For both Innistrad and Dark Ascension, I put together “Limited Overview” articles based on watching hundreds of draft matches, and I’ve done that once again with Avacyn Restored. When the format goes into release events, I enter into the 64-man drafts and watch as many matches as I can over the weekend, and then put together an article using the information I have gained from watching hundreds of decks in action. Over the past few days, I’ve seen all but two mythics in action (Avacyn and Descent into Madness), and every single rare. I’ve seen every two-color deck in action, and I’ve got a pretty good idea of what makes the format tick. You can find my previous two articles here and here.
I’ve also learned a few things since I last did one of these articles. The last time, the most popular segment was where I took some of the numbers that I had put together and discussed their implications. This time, I took that a little further, using a little bit more data and measuring some additional things. Specifically, I’ve tracked the win percentages of the different two-color combinations, as well as their overall share of the total field. I’ve also broken down the win percentages and overall share of each color. I added a measurement of the average number of turns before a game ends as well as a table showing the likelihood of reaching a particular turn.
This kind of approach has some big strengths, but also some important weaknesses. For the strengths, it is much less subjective than the typical overview of a limited environment. Most of the time, you see people making assumptions about a format based not on a lot of exposure to the environment, but based on a lot of personal variables. Furthermore, this kind of approach allows you to see the environment from many different perspectives; instead of seeing only the approach of a few pros, you get to see how people are thinking about the format in general, which gives you a better picture about how everyone else is going to interpret your actions in a draft. Most importantly, this approach has a large amount of concrete data. It takes a much broader slice of what the format is about, rather than just showing a narrow perspective based on the experiences of a few people. One of the weaknesses is that it is virtually impossible to get the thoughts of the people involved in the draft. Another weakness is that you can only measure what you can observe; you can’t see what is in people’s hands when they make decisions, and you can’t see what they picked in the drafts. But, most importantly, this reflects an environment at a particular juncture of time. Many of the principles I’ve observed will be constant through the environment, but many others will change over time.
With all that said, I hope that this will be a useful tool for figuring out the Avacyn Restored limited environment. I won’t focus very much on individual cards; there are plenty of articles doing that. Instead, I’ll try to focus on the principles that you can use to approach the majority of situations within the environment.
Speed
One of the biggest areas disagreement about Avacyn Restored deals with the speed of the format. I’ve heard some people say that this is an incredibly fast format, as fast as Zendikar, and I have heard others say that it is almost as slow as Rise of the Eldrazi. One of the important things that I wanted to track was to see how long an average game of Avacyn Restored limited lasts. In order to measure this, I kept track of the end turn of every game that I watched. In the end, I saw nearly four hundred games, and I was able to build up a lot of data.
There are two areas of worry here; first, is that many players have a tendency to concede much too early, and the second is that some games end because of factors outside the game. For the first, it is important to realize that in the majority of cases, players are conceding from positions where they have a very small ability to get back into the game. While it is unfortunate for our data collection, it does still help us figure out the overall speed of the format, since it shows us whether these players are reaching heavily unbalanced game states early in the game. In other words, players tend to concede early when the other player has a fast and overwhelming start; players are much less likely to concede early when something drastic happens in a very long game. In fact, they will often stick it out longer than they reasonably should. For the second, I did not include any games where a player timed out in the first three turns of the game. There were relatively few of these, so it didn’t hurt the data collection too much.
On to the numbers!
Turn Number
|
Total
|
% of Total
|
% of reaching Turn
|
Four
|
2
|
0.005195
|
1.000
|
Five
|
19
|
0.049351
|
0.995
|
Six
|
39
|
0.101299
|
0.945
|
Seven
|
59
|
0.153247
|
0.844
|
Eight
|
66
|
0.171429
|
0.691
|
Nine
|
47
|
0.122078
|
0.519
|
Ten
|
48
|
0.124675
|
0.397
|
Eleven
|
33
|
0.085714
|
0.273
|
Twelve
|
24
|
0.062338
|
0.187
|
Thirteen
|
17
|
0.044156
|
0.125
|
Fourteen
|
12
|
0.031169
|
0.081
|
Fifteen
|
4
|
0.01039
|
0.049
|
Sixteen - Twenty-Three
|
15
|
0.038961
|
0.039
|
|
Average
|
Median
|
Mode
|
Turns
|
9.205195
|
9
|
8
|
One thing that might surprise a lot of players is how early most games of Magic end. Though peaking at turn 8 is a little bit faster than typical, most formats tend to have a peak around turns Nine or Ten. You don’t have many opportunities to make important plays in a Magic game, which is why things like mana curve are so important. But let’s talk about specific data from Avacyn Restored.
