oraymw's picture
By: oraymw, Oraymw
Oct 17 2013 5:13am
Login to post comments

Welcome to Ars Arcanum, the MTGO stats-based limited column. In this installment, we’ve got a Theros Draft Overview. These articles are my bread and butter, and I’m really excited to show you the data. With every new set I pick up a few more readers, so I’ll briefly mention how these articles work. During release events, I’ll open up MTGO and pull up the 64-person draft queues. I then start watching replays for all of those events. As I watch them, I record various data points, such as the colors being played, or the number of turns before the game ends. I then take all of that data and do some number crunching with it, put it in some pretty charts, and do a massive analysis, which is what you are currently reading.

Before we get started, I do need to address a few points. When I first started doing these articles, it was seen as a pretty new thing. There were a few people that had tried to capitalize on the potential for using MTGO as a data collection device, but none of those people ever saw a lot of success, for a variety of reasons. However, in the past year, I’ve seen a lot of people that are approaching new sets from a statistical point of view. The problem is that a lot of these people are just doing a poor job of it, and I wanted to talk about some of those points. I really don’t want to snipe at people, since that would definitely seem petty. Instead, I’d like to explain some of the reasons why I use particular methods, instead of others. If you don’t want to read a detailed explanation of methods, then please feel free to jump down to the data.

First, I’ve seen a trend of people that using computer assistance to scrape data from MTGO. This is a fabulous idea, and I’m glad that people are doing it. However, it seems that many people are not understanding the limitations of these kinds of tools. The upside of these techniques is that they allow people to gather a lot of data in a very short amount of time. This does come with several downsides though. For example, one of the most important parts of my analysis is where I break down the win rates for decks by archetype. This means that I have to watch matches, and determine what colors those players are playing. This often includes some amount intuition, and most people’s heuristics are sloppy, which means that they are just not very accurate at determining these things. The problem is that most heuristics tend to count lands or spells that are played, and then assign a color to the deck based on what is played. As I watch games, I’m able to see both lands and spells, and then apply my understanding as a player to determine what deck the players have. There are downsides to this approach too, which is that it can be subjective, but I do have tight heuristics that I use whenever I can’t easily tell what the main colors of the deck are, and if I still can’t determine the answer, then I just cut that data from the sample.

Data scrapers also tend to use game data, instead of match data. This is a problem in several ways. First, it is much less grokkable and applicable, since players tend to think much more in matches than in games. But more importantly, this segregates data that carries over between games. This comes into play especially when we look at the colors of decks; for my data, I determine the characteristics of the deck based on the entire match, which makes it easier to categorize decks, and makes the data less susceptible to mana screw outliers. For example, if the player only draws Forests in one game and loses, but in the next game, he or she draws Forests and Islands, then I can categorize the deck as UG. But if a data scraper sees this and doesn’t combine match data, then they will record data for a UG deck and a mono-green deck.

Finally, data scrapers are terrible at gathering data for the ending turns of games. These will normally just record whatever number is shown in the box at the end of the game, which has several problems. First, it is counting games where the loser lost because of disconnection or time, which simply doesn’t help us understand the format. Secondly, it counts early concessions; when someone concedes on turn 3 because they got frustrated. This is interesting,  but it doesn’t give us an accurate idea of the speed of the format, since it skews much lower. When I gather my data, I don’t count disconnections for game ending data, and I don’t count concessions without an imminent win on the table. This makes our ending turn count much more accurate, which is helpful because it allows us to see how long games really take, which we can apply to our deckbuilding.

But the most important part of actually watching the games is in the insight that I gain into the data. A data scraper doesn’t actually see what is happening in game; I see every play that the players make, and I can tell what cards are overperforming, what cards are being overvalued, and which strategies are performing particularly well. This is what Ars Arcanum does well; it can take the data and break it down into understandable and applicable information that you can then apply to your next draft.

That isn’t to say that this approach is superior to other approaches. It takes a lot more work to gather this data. It has many more opportunities for bias. It also produces naturally smaller sample sizes, and it doesn’t take the data from as broad of a swath of events. Because of this, it is important not to take any statistical analysis of a format at 100%. I don’t even take my own data as given, because I fully understand the multitude of ways that it can fail. Instead, I would recommend that players use a variety of methods to triangulate their ideas about a format.

