Grave Matters: GP Birmingham Analysis

Another event-filled weekend, another data dump, another chance for Modern to adapt and contain the arisen menace. It may not be likely, but as someone locked into making the trip to GP Las Vegas, I feel the need to hope. Barring a sudden abandonment of Hogaak, Arisen Necropolis in Birmingham, I was also hoping against all reason for an emergency ban. Since there was no ban as expected, lets get into the data from GP Birmingham.

Day 1 Metagame

I genuinely feel spoiled by Channelfireball this time. They released not only the usual Day 2 and Top 8 data, but Day 1 as well. Thus, for I think only the second time, I can analyze the entire event as it unfolded. Only the percentages were reported, but since Channelfireball also reported the attendance numbers, I was able to convert that into actual deck numbers (plus or minus one due to rounding). I’m only focusing on those decks that represented 3% of the field or more to make the data comprehensible.

Deck NameTotal %Total #
Hogaak10.394
Jund7.770
Tron7.366
UW Control6.256
Burn5.954
Izzet Phoenix5.449
Mono-Red Phoenix4.1 37
Humans4.137
Eldrazi Tron3.633
Urza Thoptersword3.027
Other21191

Day 1 doesn’t look like anything special. As we’ve come to expect over the years, the other category is the largest by quite a wide margin. Hogaak is the most popular deck, but at just 10%; this data spread would lead you to think that the format is relatively healthy.

Day 2 Metagame

And then you’d look at the Day 2 metagame and have that illusion snatched away. This chart shows the same decks from the previous table so we can compare conversion rates. It’s not a pretty picture.

Deck NameTotal %Total #
Hogaak21.937
Tron7.112
Burn7.112
Urza Thoptersword5.910
Mono-Red Phoenix5.910
Jund 5.39
Izzet Phoenix5.39
UW Control4.17
Eldrazi Tron4.17
Humans3.05
Other11.319

With ~22% of the Day 2 field and a massive ~40% conversion rate, Hogaak completely dominated GP Birmingham. Its nearest competition was other with ~11% and a piddly 10% conversion rate. This is very clearly an unbalanced metagame. Tron and Burn are the best-represented decks after Hogaak. Burn makes sense to me since in my experience Hogaak readily bolts itself multiple times in the early turns. Tron has the sweepers to manage Hogaak, but most of its threats are so slow I’m surprised by its success. I’d guess that maindeck Relic of Progenitus is key, though I’m skeptical that it is enough.

The Top 8

The Top 8 is not really useful for judging the metagame as a whole so much as seeing how that metagame turned out. The Day 2 numbers were mostly predictive of the final results.

Deck NameTotal #
Hogaak3
Mardu Shadow1
Urza Thopter Sword1
Burn1
UW Control1
Hardened Scales1

That’s a lot of Hogaak in the Top 8. However, unlike last week, Hogaak didn’t win. Instead, it was Mardu Death’s Shadow. The reason it won is instructive. Game 1 was won thanks to Temur Battle Rage, where Game 2 was won thanks to discard spells eliminating all Hogaak’s enablers. Like all Belcher decks, Hogaak’s payoff cards only work when used alongside a specific combination of enablers. Disrupting them in a timely manner is the key.

The SCG Classic

Concurrently, there was an SCG Open in Richmond. This was a team event, so I’m not going to look into its results. Team events completely distort individual deck performance. However, the Modern Classic is another matter, and is worth inspecting.

Deck NameTotal #
Hogaak2
Jund2
Tron2
Eldrazi Tron2
UW Control2
Izzet Phoenix2
Mono-Red Phoenix1
Four-Color Urza1
Cheer0s1
Burn1

Once again, the SCG Classic presents an odd counterpoint to the rest of the data. Hogaak was just another deck. True, in a change from last time, it won the event, but that’s not analytically important right now. Based on the observed results, any deck could have won. Also again, this apparent contradiction between events highlights the mercurial nature of Hogaak. At least other busted decks showed consistent results. Hogaak is certainly absurd, but the lack of consistency keeps raising question marks and muddies the picture.

The Deck Dump

The results data tells a clear story of Hogaak warping Modern, dominating events, and generally being a huge mistake that I can’t fathom Wizards missed. However, the more specific data complicates that story. Specific events show huge deviations from expectations given the narrative, and that deviation gets wider if you dig even deeper.

In what I imagine is a first, all the decklists from the GP have also been published. Praise be to Frank Karsten! A pile of decklists doesn’t mean much for statistical analysis unless you pull them all apart to look at archetype card choices (which takes more time than I have available), but it doesn’t need to. Instead, decklist data allows me to get a look inside players heads.

In a vacuum, players  choose the cards that define their archetype and make their gameplan possible. Card selection is therefore sterile, predictable, and too boring to investigate. However, in reality players are actively testing matchups and making choices about which cards to actually run in flex slots and sideboards. These decisions reveal how they see their place in the metagame as a whole, and therefore an insight into their minds. And what I find by reading those tealeaves really muddies the picture.

The Big Question

Do the actual decklists show evidence of Hogaak’s warping of Modern? Reading through all 911 decklists would take too long for me to do for this article (and frankly be so mind-numbing I’d just forget anything I found), so instead I used a random number generator to select about 100 decks to look at. A random sample is valid for analysis as long as each member of the population had the same chance of being picked. Thus, the odds of each opinion or outcome had the same chance of appearing in the sample, and therefore should appear in the sample in proportion to the actual population; the sample should indicate whether or not players’ choices are being warped by the existence of Hogaak.

