28K Games of Modern Analyzed
A few weeks ago we published data gathered from 85K games of Standard along with an article breaking down the numbers. The response to both were great, but one of the themes in the comments was "when are you going to do this for Modern?" Well, the answer to that question is today!
Now, let me lay our our pathway forward. First, before you read anymore, jump over to the Modern Matchups page and check out the data for yourself. Second, we are going to do a short comparison between Modern and Standard, since these two data sets really illuminate the differences between the two formats. Third we will talk about some of the popular decks in Modern and break down what these numbers show about their matchups, and finally we will conclude with some short observations gleaned from the rest of the data.
Before we jump into the good stuff, just a quick note about the data itself. All of these numbers come from scraping Magic Online replays using a technology similar to the one used by bots, which has one major weakness — we can't see the cards in players' hands/decks/sideboards. This means that a color screwed Abzan deck could end up being lumped into the Jund category (or vice versa) or that what we have listed as a Jeskai Control could actually be a four-color control deck that never drew a black source for Lingering Souls or Mystical Teachings. We also don't get to see every individual card, which is especially meaningful in Modern where there are a ton of very powerful sideboard options; there are far fewer sideboard card in standard that can swing a matchup on their own like Blood Moon, Choke and Timely Reinforcements. This means that our data isn't 100 percent accurate — although it is certainly accurate enough to draw some meaningful conclusions — and is best used in conjunction with playtesting to draw conclusions.
Comparing Modern and Standard
Before we get to the decks and the matchups, let's talk a little bit about the big picture differences between Modern and Standard as formats. Probably the first thing that jumps out of the data is that Modern games are, on average, significantly shorter than Standard games — exactly two full turns shorter, in fact. While this might not sound like a lot, it is, it's tremendous, it's a tremendous gulf. These shorter games are result of having powerful, inexpensive creatures that can end the game quickly on their own like Tarmogoyf, or with help like Deceiver Exarch, along with the larger card pool allowing for faster, more broken decks (Grishoalbrand, Infect, Amulet Bloom) which can win on turn two (which brings down the average). All together this means that, as most of the pros have been saying, Modern isn't a great place for durdling around.
The other information we have is about mulligans and win percentages on the play versus the draw. While it is worth mentioning these briefly, the former will almost certainly change if the new mulligan rule is instituted game-wide after the upcoming Pro Tour. As for win percentages on the play/draw, things actually look much the same in Modern as they do in Standard, with the person on the play having a 7-8% better win percentage.
Most Played Deck - Naya Burn
Unsurprisingly thanks to its comparatively cheap price on Magic Online, Naya Burn comes in as the most played deck in our sample at 7.58 percent of the field. While the deck's overall win percentage is solid if unspectacular at 51.1 percent, by looking a bit closer at the matchups we can get a good sense of the types of decks you do and do not want to play against as a Naya Burn pilot.
The good news for Naya Burn is that it feasts on control decks like Grixis Control and almost every slow combo deck in the format; decks like Splinter Twin, Scapeshift and Tron have a difficult time keeping up with the fast clock and consistency of Burn. It also boasts a strong matchup against Infect (winning over 60 percent of the time), most likely because having 12 Lightning Bolt/Searing Blaze effects is a good place to be against a deck trying to win the game with thirteen 1/1's. If the Naya Burn player can answer the first two or three of these threats, they give themselves almost unlimited time to draw enough burn to finish the game while their Infect opponent is drawing useless copies of Mutagenic Growth and Vines of Vastwood.
On the other hand, Naya Burn really struggles with two types of decks. Firstly, anything with main deck life gain like Abzan and Anafenza's Company; Naya Burn only manages to win between 41 and 44.4 percent of the time in these matchups. While Scavenging Ooze doesn't seem to be enough to get the job done (judging by a 50/50 matchup with Jund, although this is likely aided by the life loss of Dark Confidant), once an opponent casts a Kitchen Finks or Siege Rhino, both of which are the equivalent of a Time Walk for a deck that draws about two points of damage per turn, things go downhill quickly for the Burn player. The second class of decks that really give Naya Burn a hard time are go-wide aggro decks with a fast clock. Here the prime examples are Affinity (where Naya Burn only wins 36.4 percent of the time) and Elves (42.6 percent). These decks not only offer a fast enough clock to race the Burn deck, but also put the Burn player in the undesirable position of being forced to Lightning Bolt problematic creatures instead of furthering their game plan of diminishing an opponent's life total.