The last thing I want to mention is my normal disclaimer. Many players tend to look at these articles as a blueprint to follow precisely in their drafts. That is not the purpose of these articles. They provide a very accurate look at a narrow slice of events. I tend to think of them more in the vein of topographical maps of the terrain for a battlefield. I can give you an idea of the best strategic locations, and where an opposing army will probably try to make a stand. I can tell you where the resources are, like water, or where you can hide. What I can’t tell you is where the individual soldiers will be once the battle breaks out. Likewise, this data is useful for getting a big picture view of the draft format, but it often becomes less useful as soon as you open up that first pack.

With all of that business out of the way, let’s jump into the data!


Ending Turn of Theros Draft Games

Ending Turn of Theros Draft Games Compared with Average Set

Ending Turn of Theros Draft Games as Compared with GTC and M14

For the speed charts, I’ve made two changes. First, the speed charts are written in percentages. This makes for more readable and grokkable graphs, since it is easier for our minds to wrap around smaller numbers. Second, I added the middle chart which compares the set against the average ending turns of all the sets that I’ve looked at in all of my data, which makes for a really good picture of the average draft format. Also, the third chart still compares this set with other sets, and I specifically chose GTC and M14, as two extreme cases for comparison.

When Theros was first released, I heard a lot of wildly different opinions on the speed of the format. Some people said that Theros was going to be really slow, and others really fast. When I did my spoiler analysis, I decided that there were elements pulling in both directions, which would probably end up leaving the format much more like an average format. Luckily, this is exactly what we see. Theros matches the average set very well, except for the big spike at turn 9. This is really weird. Most interestingly, this number comes because Theros shaves a few points off of turns 4, 5, 6, 10, 11, 12, 13, 14. It doesn’t trail behind the average by a lot in any of those areas, but it does so by enough to make for a huge spike on turn 9. It does make us wonder why there is a spike at that turn. I was able to isolate the turn data by archetype, and you’ll see more of that information later in the article, but I did see that the majority of this spike is made up by Green, Red, and Black decks. Turn 9 is the turn where a 17 land deck will often hit 7 mana, which is when the green decks can use the monstrous ability on Nessian Asp or Keepsake Gorgon, and the turn when red can bestow Purhporos’s Emissary. However, we didn’t see a spike in white decks, which suggests that they are usually using Heliod’s Emissary as a creature, rather than using the Bestow.

In any case, the most salient point is that Theros is a very average format for speed. It has fast decks, and it has slow decks. In fact, I think that Theros has the widest spectrum of speeds for decks that I’ve ever seen. The RW deck is probably the fastest deck I’ve ever seen in these studies, while the UB deck is one of the slowest. The point is that there are a wide variety of decks in Theros, and you really just have to draft the format as if it were an average set. The set comparison chart really serves to drive this home. Theros is easily a couple of turns slower than GTC, which was a very fast format, but it’s also definitely a turn or two slower than M14, which was a very slow format. This means that curve is important, but that you also need to plan for ways to win the late game. Seven drops are definitely a liability in the format, though they can be playable as long as they have a dramatic impact on the game. Eight drops are mostly unplayable without a lot of support in ramp spells, though it is definitely possible to pick up the kind of deck that can use those spells.

There is one important implication caused by the strange spike at turn 9. As I said, this mostly has to do with the fact that there are a lot of things that decks need to do at seven mana. This is a format that is fast enough that you don’t want to fall behind on early land drops, but it also has a lot of late game mana sinks. Most decks want to be able to reach seven mana fairly regularly. Because of this, I would say that you should be playing 18 lands as your default in this format. That doesn’t mean that you’ll always play 18 lands; decks with a very low curve are probably fine at 17, especially if they don’t have many late mana sinks. Traveler’s Amulet can often substitute one for one for a land up to the second amulet. Decks with a lot of scry or cantrips can get away with fudging their numbers on lands. But in any case, plenty of people have suggested running 18 lands as a default, and the data seem to corroborate that.

I did want to make one more point about speed, and this mostly has to do with things I’ve learned from watching games and from drafting upwards of twenty times. I’ve noticed that there are a lot of useful five drops in the format. White has Setessan Griffin, blue has Prescient Chimera and Mnemonic Wall, black has Lash of the Whip, red has Rage of Purphoros, and green has Nessian Asp and Pheres-Band Centaurs, and that is just at common. The problem is that it is very easy to overload on five drops in this set. When I first saw Prescient Chimera, I thought it would be one of the best blue commons, but it turns out that it has too much competition at the five mana slot, and so you generally need to pass it for something cheaper. That doesn’t make these cards less good; it just means that you can’t prioritize them, or your deck will be way too clunky. This also helps make Opaline Unicorn a little better, since it gets down your five drops sooner.

