Making the Right Pick

Drafting is a complex matter. While I have won my fair share of 8-4’s and premier Top 8 drafts, my drafting still needs work. My draft record on the Pro Tour is only slightly above 50% (11-10), so I am probably not the best person to comment when it comes to draft strategy. However, I do love drafting and over the years have developed the belief that there is always a right pick. I repeat, there is always a right pick! By this, I mean that there is one card that is correct to take, given all the relevant data. This article is an attempt to develop a theoretical framework for determining what that pick is, which includes identifying and weighting the various factors that determine which pick is correct. For simplicity’s sake, I will restrict this discussion to 8-man drafts (no team drafts).

Twelve factors I have identified include the strength of the cards in the pack under consideration, the mana cost of the cards in the pack under consideration, the financial value of the cards in the pack under consideration, the cards you have already taken, the cards you have already passed, the prize structure of the draft, the nature of the format you are drafting, your familiarity with the different archetypes of the format you are drafting, your drafting preferences, the players you are drafting with, what you can deduce about the strategies being drafted by the players around you, and your goals for the draft.

Ultimately, I want to focus on winning drafts. Some of these factors, however, have nothing to do with winning drafts: the financial value of cards in the pack under consideration, the prize structure of the draft and, arguably, your drafting preferences and goals for the draft. Thus, I will discuss these first, to get them out of the way.

Sometimes it makes sense to rare draft, by which I mean taking the card that is worth the most money, even though it is not the pick that makes the most sense for winning the draft. Whether it makes sense to rare draft depends on the value of the card you might rare draft, the prize structure of the draft, your goals for the draft, and the power level of the pick that you would otherwise take.

Suppose you are in a friendly RTR Block draft with no prize, in which everybody keeps the cards they draft and just plays swiss rounds for fun. In this circumstance, it makes financial sense to take a Stomping Ground over Exava, Rakdos Blood Witch, even though Exava is by far the better first pick. If your sole goal is to win the draft, regardless of financial value, then Exava is the pick. It might matter more to you that you defeat your friends than that you own a Stomping Ground, but in this circumstance Stomping Ground is clearly the correct pick, financially speaking. On the other hand, suppose you are in a more competitive draft in which winner takes all. This type of draft is single elimination, where winners take the losers’ cards until the draft winner collects all the cards. In this circumstance, it does not make sense to rare draft because your best chance of profiting comes from winning the draft. Drafting Exava gives you a better chance to obtain that Stomping Ground than actually drafting that Stomping Ground.

In a normal 8-4, however, things are murkier. For example, in Pack 3, is it worth passing up a bomb in your colors to take a 10 tix card that you can’t play? After all, the prize for winning the entire draft is only 8 packs, or around 24 tix. Making these judgments involves a cost/benefit analysis of why you are playing the draft (fun? practice? profit?), the value of the card you could rare draft, and the increase in expected value from taking the best card for your deck (increased chance of winning times the prize from winning).

In the last draft of Day 2 at Grand Prix Toronto, 2010, where I was trying to Top 8, I rare drafted Koth of the Hammer and Mox Opal in Pack 3. I drafted Koth instead of a 5-mana green creature (Bellowing Tanglewurm or Molder Beast), because I had enough 5-drops already and expected at least one of the 5-drops to table (Molder Beast came around). You might say it is stupid to rare draft when trying to Top 8 a GP, but I still defend this pick because my deck would have been only minimally better, and at the time Koth was worth about $30. Also, it could be considered hate drafting (taking cards to make sure your opponents don’t have them), which can be correct if the card is powerful enough and if the card you would take over it wouldn’t much help your deck. I still managed to Top 8 that GP, with $50 in cards to boot.

Financial concerns aside, winning the draft still might not be your number one priority. You might, for example, want to test out a draft strategy for an upcoming event. That is a reasonable excuse not to take the pick that is most likely to win the draft. It is not possible to master a format without trying various strategies, which sometimes includes drafting bad decks. In order to discern what works, sometimes you have to first experiment with what doesn’t work. Draft exploration is a legitimate reason not to make the correct pick, but this should not be done when you are in a high-level draft such as a Grand Prix or Pro Tour.