After going through my sample, I can’t conclude that they are. For the most part, they look like normal decks. For example, the 97th place Humans deck could be a pre-Modern Horizons Humans list. There’s no overwhelming dependence on graveyard hate, special anti-Hogaak cards, or other signs of a warp. Ravenous Trap isn’t a common card, but it’s not outside of Humans’ wheelhouse. It’s significant that, given the narrative that graveyard hate is essential to beating Hogaak, Chris Vincent only ran three pieces. This decision was echoed up and down the list, such as the four Grafdigger’s Cage in the 720th place Hardened Scales deck (the fifth place version only had three), the 27th place Merfolk deck, or the 104th place Burn deck. Most of the players in my sample decided that it was better to maintain their gameplan and their sideboard percentages against non-Hogaak decks than worry about beating Hogaak with hate.

For the most part, the decks that do some evidence of warping only have it weakly. The 67th place UW Control list has maindeck Surgical Extraction and a full set of Rest in Peace in its sideboard. However, that isn’t too extraordinary, because it’s not a recent change; control players had been maindecking Surgical since Arclight Phoenix became a thing, and typically run at least two Rests regardless. The 8th place UW list had the Surgicals but only three Rests.

The 101st place Eldrazi Tron list had a full set of Leylines in its sideboard and a Tormod’s Crypt. While I have no way of knowing, I’d guess that the Leylines are only there because wishing for Crypt with Karn is too slow most of the time, and without Hogaak, Fabio Aldrighetti wouldn’t have bothered. Otherwise, his deck looks like normal Eldrazi Tron. The signs of an actual warp in player’s decisions are minimal.

A Wrinkle

That is, until the actual Hogaak lists are also considered. Every Hogaak list in my sample and in the Top 8 played a full set of Leyline of the Void. They also never had less than five cards that remove enchatments in the sideboard, with a minimum of 2 Force of Vigor every time. The most common configuration was 3 Force and 3 Nature’s Claim sideboard, and 2 Assassin’s Trophy maindeck. Even if everyone else demonstrated indifference towards Hogaak, it was not doing so towards itself, and was fully prepared for an anti-Hogaak meta.

The other twist in this warp narrative is that there is no correlation between a deck’s quantity of graveyard hate and its final placing. To reiterate, the 5th place Hardened Scales deck had less hate than the 720th place version, and few differences in flex slots. Almost all the Burn lists are maindeck copies of the 6th place deck with less than three sideboard cards different. Thus, I cannot conclude that it pays off for any non-Hogaak deck to specifically target Hogaak.

Deep Dive

This is in line with my testing for GP Las Vegas. I spent many hours running a very hateful UW Spirits with maindeck Remorseful Cleric with Leyline of the Void and Surgical Extraction sideboard against various Hogaak lists. I then compared its results to a tweaked version of my MCQ list (-2 Damping Sphere, +2 Settle the Wreckage, for the curious). The normal Spirits list won more than the very hateful list, but not by enough to matter statistically. I think it currently stands at ~100/~90 in favor of the normal list. I’ve been skeptical of Leyline and Surgical for a long time, and my testing justified a lot of that skepticism.

The problem is that graveyard hate is only useful against Hogaak under specific circumstances: if you can exile their ‘yard after they’ve spent a ton of resources to fill it, but before they get any value from doing so. Removing the power cards in Hogaak and Vengevine with Surgical is good, but happens if and only if Surgical is in hand while they’re in the ‘yard. Given that you’re only ~40% to have Surgical in hand and Hogaak can churn through its deck extremely quickly, the odds aren’t in your favor. Drawing Rest in Peace later in the game isn’t optimal, but is still useful since it still exiles the existing graveyard and shuts down recursion engines. Late Leylines or Surgicals do nothing. By going for the silver bullets, I was putting a lot of cards into my deck that were dead if not drawn at exactly the right time. When things line up, it does great; when it doesn’t, I lose.

Besides, the hate isn’t all that effective against Hogaak. With Rest on the board, the recursion engine is dead, and Hogaak is almost certainly uncastable. That doesn’t stop Hogaak from just swarming the board with dinky creatures and Vengevines. Given that the hateful builds don’t actually kill Hogaak and sometimes lose to having unusable cards, I don’t think it’s worthwhile.

Conclusion for Vegas

The conclusion I draw from all this data is that I shouldn’t try too hard to beat Hogaak. My testing has shown that the fast graveyard hate is ineffective and frequently counterproductive, so I’ll be playing a more normal deck this Friday. The other thing I’ve found testing is that Hogaak needs to do a busted thing to be good. It digs through its deck better than anything else I’ve tested against, but if that doesn’t turn into a significant board presence, it can’t win a game. Given time, any deck can beat Hogaak, and it relies on getting very good flips into its graveyard to win. The deck can go off turn two 60% of the time under lab conditions, but it also has to do that at least two times in a row to win. I’m better off focusing on playing a reasonable game rather than trying to shatter their statistics.

I also expect Vegas to be a relatively small tournament. Birmingham and Minneapolis have been down from their previous numbers, and I expect the trend to hold. I think this is the result of players being turned off by the threat of Hogaak rather than the deck in actual fact. You’re unlikely to hit a single copy in the Swiss, after all. That’s what I hope, at least.

Final Hurrah

One way or another, we’re nearing the end of Hogaak’s influence. Next Monday, the Necropolis will likely be banned, and then Modern can finally start to figure out the real impact that all the new sets have had this summer. And I’ll see you then, with my lessons from Vegas.

Leave a Reply