One last note on Burn as an archetype: it seems that, at this point, Naya is the only game in town. The only other Burn deck to show up at all in our sample is Zoo Burn, which is straight R/G and plays things like Hooting Mandril and Tarmogoyf, but this deck doesn't appear to be very good, considering it only won 46.3 percent of its 432 matches. Mono-Red Burn doesn't show up at all, nor does Boros Burn. Apparently Atarka's Command combined with easy mana fixing is good enough to drive these other decks completely out of the format and makes the green splash the new normal.
The New Hotness - Grixis Control
Much like Naya Burn, Grixis Control (the second most played deck in our sample at a touch over six percent of our field) has basically pushed other Grixis builds out of the format. While both Grixis Twin and Grixis Delver still exist, both come in outside the list of 25 most played decks in our sample, which is especially interesting when you consider that over the past few months both had been in the top five most played at different points.
If you remember back a few weeks to when Grixis Control first started making waves in Modern, many people figured the rise of Kolaghan's Command would be a major problem for Affinity. An automatic 2-for-1 that doubles as main deck artifact hate seems good against an all-artifact deck right? Well, would you like to take a guess at what matchups our data suggests Grixis Control struggles with most? That's right, Affinity. In fact the little Robots manage to take down the Kolaghan's Command deck nearly 60 percent of the time.
On the other hand, Grixis Control is where you want to be against creature-based combo decks. It seems that having access to four copies of Terminate backed up by more removal and counter magic is enough to keep Splinter Twin and Anafenza's Company from going infinite nearly 60 percent of the time. This same setup gives the deck a great matchup against Infect, although it has the downside of making the Boggles matchup extremely difficult (maybe some Liliana of the Veils will fix this?) Otherwise, we see a lot of 50/50 type matchups including those against R/G Tron, Merfolk, and Naya Burn.
The 50/50 Decks - Jund and Abzan
Ever since Jund has been a deck, it has been flaunted for its ability to have game in nearly every matchup, which in theory allows a skilled pilot to shift a matchup a couple percentage points based on their ability. This appear to be true, at least for the most part: in half of the most popular matchups, Jund is favored by somewhere between 0.5 and 2 percent. The outliers here are Anafenza's Company where Jund is a significant favorite (winning more than 62 percent of the tiime) and against Merfolk or Tron where Jund only manages to win about 40 percent of the time. I'm no expert on the Merfolk matchup so I'm not really sure what is happening here, but it is notable that even with a main deck copy of Ghost Quarter and three or four copies of Fulminator Mage in the sideboard, Jund still can't beat Tron with any consitency. I'm not sure there is a reasonable way of fixing this problem without giving up percentage point in other matchups. I guess you just cross your fingers and hope you don't run into the matchup too often (which isn't an unreasonable wish, considering Tron of all flavors only makes up about 6 percent of our sample).
While I'm not sure why, Abzan is significantly better than Jund in two key matchups, gaining nearly 8 percentage points against Naya Burn and Grixis Control. As I mentioned earlier, I'm pretty sure the improved Burn matchup is the result of Siege Rhino, not to mention some key sideboard cards like Rest for the Weary and Timely Reinforcements, but figuring out the Grixis matchup is a little more tricky. My first guess would be the main difference is having access to Lingering Souls, but there could be more to it than that. Unfortunately we don't have enough information about enough matchups to say definitively which deck is better in the broader modern meta.
The "Awesome, but.." Deck - Affinity
Affinity has the best overall winning percentage out of the 10 most played Modern decks coming in a bit over 55 percent. Better yet, it has a killer matchup against the two most played decks in our field coming in at 63 percent against Naya Burn and 59 percent against Grixis Control. This isn't even considering the positive matchups with R/G Tron, Amulet Bloom, Jund, and Merfolk. So we should all just follow Frank Karsten's lead and play Affinity right? Not so fast, there is one very big problem for Affinity: Splinter Twin.
The Affinity vs Splinter Twin matchup is the single most lopsided matchup among all tier one decks. It is basically the Miracles vs 12-Post of the Modern format. In fact, Affinity only managed to beat Splinter Twin 36.8 percent of the time over the course of more than 150 matches. Considering that Twin makes up somewhere between 4.7 and 8.4 percent of the field (for some reason Twin is underrepresented in our sample compared to the Modern Metagame Page), if you choose to play Affinity at a big 15 round event you are almost starting off with 0-1, since you are likely to run into Twin at least once and probably more often at the top tables since Splinter Twin also has a solid overall MWP at 53.6 percent.
