Some times I've gotten super lucky in the not-so-distant past:

● 2/22/2015 - I won four consecutive Swiss rolls on the way to taking down a Standard Preliminary Pro Tour Qualifier. After winning those four rolls I put myself into a lock for first place in the Top 8, meaning I was on the play for all seven matches I played...the entire tournament! Winning just four rolls in a row happens a surprisingly minuscule amount of the time (half of a half of a half of a half) but, according to this article, play/draw choice gives you a substantial 54/46 edge to start any given match.
● 4/25/2015 - After sliding into the Top 8 of the Regional PTQ I finally got to play against my best matchup: Esper Control. Not only that, but I was on the play! What a time to get a lucky break, am I right?

I have not yet figured out how to affect dice rolls in a meaningful way*, but matchups are another matter.

"How to Predict the Future" is a companion piece to our mini-series on luck, beginning with Belief and Luck and continuing last week with Luck? and Topdecks. Luck is something that many players revile, Worship, or both; a seemingly uncontrollable force that does more to define the winners and losers in a Magic tournament than the skill of the players involved. It turns out that there are some potentially unexpected ways we can affect our luck, including making the best possible choices in order to take advantage of good fortune (like Kai Budde prior to the most fortunate set of topdecks ever in the history of the Pro Tour) or evaluating similarly-powered cards in light of the contexts around them in order to break card appropriateness ties.

We tend to focus on topdecks when we talk about luck but matchups actually do more to define our fortunes than every other element [seemingly] out of our control combined. We can get lucky there too.

Matchups determine the upper limits of our potential performance probabilities. Champions find ways to weather a bad matchup or so in the late rounds, and our best possible performances are of course going to go hand in hand with our best execution....but! If your matchups are bad enough, it doesn't actually matter how well you play.

Imagine you are playing the Lava Spike deck in Legacy. If your competent Reanimator opponent keeps a seven-card hand you should assume you've lost on the second turn. Same if he smiles on a mulligan to six. Your fastest kill is maybe turn four; he is going to have any number of ways to get Iona, Shield of Emeria in play before you can kill him (and then you won't be able to). The Lava Spike deck might have better overall expectation across a large Legacy tournament, but heads up it doesn't matter.

The same is more-or-less the case in Standard. Dimir Control v. Mardu Dragons isn't nearly so lopsided, but you can't play against Radiant Fountains and Bile Blights too many times and really expect to win a tournament with Mardu Dragons. You have limited reach, they have tons of efficient removal, and you probably can't budge a 'walker with all those Draconic Roars and Foul-Tongue Invocations globbing up your grip, instead of having, you know, Hero's Downfall.

We can't ultimately force opponents to pick particular decks - we only pick our own deck - but predicting what they might do helps us to pick our own deck with greater skill (and helps put us in a position to get lucky more often).

All we have to do is predict the future.

Here's the thing about predicting the future: Few people are particularly good at it. Wolves who get paid millions of dollars on Wall Street are not particularly good at it, at least for any sustained amount of time. The vast majority of managers in charge of running billion dollar sports franchises are by and large completely incompetent at it. Competing economists wage wars of letters in academic journals or The New York Times about whether one or the other are ever even right versus not looking long enough into the future (just look at what is happening in Greece right now!). But as Magic players we have a couple of different sets of rules, and a couple of different sets of tools that can help us predict the future profitably (and get, or at least look, luckier).

First, we must learn to focus on potential future outcomes instead of principles and presumptions.

"How we've always done it" is about the biggest potential barrier to progress that there is. Adhering to what isn't good any more ends up just being a massive waste of resources.

Look at Uber as an emerging technology. Maybe Uber is going to have to bend; but my guess is that the attitude of the Italian governments and French cabbies is just a waste of money and time, ultimately just a tragedy of human suffering. Imagine guys if, in the 8-track pressing union, were prank-calling MP3 encoders to beat them up in distant alleyways (and smash their iPods). They might thump a few computer nerds but in the long run it would be pointless. I can hold my entire music library in MP3s on a device smaller than a single 8-track cassette. MP3s are miles better for consumers so they will win no matter the opinion of the now long defunct 8-track industry.