Alternately, you might just love a specific strategy. Some people are wealthy, and would prefer to win, but for some reason just love drafting a mill deck. This is where the notion of preference rears its ugly head. If your goal is to have fun and play with cards you like, then by all means, draft according to your preference. Some people just love fire and can’t get enough of drafting red. To quote Alfred from The Dark Knight, “Some men just want to watch the world burn.” However, the more you draft according to preference, and not what the cards present to you as the best strategy, the worse your drafts will tend to go. If you love forcing Gruul and always do it, sometimes it will work great, but others you will crash and burn.

As an example of how the notion of preference is often invoked in drafting, consider my first pick at Pro Tour San Diego. I had to choose between Turn // Burn and Unflinching Courage. I asked many people about this pick, and the responses were about 50/50. Several players said “it depends what you like to draft.” Their idea was that if you like playing aggressive decks then you should draft Unflinching Courage, while if you like playing controlling decks you should draft Turn // Burn. I think this is dodging the question. There is a correct answer, a correct pick. In my opinion (which was shared by Cedric Phillips and Sam Black), Turn // Burn is the correct pick. Unflinching Courage opens you up to a 2 for 1, while Turn // Burn is often a 2 for 1 in your favor. Also, Unflinching Courage locks you into two colors right off the bat, while Turn // Burn is playable in either blue or red. Sure there are games where Unflinching Courage shines, but Turn // Burn does a better job setting you up to win the draft. To say that this is a matter of preference is to miss the point of discussing draft picks, which is to uncover the reasons why a card are the right pick in a given situation.

The game of Magic is astoundingly nuanced and complex, and the odds that two cards have exactly the same value for your draft are infinitesimal. It may happen occasionally, but those occasions are few and far between. When I say that there is always a right pick, I mean that this is the case for all practical purposes, and that you would do well to think of it as always being true. It is certainly conceivable that two cards would have exactly the same value. For example, at Pro Tours they remove foil cards from the draft pack and replace them with regular versions of the same card. It is conceivable, then, that a pack could have 2 of the same bomb rare with no distinguishing features between them. In such a case, the two cards are exactly equal in value. But whenever the cards are different, it is extremely unlikely that all of the factors balance out such that each contributes an identical likelihood of winning the draft.

Having discussed the reasons why it might make sense to take the wrong pick (i.e. a card that does not maximize your chances of winning), I am finally getting to how to draft with the sole goal of winning. Imagine that the future of Earth is at stake: if you win the draft, aliens will take their ray guns and leave earth, but if you lose the draft… KABOOM! In such a circumstance, financial concerns, exploration, and preference have no place. None of these are reasons for making picks when everything is on the line. The sole exception to this is when you simply are not prepared for a format. Suppose you have drafted RTR Block only three times, drafting Gruul each time, and then have to play for the fate of the world. In such a case, it might very well make sense to force an aggressive green- or red-based strategy, simply because you wouldn’t know the other strategies well enough. However, although this might maximize your chances of winning that draft, the fact that this is the case would be a testament to the player’s drafting weakness. The best drafters know a format well enough to draft any strategy, so long as it is open. As I like to put it, they know how to “find the stream.” If the best strategy coming your way is Dimir control, but you force Gruul because that is what you are used to or what you prefer, you have missed an opportunity and handicapped yourself for the draft.

Assuming, then, that you are familiar with a format and have the sole goal of winning the draft, the relevant factors are the strength of the cards in the pack under consideration, the mana cost of the cards in the pack under consideration, the cards you have already taken, the cards you have already passed, the nature of the format you are drafting, the players you are drafting with, and what you can deduce about the strategies being drafted by the players around you. For good players whose interest is victory, these are the most important considerations. These are the factors that, when accurately evaluated, will most clearly tell you what the correct pick is.

