The Most Massive Modern Metagame/Matchup Analysis Yet

I’ve been looking at the data from Modern GPs for a several months now, a saga spun across seven articles, roughly 13,500 words, and a number of infographics. It’s probably a good idea to start this most recent look with a recap of the pertinent conclusions reached earlier.

Topping the list is how little the Modern-playing populace at large is swayed by tournament results, or how long it takes them to adapt. For instance, even as early as GP Portland in December, data showed, or at least suggested strongly, that Izzet Phoenix was the best deck in the format. Back then, Izzet Phoenix was the fourth or fifth most popular deck, behind Spirits, Tron, and Burn. Its metagame share grew, but at GP Oakland in January the most popular deck remained Spirits, while at GP Toronto in February the new most popular deck was Burn. Throughout it all, Izzet Phoenix exhibited the highest win rate at the most reliable sample size, higher even than KCI’s.

Burn itself is the prime example for the exact opposite conundrum. Data showed that Burn wasn’t good in December. Somehow the deck became more popular in January anyway. Then, with the addition of Skewer the Critics and Light Up the Stage, Burn increased in popularity once again in February. Throughout all of it, Burn consistently performed well below average. But Burn is hardly alone in that camp. Spirits and Tron didn’t do particularly well either, but retained a metagame share that their win rates no longer justified.

The other constant about Modern is that only one archetype ever breaks the 10% threshold. At tournaments that don’t require an invite, it’s a given that no more than a single tenth—or ninth—of the field agrees upon a deck. This variety serves as Modern’s biggest boon and bane. It makes for an enjoyable play experience and for an enjoyable viewing experience. At the same time, it makes effective metagaming and effective data analysis very, very hard.

For example, notable Magic experts name Death’s Shadow and Dredge as some of the format’s top decks. But both decks’ metagame shares and their performance fluctuated so wildly across the most recent batch of GPs that it’s impossible to substantiate any such claim. A somewhat notable exception, Hardened Scales continued to perform well even in the face of dwindling player confidence.

Thankfully, all of that is going to change, right?

A Black-Bordered Lining on the Horizon

Modern Horizons will surely change the face of Modern. More relevant, though, and more immediately impactful, may be that the next Mythic Championship will feature Modern. For a long time, the best players in the world could treat the format much the same way they treat Legacy: a nice option, but an option. Now, for the first time in while, there’s again a need to figure things out for real, and a necessity to get to the bottom of things.

We should expect Modern to become a more serious business in the run-up to the Mythic Championship, as players will want to test for the most important tabletop event of 2019’s second quarter. A more serious deck selection process can only aid the number crunching, and vice versa.

Let’s get to it then, and look at the numbers for Grand Prix Los Angeles …

Metagame in the City of Angels

Going by creature type, the tournament was clearly misplaced. It should have taken place in Arizona’s most populous city rather than California’s. Then again, one is much larger than the other, and large samples are exactly what the doctor (Frank Karsten) ordered. The players here also were real angels in that they gave me the biggest data set of deck information I ever was able to work with, both in absolute numbers and percentage-wise.

For GP Portland I knew what decks 971 players, 55% of the total attendees, played. For GP Oakland, this was 693 and 61%, respectively. Grand Prix Los Angeles set a new record at 1,029 known decks, equaling a bit more than 65%. High absolute numbers are great because they translate to much more conclusive win rates across the always same 15+3 rounds.

High relative numbers are great because they allow us to make more accurate inferences as to the full field composition. They also produce way more matches where both players’ decks are known, which yields more accurate results for matchups between archetypes. Let’s start with the first. Here are GP L.A.’s 28 most popular decks, based on our >65% sample:

 

Finally—the Phoenix alighted on the top spire of town. This also marks the second time now that Dredge placed within the top 5, and the highest share for Mono-Red Phoenix seen to date. Burn and Tron continue to cling to the top, inexplicably, while Spirits has fallen far from its 10% perch just two months prior for good reason.

