Another month gone in quarantine. At least, I think it's been a month. Time is steadily losing all meaning to me. What even is this year, and what was I talking about? Oh yeah, it's August now, so I need to take another look at Modern's metagame. Fire up the spreadsheets and break out the data; let's see what's changing!

The banning of Arcum's Astrolabe means that I can't use the totality of July's data. Anything from before July 13 reflects a Modern that no longer exists. Therefore, I'm only going to sample post-ban results, meaning everything from July 13 until July 31. I also can only sample those events that have actually had their data published, which I think is obvious, but it's not every event that has happened. I've checked: more Preliminaries and Challenges happen every week than Wizards reports. I don't know if this is Wizards refusing to publish everything or if events are failing to fire. In any case, a craftsman must work with the tools he has.

July Metagame

In a departure from my usual method, I'm doing as close as is actually possible to our earlier-style metagame update. I can't actually do an update exactly like we did in the early days for several reasons. However, I've adapted it to the current realities and am going to keep refining the method. It's not perfect, but it's a more statistical method than I have been using.

The first and primary reason is that the old system simply isn't needed right now. It used a points-based weighting system in effort to smooth out the discrepancies between paper and online Magic. Paper results were given extra weight because they were more reliable. MTGO may be more accessible, but it also has a smaller player base. Millions play Magic, but only a small fraction are willing to maintain a digital collection. Those that do play MTGO tend to play lots of events and show up in results more often, leading to more outlier results. That just didn't happen in paper, so we added weight to the more accurate results. There are no paper events happening, just MTGO events. Thus the weighting system is meaningless.

The second reason is that even if I were to use the system, it wouldn't accomplish anything. Tiers were determined using a points system, rewarding decks that placed highly rather than those that were simply popular. That points system was designed for a world with Grand Prix and Pro Tours. We don't have those now, and coupled with a lack of paper results, the pointing system didn't actually change anything. When I did June's metagame, I tried the points and it was the same as doing it by prevalence. So I dumped the points to save myself some work and readers some confusion.

The Tier List

That said, the statistical method of determining Tiers is still valid. Take the average result, and anything that does at least twice the standard deviation above that is Tier 1. Between one and two deviations is Tier 2. From one deviation to the average is Tier 3. I recorded 55 different decks in my sample range, representing 350 results. The average number of results was ~6, with a standard deviation of  ~7. So any deck with at least six results made the tier list. 13 results makes Tier 2 and 20 or more is Tier 1. I'm doing the entire list in one table because Tier 1 is very small.

Four decks crossed my threshold to make Tier 1. Eldrazi Tron is sitting atop the metagame by a comfortable margin. This is followed by newcomer Izzet Prowess, which has blue for primarily Sprite Dragon and Stormwing Entity. Old standby Ponza and a resurgent Jund round out the top. The old guard may celebrate, for Jund is apparently Tier 1 again.

Mono-Green Tron is also back, just missing the Tier 1 cutoff. It's joined by Dredge, UW Stoneblade, and Goblins in Tier 2. It's interesting and convenient that there are only four decks apiece in the top two tiers, but that's how the data fell. UW Control and Rakdos Prowess lead a very diverse Tier 3, filled with old stalwarts and Temur decks. There's a lot to consider here, so let's pull it apart for easy digestion.

Bant 'Snow-More

First and foremost, Bant generally and Bant Snow specifically are gone. No bant decks made my Tier list. In fact, Temur Reclamation and Scapeshift are the only Tiered decks that played snow-basics and Ice-Fang Coatl. And that wasn't universal. Along with snow's decline, Uro, Titan of Nature's Wrath has taken a beating. Again, the Temur decks still ran him, but the metagame as a whole has moved on. I think it safe to say that banning Arcum's Astrolabe worked. Snow isn't completely dead, however, and both Bant and Sultai Snow did put up results. Not enough to be Tier decks, but enough to tell me that they're still reasonably competitive. Sultai did slightly better, likely thanks to Dead of Winter; without Astrolabe, players need another compelling reason to play Snow to make the spin worth it.

On the same note, the metagame has been throughly shaken up. Eldrazi Tron and Ponza are the only decks still in top tier slots from before. Big threats, acceleration, and disruption are a good combination. This isn't entirely surprising in context. Before the ban, Snow midrange decks reigned and got all the press. Eldrazi and Ponza are both decks that are good at out-muscling midrange. This may also be a factor in Mono-Green Tron's resurgence. It's a master of going over the top of decks, but struggles against counterspells. The main counterspell deck took a hit, so now Tron's back.

Colorless Menace?

On that note, Eldrazi Tron has a very commanding metagame percentage relative to everything else. This shouldn't be too surprising for the above-mentioned reasons, but there is something else to consider: Eldrazi is cheap online. Eldrazi Tron has been one of the cheapest good decks on MTGO for years. It's not the cheapest, but it's never too far out of contention, making it a very solid investment. Thus, it always sees a lot of play and does well online, which is a big asterisk on its results. However, this time I think E-Tron's position is justifiable thanks to the next entry.