You might say that the identity of the players you are drafting with is irrelevant, or at best a form of preference that has no place in a draft you are trying to win, but the fact is that some players tend to draft certain archetypes more often than others. I have fond memories of drafting Shards block at my local store in Ann Arbor, Michigan, where a group of teenagers just couldn’t get enough of Esper. These three guys always got dropped off at FNM together by a parent, always talked about how Esper was the best deck in the format, and always forced Esper, no matter what they opened or got passed. And if they were seated next to each other, they would fight over Esper. This means that, if you were seated to the left of one of these players, you would do best to draft either Jund or Naya.

I also remember sitting down for a draft at Nationals in 2008 and hearing the player to my right tell everyone to stay out of his red. Presumably, he was hot for [cardBurn Trail[/card], which was a key card in the aggressive red archetype. Somehow I thought he was joking and was aware that you shouldn’t decide on your archetype before the draft, but I was mistaken. I opened and first picked a Burn Trail, but then got cut off. Perhaps I should have just listened to the (illegal) signal from the player on my right, no matter how irrational that player’s behavior was.

A more high-profile example of this phenomenon would be Craig Wescoe’s well-known preference for white weenie decks. Particularly in constructed, Wescoe is a specialist in this archetype. It has served him well, taking him to three Pro-Tour Top 8’s (including Worlds), with a recent win in San Diego. Wescoe also favors this archetype in limited, albeit to a lesser degree. Obviously, Wescoe is a good enough player that he will not pass up an open strategy in other archetypes/colors, but if you are in a draft pod with him, you should at least be aware of his preference, especially if you are on his left. Interestingly, the fact that Wescoe’s preference is so well-known benefits him in drafting, because it often means the players on his left do not try to fight with him over white. This is a legal version of the player who announced his red plan to the table at Nationals, and is similar to the Esper strategy of the kids at my local shop. Being aware of the preferences of those around you can protect you from being cut off.

This phenomenon is fairly rare at high-level drafts, however, because most solid players are aware that you cannot plan a draft before opening your first pack. You might have some general ideas about what works in the format and what doesn’t, but every draft format has several viable strategies. The correct strategy in any given draft is primarily determined by what you open and what you are passed. It is important to remember that the players on your right will be passing you two packs, while the players on your left will only be passing you one pack. Thus, the most important part of drafting, in my opinion, is figuring out what the players on your right are drafting and not fighting with them over that strategy. If the players on your left do not properly read your signals and decide to fight with you over your strategy, so be it. You can recover from being cut off in Pack 2. However, it is very difficult to recover from being cut off in Pack 1 and Pack 3. If that means giving up on a powerful first pick, so be it.

A friend once told me that, if he opens Inferno Titan in M12, he is playing red no matter what. I think this is crazy. Even if Inferno Titan is your first pick of the draft (the notation for this is P1p1, which stands for Pack 1, pick 1), you can easily get cut from red. Suppose 4 other players at the table go into red, including 2 players on your right. In that case, sticking to red would most likely leave you with very few red playables. You would then essentially be splashing RR for a 6-drop. Yes, opening Inferno Titan is a great reason to be in red, but there are times when what is happening around you, especially what is happening on your right, means that you should give up on the Titan. This is particularly true when you first pick a multicolored card such as Unflinching Courage.

This is why it is so important to stay flexible in the early part of a draft. Consider the triple Gatecrash draft format. Suppose you open a bad pack with no good rares, no good uncommons, and only 3 good commons: Basilica Screecher, Kingpin’s Pet, and Syndic of Tithes. Which is the right pick? Many would say Kingpin’s Pet because it is the strongest card. Orzhov is attractive because the extort mechanic is so powerful. However, I think Kingpin’s Pet is obviously the worst pick here, for a few reasons. First, it locks you into two colors right off the bat. Second, it puts the people to your left into both black and white. Third, Kingpin’s Pet isn’t even that much more powerful than Screecher or Syndic. Is it really worth paying a 3rd mana of a 2nd color to give +1/+0? Is it worth that cost to give a creature flying?