The Good Win Percentage, Bad Matchup Deck - R/G Tron
Overall Tron has a reasonable 52.5 win percentage, which isn't astounding but on par with the rest of the most heavily played decks in our sample. The problem is, I have no idea where this percentage is coming from. Sure, it beats Anafenza's Company 94.74 percent of the time and Jund about 60 percent of the time (both matches less than 100 matches played, which likely skews the numbers), but otherwise Tron isn't significantly favored anywhere. While it does come out slightly ahead against Grixis Control (52.5 MWP), it's a dog to Affinity, Burn, Splinter Twin, Boggles, Merfolk and of course Living End (0.00 MWP in a small number of matches). So the next logical question is, where does Tron pick up its wins?
One of the things I realized while looking over this data is just how wide open Modern is as a format. In our sample, 32 decks are at least one percent of the field and only two decks come in at over five percent. The most played deck (Naya Burn) only makes up 7.6 percent of the meta. In Standard only 26 decks are over one percent, five are over five percent, and the most played deck (Atarka Red) makes up a full 13 percent of the field. What this means is that it is completely reasonable that Tron could have a bad matchup against other heavily played decks (overall, Tron played more than 2,100 matches in our data set, but less than 600 came against the four most popular decks in the format) but still have a positive total MWP by feasting on the rest of the field.
Quick Hits on the Other Popular Decks
- Anafenza's Company is the only deck out of the ten most played which has a losing total MWP at 49.1 percent. That fact that this is for a rather large sample of matches (over 1,600) is a bit concerning, as is the fact that several important matchups (Jund, Splinter Twin, Grixis Control and Tron) are overwhelmingly bad matchups with Anafenza's Company losing more than 60 percent of the time.
- Merfolk is basically the anti-Tron; it has good matchups against most of the popular decks but really struggles against tier two options like Living End and Boggles.
- Elves and UB Faeries were involved in the exact same number of matches in our data set (1,334), the difference here is Elves won 54.7 percent of the time, while UB Faeries only managed to do so in 48.4 percent of its matches.
- Abzan is favored against every deck it played at least 100 times. This list includes Grixis Control, Splinter Twin, Naya Burn, and Affinity.
- Bloom Titan doesn't like playing against Affinity or Splinter Twin; in over 100 matches against each it posted a MWP of 40 percent of less.
Other Notes and Observations
- While there is a persistent perception that the deck is bad, Jeskai Control actually has the best win percentage of decks making up at least 1 percent of our sample. While it didn't play enough matches against any individual deck for us to break down the good and bad matchups (only matchups with 100 or more matches qualify), an overall win percentage of 56.2 percent is quite impressive.
- Living End was only played in 450 total matches and has a poor overall win percentage of 44 percent. However the deck does boast two of the most lopsided matchups likely aided by a low sample size; it beat Tron 100 percent of the time while losing to Naya Burn nearly in almost 80 percent of the time.
- The three worst decks? Sultai Dredgevine, Jeskai Mentor and Jeskai Swans. All together these decks were played 244 times and only managed to win 54 matches, a 22.1 combined win percentage This is the equivalent of going 3 and 12 over a 15 round Grand Prix.
- Two decks posted a win percentage of over 57 percent over the course of nearly 1,2000 matches combined. The first is Bant Collected Company which is just oozing with value plays like Collected Company into Eternal Witness and Flickerwisp, get backCollected Company, cast Collected Company, flicker Eternal Witness, get back Collected Company. I'm pretty sure this is the next Modern deck I'm going to build, because, well...value. The second is Suicide Zoo, which played one-drops of four different colors and might be the best deck in the format for winning on turn two or three. While the sample size is too small to make any bold statements, if you are looking for a rogue option in Modern, both of these decks are worth testing.
- How about the most surprising total win percentage? On the positive side, Jeskai Control coming in at over 56 percent (which I mentioned before) probably tops the list. As far as bad surprises, semi-popular and interesting decks like UR Storm, WB Tokens, Skred Red, and Lantern Control coming in at between 40 and 44 percent are probably most noticeable.
- Quick shout out to one of our Budget Magic decks! Blistering Rage managed to put up a respectable 47.6 percent win percentage over the course of 160 matches. Not bad for $37 or 7 tix.
- Modern isn't the places for fringe decks, at least if your goal is to win consistently. Of the 20 least played decks, only three managed to post a match win percentage of over 50 percent. Of the 20 most played decks, 13 have MWP's of over 50 percent. If we zero in on only the 10 most heavily played decks, just Anafenza's Company won less than 50 percent of the time.
Conclusion
Anyway, that's all for today. If you haven't done so already make sure to jump over to the Modern Metagame Breakdown page to check out the number for yourself. As always leave your thoughts and opinions in the comments, and you can reach me on Twitter (or MTGO) @SaffronOlive.;
