GP Charlotte didn’t just meet expectations. It set a new bar for Modern tournaments. We saw Collected Company and Splinter Twin. We saw Cryptic Command and Kolaghan’s Command. We even saw Ad Nauseam (whoa), Nourishing Shoal (wait, what?), and Lantern of Insight (…wut), and all in the Top 16. GP Charlotte was one for the ages and the heralding event to a new chapter in Modern’s history. This tournament will shape the format for months to come.
For the past few months, we’ve tracked the post-Pod/Treasure Cruise metagame. Charlotte is the culminating event of that trajectory, and what a metagame it has shaped up to be. This article is a roundup of the metagame changes and developments we saw at GP Charlotte and how the GP results will affect the format going ahead. It won’t be a GP-wide retrospective: you can expect that later in the week. But it will separate the important format movements from the flashes in the pan, analyzing the day 2 metagame, the T8/T16, and comparing those to the metagame results from before the GP.
Metagame Comparisons: Overall, GP Day 2, GP Top 8
When reviewing a GP, it’s important we don’t just focus on Top 8/Top 16 numbers, or even day 2 numbers. In almost every instance, these numbers alone do not either reflect the metagame going into the GP, or predict the metagame after. GP Charlotte’s day 2 metagame, and its top decks, will likely be no exception. The archetypical example of this was Abzan around the February Pro Tour. The PT metagame saw 25%+ of the field on Abzan. But after the PT ended, Abzan’s shares declined precipitously. To adjust for all the factors at play in metagame development, we need to compare the current GP numbers to the metagame stats from before the GP. We also need to extend this comparison into the Top 32, using deck prevalences there as another indicator of success. This method helps us assess whether an observed effect happened because of a particular event or was already happening before that event. It also helps us decide whether an effect was part of a longstanding trend or a one-time anomaly. Finally, it lets us compensate for a lack of day 1 statistics: we don’t know the day 1 to day 2 conversion rate, but the metagame-wide trends help us correct for that.
For pre-Charlotte data, I’m using the 4/1-5/1 metagame numbers and a slightly modified version of the 5/1-6/1 Metagame Update numbers. These stats include all events leading up to the GP, including SCG Columbus and a variety of events from the week leading up to Charlotte. As with all Nexus data, numbers are taken from the Top Decks page. For the GP day 2, I’m just using a modified listing from the Wizards page, reclassifying some decks to better align with our site’s data tracking (e.g. “Merfolk” being separate from “Blue-White Merfolk”). Finally, I’ll look at the top 32 decks of the event. There is a huge data pitfall here which we need to acknowledge up front. The 16th-32nd players all had 36 points, but so too did the 33rd-52nd players. We are missing those 16 decklists, however, despite their finishes being basically the same. So our dataset is skewed towards players with slightly higher breakers.
The table below shows all day 2 decks with 8+ appearances (i.e. all decks with above average appearances on day 2). It then gives you prevalence for those decks in four periods: 4/1-5/1, 5/1-6/10, the GP day 2, and the GP top 32. Finally, the last column gives the slope of the best-fit line through these datapoints (the slope coefficient, M, of the linear regression, for you stats folks). This measure “M” gives us a sense of how consistent the deck trends were over time.
Metagame Trends Over Time
|GP Day 2|
The table is auto-sorted on GP day 2 prevalence, but you can play around with the arrows at the top to sort it on other values. In interpreting M, remember that larger values indicate a consistent increase since April. Values around zero suggest minimal change, and values in the negatives suggest a downward trend.