The pick, then, is between Screecher and Syndic. If you take Screecher, you should plan to stay out of white because you are putting both people to your left into white, while if you take Syndic you should plan to stay out of black because you are putting both people to your left into black. While the colors of the people to your left are not the most important concern in drafting, they are worth considering because they will pass to you in Pack 2. If you get passed great Orzhov in Pack 1 however, you would do best to just take it and ignore the people on your left. If you take Screecher and get cut from black, or if you take Syndic and get cut from white, you should probably move into a different strategy altogether, perhaps Gruul or Simic.

But what about the pick between Screecher and Syndic? Which is correct? If there is other good white in the pack, but not other good black, then you take Screecher to minimize the possibility of fighting in Pack 2. If there is other good black in the pack, but not other good white, then you take Syndic for the same reason. In our situation, however, there is nothing else to consider, so it is a straight-up choice between Screecher and Syndic. Assuming you know nothing about the tendencies of the players around you, there are two relevant factors for figuring out this pick. First, you should ask which is better, a 2/2 or a 1/2 flier. I think a 2/2 is better, which leans in favor of Syndic. Second, you should ask which color is better/deeper in this format. In Gatecrash, the white commons were better than the black commons. Moreover, if you are trying to go black and stay out of white, it puts you in Dimir, which was notoriously bad unless you were the only Dimir drafter at the table. This concern also leans in favor of Syndic. Thus, Syndic of Tithes is the correct P1p1 in this situation.

The only caveat is that, since it was generally known that white was a strong color in Gatecrash, you should expect more players to draft white, which might leave Dimir open. However, given that white is a deeper color, white drying up on you would probably require 3-4 other white drafters at the table, mostly close to your right. Even if the pod adjusts appropriately to the strength of white in Gatecrash, it should simply balance out the availability of the strategies, which then brings us back to the tiebreaking question of whether it is better to have a 2/2 or a 1/2 flier. So, even assuming the pod adjusts appropriately to the strength of white, Syndic is still the correct pick. To pick anything else is to misunderstand the format and make a drafting error, which not only damages the quality of your draft deck, but also improves the quality of others’ draft decks.

Drafting is, perhaps, the most fun part of Magic. It can also be incredibly challenging. It is so complex that providing a comprehensive draft theory is nigh impossible. I have tried to detail the abstract considerations, but interactions are often so card-specific that nothing general can be said about them. For example, if you are drafting blue-black mill in RTR Block, you might want to take Mortis Strider. This depends on how many Sage’s Row Denizens you have, and whether you have a sacrifice outlet such as Corpse Blockade. And then, for Pack 3, having a Mortis Strider increases the value of Launch Party. These potential synergies should be kept in mind throughout the draft, which requires a significant amount of memory (at least IRL, where you are only allowed to view your previous draft picks in between packs).

Drafting is also very format-specific. Thus, if you are preparing for a big event, it is important to practice drafting that format. You must learn the various archetypes and get a sense for what is overdrafted or underdrafted. Reading format-specific articles can help, as can experimenting with different archetypes on your own. You must develop a sense for what the player(s) on your right are drafting, so that you do not get cut off and put yourself at a disadvantage, and the only way to do this is to learn to properly evaluate the power of the cards in that draft environment. And, of course, you must also learn what removal and tricks to watch out for while playing the matches. Remember, drafting your deck is only 1/2 the battle.

 
  1. I think they replace foils with random commons, in the bigger tournaments. A replacemant with the same card, makes no sense.

    Nice article! But I think there are sometimes more then one “right pick”!
    Not only the first pick matters, most of the time it is more important to take the right second, third, forth pick.

    Sometimes all the theory doesn’t work, because not everybody drafts with the same theory. And the card values are extremly diffrent from player to player…

  2. Thanks for correcting my misunderstanding of the foil replacement process Random. For that point, the same logic works if the best pick happens to be a very good common that there are 2 of in the pack. Alternately, you could just consider a regular pack where you open a foil and a regular card, although some might say the foilness makes it a better pick.

