing is basically losing, it could lead to a decrease in the deck’s winrate. Finally, being the most popular deck in the room often means that the average caliber of its pilots is lower than for more niche decks, as lesser experienced players will often gravitate to a “safe”, popular choice. Because of that, we can attribute some part of Jund’s underperformance to the level of play rather than the power of the deck itself.
• Gardens feels like Jund with extra steps and I don’t see a reason to not play Writhing Chrysalis and Krark-Clan Shaman. Sure, the manabase is better, but the raw power of the deck goes significantly down. Maybe one day I’ll understand the difference, but it’s not today.
• Balustrade Spy combo had a poor showing at the Paupergeddon. I was playing the deck for some time, so I know it has a great potential – having a one card combo that costs only four mana is scary. The biggest problem for this archetype is that no one knows what’s the best build of it. There are five land, full combo versions, six lands with Chrysalis, and I’ve also seen Spy/Wall and even Spy/Elf hybrids! In my opinion the Spy hivemind needs more time to establish its position and decide which way to go is the best for the archetype. Once they settle on one, best build, the win rate should go higher. The results from Trials and Top Pauper Player tournaments show that the Wall build has some potential, so I would start there. 

Disasters:

• Rakdos Madness was one of the biggest surprises of the tournament. Sadly, it turned out that it lacks something to keep up with the format. That was very surprising for me because it had decent results in Pauper Challenges in the last few weeks. Maybe that’s why it underperformed? It was popular, so the rest of the metagame dedicated enough sideboard slots to combat it. For instance, I saw a lot of Weather the Storms this weekend. Maybe all that Madness needs to do is to wait a few weeks for the metagame to shift a bit?
• According to the official metagame breakdown, RB Madness was the worst performing deck. But now we can clearly say that Affinity was the biggest loser of the Paupergeddon. It fell from 3% of the metagame share (which was below expectations in the first place) to basically nonexistence in day two – we assumed it was something around two copies based on statistics, but in reality, there’s a chance it got not two, but one, or maybe even zero players in day two! That’s an absolute disaster. It also didn’t perform well in any of the side events. It’s even worse when you realise that online data showed that it had over 5% of the metagame share and had decent results on MTGO. In my opinion it turned out that Affinity is a worse Jund, so a worse version of a deck that didn’t perform that well either. It’s easier to hate out, and Myr Enforcer is much weaker than Writhing Chrysalis. I just don’t see a reason to play Affinity when there’s a deck that does the same thing more effectively.
• Jeskai Ephemerate is in a similar position to Affinity. It got some representation on day one and it shrunk drastically. I’m not an expert of blue value decks, so I won’t even try to speculate why it happened. Maybe, just like in the Jund’s case, the ability to easily end up with a draw was the reason it didn’t end up converting into the day two? Or maybe it’s just a weaker version of Familiars and Caw Gates, just like Affinity is a worse Wildfire?

Differences between Paupergeddon and online meta – why did they happen? How to avoid making the same mistake while predicting the metagame in the future?

While reading my opinions about the performance of each archetype, you could ask yourself the following question: why the online meta was so disconnected from what happened at Paupergeddon? The format should technically be the same both online and in paper, so why such strong differences appeared? In my opinion, there’s a few solid explanations:

• MTGO meta gets affected by the consequence of online gaming and creation of metagame echo chambers, as well as the illusion of popularity of some decks. For instance, good MTGO players who often grind for a living are playing a lot and have decent results with the one or two decks they pick and keep playing. Since they know how to play, they have above average results and because of that, they get posted online multiple times every week – even though they’re the only one playing them. The perfect example would be LuffyDoChapeuDePalha. They played Grixis Affinity even before the Deadly Dispute ban and they keep playing the deck nowadays. And from their perspective, there’s not a reason to switch decks if they can keep winning with Affinity even if it’s not the best performing deck in the hands of other players. Since Luffy plays a lot, they get posted much more often than the average player and it has major consequences on how the deck is presented on MTGGoldfish. Maybe that’s why Affinity ended up being significantly less popular at Geddon?
• Because of the endless wave of results from leagues and challenges, and the existence of rentals, players often just copy-paste winning decklists. This leads to a stale metagame – outside big events like Geddon we don’t see nearly as much innovation as we should. For example, maybe Familiars should be ‘found’ much earlier, but we as a community were too lazy to actually properly test it in the new metagame?
• Some decks are hard to play online (High Tide, Altar Tron, Familiars, etc.) and because of that they are underrepresented on MTGO. Only people working at Wizards and Daybreak know how often decks that have to play solitaire are losing to time. The data from Paupergeddon, even though it’s limited, should encourage people to revisit Altar Tron and High Tide and learn how to click quickly – both decks are looking good and if you can manage to end games in time, you could get an edge.
• Since Pauper’s vital part is the strong community based around local game stores and organizers, there’s a decent group of players who don’t touch MTGO and rely solely on regular playtesting. There’s a chance that’s a reason why home brews or less popular decks can still do well in such big tournaments. Who would have thought that Caw Gates can win the whole thing? So if your deck is at least decent, you know what you’re trying to achieve and find right tools for the expected meta, you can to some extent ignore the online trends and do your thing. 
• There’s a chance that we try to analyse a very small sample when in reality we shouldn’t focus on it too hard. Even if Paupergeddon Lucca had over 1000 players, if we divide it into thirty or more archetypes, the sample size on some of them would be low enough to question if some decks are really good/bad, or if variance was the key factor in their performance. Screw, flood, mulligans, and bad matchups happen to everyone, and there’s a good chance that if we held another Paupergeddon with the same decks and same players next week, the results would vary significantly. Because of that, I’d treat the data we’ve got more as a suggestion as to which decks are worth trying and which are better left on the shelf for now, rather than the absolute truth.

The future of Pauper

I’ve been playing competitive Magic for over twelve years, and I have a hard time recalling another such cool and balanced constructed format. It reminds me of Modern from the good old days – when you could pick one of the fifteen or even twenty archetypes, master it and win a big tournament. You can play anything: from fast aggro decks, through different types of midrange, tempo, ramp, to combo decks – both creature based and the true solitaire decks of yore, that are few and far between in contemporary Magic. The fact that such diversity exists while nothing breaks the magical 55% win rate is a sight to behold, even if Faeries are close according to what Gavin Verhey said in the last B&R announcement. Paupegeddon showed us that despite the format being very wide already, there are still multiple archetypes that are waiting for their breakout performance – it is solely on us to try and push Pauper’s limits even further. I hope that my analysis will encourage you to not blindly copy whatever won the last Challenge and test underdeveloped decks a bit more – who knows how many hidden gems still wait to be discovered? I know I’ll spend the next few months trying to find out, so wish me luck!