At a tournament recently a long-time reader told me how much he appreciated some of the more number-crunchy articles that I’ve done. “A lot of people will tell you how they feel about a format but I really like it when you crunch numbers. Empirical evidence is such a good thing.”
I think of my outlook as a little different than that specifically. What I’m generally interested in is rigorousness. I like having a sense for how things are really operating. I don’t have much patience for theory that isn’t useful (and perhaps less patience for good theory misapplied). I don’t have much patience for empirical claims that have loose evidence (and far far less for claims that are actually false empirically).
Long ago I wrote an article called Good and Bad Magic that was about theory empiricism and the problems of “results-oriented thinking” (i.e. success) in an attempt to quantify what we need to be focusing on if we’re interested in understanding the game we play. It is one of those articles that I’m quietly proud of and I think that it is well worth everyone’s time to sit down and read. One of the things that it gets to is that while it’s true that the best way to get good at Magic is to simply play again and again and again (and again and again) the way we think about playing is important as well. It’s also practical. If for example we think that we know how certain matchups work or that certain decks have quality X it’s important to know whether or not that is true.
The best example of this in many ways is Faeries. In Block season I played Faeries in one PTQ and quickly realized that it was both the best deck in the format and a deck that I was not going to sleeve up again. Its performance simply couldn’t be denied. You could look at the top tables of the PTQ and even if there were a lot of people playing Faeries there was still a huge representation of them at top tables even given their numbers in the field. Further while some people would dismiss the mirror as a mere coin flip I’d watched enough games to know that this wasn’t exactly accurate. I watched Sam Black lose one match in a nail biter and could tell that a lesser player would have just been smoked. I watched him dispatch players in the mirror match with ease. I know that watching him in all of the games I couldn’t have pulled off the wins he was getting. It was simply a matter of skill with the deck; he had a much greater degree with it and I did not. Even if someone might retort that this only meant Sam would have a 55% matchup in the mirror (I’d put it at closer to 60% at a PTQ level) there were two things this made me consider: (a) how different is that from most Magic these days now that we live in the time of the Magic Hive Mind and (b) I knew that at that point in time I didn’t have the time to practice that I would like and most competent Faeries players would have me on the 40-45% side of the matchup if they were paired against me. I dodged the fight and played the less powerful (in my opinion) Merfolk. I cruised to 2nd place with my deck before losing to Justin Meyer’s teched out Faerie deck (main deck Oona to match my main deck Oona). I didn’t play Faeries the best deck but at least my reasons (constraints on time to get as competitive with it) made the choice reasonable.
Fast forward to the moment that Faeries fell and fell hard:
There were still people who were hard-core Faeries defenders at this point. They were crazy but they were out there. Playing a good deck that got to include 4 Volcanic Fallout and main (or side) 4 Great Sable Stag was just too much for the little winged menace. You can see this reflected in the Top 8 of U.S. Nationals where thirty-one Volcanic Fallout and thirty-two Great Sable Stag were played. As the metagame developed after U.S. Nationals it became more reasonable to play Faeries under one circumstance: you assumed that people wouldn’t bring the hate that they needed to beat Faeries because they assumed that the deck was dead. This is much like the ol’ Mind’s Desire game: if people wanted to beat you you would lose with your Extended Mind’s Desire deck of old. If they hadn’t remembered to beat you you would crush them.
More interesting is the case of Boat Brew which managed to vacillate between being one of the worst decks in the format (Regionals this year) to one of the best (SCG Indy $5K). If you can base this belief on data you’re going to get a lot further than someone who just gut-checks it and thinks that they have a handle on things.
The problem with the gut-check is that Magic is a dynamic game. Truths change. Icy Manipulator can go from being fantastic for a year or two to terrible for almost ten years to fine to poor. In certain environments Tarmogoyf ain’t that hot (I’m talking to you Time Spiral Block). If we just simply continue to believe that say Elves is good you shouldn’t be surprised when someone hands your ass to you because the environment has moved on.
Unlike the last time that I did a metagame analysis of the Pro Tour data the format in question is going to be a live one. Shards of Alara Block as much as I liked it does not have much of a life span past the Block itself except perhaps as the core for the new Standard post-Zendikar. We might like knowing that Jim Davis had the best performing deck of the Constructed portion of the format but ultimately the data is more interesting than it is useful. This is not the case for Extended which is a much more living and breathing format.
If you weren’t already aware there is a big difference between the Top 8 as listed in the standings at the end of the Swiss and the actual Top 8 of solely the Extended format.
Here are the full Swiss Top 8 and archetype played:
1 – Tsuyoshi Ikeda – Zoo
2 – Evangelos Paptsarouchas – Hypergenesis
3 – Yuuya Watanabe – Dredge
4 – Martin Juza – Zoo
5 – Naoki Shimizu – Dredge
6 – Hunter Burton – Zoo
7 – Brian Kibler – Rubin Zoo
8 – Paulo Vitor Damo da Rosa – Dark Depths
Or put another way:
Guess how many of these people would not have been in the Top 8 if we only account for the Constructed portion of the event? FOUR.