    And I completely agree that the 2nd, 3rd, 4th, etc. picks are also important. Most of what I said applies to the other picks as well (for example, when to abandon a color or when looking for synergies with cards you have already taken), it’s just easier to use first picks as examples.

  3. Saying that there is always a right move in a game like Magic is even less correct than saying that there is always a ‘right move’ in chess. If you believe that it’s possible to ‘solve’ Chess, then that’s probably true that a particular move will always be best, but the strategy of the game goes far deeper than it’s possible for a human brain to analyse (especially under timed conditions), so is overall *not* a helpful way to look at things.

    Imagine those situations where at an FNM you are considering whether to bluff a combat trick. You must ask yourself if your opponent is good enough to see through it, bad enough to not even notice the bluff, or likely to make the ‘correct’ response. A ‘correct’ reading of your opponent is the optimal play, but unless you are able to construct a perfect model of his brain and thus predict his actions, you cannot know how he will act, and thus what the correct decision is. This decision must be made on such a huge array of subtle and often subliminal information that it’s best described simply as ‘instinct’.

    In a game like Chess, this idea of perfect play might be more approachable, but due to the nature of draft (ie, reading signals) it’s really not appropriate to Magic. Questions like “Is my neighbor is sending me a certain signal”, whilst having a right answer, have an answer that is not, in reality, knowable for certain, and is based on too many factors for a player to do anything other than take a ‘best guess’.

    You say that the likelihood of picks having identical values is infinitesimal, but it is generally agreed that some picks are very, very close, and our ability to distinguish between any picks has limits placed on us by the nature of the game and the nature of the mind. It seems obvious that often picks will be closer to each other than we are able to discern, even theoretically, and thus it is completely meaningless in practice to call one ‘right’ and one ‘wrong’

    Until someone builds a supercomputer that can play Magic perfectly (by which I mean it is impossible to construct a scenario that it would lose to) and we can study its processes and actions, the better way to learn to draft is to take a number of general principles to heart, and then learn, through theory and practice, how they should be balanced against each other.

  4. @Fishy I disagree, I think this approach is valuable. There was great Limited Resources podcast about this a while back that formed the way i think about this.

    Intuition that comes from practice is very important to playing at high level, but high level players should be able to articulate the precise and rational reasons for their choices when asked. You have to pay your dues by learning the tendencies and doing all the math. Going with your gut from a position of experience and knowledge, by applying patterns you’ve seen and making instant correct decisions subconsciously, will bring rewards. Going with your gut without that foundation will bring ruin.

    It’s true that the best pick is uncertain to us, but that doesn’t change anything.

    The benefit of insisting there is necessarily a best pick, is to promote the habit of finding an actual reason to take card A over card B (time permitting, in hindsight after the draft if nothing else). This leads to improving your play.

  5. @raisins I agree with everything you said about having good reasons for picks, but it’s not really what I was saying. Nor what the author was saying. Of course you want to make picks based on reason rather than gut, but the idea that you can come anywhere close to making ‘correct’ picks is ludicrous. Just trying to keep track of what information you’d need to calculate to make a probability estimate of a certain picks is astronomical.

    You’d have to start with some kind of mathematical certainty about which decks were better or worse than others. This alone would be such a massive investment of computing that it’s not worth considering. Then you’d have to weigh each pick in accordance with maximizing your EV on deck building. Even if you had a perfect list of every possible deck in order, making picks to maximize chances of getting the best deck possible would require a super-computer.

    Let’s assume we can get this far though, cos this is where the calculations involved start going from merely astronomical, to exponentially bananas.

    Pick 2, you’d have to take into account what you thought your neighbor picked, and balance that against the cards in a vacuum. By pick 3, as well as considering your pick, the information that you have about your neighbor’s pick, you’re also calculating extremely niche and vague probabilities that affect your calculations based on what the person 2 seats down might or might not have taken. (Let’s not forget that we’re not only trying to draft the best deck, we’re also trying to maximize our chances of simultaneously stopping our opponents from getting good decks.)