First, the two most frequent end-turns were Eight and Seven, respectively. That is a relatively fast format. For example, a deck with 17 lands will hit six mana about 50% of the time by turn seven, and about 61% of the time by turn eight. So, for around 40% of your games in Avacyn Restored, you will never hit six mana. For more perspective, this means that games end with players having drawn around 14 to 16 cards. Typically 5 to 7 of those are lands, and each player will have played around 5 to 7 cards from their hand. This means that most games of Acacyn Restored end with both players have 1 to 3 cards in their hand, or even more if they got cards with something like Borderland Ranger or hand something bounced by Mist Raven. This format isn’t as fast a Zendikar, but it is still a pretty fast format overall, which is probably not what most people expected when they looked at the format initially.
Second, we see that the third most frequent end-turn is Ten, and that the numbers of later turns tapers off more slowly than we would expect from a fast format. Normally we would expect a steeper drop off a couple of turns after the peak, but we see a lot of numbers for games dragging on fairly late. What this means is that for the percentage of games that manage to get to where each player is topdecking on a stalled out board, those games tend to be difficult to end. When a player lands a Seraph Angel and locks down the board, but isn’t able to decidedly close out the game, the game is probably going to go on for quite a long time. What you expect in a fast format is that either the fast decks win quickly, or the slow decks take over the game around turn 6 or 7, and then put down a big threat that ends the game in 3-5 turns.
There are three reasons why we see a drawn out end game. First, is the explosive nature of the red aggressive decks. It is more difficult than normal for a deck to stabilize against these decks. When they finally manage to slow down the red assault, they still have to worry about getting blown out by a Thatcher’s Revolt, which means that they have to leave back more blockers than they would otherwise. This adds a few turns to their clock after they stabilize the game. Second, is that the aggressive Green decks are using soulbond creatures, which can be large even into the late game. This also means that they have a larger number of potentially game ending threats; anything that is paired with a Druid’s Familiar can potentially win the game. Again, this means that they control decks have to wait a little longer to turn the stall into a win. Thirdly, there is a decent percentage of slow White and Black decks that tend to lose against the fast decks in the format, but when they go against the mirror, the games tend to be very long and drawn out.
What does all this mean for a player? First, you need to realize that the majority of games in Avacyn Restored are over very quickly. You simply do not have time to mess around. It is vitally important to have a good curve and to mulligan hands that aren’t able to do anything until late. Value two and three drops highly. Second, value highly those cards that are useful in both the early and late game. Mist Raven is an incredible card in the format for specifically this reason. If it comes down early, it can simply close out a game; but if it comes out late, it will probably take out a key blocker and then present an important evasive threat. Trusted Forcemage, and Soulbond creatures in general, are so powerful for exactly this reason. Third, avoid expensive cards. I know that Griselbrand looks very tempting. But it should be noted that in all the games I have seen, it has never been played for a win. Consider that even if you are lucky enough to get the 27% chance of reaching turn 11, you still only have a 47% chance of hitting eight mana, let alone eight of it being black. Gisela is a seven drop that is worth playing, but even something like Spirit Away is simply not as good as it looks. For reference, I would rather have a Mist Raven or Trusted Forcemage than a Spirit Away in almost every deck. That isn’t to say that those cards are unplayable; they can be very useful to close out a game against the slow White or Black decks, but I err towards leaving them out of the maindeck and siding them in for those slow matches.
Colors
In every draft format, everyone has an opinion about what color is the best. I remember drafting Rise of the Eldrazi and hearing from one group of players that Green was the best color and from another group that it was the worst. Most people form these opinions based on a limited amount of personal experience; now you are going to get a chance to look at this set from a more empirical perspective.