Now that we have a good idea of the speed of the format, let’s go into the archetypes.


Popularity of Two Color Archetypes in Theros Draft

Popularity of Colors in Theros Draft

The above two charts show the popularity of decks in Theros. The first chart shows the popularity of the two color archetypes, and the second chart shows the popularity of individual colors. For archetypes, we see two major tiers: Azorius, Simic, Golgari, Dimir, and Boros make up the top five decks, and they are all pretty close together in popularity, while Gruul, Selesnya, Rakdos, Orzhov, and Izzet make up the second tier. The color chart shows us something that we probably could have guessed quickly from a cursory glance of the archetype chart, which is that blue is the most popular color, while red is the least popular.

In the top five decks, there are three blue decks, two white decks, two green decks, two black decks, and one red deck. The most interesting thing to me is that four of the top five most popular decks also happen to be four of the five slowest decks in the format. Dimir, Golgari, Azorius, and Simic all fall on the slower end of the spectrum in Theros, which makes sense when we consider that blue is the slowest color while red is the fastest color. I should note that neither Azorius or Simic are particularly slow; they would definitely be categorized more as tempo decks, and not as control decks. But when we take a look at the comparative speed charts for the archetypes, we’ll see that there are several decks that are faster than either of those.

This raises an interesting question. If a large part of the field is playing slightly slower decks, is it better to play a faster deck, or a slower deck? This is a difficult question; when a lot of slow decks are popular, it can often mean that people have to skew their decks more towards winning the long game in order to compete. This often provides an opening where fast decks can gain an advantage. Likewise, a format with lots of fast decks often provides a good opportunity for a slow deck to overperform. As I’ve said many times in Ars Arcanum, it’s more important to choose the best deck for a particular metagame than the most objectively powerful deck. However, up to this point, we just don’t know whether this is the kind of format that will reward this type of strategy.

There is another very important point that I need to make about popularity in Theros. The data for colors is a little bit misleading, and it could lead people to misunderstand the relationships of the colors in the format. Yes, blue is the most popular color, in that it shows up the most often in the most popular decks. But it should be noted that while blue is strong in Theros, it is mostly strong as a support color. It is definitely possible to have a blue focused deck, but blue usually acts like the salt of the format; it brings out the flavor in other decks. This is because blue’s best commons are flexible “swiss army knife” cards. Blue’s best cards are easy on the mana requirements, and they are strong in a variety of different strategies. Voyage’s End and Griptide are great in an aggressive deck, because they allow you to set up a good board position early on, and then ride your tempo advantage to a win. Normally those two cards would play worse in a control deck, but in Theros, the control decks need to be able to deal with very large creatures that have had a lot of resources invested into them. Hitting a 6/6 Wingsteed Rider or an 8/9 Nessian Asp with Voyage’s End will often give the control deck enough of a breather to set up a stable board position. Likewise, blue’s two best creatures, Vaporkin and Nimbus Naiad, are great for any deck. In an aggro deck, Vaporkin comes down early and keeps pressuring your opponent’s life total, while Nimbus Naiad lets you make a big creature fly to close out the game. But in a control deck, they provide the evasive creatures that can win the game while your other creatures hold down the fort.

Blue is the most flexible color, but you don’t have to have a lot of blue in your blue decks. If you’ve got a heavy black deck with a few flyers and bounce spells, you’ll be in great shape. Likewise, a WU deck will want to focus on lots of great white creatures, and then support those creatures with blue spells to create a tempo advantage or to use evasion to close out the late game. The flavor of a U/X archetype is determined much more by the color with which blue is paired.

Conversely, red is a color that forces its companions to adopt its strategy. The four red decks in the format are the four fastest archetypes, and none of them want the game to go very long, because they run out of steam too quickly. Red decks are focused and aggressive, and they want you to reduce your opponent’s life total as quickly as possible. R&D has mentioned that they attempted to make a UR deck that was more controlling, but it turns out that that attempt was unsuccessful, because the UR decks that are good are the ones that want to attack early and often, and then close out the game with blue bounce spells. So, while red is less popular overall, this is mostly because it is so parasitic, and it requires dedication to a certain strategy.