If we don’t take into account the performance of the decks in and of themselves separated from the total records of their pilots we do not get a full picture of how things actual turn out in the format. Mixed format events are very difficult to effectively track essentially because you continue to play within the Swiss of your record and yet it isn’t counting solely the format. This means that you are no longer sifting things to the top or the bottom based on the weight of the deck in the hands of the player but are instead giving heavier weight to play skill in a big way. Six rounds is a long time. Personally I understand those people that love the mixed formats for just that reason – they love seeing the best players win. I don’t like them partly because I am a worse player than the best players sure but mostly because I’m interested in the truth of formats and which decks are truly the best.
If we want to examine the lists to get the Extended Top 8 what we get after ten rounds of Swiss is this (with MAO = modified average opposition i.e. strength of opponents records in Extended only):
1 – Hunter Burton – Zoo – 28 points
2 – Brian Kibler – Rubin Zoo – 27 points (MAO 187)
3 – Ben Rubin – Rubin Zoo – 27 points (MAO 151)
4 – Tsuyoshi Ikeda – Zoo – 25 points
5 – Paulo Vitor Damo da Rosa – Dark Depths – 24 points (MAO 180)
6 – John-Paul Kelly – Cogs Control – 24 points (MAO 175)
7 – David Ochoa – Dark Depths – 24 points (MAO 171)
8 – Matej Zatlkaj – Dark Depths – 24 points (MAO 166)
9th-13th (tied for 8th) – Koutarou Ootsuka Dredge (MAO 159); Kazuya Mitamura Dredge (MAO 159); Evangelos Paptsarouchas Hypergenesis (MAO 158); Nicolay Potovin Zoo (MAO 154); and Samuele Estratti All-in-Red (MAO 148).
Or put another way:
4 Zoo (2 Rubin Zoo)
3 Dark Depths
1 Cogs Control
This is a far cry from our earlier picture which both underreports the power of Dark Depths and Rubin Zoo. Earlier this week Patrick Chapin was beaming with pride about the performance of the deck and when you measure it against objective reality it seems absolutely clear that he has every reason in the world to be so proud. The other player who ran the deck (Matt Sperling) was tied for Top 16 within the Extended portion of the event. While it is possible that other people played the deck and crashed and burned but based on what little that I can see from the analysis of archetypes supplied by Bill Stark Sperling Rubin and Kibler were the only 3 people playing the deck and they all placed in the Top 16 of the format. That is unequivocally insane. Wow. Kudos kudos kudos. I’m sure that Kibler will address the specifics of this more in his own article but seriously this is an amazing stat.
Stark’s archetypical analysis is at once incredibly useful in the way that it chooses not to categorize certain archetypes but also incredibly difficult to use if you’re trying to amass a good sense of the archetype. To claim for example that 53 people played Zoo hardly begins to tell the whole story. Once you re-classify some more interesting data emerges. Zoo variants for example were played by 95 players and made up nearly 25% of the metagame. Here is all of the data:
I folded in a lot of the archetypes into each other. All Zoo variants for example are listed together. All decks that ran Cryptic Command and Tarmogoyf are as well. Where an archetype could be meaningfully collapsed I collapsed it down.
The categories of data are as follows:
Day 1 – The counts on decks that were played.
18+ Match Points – The count of decks that achieved 18 or more match points.
Top 32+ - The 40 players with a match score that tied them for Top 32 or better.
Top 16+ - The 21 players with a match score that tied them for Top 16 or better.
Conversion % - The percent of decks from the total of the archetype who achieved whichever level of success listed.
The percentages of the total in yellow represent the average success you would hope to achieve if you just showed up and were given a random deck from the tournament. The selections marked in red either met this level of success or were statistically close to it.
The first thing to note is that a number of archetypes were terrible failures. All Blank-Level Blue variants Affinity Martyr Bant Death Cloud Doran and Tezzerator decks were statistically not a great call to make with very few positive results. You would have done better at all levels of measure to randomly swap decks with someone in the room. Similarly various other decks (listed in the first italicized parenthesis) had dismal results though very few iterations to really count. Archetypes that aren’t listed (the “Others”) just generally didn’t do so well. Essentially no particularly new successful archetype was found only by a very few (less than 1%) that couldn’t be wrapped into other archetypes.
A small few (listed in the second italicized parenthesis) managed to place someone in the Top 32 but were played by so few people that their presence marked in red on the chart must be given the extra note that they are too difficult to statistically place much weight on. Similarly other archetypes had some marginal success but only found in outliers. A single player of B/G Dark Depths Cogs Control and Hypergenesis were solely responsible for the positive notation of those archetypes. Hypergenesis has perhaps the most interesting of these deck stats to examine. Given the relatively large number of people that played it it remains to be seen whether the solitary person who managed success with it was merely a lucky outlier or simply had a superior build to those who had otherwise lackluster performance. Further exploration of the archetype will give us answers there. This hints that maybe all of these archetypes should be explored more but doesn’t give solid data on the quality of the archetype as a whole.
