The Decision Game Of Magic

Jadine Klomparens doesn’t see Magic as a game of cards. It’s a game of decisions! Get her insights into the decisions Magic players face!

There are many different actions in a game of Magic. We draw cards, cast spells, trigger abilities, and attack with creatures. We block, shuffle, sacrifice, mulligan, and concede. And yet for all of that, the only thing we ever really do is decide.

Magic is a game of decisions. No matter the format, no matter the stage, all you ever must do to play good Magic is make one correct decision after another until the game ends. Of course, as we all know, that’s easier said than done. Magic is a very hard game.

I was first introduced to the idea of Magic as a decision game through, of all things, poker. There was a time in my life when I was very interested in getting into poker and started reading different books about the game. I never actually started playing poker, but I found that a lot of the theory I was reading could be applied to Magic.

The decision stuff specifically I found in a book by Annie Duke called Decide to Play Great Poker. I was enchanted by the idea that, in her words, “Poker is a game of decision-making under conditions of uncertainty.” The game is all about making decisions, and the problem is that you do not have all the information necessary to figure out what the outcome of those decisions will be. This is a description of the game of poker, but it describes Magic just as well.

A few months ago, I was part of a conversation after a tournament discussing what the win rate of the theoretical perfect Magic player, someone who would always make the best possible play in every scenario, and who also had access to every piece of relevant information, would be. Said information would include not just the opponent’s hand, but also the composition and order of both players’ decks at all times.

Obviously, you can never be as good as this theoretical Magic player. No matter how good you are at reading your opponent, you’re not going to know the order of the libraries. But if you were somehow to acquire that mystical power along with the ability to mentally process all that information and arrive at the correct conclusions, you would almost never lose.

As I put it in that conversation, this perfect Magic player would be able to make terrible plays and have them be right. They would keep no-land hands because they could see that they’d hit five lands in a row off the top of the deck. They would never be surprised by the opponent topdecking another copy of the spell they had just Thoughtseized away.

They would attack into four open mana with all creatures every time the opponent didn’t have Settle the Wreckage and would hold back appropriately every time they did. And, perhaps most illustratively, they would be able to play Lantern Control without the card Lantern of Insight.

Their Codex Shredders would establish the Lantern “lock” without needing the information provided by Lantern of Insight. They’d be better than that, even, able to look past the top card and know when aggressive milling will get the opponent to a run of lands or when milling a problem card isn’t a good idea because the next card is even worse. Honestly, I have a hard time believing that this perfect Magic player would ever lose a game.

The Difference Between Correct Decisions and Good Plays

Let’s leave our perfect Magic player alone for a minute to bask in the glory of winning their 100th Pro Tour and consider a different scenario.

You’re playing against a U/W Control opponent who has nine total lands on the battlefield, of which only an Island and a Plains are untapped. They have one card in hand, and only one life point remaining. Unfortunately, you know their top card is Approach of the Second Suns, and they’ve already cast that spell once this game.

You draw Shock for the turn, joining the Goblin Chainwhirler in your hand. Either will kill your opponent, but you only have three Mountains on the battlefield, so you can only cast one of them. You know that the only relevant cards in your opponent’s list are two Negates and one Essence Scatter, you haven’t seen either so far this game, and you’re confident that the card in their hand is one or the other. What’s the play?

The best play is to cast the Goblin Chainwhirler, because your opponent is more likely to have one of their two Negates than the single Essence Scatter. So you cast your Goblin Chainwhirler, and your opponent casts their Essence Scatter. You made the best play but the wrong decision, and it cost you the game.

This is a bit of a definitional thing, so my apologies if this isn’t how you’re used to using these words, but the distinction is important. I’m using the phrase “correct decision” to mean the choice that the theoretical perfect Magic player would make based on information that you probably don’t have access to. Correct decisions win you games, but it’s impossible to always make them. All we can do is make the “best plays,” the ones that have the highest likelihood of being the correct decision.