Overall, this might be the set that is best balanced for popularity that I’ve seen. There are a wide range of popular decks, with a variety of strategies, and the differences in numbers seems to have more to do with the strategies of the colors, and not so much with their power levels. Players have yet to find a truly dominant deck, and so they aren’t leaning that hard in one direction. Some players will swear by particular archetypes in Theros, but it seems that most players are tending towards being more open in their deck choices.

In our next section, we’ll take a look at the win rates of these decks, and I’ll start to break down a few particular strategies.


Win Rate of Two-Color Archetypes in Theros

Win Rate of Colors in Theros

Here’s where we get to the real meat and potatoes of the article. In the first chart, we see the win rates of the two color archetypes in Theros. I chose to continue to call them by their guild names since people should still connect with Ravnica. If not, the chart is still handily color coated. If you are new to Magic for Theros and you are colorblind, then I apologize. The second chart shows the win rates for each color in Theros.

I should make one important point right off the bat; this is a very well-balanced format. Our top deck clocks in with just under a 59% win rate, and we have five decks with win rates of 50% or better. We do have one low performer with Gruul at 41.38%, but all of the other decks were still able to make a respectable showing. On the color chart, we see that black, blue, red, and white are all neck and neck, without a clear frontrunner, and green falls behind the pack a little bit, but not enough for me to advocate avoiding the color. This just means that we need to be more particular with our green decks. We also see that three of the top five decks have blue in them, but as I mentioned in the last section, blue is just a very strong support color in this format, and it’s real strength is in bringing out the best qualities of the other colors.

In this section, I’m going to focus on three particular topics. There is a ton to talk about in Theros, especially with this data, but it just isn’t possible to go into all of the topics full detail. The three topics I’m going to cover are red decks, the WU deck, and the black devotion deck. I haven’t shown you the stats for that deck yet, but I will definitely be getting to it by the end of the article. I would love to go into more detail on blue as a color, or on the RW deck, or on ways to make green work well, but I’ve chosen to focus my articles on the strategies that the data shows are overperforming.

Our first topic is red decks. The top two performers in the format are RB and UR. The UR deck is about 3 points higher than Azorius, which is within range of our margin of error, but both decks being at the top leads us to believe that red aggressive decks have a powerful position in the format. I should note that these two decks also happened to be two of the least drafted decks in the format. In the past, I’ve repeatedly made the point that the data does not show that less popular decks suddenly pick up a large amount of wins, but that this tends to depend more on metagame factors than underdrafting colors. However, in my over/under drafting article, I did make the point that underdrafting can often count for around 3 to 5 points of a decks win rate. Since the win rates of the decks in Theros are so close, it is easy to imagine that drafting these colors more heavily could easily cause shift in the win rates of the best decks.

The main strength of these decks is that they are plugging a metagame hole. While the most popular decks in the field tend to be slower decks, these red-based aggressive decks are able to punish any deck that gets off to a slow start. The best commons in red are definitely Lightning Strike and Rage of Purphoros. Both of these cards are among the best removal spells in the format. They are easy to cast, and they get blockers out of the way so that the red decks can close out the game. Lightning Strike can just go to the face to win the game, but Rage of Purphoros is also very strong since it can remove a blocker and then set up more live draws when the red deck no longer needs land. But you don’t need me to tell you that top tier removal is good, so let’s look at the other important commons in the red deck.

The next best common in the color is deceptive. Normally Auras are pretty bad, since they provide your opponent’s with an opportunity for a two-for-one. However, Dragon Mantle avoids most of these problems. It only costs one red mana, so it is easy to slot into your curve wherever you need it, and this is important because it allows you to easily play around removal. Then it automatically replaces itself, so you don’t have to worry about removal on their turn. But these two things wouldn’t normally make a card strong. Dragon Mantle, however, comes with two very important bonuses. First, Dragon’s Mantle is part of a very small group of one mana heroic enablers. But even more importantly, it’s the only member of that group that provides a permanent bonus. Just being red, a card that targets, and a cantrip is enough to make Dragon’s Mantle good, but it is amazing how relevant the firebreathing ability is. This is an ability that we’ve seen many times in Magic’s history, but it is often limited by the bodies on which it is found. The ability is certainly powerful, but it really shines when you can put this ability on any creature. A Deathbellow Raider suddenly becomes a powerful and resilient threat. Ill-Tempered Cyclops allows you to dump all your available mana into a trampling haymaker in the late game. Spearpoint Oread becomes impossible to block. And Two-Headed Cerberus becomes game warping. This gets even better when paired with blue or black, since you’ll have plenty of evasive creatures that become much more threatening with firebreathing, as well as a few high toughness creatures like Gray Merchant of Asphodel or Wavecrash Triton that can force your opponent to trade off two creatures for your one.

