I recently spent a weekend in Lille, where I learned more than a little about Guilds of Ravnica Standard. In fact, the town played host to the largest tournament in the format to date. During the event, Frank Karsten and I did strive to put as much information as possible into the official coverage. But proper data mining takes time, and that’s the raison d’être for this article.
I will focus on two archetypes in particular: Izzet Phoenix and Golgari Midrange.
Izzet Really? Yes, It Is!
The Top 8 doesn’t quite show it, but Izzet Phoenix was the third most successful archetype of the event. Two copies ended up in the Top 10 and a total of seven in the Top 32. Even the ubiquitous Golgari only entered nine copies into the Top 32, and Jeskai Control had seven as well, albeit three in the Top 10.
The first version of Izzet Phoenix that registered on my radar was the one Jérôme Bastogne used to win a trial on Friday. Not only did the name ring a bell—Bastogne was a finalist at the 2016 World Magic Cup—the deck was designed to maximize Izzet’s penchant for chaining instants and sorceries. To take something—anything really—to the extreme and see where that gets you is often a worthwhile endeavor in Magic, if only because nothing else generates as much insight.
With 29 instants and sorceries, a minimum of creatures, and two Firemind’s Research, it was easy to fall in love with this main deck, and the sideboard featured some idiosyncratic ideas too.
I was happy to see an extreme version like this succeed, and for a long time it seemed as if my emotional investment wasn’t misplaced. While Bastogne failed to make Day 2, fellow Belgian Alexandre Darras took the same list—up to the 14th sideboard slot—on a winning spree in the main event. The former GP champ didn’t lose a match until round 12!
Unfortunately, Darras’s second loss came in round 13 and his third followed. Time and again, I watched Izzet players in the feature match area struggle to find enough Phoenixes within the top half of their library. If all or most of them hid in the bottom half, there was little a Phoenix-reliant version could do.
Of course, when it comes to blue-red decks, you can always count on the Vieren brothers, also from Belgium, to come up with something. Indeed, both eventually surpassed Darras with matching 12-3 records and a matching 75. Their take on the archetype included a few more alternative payoffs for instants and sorceries and fewer actual instants and sorceries:
Pascal and Peter Vieren
This appeared to be a step in the right direction, but ultimately several more steps was what proved best. Arne Huschenbeth placed in the Top 8 and Raul Porojan finished 10th with a 12-2-1 record and a version that relied less on Phoenixes than on Drakes.
It was the polar opposite of Darras’s and Bastogne’s build: a minimum of instants and sorceries but a total of 14 creatures.
As a reporter it’s always nice to find your initial hypothesis disproved—then there’s a definite story to tell. And this particular story, of Drakes performing well in a field full of Golgari and Jeskai Control, checks out. Both foes are limited in their capacity to deal with flyers or 4-toughness creatures, and the above, with its Dive Down and Maximize Velocity, exacerbates and preys on such weaknesses beautifully.
Admittedly, while there’s strong evidence that this is the way to take Izzet Phoenix going forward, there’s no actual proof yet. The sample size remains too small.
Golgari Data Farm
Proper data mining takes not just time. It also takes vast amounts of data. If Magic has taught me anything, it is that reality is a woefully imperfect approximation of mathematics. To reverse engineer truth from results you need a lot of results. Luckily, Golgari Midrange did put in a lot of results over the weekend.
At the same time, you should note that the archetype didn’t live up to the hype or fear that the news of its 40% metagame share on Day 2 elicited. It never reached the number again, not by any other metric, whether it’s spots in the Top 8 (25%) or Top 32 (28%). Even among the Top 84, which encompasses everyone with an 11-4 record or better, its share barely came to 35%. This is not the new Red-Black Chainwhirler or Temur Energy.
While we were doing the metagame breakdown, I noticed that lists differed even in key elements. Yes, most people seemed to agree on the land base, on the inclusion of Ravenous Chupacabra and Golgari Findbroker, on a full playset of Jadelight Ranger, and finally found two to four to be the correct number for Find // Finality.
But already on something as elemental as the question of Llanowar Elves or no Llanowar Elves there appeared to be a divide among members of the Golgari Swarm. The same with the selection of planeswalkers. Carnage Tyrant too was all over the place, though not in the sense that it was universal. Rather, it showed up in zero, one, two, three, and four main deck copies.