Color
|
White
|
Blue
|
Black
|
Red
|
Green
|
Wins
|
59
|
77
|
33
|
62
|
77
|
Total Games
|
128
|
136
|
82
|
124
|
137
|
Average Wins
|
46%
|
57%
|
40%
|
50%
|
56%
|
Of total
|
42%
|
44%
|
27%
|
41%
|
45%
|
I have heard a lot of pros arguing about whether Blue or Green is stronger in Avacyn Restored, and there is a reason for that argument. After my observations, it seems like Blue and Green are in a virtual deadlock. On this chart, Blue has a slight edge since it has been played in one less game, but that could easily change by watching a few hundred more matches. Both of these have a very significant lead over the other colors in the format, and I think it’s safe to say that they are the only real contenders for best color. Also, not only is black the least played color by a huge margin, but it is also the least winning color by a huge margin. Right now, it is being played in barely more than 2 decks per draft, but apparently that is still overdrafting, since the color is so weak. Perhaps if black is only drafted heavily by one person at a table then it is strong, but as soon as more people jump in, its value drops dramatically.
Why are Blue and Green the strongest colors?
In my opinion, Blue is a little bit better than Green overall. This stems from a few very important factors. First, is the card whose art was featured at the beginning of this section. Mist Raven seems like it is a very strong card, and it would be in virtually any draft format. In Avacyn Restored, it is even better. If you think that it is the strongest common in the set, not only are you right, but you are probably underestimating it. Mist Raven is better than almost all of the uncommons in the set, with the exception of Druid’s Familiar and possibly Wolfir Avenger, though it is possible that Mist Raven is just better. It is also better than almost all of the rares in the set, with only a few notable exceptions, such as Deadeye Navigator or Wolfir Silverheart. This is a fast format, and being able to take a creature off of the board is a huge bonus, but the fact that Mist Raven also comes with a crucial body in the midgame are so powerful, but there are factors in this set that make Mist Raven even better. First, there is hardly any removal in the set, so useful cards that can take a creature off of the board or break up a Soulbond pair are even better than they would be normally. Second, there are a lot of ways to reuse Mist Raven, such a Flicker effects or Peel from Reality. Getting multiple bounces out of a Mist Raven is just incredible. Finally, Soulbond is the central mechanic of the set, and it means that any creature becomes more important. I frequently find myself playing 1/1s for 1 because having a creature to pair with is powerful in this format. Being able to put another useful ability on what is already a 2/2 flyer really puts this over top. As good as you think Mist Raven is, it is probably better.
But one common is not enough to make a color the best in the set. In addition to having Mist Raven, blue also gets the deepest pool of commons. Nearly every common in Blue is better than it looks. Wingcrafter has ended up being an incredibly powerful threat by over-performing in just about every deck, except possibly WU. Spectral Prison originally looked terrible, but it turns out to be quite strong. Even something like Geist Snatch is surprisingly good because most of the strong cards in this format are creatures, mid-range tempo cards are vital to the blue deck, and a 1/1 flyer is very relevant in the world of Soulbond. Ghostly Flicker is almost always a two-for-one, and Crippling Chill often feels like removal. Even something like Alchemists Apprentice does extra work by being a human for the red and white decks or a 2 casting cost creature for the black decks.
And perhaps this is what really makes Blue so strong in this format. Not only are the cards strong, and not only is the color deep, but the cards are deceptive, especially to inexperienced players. When a player gets a Mist Raven, they might not realize that it is probably just better than the “bomb” rare that is also in the pack. It is the same with all of the other cards, and this could cause people to pass Blue cards when they should be taking them. Wingcrafter looks mediocre, but once you’ve seen the kind of work it does, you realize that it is very strong. Weaker players are probably passing these cards, but the stronger players recognize their strength more easily. Not only are the cards well placed in the format, but they are also being taken by the best players, which leads to Blue being exceptionally strong in the format.
If Blue commons are so strong, then why is Green neck and neck with it in wins?
Green, too, has a breakout common in the form of Trusted Forcemage. In the past few years, there have been fewer board dominating common creatures, but I remember the days of Timberwatch Elves and Kabuto Moth. Those cards were easily first-pickable in their day, and they were often much more powerful than they looked at first. Trusted Forcemage is a card in the same tradition; it gives a persistent bonus to creatures and it gives an overall bonus of greater than +1/+1 when you count both creatures. Being able to spread this pump over two creatures is exceptional, though it does have a few weakness. The best thing about it is that it is a creature that can still attack and block while giving out its bonus. This is definitely the second best common in the set, but it is not what makes green so powerful in the format. That honor goes to Wandering Wolf, Timberland Guide, and Nightshade Peddler. This is a format where two drops are a rarity, but Green not only gets three of them, they are also all very strong. Getting a high density of these cards allows a player to set up a strong board position in the first 4 turns of the game, followed by a stream of powerful tempo plays and synergies to keep up the pressure.
