facebook

CoolStuffInc.com

CSI 22nd Anniversary Sale ends Sunday!
   Sign In
Create Account

Lies to Make You Better at Magic

Reddit

An entire genre of article exists to try to make people better at Magic without giving the reader direct tips about playing or deck-building; they instead try to change people's mindsets going into tournaments and play-test sessions so that the reader improves more quickly. These articles aren't necessarily bad. They've been responsible for plenty of people deciding to buckle down and improve their games.

This is not one of those articles. I cannot say that this will make you a better player. In fact, it's more likely to make you worse. But it will be more accurate than those articles because it won't lie to make people improve.

Say you're preparing to go to a tournament. You've tested with your team, and after playing ten matches (full ones!) against different gauntlet decks, you managed a winning record against most of them. The testing results, combined with your confidence in the deck, makes you so confident that you think if you don't win this thing, it was because you screwed up somewhere along the way.

You don't win the tournament, of course. You don't even make Top 8. So, what went wrong?

Play-Testing Won't Tell You What the Best Deck Is

. . . unless you play a massive number of games.

Let's look at those play-testing results:

Aggro versus control: 9–1 in matches (9–1 Game 1, 10–6 sideboarded)

Aggro versus combo: 3–7 in matches (2–8 Game 1, 7–9 sideboarded)

Control versus combo: 5–5 in matches (5–5 Game 1, 6–5 sideboarded)

What conclusions would you draw here? Well, the control deck is obviously worthless, since it gets crushed by aggro and doesn't do better than fifty-fifty against combo. The aggro deck looks pretty decent, but it has a terrible Game 1 against combo. If these were the only three decks in the format, any reasonable person would play the combo deck, since it doesn't have a bad matchup.

Does that all make sense? People have certainly chosen decks and done well without playing thirty full matches of Magic, and that's a lot of work that went into it.

The problem is that I generated those match results by flipping a coin for every game. Random.org makes this pretty easy.

The truth is that, unless you're spending all your free time playing the exact same decks against each other, you're not going to find out what beats what a percentage of the time. Every matchup percentage you see in articles is a guess. Just to determine that a specific matchup is within ten points of fifty-fifty would take a hundred matches (and even that's only 95% certain). Keep in mind that this is all perfectly best-case-scenario stuff, too: equal play skill, the decks never change, the plays never change. In reality, by the time you figure out whether a deck is good against a specific target, the metagame is two steps evolved already.

Small sample sizes won't just make you think you have a great matchup when it's actually middling or bad. Have you ever built something the night before a tournament, taken it to a play-testing session, beaten everything in sight, and thrown out the deck you've been testing for weeks? Have you tried out a new sideboard card, had it win both games you drew it, and made it a four-of on the spot? Anything that seems good purely based on how it performed in a small number of situations needs to be viewed skeptically. And “small” is probably a lot bigger than you think.

If no one can determine whether a deck even has more than a 50% matchup against anything else, why does anyone bother play-testing? It makes you better. By getting blown out by a trick in one specific game (that was, statistically speaking, close to irrelevant), you learned how to play against it better. Even if you don't end up going to the tournament at all, you've practiced a bit with that archetype, so the next time a similar deck rolls your way, you'll be prepared.

As I've written previously, there's no secret to how people get good at Magic: They play way more Magic than most people think is possible, and they do it carefully, against good opposition, surrounded by people who will critique them for everything. This is how the guy who rolled out of bed and was handed the world's most complicated control deck still managed to do well at your PTQ: He's been playing five hours of Magic a day in different formats, so he play-tested before the format even existed.

The good news? Because the correct deck is so impossible to determine, all you need to do is choose something that's been doing well in other tournaments (sample size matters there, too, obviously) and practice with it.

You're Not Going to Choose the Best Deck Anyway

. . . you're going to choose the deck you like the most.

PVDDR, Brad Nelson, and Conley Woods are all good players. Preparing for tournaments, though, they'll likely be drawn to different decks: control, aggro, and unpredictable, respectively. I'm sure that they're all convinced that they're choosing the deck that gives them the best chance to win the tournament, but each will usually play a deck that fits his style. We all do it to some extent or another.

