It takes either large samples or an overwhelming tendency to determine which deck is favored in a matchup. Luckily, we find ourselves right in the middle of a stretch of four Modern GPs. Players in Barcelona and Minneapolis together completed about 9,000 matches, so large samples exist.

This article lists all encounters between two archetypes where the aggregate record displayed a sizeable advantage for one of them. To be precise: results at least as extreme as the following would occur with a chance of less than 6% (p<0.06) if the winning deck didn’t have an actual advantage.

Modern Winrates in the Summer of ‘Gaak

Combined Results from GP Barcelona and GP Minneapolis

First up is the archetype that completed the most matches across the two GPs: Hogaakvine, which I’m going to simply call Hogaak going forward. (Hogaak Dredge is subsumed under Dredge. More accurately, it has taken over Dredge.)

Hogaak, Arisen Necropolis

Hogaak only was the fifth-most popular deck at Grand Prix Barcelona, but still its pilots accumulated the highest number of total wins in the tournament. This is an extraordinary feat and it won’t remain the last unusual statistic regarding Hogaak …

Hogaak …

  • won 65.9% of 132 matches versus Jund (p<0.001)
  • won 66.7% of 120 matches versus W/U Control (p<0.001)
  • won 92.9% of 14 matches versus Jeskai (p=0.001)
  • won 62.6% of 99 matches versus Tron (p=0.008)
  • won 77.3% of 22 matches versus Infect (p=0.008)
  • won 58.9% of 180 matches versus Eldrazi Tron (p=0.01)
  • won 90% of 10 matches versus Red Prison (p=0.011)
  • won 75% of 24 matches versus R/G Valakut (p=0.011)
  • won 100% of 6 matches versus Temur (p=0.016)
  • won 88.9% of 9 matches versus Mardu Pyromancers (p=0.02)
  • won 59.5% of 111 matches versus Burn (p=0.029)
  • won 58.9% of 124 matches versus Izzet Phoenix (p=0.029)
  • won 100% of 5 matches versus U/R Thing (p=0.031)
  • won 100% of 5 matches versus Zoo (p=0.031)
  • won 100% of 5 matches versus Faeries (p=0.031)
  • won 81.8% of 11 matches versus Ponza (p=0.033)
  • won 76.9% of 13 matches versus Merfolk (p=0.046)

Usually, the list would run from most compelling positive record to most compelling negative record. This one doesn’t, because Hogaak didn’t have a significantly negative record against anything at all.

Nevertheless Hogaak’s list—of significantly positive results—is absurdly, almost comically long. The deck enjoyed a sizeable lead in direct encounters with most of its main competitors (Eldrazi Tron, Izzet Phoenix, U/W Control, Jund, Burn, Tron), and it beat a lot of fringe archetypes alike. I’ve been looking at tournament data for a while, and tier zero never seemed so apt of a description.

For some potentially practical information, let’s turn to the strategies that managed to hold up the best against the free 8/8s and 4/3s. Red Prowess went 61-51 and Devoted Vizier went 20-14 versus Hogaak. Neither record constitutes evidence of a favorable matchup, as you could expect to win this much or more in one out of every five attempts at a literal coinflip contest. Humans went 69-54, a record about half as likely to be the result of random chance.

The deck with the strongest performance against Hogaak turned out to be Twiddle Storm. In total, Nathan Wells went 12-3 with it, and ended up as the single 36-pointer to miss the Top 16 on tiebreakers (no byes). The amazing coverage by Sam Ihlenfeldt caught the deck anyway.

Wells beat three Hogaak opponents over the course of the tournament. The only other player in the field to Twiddle Lotus Field didn’t do as well overall, but defeated one additional Hogaak deck before exiting at 3-3. A total of 4-0 doesn’t reach the same level of significance as the pairings listed above. It corresponds to p=0.063. But it’s the best any archetype was able to muster.