...Or Red's Prowess?

An unexpected entry is Izzet Prowess. Prowess doing well is no surprise; 2020 has been a good year for tiny hasty red people. That Izzet supplanted Mono-Red to such a degree is surprising. At least some of that can be put down to deck-of-the-month syndrome. It's new and plays new cards, so there's excitement for the deck and higher adoption as a result. However, Dragon and Entity genuinely seem like upgrades to the deck, so this might be no accident.

Still, to link back to E-Tron, Prowess is simply the most popular archetype on MTGO right now. I recorded 53 different Prowess decks, the most popular being Izzet and Rakdos. That means Prowess represents 15% of the metagame, far surpassing any other strategy. This in turn helps explain E-Tron's prevalence. Prowess is explosive thanks to all its one-drops. Etron is really a Chalice of the Void deck, and Chalice is a nightmare for Prowess. Izzet can fight Chalice better thanks to Dragon's buffs being permanent, but like any velocity deck when the motor spins down, Prowess just sputters. So long as this Prowess saturation continues, Etron will be very well positioned.

Also, UW Control is Back

The final thing to note is that UW is back. It has been hanging around for some time, but the Bant decks kept overshadowing everyone else thanks to all the now-banned cards. With Bant finally suppressed, UW is rising again. Not only that, there's a choice of strategies. UW Stoneblade was the more popular deck in July, but straight UW Control was a fine choice. Worth noting, if I'd combined the two, then UW would be Tier 1. Don't sleep on this deck.

Worth noting: UW Control has a pretty narrow range of spells, while Stoneblade is all over the place. There is a standard package of Cryptic Command, Path to Exile, Mana Leak, Opt, Jace, the Mind Sculptor, and Teferi, Time Raveler across all flavors of UW. The distinction is win conditions. Control tends to have few creatures and relies on planeswalkers to win. Stoneblade, naturally, plays more creatures, but there's no consistency. Some are closer to UW Tempo with Restoration Angel and value creatures. Some are closer to control and use Shark Typhoon. The central interaction package is proving to be very flexible for however you choose to play UW.

The Weekly Discrepancy

However, this metagame only emerges when you look at the aggregate data. On a weekly basis, things are not so clear. And E-Tron starts to look like an outlier. And other decks start to look more like real contenders. This is the problem with only using MTGO results: there are wild fluctuations, and flavor of the week is very, very real. As a result I want to reiterate that my results only apply to MTGO, and there is no "real" Modern metagame right now. When paper eventually comes back, there will be more stability and reliability to the metagame results.

Week 1

Case in point: consider how things looked during the first week post-Astrolabe. There was no data to go on, and so it reflects the rawest take on the new Modern. And the top tier was figured out, even if nothing else was. I'm going to include the full list of every deck for each week. This is partially to settle any questions about where X deck ended up, but also so that everyone can see the raw data I created the tier list from. And if they want to challenge my work, they have the data to do it with.

The top four decks the first week are July's Tier 1. They're not in the same order, but they're all there. And on top by a decent margin considering the overall sample size. Past that, the whole of Modern is in flux. The mid-level decks are nothing like the final Tier 2, and there's a wild assortment of singletons. Which is about what I'd expect of a brewers metagame post-ban.

Week 2

The week of July 19 was the week that made E-Tron the top deck in July. However, it did so in a way that makes it look like an anomaly.

That is a huge spike in placings, from 7 to 21. Especially when looking at the other decks, which are all doing about the same as they did the previous week. Week 2 had more results reported than the other weeks, but that doesn't seem to have been a factor in E-Tron tripling its presence. I don't have a means of explaining this spike. Mono-Green Tron was also up, so it may have been a good week for Big Mana. The catch is that Amulet Titan plummeted from 4 to 1, so if it was Big Mana's time, it was only for the colorless crowd. This smells like a popularity-driven spike, and therefore an outlier and not indicative of anything. However, I have no way to verify if that's true.

Week 3

That week 2 spike looks especially suspect when moving to last week's results. And then adds more side-eye to the conversation when looking at Izzet Prowess.

E-Tron is back to previous levels, while it was Izzet Prowess specifically that spiked. No other Prowess deck spiked, so it was something to do with the Izzet version that week. The spiking of the two decks that ended up being the top of the metagame is particularly suspect when looking through the rest of the data. Everything else that eventually made Tiers 1-2 was very consistent. Were it not for those spikes, the top 2 decks wouldn't have been on top by the margins they were and Tier 1 would have been a very level playing field. Thus, I regard the top decks with suspicion.

What it Means

There's no bones about it, Prowess variants were the most popular decks in July, followed by Eldrazi Tron. E-Tron is suspect for many reasons, but I can be unequivocal about Prowess. Even without Izzet's spike, the archetype would have beaten everything else by a good margin. That's what I would prepare against first and foremost. Outside of that, Modern looks like it's in a good, diverse place again. However, we'll see how it develops.