The low end of Green is lower than in Blue. Where Blue got a diversity of playable effects at the bottom of its playable list, Green got three low powered tricks that all compete for the same narrow deckspace with Joint Assault, which is just better than all of them. On top of that, it gets both Lair Delve and Grounded which are just too mediocre to be worth a deck slot.
Where Green makes up for this is in the uncommon slot. Green has the strongest suite of uncommons in the set, with Druid’s Familiar, Wolfir Avenger, and Blessings of Nature headlining. Each of these cards is incredibly powerful and sets a player up for a strong Green deck. But on top of that, Green gets several other strong Uncommons such as Gloomwidow or Howlgeist. Having such strong uncommons means that Green decks can occasionally be much stronger than Blue decks, but if cards from other colors show up in the uncommon slot, then the Blue deck will have the edge. Overall, it ends up being about even fort both colors.
So what is a player to do when Green or Blue are just not open?
Red is stronger than White, particularly because of its potential explosiveness. A perfectly crafted Red deck will beat anything in the format. When you drop Thatcher Revolt for a surprise 10 damage on Turn 5, it is just hard to win. I’ve heard several players say that Red is the strongest color in the format because of this explosiveness, but what they are missing is the problem Red has with sequencing. It depends on dropping a powerful creature on Turn 2 in order to get all those explosive starts that it needs. But Red only has two aggressive 2 drops, only one of which is common. Kruin Striker is incredibly powerful, but it takes a little bit of luck to draft enough of them to put the Red deck together correctly. If you end up missing some part of your important sequence, the Red deck tends to lose steam a little bit too quickly, which is why it puts up so many fewer wins than Green or Blue. But, it is still a powerful option when one of those colors is not open.
Your other best option is White, though it is a little bit of a gamble. The problem with White is that besides Seraph of Dawn, it just isn’t doing anything powerful at common. Also, unlike Blue and Green, White simply has a bunch of mediocre commons that down really combine to do something powerful. It is much better as a support color than a main color. With that said, it is possible to build a white deck that uses early cards like Angelic Wall to set up a very powerful late game. To make this more useful, White also has a powerful suite of bomby rares. Picking one of those early can be a strong option, and Seraph of Dawn is the best card in the format for getting to the late game.
Archetypes
The other main focus of my observations was to see which color combinations were most popular and which were most powerful.
Archetype Popularity
Win Percentage by Archetype
The most popular decks were WU, UG, RG, and RW, in that order, with a sharp drop off from RW to UR. Essentially, people have seen that both Blue and Green are powerful, and it seems like they are settling into decks that try to take advantage of that power. The other most popular deck is RW, probably because putting together an explosive RW deck seems very strong, especially when you can attack for 25 to 30 damage on Turn Five with a Goldnight Commander and Thatcher Revolt. UR is misplaced in popularity compared to its win percentage, probably because people haven’t figured out how to draft an aggressive UR humans deck yet. After that, deck popularity pretty much matches win percentage with GW, BG, BR, UB, and WB filling out the bottom ranks.
The two most interesting things from this entire study come from the win percentage chart. First of all, we see that UG is the decisive leader. If you wondered what the strongest deck in the format was, you don’t have to any more. UG leads the next closest deck by a whopping 18 points. This means that you will win about 1/3rd as many matches with UG as you will with even RG, which is a huge difference.
Some people might not understand exactly what it means for UG to have a 73% win rate. For some perspective, consider that the best deck from my study of DII managed to get about a 60% win rate. In this format, UG is incredibly good, more than most decks have been in any limited format I have ever seen. It is comparable to how strong BR was in triple Zendikar. It seems like your best strategy is to go into the draft with the intention of drafting UG.
The next interesting tidbit is that there were only three decks that managed to obtain higher than a 50% win ratio. This is equally abnormal, and largely a result of UG beating all those other decks so handily. But what really strikes me is the nature of those other two decks; RG and UR. When I first took a look at the format, I predicted that RUG-centric decks would be the leaders in Avacyn Restored, and it looks like my prediction is spot on. What makes this exceptionally remarkable is that it is easy to start out trying to draft UG, and then substituting Red for whichever color is getting cut too hard by the person on your right. At this time, it seems like that is simply the best strategy in the format.