Eldrazi Tron …

  • won 67% of 100 matches versus Jund (p<0.001)
  • won 61.5% of 96 matches versus Burn (p=0.016)
  • won 58.9% of 124 matches versus Izzet Phoenix (p=0.029)
  • won 80% of 10 matches versus Storm (p=0.055)
  • won 57.3% of 131 matches versus W/U Control (p=0.058)
  • won 73.3% of 15 matches versus Scales (p=0.059)
  • won 41.1% of 180 matches versus Hogaak (p=0.01)
  • won 15.4% of 26 matches versus R/G Valakut (p<0.001)

Eldrazi Tron’s list is half as long, includes three winning records (against Storm, White-Blue, and Hardened Scales) that aren’t quite as reliable, and even features a 4-22 run-in with Red-Green Valakut. Eldrazi Tron was a successful deck at the GPs in Barcelona and Minneapolis, but its results didn’t come anywhere close to what Hogaak did.

Izzet Phoenix …

  • won 71.4% of 21 matches versus Affinity (p=0.039)
  • won 69.6% of 23 matches versus Devoted Vizier (p=0.047)
  • won 40.2% of 102 matches versus W/U Control (p=0.03)
  • won 41.1% of 124 matches versus Hogaak (p=0.029)
  • won 41.1% of 124 matches versus Eldrazi Tron (p=0.029)
  • won 36.4% of 66 matches versus Tron (p=0.018)
  • won 36% of 86 matches versus Burn (p=0.006)
Arclight Phoenix

This former apex predator, by contrast, was not a successful deck. Note how its positive results are neither as significant nor as numerous nor as relevant as its negative results. Izzet Phoenix lost against five of the other seven major archetypes and went about even against two of them.

White-Blue Control …

  • won 91.7% of 12 matches versus Neobrand (p=0.003)
  • won 59.8% of 102 matches versus Izzet Phoenix (p=0.03)
  • won 42.7% of 131 matches versus Eldrazi Tron (p=0.058)
  • won 37.9% of 66 matches versus Red Prowess (p=0.032)
  • won 0% of 5 matches versus Hollow One (p=0.031)
  • won 35.5% of 110 matches versus Jund (p=0.001)
  • won 33.3% of 120 matches versus Hogaak (p<0.001)

Similar proved true for this deck, although it at least won against Izzet. Still, the deck doesn’t look well positioned in the current metagame.

Jund …

  • won 71.4% of 84 matches versus Humans (p<0.001)
  • won 64.5% of 110 matches versus W/U Control (p=0.001)
  • won 82.4% of 17 matches versus Spirits (p=0.006)
  • won 100% of 6 matches versus X Rack (p=0.016)
  • won 100% of 5 matches versus Faeries (p=0.031)
  • won 68% of 25 matches versus Devoted Vizier (p=0.054)
  • won 12.5% of 8 matches versus Mardu Pyromancers (p=0.035)
  • won 16.7% of 18 matches versus R/G Valakut (p=0.004)
  • won 30.4% of 69 matches versus Tron (p=0.001)
  • won 33% of 100 matches versus Eldrazi Tron (p<0.001)
  • won 31% of 87 matches versus Burn (p<0.001)
  • won 34.1% of 132 matches versus Hogaak (p<0.001)
Wrenn and Six

Jund, the allrounder, exhibited an exceedingly even spread of positive and negative matchups: the same number of either and a mix of major and minor archetypes among both.

Humans …

  • won 87.5% of 8 matches versus Merfolk (p=0.035)
  • won 60.9% of 64 matches versus Burn (p=0.052)
  • won 28.6% of 84 matches versus Jund (p<0.001)

Usually, though, when calling an archetype an “allrounder” what is meant is that the deck has very few lopsided matchups. That doesn’t seem to be true of Jund right now, but it evidently is for Humans.

Burn …

  • won 69% of 87 matches versus Jund (p<0.001)
  • won 64% of 86 matches versus Izzet Phoenix (p=0.006)
  • won 100% of 6 matches versus Grixis Control (p=0.016)
  • won 72.2% of 18 matches versus Affinity (p=0.048)
  • won 68% of 25 matches versus Thopter Foundry (p=0.054)
  • won 39.1% of 64 matches versus Humans (p=0.052)
  • won 40.5% of 111 matches versus Hogaak (p=0.029)
  • won 38.5% of 96 matches versus Eldrazi Tron (p=0.016)
  • won 9.1% of 11 matches versus Hatebears (p=0.006)

Burn completed about 100 fewer matches than Humans across the two GPs, so Humans had a much higher chance to find significant success or failure in more matchups. Nevertheless Burn sports a longer list, which really goes to show how much of an outlier the Humans performance was.