    By pick 8 you’re looking at packs and calculating the weighted odds across hundreds of thousands of branches of a decision tree involving not only the ideas of the perfect deck and the perfect pick, but also across the decisions of 7 other players. By the end of the first pack, these trees have more nodes than there are atoms in the universe.

    Richard Garfield himself discussed these issues in his famous talk on Luck vs Skill in games (it’s on Youtube). He points out that, in games, a level of complexity beyond what the human brain can comprehend manifests as luck. There are steps you can take to maximize your chances, but the ‘correct’ approach is to recognize that the complexity is far beyond you, and that at some point you have to admit that you are making a ‘best guess’.

    You should recognise what factors can increase your chances of guessing right, but approaching it as a case of right/wrong will lead you astray much more than it will help you, since it’s just not the case.

  6. The article is pretty light in terms of content. Yes we all know that prize structure and EV issues make slamming a sphinx‘s revalation when it is out of your colors the right pick except on the pro tour.

    The article tries to form a theoretical framework and fails at that. Instead we are simply inundated with examples of why certain picks are right. Discussing turn/burn and unflinching courage does nothing to advance the discussion. Many pros feel that the card is powerful enough to justify risking a two for one and going into GW. Some disagree and we go round and round. We are no closer to uncovering general principles to evaluate picks.

  7. @Fishy

    Also the `atoms in the universe` argument sounds impressive but in practice is not useful or even an accurate description of a draft. Nobody is going to p1p1 mending touch or a whole host of other cards. The vast majority of decisions are so bonkers that it does not make sense to analyze them even if we had the computing power.

  8. Fishy, you might be confusing what philosophers call “epistemic” vs “metaphysical” issues. Epistemic in this context refers to what we can know, while metaphysical refers to how things are. The point of my article is not that we can always know the correct pick , but that there is a correct pick that we should try to discern. Even if we the complexity is beyond us, that doesn’t disprove my point. The idea is that we should move past the notion of “preference” whenever we can’t figure it out, and instead habituate an investigation of the reasons why some picks are better than others. Usually packs can be narrowed down to just a few cards under consideration, and having a framework in mind (i.e. a list of relevant factors) is helpful toward that goal.

  9. THis is an bad article! Jus take tha money > tha bombs > tha removal > other stuff.

    Thats the entire flowchart in one sentense for ya’ll.

  10. First of all, LMFAO at the three tries to spell “Sentence.” That made my week and it’s only Monday.

    I am relatively new to Magic and I thoroughly enjoy drafting online and at FNM. In fact, I just finished 3-1 for the first time last week (lost my first match when I got mana screwed and they top decked the perfect card).

    Anyhow, I brag I brag. lol… from my limited experience in drafting, I find I am having more success the more I simplify the process. Yes, last week getting a second and third pick Ascended Lawmage made my decisions easier going forward, but it also told me what the persons to my right was NOT doing. They were not in Azorious. I always heard of signals, and everyone trying to figure out what people were doing, that no one had ever told me it was far easier to figure out what people were not doing. I didn’t know if they were in Gruul or Orzhov or Simic. I had no idea what they WERE taking. But I most certainly could tell what they were not taking and I loaded up on every blue and white card I could get my hands on when there were no gates or gatekeepers (so much fun to play I just can’t help myself).

    It was… an epiphany for a new player like me. Find a simple thing to focus on, then use that to build on. For me, it was what good cards are not being taken and letting that guide me to the colors I will have the best chance to get. That may be the point of this whole article, and I may have totally missed it and all along been trying to figure out how people could figure out what I WAS taking instead of what I wasn’t. But I thought I would throw in a fairly new player’s perspective.

  11. Hey I wasn’t sure how to contact you Dustin, but I saw your balancing suspend modern deck and was wondering how much testing you had done, and what variations you have tried.

    I was thinking about changing the mana base / cascade /suspend options around and was wondering what matchups you brought in boom/bust.