I was interested especially in whether or not it was possible to discern trends that separated the more successful from the less successful versions. So I did. I separated all Day 2 Golgari decks into five tiers:
- Decks that failed on Day 2 and ended up with 18 to 24 match points
- Decks that were on the cusp of a relevant record at 25 to 28 points
- Decks whose pilots went 10-5, 10-4-1, or 11-4—that is 30 to 33 points
- Decks that made it into the coverage by virtue of reaching 34 to 36 points
- Decks that made it to the semifinals for an overall record of 13-2-2, the equivalent of 41 match points
These thresholds are debatable. I had to come up with something, and specifically I had to come up with a division before I started aggregating data. Another division might show clearer trends, but choosing categories based on preferred outcome would have been cheating.
Then I tallied up the main deck numbers for certain cards. Lots and lots of numbers, many next to French card names. Finally, I worked out how big a share of main decks included how many copies. I did this for all Day 2 lists in total, as well as for each of the five categories listed above.
Overall, 87% of Golgari main decks on Day 2 included some number of Tyrants. Although there’s some fluctuation in the details, the two lower-placing tiers of the field remained below par, while the higher tiers went increasingly above. The trend was even more pronounced for three-plus Tyrants. Only 17% of 18–24 point finishers chose to run that many, but 21% of those ended up with 25–28 points, 32% of those with 30–33, and 36% of those with 34–36.
Of course, anyone who looked at the Top 8 decks could have come to the conclusion that two to three Tyrants led to the best results at this event. But at the very least, the additional data lends further credibility to the notion that “the correct maneuver is usually to deploy the giant, implacable death lizard.”
It is far harder to glean a clean recommendation from this, but it does serve as a reminder that none of these statistics are to be seen in isolation. The numbers for 6-drop Vraska went down toward the top of the field because these players needed to make room for Carnage Tyrant.
Another point to consider is how much stock to put in the decisions made by a few compared to those made by many. Doom Whisperer is an uncomfortable example in multiple ways. For one thing, the bars for this card definitely get darker higher up in the standings. Except for the very top, but there were only two players who decided to break the trend, right? Well, except it wasn’t a big trend to begin with because the 34–36 point range wasn’t crowded either with its eleven players. In the end, you might as well look at the 24 players on 25–28 points, see that even among them 61% didn’t care for Doom Whisperer, and follow their lead.
Midnight Reaper, meanwhile, is the poster child for a card whose appreciation didn’t change appreciably across tiers.
The most notable thing to say about Vraska, Golgari Queen and Vivien Reid is to which extent the top finishers were in agreement. Javier Dominguez and Christoffer Larsen both identified one and three as the correct numbers. 91% of 34–36 pointers chose the same or deviated by one copy of Vraska. 73% of them had the same or one fewer Vivien.
Speaking of strong convictions: The words “Llanowar Elves 2 strong!” appeared on Christoffer Larsen’s Top 8 profile no less than three times, successive iterations in the original punctuated with an ever increasing number of exclamation points. Curiously, next to Javier Dominguez, it was the worst finishers of whom the largest share agreed with Larsen. Other than that, the graph shows a clear trend: valuing Llanowar Elves highly went hand in hand with placing highly in the tournament.
Earlier I mentioned the negative correlation between Carnage Tyrant and Vraska, Relic Seeker. There was no indication that the same was true for the relationship between Llanowar Elves and Druid of the Cowl. True, some people used the Druid in place of the Elves, notably Guy Duparcq in 34th place.
But more often Druid of the Cowl acted as a supplement rather than a replacement. Pro Tour Ixalan quarterfinalist Christian Hauck finished 12th with just such a build.
While doing the metagame breakdown, Frank and I discussed whether or not the above should be considered Golgari Ramp rather than Golgari Midrange. Alas, it proved too hard of a determination where to draw the line on such a slippery slope.
I also ran the numbers for some other cards. In addition to Hauck’s, Karn, Scion of Urza showed up in eleven decks. 8% of the 25–28 pointers as well as 7% of the 30–33 pointers ran one copy and the same shares of each ran two. 10% of Day 2 lists included one The Eldest Reborn, and another 7% included a second copy. The Saga mostly played in the 18–28 pits, though. Izoni, Thousand-Eyed got similar numbers, only lower.
Unless I missed a “Carcasse fongiforme” somewhere, absolutely no decks with one, two, or three Molderhulk made it to the second day. But Tommaso Magnoli went 9-6 with four!
If anyone ever tells you that they’re tired of Golgari decks, I suggest beating them with Hulk. Hulk 2 strong.
Until next time, I hope you’ll always find some Arclight Phoenix in the top quarter of your library.