Tron …

  • won 69.6% of 69 matches versus Jund (p=0.001)
  • won 63.6% of 66 matches versus Izzet Phoenix (p=0.018)
  • won 69.6% of 23 matches versus Thopter Foundry (p=0.047)
  • won 15.4% of 13 matches versus Infect (p=0.011)
  • won 37.4% of 99 matches versus Hogaak (p=0.008)

I’ve read statements from notable Tron players who consider Red Prowess to be one of the poorest pairings possible. The Grand Prix sample doesn’t quite agree. 18-27 isn’t great, but it’s neither significant nor unwinnable.

Red Prowess …

  • won 72.7% of 22 matches versus Devoted Vizier (p=0.026)
  • won 62.1% of 66 matches versus W/U Control (p=0.032)
  • won 87.5% of 8 matches versus Amulet Titan (p=0.035)
  • won 72.2% of 18 matches versus Spirits (p=0.048)

Red Prowess joins Hogaak as the only other of the major archetypes with no apparent bad matchups. Note though that at least half of its good matchups resided more on the fringes of the metagame.

Thopter Foundry …

  • won 100% of 5 matches versus Mardu Pyromancers (p=0.031)
  • won 32% of 25 matches versus Burn (p=0.054)
  • won 30.4% of 23 matches versus Tron (p=0.047)
Thopter Foundry

Whereas 1,881 matches involved Hogaak and Vengevine, just 558 involved Thopter Foundry and Sword of the Meek. The lack of data begins to make itself felt. The deck’s 24-14 record against Izzet Phoenix, its 30-22 against White-Blue, its 18-11 against Red Prowess, and its 8-3 against Dredge all don’t meet the criteria for significance, but are cause for optimism. The Thopters even won 42 of 82 matches versus Hogaak.

Dredge …

  • won 0% of 5 matches versus Goblins (p=0.031)

Dredge on the other hand really did go roughly even against the vast majority of archetypes. Once upon a time, Dredge was known for extreme matchups. Another truism to reconsider in a world where everyone packs graveyard removal.

Devoted Vizier …

  • won 32% of 25 matches versus Jund (p=0.054)
  • won 30.4% of 23 matches versus Izzet Phoenix (p=0.047)
  • won 27.3% of 22 matches versus Red Prowess (p=0.026)
Devoted DruidVizier of Remedies

Devoted Druid devotees only completed 360 matches in total across the two GPs. Given larger samples, it’s virtually certain that additional performances would reach the benchmark for significance. This includes promising results such as Vizier’s 20-14 run against Hogaak or its 12-7 against Humans. Alas, it’s just as certain that some results would even out over the long run, and it’s impossible to know which belong in the former and which belong in the latter category. As it stands, Vizier bears the unfortunate distinction of having no apparent good matchups.

Red-Green Valakut …

  • won 84.6% of 26 matches versus Eldrazi Tron (p<0.001)
  • won 83.3% of 18 matches versus Jund (p=0.004)
  • won 12.5% of 8 matches versus Infect (p=0.035)
  • won 25% of 24 matches versus Hogaak (p=0.011)
Valakut, the Molten Pinnacle

Affinity …

  • won 27.7% of 18 matches versus Burn (p=0.048)
  • won 28.6% of 21 matches versus Izzet Phoenix (p=0.039)

Infect …

  • won 84.6% of 13 matches versus Tron (p=0.011)
  • won 100% of 5 matches versus Mill (p=0.031)
  • won 87.5% of 8 matches versus R/G Valakut (p=0.035)
  • won 22.7% of 22 matches versus Hogaak (p=0.008)

With this, we’ve reached the end of the largest matchup data collection ever. It was exhausting as well as it is exhaustive. That means, if an encounter didn’t get its own bullet point above, then the record was more even than 5-0, 7-1, 8-2, 10-3, 11-4, 13-5, and so on.

TL;DR

Hogaak Summer