Another important bit of information is the underperformance of the popular WU and RW decks. Both decks put up a lot of numbers in the played column, but failed to put up a reasonable number of wins. Like I said earlier, RW can be addictive in its explosive power, and a lot of players will see that deck do incredible things and think that it what they need to do. The problem comes in sequencing; it is just too hard to get everything in the exact right order to make this deck consistent enough to be a top contender. However, I predict that it will continue to be drafted to highly. As for WU, I think that it is simply a result of people marrying their White first picks too early. There are a lot of bombs in White, but there is not a common base to support them, and these decks tend to be unsynergistic messes that win when they draw their bombs, but lose otherwise.
Conclusion
I had hoped to say a little bit more about the format, but I’m already past 4000 words, and there simply isn’t space to say everything in one article. However, I hope that this will give you a sense of where the environment sits currently, and how to best take advantage of the format. Let me review my main points.
1. Avacyn Restored is a fast format. You don’t have time to mess around.
2. The games that don’t end quickly tend to be very long and drawn out. You should value highly those cards that are useful in the early game and allow you to punch through in the late game.
3. Blue cards are deceptively powerful. Mist Raven is probably better than that “bomby” White rare in your pile. If it’s a close pick, take the Blue card, because it’s probably better than you are giving it credit for being.
4. Green leans a little more heavily on its uncommons than Blue, but it is still tie for the best color in the format.
5. Red makes a great substitute for either Blue or Green if one of those colors is being drafted too highly.
6. Don’t be deceived by the RW deck’s occasional explosive games. You need to remember that the deck will often stumble on sequencing and just lose to a Mist Raven or Trusted Forcemage.
I am really enjoying Avacyn Restored as a format. Soulbond is a deep limited mechanic that changes card evaluations all through a pack. There isn’t much removal, which means that it is more important to have sharp card evaluation skills. Hopefully this little article will let you get an edge in the coming weeks.
22 Comments
This was very informartive and well-written. A really neat asset for drafters everywhere.
It's refreshing to see green be an actively good color. I can't remember the last time it wasn't in the bottom two. Sure there's always outliers like Overrun decks, but in general green has gotten shafted time and time again.
Green was really good in Innistrad! It was also decent in Scars block and in RoE, as well as Shards of Alara. I think that WotC has been pushing the power of green a lot in the past few years, but that is mainly because it has been so bad historically.
Also, I think that Green is harder to evaluate when it is powerful, so people tend to undervalue it for a while. Green in RoE, for example, was so powerful partly because of its fixing, which allowed you to splash all the removal from the other colors.
I also liked the article a bunch. I see a lot of mono-black, and I see it doing well.
Mono Black has performed very well, but I didn't have a large enough sample size for it to be statistically relevant :( I only saw four of those decks, but they won 3 matches. But it seems like mono-black is the only way to go with Black, but it depends on getting lucky enough to get passed a sick deck.
Wow, the hard work that went into preparing this article really shows. The statistics on deck archetypes and color wins was fascinating, and this article really settles the questions of how fast the format is, and how badly does black suck. This information will be invaluable in my future drafting of this set. All that said, I don't agree that this is a great limited format. High speed, likely stalls, and a higher than usual level of randomness due to miracles and poor removal does not sound like a good format to me.
I can understand why someone would not like this draft format. It is significantly different from most other draft environments.
With that said, I really like it, but that is because I love tempo oriented environments, and I like it when removal is bad, because it forces you to make tougher decisions, instead of just take the best removal card.
First, great article, very comprehensive and spot on analysis of AVR drafting.
While I agree that weak removal or less removal makes card evaluation more difficult, the real problem with this format is the disparity in card strength.
I've gotten used to bomb rares (I'm not a fan) but bomb uncommons annoy me, and the disparity between Druid's Familiar and every other card is insane. Likewise, blue decks are basically as good as the number of Mist Ravens they have, and white with Seraphs.
Seraph might not be Pacifism but I seriously doubt people are picking any other white common over it. Green is the only color with interesting decisions beyond trusting in the forcemage, and the color is so overdrafted usually your decisions are made for you.