Looking over the table, we can loosely divide the decks into three categories. First are the decks with upward-trending slopes, ones that increased consistently in all periods of time: Abzan Company is an obvious winner here, with consistent upward movement since April. Second, we see decks clustering around zero, which maintained their status quo both before and during the GP. This includes format staples like Burn and Affinity, as well as decks people metagamed against like Amulet Bloom. Finally, we see the declining decks, ones that were falling out of the top-tiers before the GP and only had their decline finalized during Charlotte. Abzan and Infect are big here, two decks that seriously underperformed at the GP and have been underperforming for months. That said, this table is clearly missing some nuance. Grixis Twin did not have the best T8 performance, even if it sent a lot of players to day 2 and even if it had a big metagame share going into the event. So why is it ranked near the top of the table when a deck like UR Twin did better? Or what about Burn, a deck everyone prepared for but still managed to succeed in the face of so much hatred?
The problem with the above table is that it doesn’t place enough weight on the GP results themselves, which we know are going to heavily effect metagame shares going ahead. This means we need to make some adjustments.
If there’s one thing we can count on in Modern it’s the hype factor. Decks that get in the T8 tend to be better received than decks in the T16, even though there is often little statistical difference between the two. They may legitimately be better, but even if they aren’t, the community tends to put heavy emphasis on T8/T16 finishes more than just day 2 metagame share. Here’s one way we can try and account for the “hype” factor. In this next table, decks in the T8 are weighted twice as heavily as decks in the 9th-16th, and decks from 17th-32nd are weighted half as much as the 9th-16th decks. I then take the average between this adjusted T32 score and the old day 2 prevalence to get a “hyped” day 2 share. This reflects player interpretation of the day 2 numbers, not the actual numbers themselves. If a deck does poorly in the T8 despite sending lots of players to day 2, it looks bad and people undervalue its day 2 performance. But if it overperforms, people take notice and overvalue the day 2 stats. The table below reflects these adjustments. Note the new slopes with the hype adjustments.
Metagame Trends Over Time: “Hype” Adjusted
|GP Day 2|
To some extent, “hype” is a bit misleading here. It’s not just that the decks are interpreted as doing better because players are irrational hype maniacs (which, to some extent, they are). Rather, a deck’s overperformance in the T8 can be a legitimate sign of strength, which draws players to those decks and gives context to their day 2 shares. By contrast, a deck that sent a lot of players to day 2 but then failed to send many to the T8/T16 would be questioned by the community.
This table gives us a slightly different metagame picture than the raw numbers did above. Here, the clear winners are Abzan Company, Naya Zoo, and UR Twin, all of which exceeded expectations and put up some huge finishes. You can expect these three decks to see a lot of play in the coming months, and you would be wise to prepare to beat them. Grixis Twin is also still a winner here, although the table is probably overstating its importance in the metagame going forward. Some Grixis players are likely to return to UR Twin, which will further decrease its metagame share. That said, there are certainly pilot and deckbuilding factors at play in the difference between the decks: it’s no surprise that the untested Grixis Twin list didn’t do as well as the time-proven UR Twin build. That list also probably had more inexperienced pilots on it, all of whom wanted to play the next best thing. Based on this, I don’t expect Grixis Twin to fall off the map completely. It will remain a 5%-6% deck in the format in the months to come, poised to make a big comeback if it gets a big finish. But UR Twin will still remain the Twin deck of choice at over 8%-9%.
Some losers in the hype-adjusted table are Grixis Delver and Jund. These decks did not do as well as they could have done, despite lots of hype surrounding their strengths. That said, it’s critical to note that Abzan did absolutely horribly compared to both these decks. Abzan was already at the bottom just looking at the raw numbers, but the hype-adjusted table really shows how far the deck has fallen. This is a critical datapoint in understanding the underperformance of Jund. Based on the GP, players won’t look at Jund and see a bad deck. Rather, they’ll look at Abzan and see a worse deck, steering them towards Jund as the BGx deck of choice. So although Jund didn’t live up to the hype, Abzan failed much harder. This suggests Jund’s share will actually continue to rise in the coming months, or at least stay where it is as Abzan’s continues to fall. Expect Jund to be at around 7%-8% and Abzan to fall into the same range (or lower). The metagame just isn’t great for the BGW mages, and we now have multiple months and events to confirm that. As for Grixis Delver, there was always a good chance that its metagame share was artificially buoyed by MTGO hype. Now that Delver has been discredited at the GP, expect its share to fall into the 4%-5% range as Grixis Twin (and Grixis Control) become the Grixis options of choice.