There're a few factors at play here—the first is what's known as the halo effect (once again, I'm paraphrasing research, essays, or ideas by Daniel Kahneman, though not everything is necessarily his originally). When we see something that we like, we're going to ignore its negative aspects and focus on its positives; conversely, for something we dislike, we'll see only the bad in it. If a player likes a deck, he'll overestimate its chances and underrate decks he dislikes, thus giving himself what sounds like a logical explanation for choosing his own pet deck.

Then, once he plays with it, confirmation bias starts to creep in; any match he wins is proof he's made the right choice, while any loss is either forgotten or written off to bad luck (which we'll get to later) or a mistake in play that he won't make again.

After that, his perception of the field will be twisted: Any deck that he consistently beats will be seen as unplayable, and anything that beats it starts to seem good. This is because, from that player’s point of view, he sees certain decks winning or losing constantly, and that must be all the information necessary. (Behavioral economics nerds will sometimes call this, “What you see is all there is.”) No one is going to switch decks to one that he sees losing every time it’s played. If it makes Top 8? Well, the pilot must have been unbelievably lucky.

So, does this mean you should always pick the deck that won the last tournament, or make a list of heavily played decks and roll a die? Of course not—there are reasons to play the deck you like. Specifically: You like it! This isn’t some Magic-is-a-game-so-have-fun, Timmy thing—if you enjoy playing a deck, you’ll work harder at improving it and playing it against the field, and you’ll talk about it with other people more. Just make sure it’s at least a reasonable choice, and be wary of biases you have for or against specific cards.

But back to that tournament you lose. You've read all the articles; Magic is a skill-intensive game. There were hundreds of choices you didn't see. If you lost, you must have screwed up somewhere. Right?

A Large Percentage of Magic Games Are Decided by Luck

. . . you just need to get over it and maximize your chances.

The evidence that Magic is a skill-rewarding game is everywhere; ChannelFireball team members will routinely take a quarter to half of Top 8 slots at major tournaments. The same players do well over and over.

But what about the team members who didn’t Top 8 that event? What about the name pros who did well at the last tournament but not this one? We tend to overlook the cases in which incredibly good players don’t do very well. If you total together ChannelFireball team results at any big tournament, the composite record will usually be somewhere between a 60% and 80% match-win rate. Archive statistics on this sort of thing can be hard to find, but a player who wins 70% of his matches on the Pro Tour is Player of the Year–caliber. Of course, this is because he doesn’t win 70% of his matches at every tournament; he’ll go on what seem like hot and cold streaks, just like any other chance-influenced outcome.

The point, though, is that if you show up to a tournament with a 70% match win rate against each and every person, that would make you among the best players there. It also means you have just worse than a fifty-fifty shot of starting out 2–0. And after six rounds with a 70% win rate, your chance of going undefeated drops below 10%. To win a tournament, a player needs to be incredibly lucky, because Magic still involves a large amount of variance. When people talk about variance, it's usually to explain a loss. This is often true. But it's just as true when someone does well.

But wait, every article about how to become better at Magic says to never, ever, ever blame luck, because doing that will make you awful forever, and you’ll never see your mistakes! That’s completely true. Win or lose, every play you make should be up for second-guessing after the fact to see if you would have won more often by doing something different. Eventually, you’ll find the lines of play that push you from a 50% win percentage up toward the elite of winning 70% of matches. The fact that a specific game was decided by luck doesn’t change the fact that you could have shifted the odds slightly. The question to ask after the game isn’t, “How could I have won this game?” That will lead to hindsight biases caused by knowing what cards were in your opponent’s hand, what ended up being on top of your deck, and so on. Instead, ask, if you’re facing a nearly identical situation in the future, what should you do then?

 


Anyone using the information in this article to make excuses is getting the wrong message. It might be impossible to determine the statistically correct deck to play, but you still need to play-test to get a feel for how the interactions work and deduce how those hundreds of games probably play out. There’s nothing wrong with playing a deck because it’s your style, but you still need to watch out for biases that’ll make you play something that’s completely unworthy of attention—or play bad cards because you happen to like them. Magic might have a huge amount of variance, but you still need to scrape and claw for every edge you can find to improve.

Jesse Mason

killingagoldfish.blogspot.com

@KillGoldfish

Sell your cards and minis 25% credit bonus