Win Rates in the City of Angels

It is customary to look at the rates at which archetypes advanced to the second day. But actual win percentages must provide better insight, right? So let’s bypass the one and advance straight to the other:

Above I mentioned that Izzet Phoenix showed one of the highest win rates at previous GPs. This trend didn’t continue in Los Angeles, and it seems that the metagame has indeed begun to turn somewhat hostile. We see evidence in the huge gains for Dredge, which is said to be one of Izzet’s worst matchups. But when evaluating these numbers, we also have to take into account that a larger population gravitates to more average performances almost by default. The main stream usually finds itself somewhere in the middle.

Exceptional results, on the other hand, are the domain of the few rather than the many, and we need to take theirs with a grain of salt. For example, the percentages for Hardened Scales, Amulet Titan, Whir Prison, Infect, Humans, and Eldrazi look a little too good to be true because they are way more likely not to reflect actual truth. Some of them may be real, but the sample sizes—of 215, 208, 83, 88, 257, and 139 matches, respectively—don’t hold up to Izzet Phoenix’s whopping 806 matches.

That doesn’t mean we should ignore these results. Whir of Invention Prison has shown quite a bit of promise before. Humans and Amulet Titan were the two decks that advanced to Day 2 at the highest rates in Toronto. And Hardened Scales posted a positive record at each of the past four GPs now. All of them are evidently great, but Izzet Phoenix still has the results too to maintain its position atop the metagame. I wouldn’t expect this to change for a while.

Notable Matchups in the City of Angels

I could tell you that Vampires went 0-2 versus Affinity or that White-Black Tokens split two matches against Burn. Or I could tell you about that one time when Zombies faced off against Grixis Control, and Zombies lost. I just mention this so you know that I dug deep into an enormous cavern echoing with white noise to mine the nuggets of relevant information.

Relevant, I have decided, were aggregate results where one deck beat another in at least 80% of four to ten matches, 70% of 11 to 20 matches, or 60% of more than 20 matches. I’m sure one could calculate a corresponding confidence level to make the approach look more scientific. But I don’t think there’s much benefit to it other than to obscure the undeniable fact that it is a bit of an arbitrary determination. Your mileage may very well vary.

Whir Prison went:

  • 6-1 versus Izzet Phoenix
  • 5-1 versus Green Tron

I’m starting at the bottom of the metagame here, for two reasons. For one thing, it is amazing that a deck piloted by as few as nine out of 1,029 players even generated any notable results at all. Such is the beauty of knowing 65% of the field.

The second reason is that these results are legitimately interesting. People have been talking of Whir’s excellent Phoenix matchup at length, although it’s always nice to see evidence. More surprising is its record against Tron. I only saw one Tron/Whir feature match during the excellent video coverage of GP L.A., and I would imagine that this served as an introduction to the matchup for many.

Between Oblivion Stone, Ulamog, and World Breaker it was a decisive victory for UrzaTron then. I was rooting for the quirky underdog deck (underdeck?) and kept thinking: Please, say it isn’t so. Well, turns out, it may not in fact be so. The above suggests that we witnessed a potentially singular outlier. The deck can tutor for Damping Sphere, Tectonic Edge, and Crucible of Worlds, has multiples of Sorcerous Spyglass and Welding Jar, and Chalice of the Void takes care of Nature’s Claim. I believe 5-1 may be more realistic than the 0-1 we knew about previously.

Switching to the top of the metagame, Izzet Phoenix went:

  • 10-1 versus Zoo
  • 7-1 versus Elves
  • 16-4 versus Spirits
  • 4-1 versus Ad Nauseam
  • 11-4 versus Red-Green Valakut
  • 15-7 versus Red Phoenix
  • 17-10 versus Affinity
  • 21-14 versus Death’s Shadow
  • 10-16 versus White-Blue Control
  • 1-6 versus Whir Prison

Possibly the most notable information we can derive from this list concerns what is absent from this list. Where is Dredge? I know that Dredge beat Izzet Phoenix 16-3 in 19 matches across Port- and Oakland. This time, however, we only have a record of Dredge going 29-21 against the deck. That’s still solidly favorable, but 58% isn’t quite the same as 84%. It seems that Izzet players managed to turn a hopeless pairing into a winnable pairing through improved sideboards, main deck Surgical Extraction, and potentially a better understanding of the matchup.