Flexibility in Magic goes the same way. You might start out with a nice piece of technology; maybe a deck with a large number of free value-adding synergies. But as time and the metagame moves you have to be able to slough off part of -- if not your entire -- strategy or no amount of luck will save you.

A good example is this recently reprinted card:

The last time it was in Standard that card was among very good friends. Goblin Warchief and Siege-Gang Commander made it both powerful and fast in a way the 2015 version probably won't be.

But the best Goblins deck designers figured out to cut Goblin Piledriver at Japanese and English Nationals, separately.

The problem was that, as powerful as it sometimes was, Goblin Piledriver was never going to be tough; that meant it was mostly just dead. Or at the very least never getting through. "Everyone" could block an x/2 creature, and attacking with enough Goblins to make one big was going to invite tragic blocks on all your other guys.

Goblin Piledriver was a card that - if it were in your deck - you might be loath to side out. It was the definition of a deck darling and, as a proactive card, untouchable in a way Draconic Roar or Foul-Tongue Invocation isn't today. It would be more like randomly siding out Seeker of the Way.

So when you were looking over the shoulder of a Goblins player with an opponent on five but no way to bust through you might attribute his topdeck of a late game Piledriver as bad luck. When in fact it, in fact, was completely under that player's control. It wasn't luck at all!


Cutting the card for Goblin Goon, Electrostatic Bolt, or even Shrapnel Blast would almost mathematically create different opportunities for good fortune.

What do you think about this one?


Automatic four-of for as long as it has existed, right?

I fear that not enough players have yet considered the number of liabilities this card can create. Like, how bad does Courser of Kruphix become if it is your only enchantment? Against attackers Courser of Kruphix players are usually investing time, mana, and planning for future turns every time they tap three. Sometimes they even forego the opportunity to play legitimate four drops to get a freebie card and life point! That doesn't make Courser "bad" necessarily (Abzan Control plays a more resilient Courser than Sultai Control for instance, while Five-Color GB Dragons plays the odd three there is certainly a range), but in a world of Dromoka's Command it is something that must be consciously contextualized.

How about with this new addition?

Courser of Kruphix has never been Jace, the Mind Sculptor across the board, but it has generally been reliable defense against small red men. What incentives does Exquisite Firecraft create? What synergies does it uncover? How reliable of a blocker does Courser of Kruphix remain when a red opponent starts playing eight cards that deal four, and efficiently?

Magic is a game of relatively small n. The smaller the n of a game the more meaningful individual events seem. Commentators go absolutely nuts over every goal in a futbol match and the same is true for crunch-time minutes in the NBA playoffs (though the reality is that baskets scored in the middle of the second are worth the same as buzzer beaters).

When we talk about even substantial tournaments with memorable results we can be talking about maybe a dozen opportunities to check our predictive ability, at least considering only a single Magician.

But this, too, can help us to get lucky!

Imagine a hypothetical format with four potential expected strategies.

Further imagine two different decks (that for our modeling purposes are outside the above stated parameters). One deck (A) has a 50% initial expectation against all four expected decks. The other (B) has 99% initial expectation against two and 1% against the other two.

Which is the sounder deck choice?

If your goal is to win the tournament, one is always better than the other. But which?

Imagine you played a four round tournament and you played against each deck once. Your expected outcome would be 2-2 regardless of which deck you chose.

2-2 does not win the tournament.

In fact your long run w/l with each deck is about the same.

And given the same distribution Deck A will 4-0 about the same amount of time that I will win all four of my Swiss dice rolls (or, not very often).

But, what if the four decks do not occur at an exactly 1/1/1/1 clip? Sometimes it is 0/0/0/4, or 0/0/2/2, say. Deck A never changes in expectation, but Deck B gets way better or dramatically worse.