In the Chainwhirler / Shock example, we have two options. If they have Negate in hand, we should cast Goblin Chainwhirler. If they have Essence Scatter, we should use Shock. They are more likely to have Negate than Essence Scatter, so the best play is to cast Goblin Chainwhirler, but in this example even the best play only has a 66% chance of being the correct decision.

The work of being good at Magic is the work of learning to pick from a sea of options the one that is the best play, the one that is the most likely to be the correct decision. The frustrating part of Magic is that being good at it has nothing to do with making the correct decisions. No one besides our theoretical perfect player can play Magic by seeking to make correct decisions. We mortals must be good at Magic by merely making the best plays we can, and in doing so accept that our choices come with a fail rate.

This is why we teach that you should never look at the top of your deck when mulliganing and why results-oriented thinking is harmful instead of helpful. These things concern themselves with what the correct decision was, not what the best play was. The two have nothing in common. If you want to know if your mulligan was correct, you’d be much better off sitting down and coding a simulation than looking at the top of the deck.

“Correct decisions” as the choice you would make with perfect information and “best plays” as the ones that are the most likely to be the correct decision are the definitions I’ll be using throughout this article, but please note that there is some simplification going on. I’m leaving magnitude of results out of the discussion and ignoring situations where a choice is 85% to marginally improve your situation and 15% to end in disaster. This kind of risk assessment is an important part of Magic, but only needlessly complicates this discussion. It’s a factor in finding true best plays, but let’s ignore it for right now.

Creating Easier Decisions

So far, this has all been concepts and ideas that competitive Magic players are familiar with. We all know that you can’t win every game, that sometimes you play around their twenty-outer but they had their one-outer and so you lost a game you felt like you had no business losing. These things happen. They come with the territory of being a Magic player. We know and accept that all we can do is make the best plays and be content with winning more than we lose.

Well, all we can do is try to make the best plays. Magic is a very hard game. Making the best plays is what we try to do, what we’re constantly getting better at and working on, what much of the content we consume daily is helping us learn about. It’s so hard to be good at making the best plays that it’s no wonder that we forget that some best plays are better than others.

That is, some decisions are easier than others. Consider the Essence Scatter / Negate example from earlier. What if, instead of knowing that your opponent has two Negate and one Essence Scatter in their deck, we know that they have two Negate and zero Essence Scatter?

Obviously, you cast your Goblin Chainwhirler and win the game. This decision is easier than the earlier one, but not because it’s a simpler conclusion to reach. The math behind both scenarios is trivial, neither best play is very difficult to find. No, this decision is easier because the best play is more likely to also be the correct decision. With two Negate and one Essence Scatter, playing Chainwhirler is 66% to win the game. With two Negate and zero Essence Scatter, it’s 100% to win the game. Casting Chainwhirler is the best play in both scenarios, but that same best play is much better in one of the scenarios.

Easy decisions are great because they reward us for our skill in discerning the best plays in games of Magic. We want our games to be as full of easy decisions as possible, which raises the following question: is the ease or difficulty of the decisions in our games something we have influence on?

To gain some insight into that question, let’s take a look at a slightly simplified version of a scenario I was in during last weekend’s Unified Standard Regional Pro Tour Qualifier, where I was lucky enough to have teammates that allowed me to play with the same B/U Midrange deck I played at #SCGINVI the day before.

My opponent was playing U/W Control with Approach of the Second Sun. We’re deep into Game 2, and the battlefield is a tad cluttered. I have The Scarab God and seven open mana, including a Field of Ruin. My opponent has two Knight of Grace that have been holding back my offense all game and that I have not found any way to take off the table, but The Scarab God is now ready to build a battlefield beyond the Knight of Grace’ ability to stave off over the next couple of turns. Things would be looking good, if not for the Approach of the Second Sun my opponent just resolved.

Because of the two copies of Knight of Grace, I don’t believe I will be able to finish the game out before my opponent draws their Approach again if they have anything that will let them get through their deck faster than a single card a turn. As you can see from my list, I only have three copies of Negate that can stop my opponent from winning the game again when that happens, and one of them is in my graveyard.