Good information. It's particularly useful to see this when I hear so many people saying the new set is slow. It CAN be slow if the right decks meet, but it's almost always over quickly.
I started off enjoying the set. I thought I would like it from a flavor-meets-mechanics viewpoint before it came out, then enjoyed it for the first week. But I'm starting to like it less and less, despite the fact that I, too, enjoy tempo quite a bit. I can understand the enjoyment of the lack of removal, but I think removal vs. power is unbalanced in this format. There are so many powerful cards that hit the board and just win the game, some as innocuous as Champion. If it just granted itself the unblockable ability or even granted the ability to each creature on your team based on each creature's power, it would be extremely powerful but ok. The ability as is just wins games, especially if you get it on turn three or four, unless, of course, you're already being blown out. It is an example of a card people just look at and don't think it's anything special and there are dozens of them in the set. But when they come down, you win. And there is almost no removal for it, only a handful of cards that can deal with it and most of them you have to have in hand already and use it pronto or you lose. To me, that's not a fun format. If I wanted to play a non-interactive game, I would go lose to a Legacy Storm deck. I think the power level of the set is extremely high from top to bottom. This set is the first where I really see how far Wizards has turned power creep to power ramp. Some people disagree because there are some "unplayable" commons. Go from top to bottom and look at the power level. It's nuts.
Not my favorite format, but not the worst either.
Excellent article, the best limited article I've read in quite some time. I hope you create more content like this.
I must say, I absolutely loved your in-depth statistical analysis of power levels in this set! This article was truly great, and I feel better equipped going into AVR limited with this knowledge.
Thank you, keep up the great work. :-)
Great research and great information. Congratulations by your job ;)
Great article, I love reading about breakdowns of creatures by color, and the toughness vs removal question. Keep up the good work. I had a couple questions, did you watch multiple rounds in the same draft, and how many different drafts were you able to watch?
This data comes from six 64-man drafts. All of the data comes from Round 1. I did this because if I took the information from later rounds, then the bad decks would be played much less, because they would have been weeded out.
There is an inherent problem with taking data only from round 1. Specifically, you don't see which decks are winning in rounds 2 and 3. It is entirely possible that RW is very good against UG, and that RW will therefore give you a better chance of going 3-0, but also a higher chance of losing round one. Or that could be completely false. I have no idea, and you'd have to do a much different study to find out that data.
Great, I was wondering if it was from just round 1, and think that was the best way to go about this study. There is just too much data to sift through to get that full picture from round 2 and 3, and while I'd love to see this further analysis, it is probably too much for one researcher to do on their own in a weekend. Great article, I hope to read similar articles from you in the future, and maybe I should just go out and collect my own data sometime, I'd enjoy seeing how you set up your data set.
wonder what would happen if you did a study of just round 2+?
This would be a perfectly reasonable study to do. The main problem is that you need close to 200 matches to get a decent sample size for this study. That means you need to watch twice as many events. There are only so many 64-man drafts that fire in a weekend. I watched every one that I could in the above study.
But it would be a reasonable follow up study to the one that I did here. You would also need to do them close enough together that the data from your first study isn't completely obsolete by the time you do your second study, so the window has probably passed for me to do one like that for AVR.
The benefits are that you'd have a much better idea of what kind of decks are winning in the later rounds. That kind of data is really useful to have alongside the first set of data.
Besides that, an article like this takes a lot of friggin' work.
Great idea for a article. Hope you keep this coming for sets in the future. If you want to see more people gushing about how good your article is, I posted a link to it over on the MTGSalvation Limited forums and there is a lively discussion going on the results you found.
http://forums.mtgsalvation.com/showthread.php?t=420595
Awesome analysis :)
So great to see real numbers. It's like science!
Suggestion for next time: Include error bars (standard deviation).
I do appreciate this comment. Let me explain this part... I'm not a math person. I'm a professional writer. I'm very good at math, but I've never learned much about math.
What I'm doing in this article is pretty simple math, as you probably noticed. However, I am currently working my way through learning some stuff about statistics on my own. You'll see extra things popping up in this article as I learn more about how to do the math.
So... expect standard deviation at some point! Not sure when I'll figure that out though :)
Thanks for taking the time to make this. It was awesome from your part to share it.