The next two important commons for red decks are Ill-Tempered Cyclops and Spearpoint Oread. The red deck really doesn’t want to play a long game, so it is important for it to have cards that can come down early and impact the board. The strength of these cards is that they give the red deck the reach it needs to punch through that last little bit of damage from turns 7 to 9. By that time, your opponent’s will have started to build up some defenses, but these two cards allow you to keep up the pressure once you get to six mana.

Minotaur Skullcleaver and Two-Headed Cerberus are also key roleplayers in the deck. The Skullcleaver is much better than it looks, because it allows you to hit with a potent life point swing on turn three, but after that, it provides a body for your different aura effects. Since one of the best ways to deal with bestowed creatures is one of the blue bounce spells, Skullcleaver provides some interesting resiliency. Two-Headed Cerberus is another card that performs a lot better than it looks. I like to think of it as red’s other heroic card; you really want to be putting an aura on this. If you can’t, it still becomes a pretty good target for Titan’s Strength, and it is the absolute best target for Dragon’s Mantle.

There are two more cards that I feel need an advocate in this format. The people that have been playing a lot of red probably won’t be surprised, but when I talk to other people, these creatures tend to get pretty short shrift. Those cards are Akroan Crusader and Deathbellow Raider. First, the Crusader is actually a key part of the red deck. This aggressive deck really wants to get creatures out as early as possible and then start suiting them up if possible. Akroan Cruader provides a relevant one drop that can turn on your Ordeals, Dragon Mantles, Scourgemarks, and Fate Foretolds. At the same time, the Crusader gives you a little bit of extra value when you play these cards, which makes your tempo spells and removal even better. Having a swarm of creatures is very useful when you get to the point of the game that you start dumping out your Griptides and Rage of Purphoros. Deathbellow Raider is another key card in the red deck. It is obviously very strong in the RB deck, where it provides a powerful early game threat that is still resilient in the late game, and that wears auras very well. But it is also deceptively powerful without black. A 2/3 for 2 is able to attack into almost any opposing board for the first few turns in this format. Then, when you hit five mana, it isn’t that hard to suit up with an aura. A Baleful Eidolon turns it into a very difficult to block 3/4 that will soon be able to regenerate, while both Nimbus Naiad and Spearpoint Oread allow your Deatbellow Raider to basically attack with impunity. Both of these cards require you to build your deck to support them, but they are very powerful once you have done that.

The last thing that I want to show you about red is an example of one of the red deck’s Ending Turn charts.

Winning Turns of RB decks in Theros as Compared Against the Format

This chart is a very clear illustration of how the red decks play out in Theros. The deck has a huge spike at turn seven, with a decent number of wins on turns eight and nine. But then the graph drops off precipitously. This highlights a key point about red in Theros; if you aren’t able to win by turn 9, then you will probably just lose. This is a deck that needs a lot of focus to work. You want to draft a lot of aggressive creatures and a lot of ways to support those creatures, and you have to be very careful not to overload on expensive cards. You should go light on the five drops, and anything that costs more than five mana should be looked at very carefully before you play it.

The next deck that I want to talk about is the WU deck. My instincts tell me that this is the best of the two-color archetypes, since it is the most popular deck and is only a few points behind on win rate against the two least popular decks. I can’t say that with complete certainty, but it is definitely high on my priorities when I go into a Theros draft. I also feel that this is one of the most misunderstood decks in the format, and I would like to take a little bit of time to set the record straight for WU.

First, it seems that most people are thinking of the WU deck has a very aggressive white heroic deck. This is a little bit off, because the WU deck isn’t actually particularly fast. It certainly has the capability of very aggressive starts, and it can back up those starts with blue tempo cards, but it actually tends to work out best when it is built as a traditional WU skies deck.