Most important then are these archetypes which managed positive conversions in several measures and were played by enough people to have at least a small confidence in their quality:
That is all of the decks that managed to find success in the hands of more than just one person and also managed to beat the “swap a deck with someone random” test. It’s not much. Interestingly the Zoo list is only just barely on that cusp.
Remember that with Zoo Rubin Zoo has been folded into the results. Rubin Zoo has a ridiculous 100% conversion on all measures for its 3 players. With that removed non-Rubin Zoo becomes:
In other words the real decks to play for the event (out of what was played) were in order:
Of course there are some major caveats to be had with such a strong advocation of Rubin Zoo. The three players playing it most certainly are on the “upper end” of the skill spectrum. Even the player with the shortest resume Matt Sperling is a very talented player in his own right and certainly ranks far above the average Pro Tour player at least in my estimation. Rubin and Chapin rank in the best players of Magic’s history. Certainly that weighs in the measurement at least in some way – other archetypes were certainly also played by heavyweights but they weren’t also played by players new to the Tour. Secondly the way that Rubin Zoo played out was largely a total surprise to the field. Fighting Rubin Zoo and fighting another Zoo deck are very different challenges and a fair number of people certainly weren’t as prepared for Rubin Zoo in the way that they were prepared for other Zoo lists. That is definitely going to change now.
Dark Depths bears a lot of noting simply because its number if we exclude the top-heavy nature of the Rubin Zoo players would make it the best performing deck of the tournament. This says a lot for an archetype which is basically still in its infancy. I predict we’ll see this archetype get honed even more once everyone gets their mitts on it and MTGO runs its course. At the same time the hate for the archetype will certainly step up as well.
Burn is somewhat of a surprise. Charles Gendron Dupont was the archetypes top finisher finishing 16th place (if we count only Extended).
3 Blinkmoth Nexus
2 Darksteel Citadel
4 Great Furnace
3 Teetering Peaks
4 Goblin Guide
4 Hellspark Elemental
4 Keldon Marauders
4 Spark Elemental
4 Lava Spike
4 Lightning Bolt
4 Molten Rain
4 Rift Bolt
3 Shrapnel Blast
The Teetering Peaks and Goblin Guides are especially exciting for what this deck is attempting to do. The Molten Rain make the deck a smidge closer to a RDW-style configuration but also just serve to try to give the deck just an extra moment to bring the opponent to zero.
Dredge might have been a more noteworthy deck statistically if so many people hadn’t chosen to play it. As it was it was one of the few archetypes that performed admirably and remind us that the archetype is alive and well.
Looking at the format objectively is important. We’re going to be living in the Post-Austin Extended for quite some time. With Zendikar hitting MTGO soon the online world will catch up with paper Magic and we’ll all be seeing the results of Austin every time we start a match.
I hope you enjoyed this look at the format*. Until next time!
A quick word of thanks to Charles Gendron Dupont and Josh Utter-Leyton for helping me figure out the two missing archetypes from the top 79 players. It was appreciated.
* Here is the “Extended Only” Top 8
8 – Matej Zatlkaj – Dark Depths – 24 points
7 – David Ochoa – Dark Depths – 24 points
4 Dark Confidant
4 Vampire Hexmage
3 Vendilion Clique
3 Beseech the Queen
4 Chalice of the Void
4 Chrome Mox
2 Engineered Explosives
4 Muddle the Mixture
2 Threads of Disloyalty
6 – John-Paul Kelly – Cogs Control – 24 points
3 Baneslayer Angel
4 Trinket Mage
3 Vendilion Clique
3 Chalice of the Void
4 Chrome Mox
2 Cryptic Command
3 Engineered Explosives
1 Expedition Map
1 Pithing Needle
1 Pyrite Spellbomb
3 Spell Snare
4 Thirst for Knowledge
2 Vedalken Shackles
2 Wrath of God
5 – Paulo Vitor Damo da Rosa – Dark Depths – 24 points
4 – Tsuyoshi Ikeda – Zoo – 25 points
3 – Ben Rubin – Rubin Zoo – 27 points
2 – Brian Kibler – Rubin Zoo – 27 points
3 Baneslayer Angel
4 Knight of the Reliquary
3 Noble Hierarch
3 Qasali Pridemage
4 Wild Nacatl
1 Elspeth Knight-Errant
4 Lightning Bolt
2 Lightning Helix
4 Path to Exile
4 Punishing Fire Sideboard:
3 Ancient Grudge
3 Blood Moon
3 Ghost Quarter
1 Hallowed Fountain
1 Kataki War's Wage
4 Meddling Mage
1 – Hunter Burton – Zoo – 28 points