There are a variety of other conclusions you can draw from the data. Infect continues to fail and will probably fall even further after Charlotte. RG Tron and Amulet Bloom were overhyped going in and players over-prepared for them at the event themselves, but are likely to remain viable based on their pre-trends and their consistency even in the face of hate. Merfolk may not have done that well at the GP, but that deck has always enjoyed a modest metagame share and will continue to even after Charlotte. In all these different cases, the key is to check the hype-adjusted scores against our general understanding of the decks. That’s how we can predict what will come next.
One final point on this: expect a lot of Collected Company in the months to come. This was the big breakout success story of the GP, and players are going to take notice and go even crazier around this card than they already are. We will also see more players take this deck seriously, which will temper the deck’s rise as players build sideboards (and even maindecks) to crush creature-powered Company decks. Company decks are not going to hit some insane, Deathrite Shaman level of prevalence as the format evolves. But they will keep rising and we can expect to see at least 1-2 of them in the tier 1 decks. Elves and Abzan Company are poised to each be 7% or so of the metagame. The other tier 1 decks will also start to take shape around these decks, so expect more stuff like Tron and Amulet Bloom. I predicted a Company success in my GP Charlotte article last week, and I’m excited to see this deck do so well. Company was exactly what a variety of decks needed, and Modern is a more diverse format after its inclusion.
Processing Weird Decks and Outliers
Before wrapping up, I want to say a brief word about “weird decks”. In my last article, I said to watch out for weird decks at the GP and boy, did Charlotte not disappoint. Who would have thought we would see not only the Ad Nauseam combo deck, which is at best tier 3, but also the Griselbrand/Nourishing Shoal combo (a new version of a tier 4 deck), and the Lantern of Insight/Ensnaring Bridge prison deck (basically a kitchen table deck before Charlotte). Elves is also a possible “weird deck” inclusion, but less so because it at least had MTGO performances going into the GP. When dealing with decks like these, it’s hard to know whether they are the real deal or if they are just temporary successes that no one else will replicate. Statistically speaking, I always lean towards the flash-in-the-pan interpretation. Fabiano’s Sultai Control never took off after SCG Baltimore, which suggests a much more niche control deck like Fateseal/Top Control couldn’t either. But then again, never underestimate hype or a player’s desire to try something new, which are two factors that could drive these “weird” decks to succeed. Also, these decks might legitimately be powerful independent of hype.
I’m going to discuss all these decks more in articles later this week, so for now I’ll just say this. Rogue decks can definitely succeed in Modern and it’s good to get ahead of those trends. But not all rogue decks turn out to be good, and if I had a dollar for every time a rogue player has told me their deck has somehow broken/solved the metagame, I could have bought Pascal’s Tarmogoyf. The key to separating the five-minute-spotlights from the real-deals is in assessing deck power and deck weakness. If a rogue deck has glaring weaknesses (e.g. folds to graveyard hate), then it’s probably worse than all the higher-tiered decks that also “fold” to graveyard hate, such as Living End. Similarly, if a rogue deck doesn’t play powerful cards, then it’s probably worse than the higher-tiered decks that are playing powerful cards, such as Amulet Bloom. So in assessing these decks, don’t look at them in a vacuum. Compare them to existing decks and see how they stack up.
Modern after GP Charlotte
We’ll have a lot of post-GP analysis in this coming week, so stay tuned for more reflections and content on Modern’s coolest event in years. With so many decks and metagame trends at stake, there’s just so much to discuss. Charlotte turned a page in Modern history, and I fully expect the format to be forever changed as we go forward. But as is often the case in Modern, our format will still have a number of the same old decks doing the same old thing, which is exactly where we want to be as a nonrotating format. It’s a good combination of innovation and consistency, all harmonizing together in a diverse format where anything is possible.