Throughout this article, I’ve been talking about a speed chart that ranks all the decks. I’ve categorized all the deck archetypes according to the turns on which they win, and made a chart that shows all of those decks in order of their average winning turn. I’ve also made charts that show the progression of those decks against the format, and I’d like to bring out those charts now:

Average Winning Turns of Two-Color Archetypes in Theros

Winning Turns of WU in Theros as Compared with the Format

As you can see, Azorius is actually not a particularly fast deck. In fact, it is the fourth slowest deck, and it is even slower than two of the blue decks. WU is definitely faster than the black based attrition/control decks, but it would be a mistake to draft Azorius as if you were drafting an aggro deck. Azorius’s own chart is even more helpful. While we do see that seven is the highest number on the chart, which shows us that WU is more of a tempo deck, we also see that it is flatter than the aggro decks, and that it has plenty of game all the way until turn thirteen, and even performs better than the format on average from turns ten to thirteen.

With that said, WU is definitely a heroic deck. Wingsteed Rider is probably the best common in the deck. In pack 3, I would take Rider over any other common in the deck. In pack one, it will often be better to take Voyage’s End or Griptide, simply because those cards are such great cards in any archetype, but once you are solidly WU, there are very few cards better than Wingsteed Rider, even at rare, though Thassa’s Emissary and Sea God’s Revenge certainly count as uncommons that I would take over the rider.

The basic strategy for the WU deck is to gum up the ground with blockers, and then win the game through the air with your flyers. Wingsteed Rider, Vapor Kin, Nimbus Naiad, and make up the bulk of your flying force, and they are your main ways to win the game. Playing auras on these creatures is also quite good, since it allows you to make a powerful threat that closes out the game quickly while your opponent tries to deal with your defenses. Your premiere defensive cards are Coastline Chimera, Omenspeaker, and Wavecrash Triton, though Scholar of Athreos is also a very solid conclusion, especially if you splash a little black. Setessan Battle Priest has also been better than expected in this deck, since it is a great way to hold the ground early on, and it wears Fate Foretold and Chosen by Heloid very well. The best part about the Wavecrash Triton and the Coastline Chimera is that they are particularly good wearers of the white auras and the two mana cantrip auras. Either of these cards becomes incredibly hard to deal with when they wear either a Hopeful Eidolon or an Observant Alseid, letting you get in an attack while still maintaining defenses.

Once you’ve gotten to the late game, the WU deck mostly just wants to keep Nessian Asps off the board while it wins the game, which is where Voyage’s End, Griptide, and Sea God’s Revenge come in handy. All of these cards are perfect for closing out the game, and allow you to keep the clock up with all of your flyers.

The best part of this deck is that it is truly an aggro-control deck. It can land early aggressive creatures and push those creatures through over and over with bounce creatures, but it can also assemble a great defense to stay alive while the flyers win the game. This versatility is what makes it one of the best choices in the format, since it can compete against the hyper aggressive red decks and also against the defensive black decks.

This brings me to the last deck that I’m going to discuss, which is the black devotion deck. I won’t go into it in too much detail, since it’s been covered at length in other places, such as in Marshall Sutcliffe’s Limited Information article. Marshall called this strategy the strongest deck in the format, and I really wanted to see if that was the truth. I should mention that I think that “mono-black” is a misnomer for this archetype; it is true that these decks want to be very heavily black, but they don’t necessarily need to be mono black. Normally, the monoblack deck relies on cards like Tendrils of Corruption, Corrupt, and shades, and so those decks normally need an extraordinarily high swamp count, and any non-black card you add to the deck just makes it worse. But in Theros, you just want a lot of black symbols. This means you want a heavy black component, but it is perfectly reasonable to support the black cards with a powerful suite of non-black cards.

This is exactly the kind of deck that is difficult for me to gather data on accurately. Since I can’t see people’s drafts, their decklists, or their hands, I can only gather data on what actually comes into play. For a deck like this, I have to base my data on whether decks are playing the black devotion cards. That is a complicated thing to do. Those cards tend to be powerful, and playing powerful spells tends to win games, so that necessarily skews the decks toward a higher win rate. In order to measure this as well as I could, I decided to set a relatively low bar for the devotion decks. Specifically, if a deck played two separate cards from a combination of Gray Merchant of Asphodel, Disciple of Phenax, Abhorrent Overlord, or Erebos, God of the Dead. So, if I saw two separate copies of Gray Merchant, or if I saw a combination of Gray Merchant and Disciple of Phenax, or if I saw two separate Disciples, or if I saw a Disciple and an Abhorrent Overlord, all of those decks would count as a black devotion deck. Since these cards are also all pretty strong, we should take the data for this deck with a little bit of a grain of salt. I think a conservative estimate would be to shave about 10% off the wins for the archetype.