In other news, the above lists eight positive and just two negative records, although the picture is tainted a bit by the fact how many of them are fringe decks. Against a lot of the major players Izzet Phoenix fared about even, for instance 25-29 versus Tron, 15-16 versus Burn, and 14-14 against Rock.

Green Tron went:

  • 4-0 versus Lantern Control
  • 5-1 versus Merfolk
  • 4-1 versus Red Prison
  • 6-26 versus Burn
  • 6-15 versus Death’s Shadow
  • 1-5 versus Whir Prison
  • 0-4 versus Infect

In addition, Tron beat White-Blue Control 7-3 and lost to Storm 2-5. Neither matches the strict criteria outlined earlier, but both matches results from earlier GPs.

Burn went:

  • 6-1 versus Elves
  • 26-6 versus Green Tron
  • 15-4 versus Spirits
  • 4-10 versus Affinity
  • 9-14 versus White-Blue Control
  • 4-10 versus Red Phoenix
  • 1-5 versus Eldrazi
  • 1-6 versus Bogles
  • 2-15 versus Rock
  • 0-7 versus Amulet Titan

Death’s Shadow went:

  • 5-1 versus Red-Green Valakut
  • 5-1 versus Storm
  • 15-6 versus Green Tron
  • 14-21 versus Izzet Phoenix
  • 3-9 versus Affinity
  • 3-11 versus Dredge
  • 1-4 versus Eldrazi
  • 1-4 versus Red Prison
  • 2-10 versus Spirits

Dredge went:

  • 4-0 versus Death and Taxes
  • 6-1 versus Red Prison
  • 11-3 versus Death’s Shadow
  • 14-4 versus Red Phoenix
  • 3-9 versus Amulet Titan
  • 1-4 versus Ad Nauseam
  • 0-4 versus Devoted Vizier

The Rock had the lowest number of clear matchups relative to the number of its pilots:

  • 15-2 versus Burn
  • 7-2 versus Spirits
  • 1-5 versus White-Blue Control

Spirits went:

  • 5-0 versus Death and Taxes
  • 5-0 versus Devoted Vizier
  • 10-2 versus Death’s Shadow
  • 5-1 versus Affinity
  • 4-1 versus Merfolk
  • 7-2 versus Rock
  • 4-15 versus Burn
  • 4-16 versus Izzet Phoenix
  • 1-6 versus Eldrazi

White-Blue Control went:

  • 7-0 versus Eldrazi
  • 5-1 versus Rock
  • 16-10 versus Izzet Phoenix
  • 14-9 versus Burn
  • 1-5 versus Hardened Scales

Red Phoenix went:

  • 4-0 versus Red-Green Valakut
  • 5-1 versus Affinity
  • 7-15 versus Izzet Phoenix
  • 4-14 versus Dredge

Humans went:

  • 4-0 versus Ad Nauseam
  • 4-1 versus Hollow One

Affinity went:

  • 5-1 versus Red-Green Valakut
  • 9-3 versus Death’s Shadow
  • 10-4 versus Burn
  • 10-17 versus Izzet Phoenix
  • 1-4 versus Amulet Titan
  • 1-5 versus Red Phoenix
  • 1-5 versus Spirits
  • 0-5 versus Hardened Scales

Red-Green Valakut went:

  • 4-0 versus Jund
  • 4-0 versus Red Prison
  • 4-11 versus Izzet Phoenix
  • 1-4 versus Jeskai Control
  • 1-5 versus Affinity
  • 1-5 versus Death’s Shadow
  • 0-4 versus Red Phoenix

And that’s it. There were no more results reaching the same level of unambiguity that aren’t already included above.

TL;DR

Share this

Discussion

Scroll to Top