Luckily, I have a copy of Torrential Gearhulk that can use that Negate in my graveyard to stop the Approach, but leaving six mana up every turn is going to drastically slow down my ability to close out the game, as there aren’t enough creatures in the graveyards to let The Scarab God develop my battlefield at instant speed indefinitely.

All of this is to set up the following question: should I use Field of Ruin during my opponent’s end step and shuffle that seventh-from-the-top Approach into some other position?

Decision-wise, using the Field of Ruin is the correct decision if the Approach of the Second Sun ends up further down in the deck than seventh. Statistically, this is more likely than not to happen, so using the Field of Ruin is the best play by the definition we’ve been using. It’s the play I made at the time…and it is a play that I now think that I shouldn’t have made.

Using the Field of Ruin may have been the best play in that exact spot, but it also made the rest of the decisions I would have to make that game much harder. Essentially, the game boils down to me wanting to end the game as quickly as possible. Realistically, my opponent’s only path to victory is to find their Approach of the Second Sun before I’m able to win. My hand and the game state were such that the only way I could progress towards winning was to tap out of my Approach protection on a single turn. If they find the Approach on the turn I do so, they win the game.

U/W Approach right now commonly plays two copies of Approach, and that was the number of copies I thought my opponent had. Before I used the Field of Ruin, I knew exactly where one of them was: seventh from the top. By shuffling my opponent’s library, I effectively doubled the chances that they would have a copy of Approach on the top of their deck.

Shuffling with Field of Ruin increased the uncertainty of the situation and decreased the efficacy of my plays. I still had to tap out to close out the game before my opponent could climb back into it, and my play greatly increased the chances that doing so would lose me the game, despite also being very likely to have strengthened my overall position.

Identifying Inflection Points

There are lots of ways to make the decisions in your games easier, far more than I could ever describe in a single article. Instead of going through and discussing as many of them as I can, I want to share my method for finding the decisions that are the most worth thinking about.

By definition, the hardest possible decisions are those where the best play is 50% to be the correct decision. I call these decisions where the delta between the two options is essentially negligible “inflection points.” Whenever I become convinced that a spot I’m encountering frequently is one of these inflection points, I make a note of it and think about what I can change to skew the percentages in favor of one decision over the other.

I’ve played more with Modern Jund than with any other deck, and when you play a deck that much, you start to get a feel for the types of hands you’ll draw. I’m of the belief that the hardest decisions Jund faces in mulliganing are the above-average two-land sevens on the play. Stalling on two is quite bad for Jund, so average two-land hands are definite mulligans, while the best two-land hands go Thoughtseize into Tarmogoyf and have a high enough power level to be worth the risk.

The two-land hands in between average and excellent pose difficult decisions, because I believe keeping them hovers around 50-52% to be the correct decision. Such hands are among the factors in always choosing to play 25 lands in Jund rather than 24. Not the only factor, as bumping the needle on this decision from 52% to 54% would hardly be worth changing an entire card in the deck, but it’s a nice fringe benefit.

Another example of a common inflection point is role assessment in mirror matches. When both players are trying to do the same thing, knowing whether you are supposed to be the aggressor or the defender depends a great deal on not only the cards you’ve drawn, but also the cards your opponent has drawn, not to mention the cards that both of you will draw as the game plays out. You can make the best judgment possible based on the information you have, but the likelihood of your best play being the correct decision won’t be nearly as high as you would like.

Alternatively, you can skew yourself heavily towards a single role:

Aaron Barich is your Season One Invitational Champion, and he won with a Mono-Red Aggro deck that does not have to guess what role it wants to take in the mirror. If the aggressive stance looks at all good in a certain game, Aaron knows that his deck is full of reach to finish the job if the unknown information doesn’t pan out the way he hoped. This approach to mirrors won’t always be right, but it’s an effective way to make your decisions easier.