With that said, the black devotion deck managed to put up a 72.3% win rate (without accounting for that 10% guess). This is an absurdly high win rate, but if we shave off our conservative amount, we still come up with a win rate of above 60%, and it’s still possible that the win rate for black devotion decks is around 65% - 68%. I mean, theoretically it is even possible that the win rate could be higher. With this kind of win rate, I have to say that I concur with the opinion that the black devotion deck is the strongest deck in the format. With that said, it’s also a difficult deck to get into, since so many players are prioritizing those cards. I would say, though, that P1P1 I would take Gray Merchant over any other common in the set.


Phew. We’re now up against 7000 words. Thanks for sitting with me through this wild ride! Hopefully this information will help you succeed in Theros. In conclusion, here are the salient points from the article:

  1. Theros as a very average speed as compared with other formats. It has a wide variety of possible deck strategies.
  2. Most Theros decks should default at 18 lands, though it is possible to fudge on that number in an  aggressive red deck, or if you have a few Traveler’s Amulets or Voyaging Satyrs.
  3. There are an overabundance of decent five drops in the format, which means that you can’t really prioritize them unless they are extremely good (think Keepsake Gorgon or Gray Merchant).
  4.  Blue is the most popular color and red is the least popular color. This is mostly because blue is a very powerful support color that brings out the best in other colors, while red requires a heavy commitment to an aggressive strategy.
  5. Currently, the most popular decks in the format are slower decks, which provide a good window for aggressive decks to excel.
  6. The win-rates in Theros are very flat, and all of the archetypes are viable, depending on what cards you end up getting passed.
  7. Aggressive RB and UR decks are currently the winningest decks in the format. Part of this is because they are underdrafted, but it’s mainly because of the nice way that they take advantage of the current metagame.
  8. WU is probably the best of the two color archetypes, though it is a slower deck than many players realize, and functions best as a traditional WU skies deck.
  9. Black Devotion is probably the strongest deck in the format, though it is a little bit difficult to get that deck regularly. There are few cards that you should take over Gray Merchant Pack One Pick One.

As always, you can follow me on twitter @oraymw for updates about articles. I’ve also put up a Tumblr account at http://oraymw.tumblr.com/ where I post links to my articles. You can go there and subscribe to the RSS feed, and then you’ll be able to get updates whenever a new article goes live.

Finally, I encourage you to check out the podcast that I do with my buddy Zach Orts, which is called All in the Telling. In it, we look at stories from a professional standpoint in order to get a better understanding of why they are important to the human experience. But mostly, we just talk about what makes awesome stories awesome.

Ars Arcanum Archive




Amazing by mtgotraders at Thu, 10/17/2013 - 21:49
mtgotraders's picture
Once again Matthew delivers! Thanks for doing these Matthew.
Great Job! by Orgion at Fri, 10/18/2013 - 09:03
Orgion's picture

Cracking article as always. Played just over 20 odd drafts too and your research matches my current experience, with one difference though in that I've not been able to draft a successful RB deck at all nor have I seen one in action. I'll admit that I don't generally start out trying to draft RB so maybe that's the reason. Still given that you say they have a good win percentage, I'd love to see what a successful one looked like.

Overall, reckon you were right on the money with your previous article, this is definitely seems a tempo orientated environment. Once someone starts making a board presence it feels like you have to be doing something to the board every turn; missing a land drop really hurts. The format feels very much like Avacyn Restored, even to the point that RG is a weak deck archetype. Do you think it's because R and G contain the most common monstrous creatures and so are more susceptible to being 'set back' in tempo by bounce and other removal spells? As with RB I'd love to know what makes a good RG deck.

You sound like you've got lots to say still, so please, please give us another article!

Oh, and just one other thing, I know you said your research tended to smaller sample sizes. I'd be interested in knowning the actual size of your samples when you report your stats, maybe as a footnote to the charts perhaps?

Anyway, as I started saying - really good article! thanks for taking the time.

Here's a version of the deck by oraymw at Sat, 10/19/2013 - 01:23
oraymw's picture

Here's a version of the deck that went 3-0 (6-0 in games):


Neato, this goes exactly with by Doctor Anime at Sat, 10/19/2013 - 14:16
Doctor Anime's picture

Neato, this goes exactly with what I've been saying. Awesome breakdown of the stats and analysis.