What This Analysis Involves(Feel free to jump straight to the next section if you just want the data)
For every GP and PT, Wizards uploads pairings, results, and standings for each round. These pages are used by many people to do many things, one of the more interesting being Ajlvi and his MtgEloProject (a fantastic site, well worth checking out). Using these pages we can reconstruct the tournament from the ground up; who played who each round, and who won. All we need to do next is record what everyone played and we can generate some quite interesting pieces of information. Unfortunately, this last part is a bit harder than it sounds.
There are 3 main ways in which we can fill in the archetypes played by everyone in the tournament. Firstly, we can use the published deck lists. Wizards published the standard list for everyone who went 6-4 or better this year, which ended up being a good chunk of people. Next up, there's the people they put on coverage to do deck techs/feature matches. I watch all the coverage anyway, so it's pretty simple to just take notes as I go recording what various people are on. Finally, and most time consuming, there's trawling through twitter/tournament reports to find people who post what they played, and more importantly, what their opponents played each round. This often lets me fill in another 25% or so of a given field. For this PT, I've been able to find archetypes for 203 of the 510 players who registered, giving me a relatively healthy 40%.
Why bother with all of this, you might ask, surely you can just look at the day 2 conversion rates that Wizards posts? Well, although these conversion rates are sometimes useful to look at, they can be easily effected by various extraneous factors. The presence of draft at the PT is major factor in making day 2 or not, and as such this metric isn't as useful as one would hope. This PT especially shows this off, as only 4 of the 8-2 or better standard players were on one of the white aggressive archetypes, despite so many of the pilots having good records deep in the tournament. This was mainly due to people like Team CFB going 12-0 in draft day 2 and only just scraping by with their standard deck. Conversely, the two 10-0 standard players only went 2-4 and 1-5 in draft, reducing the visibility of their performances with Mono U and Drakes respectively.
Another thing that collecting this data lets us do is look at specific matchups and see how they broke down on the weekend. Although this only really works for the decks with a larger metagame share due to the small sample sizes, it's still a useful and interesting datapoint.
|Apparently there was a time when people sold this card for less than $5... (Art by Slawomir Maniak)|
Just Give Me The Data Already
Ok, ok, here's the win percentage table that you get from all the players for whom I know the archetype:
So, what are you looking at here? Each row shows the result for one archetype. You have the name, the win percentage of that archetype against only other people I know the archetype for, the percentage of matches that ended in a non intentional draw, and the total number of recorded matches for that archetype.
Another important thing to look at is the metagame breakdown of the 203 people I've recorded archetypes for:
What Does This Data Mean?
A few interesting things stand out when we first look at the data. The main ones are as follows:
- Control performed incredibly well, outperforming almost every other deck (although this does come with the usual small sample size caveat).
- The WR decks playing Heroic Reinforcements actually performed quite poorly, with a win rate of only 43.31%.
- On the other hand, the Wr decks, many of which were relying on some form of Healer's Hawk/Ajani's Pridemate tech performed much better.
- GB decks are almost as middle of the road as you can get.
- Red decks did not perform especially well, despite having access to Chainwhirler to fight the White aggro decks.
- Mono U was relatively poor, despite having one of the two 10-0 players.
- Izzet Drakes performed very well over a large sample size.
Before we go into more detail about some of these, It's important to think about how close these win percentages are to a "true" matchup percentage of any given deck against the field. For those of you who like to read Frank Karsten's articles, the 1 sigma interval on the archetypes with 100+ matches recorded is around 4% or less, whereas the archetypes with around 30 samples are closer to an interval of 10%. The actual calculation of these is rather involved, so I'll leave it for another day.
For those of you who prefer to ask "that one friend" to build the perfect manabase for you, what this means is that for the smaller (~30) samples, the results should mainly be treated as guidelines and approximations, but the larger samples can be given more attention. For those, the "true" win rate of the deck were you to run it against the field an arbitrarily large number of times will most likely lie within the quoted percentage plus or minus 4%.
Another important thing to consider with this data is the mirror match. Every time a deck plays itself it adds one to the number of matches won, and one to the number of matches lost. That is, it will always have a 50% win rate against itself. This means that the larger portion a deck makes up in a given field, the closer to 50% its overall win rate will be. To think about this in the extreme case, if an entire metagame is made up of one archetype, then the win rate of that deck will always be 50%. This only makes WR's low win rate more shocking, as it is pulled closer to a 50% win rate against the field by virtue of making up 16% of the recorded field.
A Closer Look at Some Archetypes
|Probably best to hold out for some other heroes. (Art by Scott Murphy)|
If we first take a look at the archetype matchups of WR, we see quite a sad story:
WR comes up short against every single other pillar of the format, especially falling off against the leaner Wr variants. This makes sense due to the specific Healer's Hawk tech employed by many of those on Wr builds. There really isn't much going for Heroic Reinforcements at the moment, and I'd advise steering clear.
GB Likewise looks to have been relatively poorly positioned (although it is the type of deck that can be altered to remedy this moving forward). It fared poorly against most varieties of control, and of the other main archetypes was only successful against the WR builds that I expect to be superseded by the Wr variants. Having a poor showing against Izzet Drakes is also not good news for Golgari Mages, but again, GB is a deck known for its ability to adapt to changing metagames, especially with the sheer number of playable cards the archetype has access to in the current standard format.
The performance of Izzet Drakes may have come as a surprise to many, as most of the deck lists look, quite frankly, awful. However, the results of the PT have shown that it is indeed a real deck capable of quite powerful and varied game plans, and Yuuya Watanabe in particular showed off some nice innovations with the addition of Entrancing Melody. The deck had strong results against many of the other major decks of the format. I expect Drakes to remain a strong choice in the next few weeks to a couple of months, but it may fall off if people are able to build GB and Control in a way that is able to successfully fend off Arclight Phoenix.
Here I've grouped all control variants (Jeskai, Dimir, Grixis and Esper) into one category, as the individual decks didn't quite have a large enough sample size to produce meaningful results:
The combined control decks had an overall win percentage of 57.75% over 258 matches (with 0 draws!). These decks seem to have the tools to deal with most of the aggressive decks in the format except for those with Ajani's Pridemate which grows slightly out of range of the burn removal that is used by most of them. Despite this, the control decks were able to come out ahead of every other main archetype they had to face, and are likely one of the stronger options available to you moving forward.
Overall, If I had to play in a standard event tomorrow I'd start with either Drakes or a control variant, likely Dimir. Drakes seems to be the most powerful thing you can be doing at the moment, and control seems to have the tools to fight almost anything you can find on the other side of the table. I'd probably be inclined to make sure to have access to a fair number of Ritual of Soot in my control decks to ensure game against the Ajani's Pridemate plan, and also something that can deal with Carnage Tyrant, perhaps a Nightveil Predator or 3.
I hope you've found this analysis interesting and useful. If you've got any questions or comments please feel free to contact me here or via twitter, @ArcKayNine. I've also got some modern stats coming up soon hopefully, as a collaboration with some other people from around the mtg internet, so keep an eye out for that.
May all your Tormenting Voices discard Arclight Phoenixes.