A common thing you will read is that players choose decks that are [some number that is essentially 50%] "against the field." That is rarely the best choice for the average player. In a recent interview Jon Finkel stated that many players are "too risk-averse" and "too afraid of randomness." Imagine instead of a certainty of a 50% ratio you embraced sometimes having only 1% starting likelihood of winning but sometimes having a massive 99% advantage. If you are good at predicting when Deck B's bad matchups are unpopular you can dramatically improve your results. Even when you are not good at it (you pick Deck B both when you should and shouldn't) your results will not really be worse than always picking Deck A; in fact, they might, surprisingly, be better.

Remember when I said Magic has its own rules?

Forgetting about the fact that there is some skill to predicting when your good v. bad matchups will show up, perceived results math will tend to favor Deck B. That is because 0-2, 0-3, 0-4, 1-2, 2-2, and even 3-1 are all basically the same record if you are only interested in 4-0.

Put a different way, no one cares about Deck A's 2-2. That's just what we expected of Deck A! All of Deck B's 0-4s are exactly the same as Deck A's 2-2s because we are solving for 4-0s.

Because we can't get better than our starting values by ourselves, when Deck A exceeds expectations it will usually be because the opponent donated percentage with mistakes or mulligans. On the other hand, Deck B's wins will consistently come hand-in-hand with prediction and preparation ("I picked this deck because it is good against that kind of deck, which I expected"); and its losses also with prediction ("I knew I would have trouble with that kind of a deck, but that was a risk I chose to accept.")

How lucky!

If I can only choose "always Deck A" or "always Deck B" I think I'd rather choose the latter. But part of the fun of Magic is that you can change decks from tournament to tournament. I think the best strategy is to be Deck A sometimes but Deck B more often than you might guess.

When to pick Deck A:

● When you are either one of the best players in the world or in some other context where it is more likely your opponents will Donate points to you (via mistakes) than you will Donate points to them. So… If you're the best player at your LGS, Deck A will often make for the better FNM choice.
● Most long tournaments - If you are playing in a tournament of sufficient size that you will almost necessarily play one of your bad matchups one too many times for comfort, you can't really play Deck B. At Grand Prix New Jersey last year I was playing for Day 2 with the burn deck and burn master Patrick Sullivan instructed me not to choke. Does it count as choking when my opponent was Reanimator and played turn two Iona both games? I played the tournament very well but just happened to run into Rest in Peace + Energy Field and turn two kills in three of the nine rounds. A steadier deck might have been a better choice.
● Any time you think Deck B would have poorer-than-average matchups.

When to pick Deck B:

● When your opponents are unlikely to give you any free points. Imagine for a second you are playing Owen Turtenwald heads up, one match, for a million dollars. Would you pick the deck that can "beat anything" (but also lose to anything, exactly half the time), or the deck that has a great matchup half the time? The fact is, it is much more likely you're donating points to Owen than he is donating them to you (meaning he is likely to win close games). The realistic expectation is much lower than 50% for Deck A than is advertised, at least for most players. Deck B will retain good chances in some of the matchups, no matter what.
● When you are capable, but liable to give up some percentage a good amount of the time (same reason as above, if less pronounced).
● When you can actually make a good prediction of the metagame, so you know your chances of playing against your good matchups is heightened relative to average.

Now of course in real life Magic offers us more than two choices; and there isn't really one deck that goes halfsies against everything and another that either always wins or always loses. These are just constructions meant to illustrate some points. Randomness can be good for your win percentage, and the same is true for getting out of a comfort zone based on certainty. Both tools can help improve your luck.

You'll also note that while our Deck A and Deck B are a little extreme, there are some pretty decent analogues in real-life Magic. My Five-Color Monoblue Dragons deck is a good example of a deck that struggled against Abzan Aggro but was fantastic against the then-super common matchups of Esper Dragons and Monored Aggro. I tend, personally, to do better when playing decks with more extreme perspectives on the format. For a player like me, who can sometimes make a key mistake on the table, getting a lot of the work done ahead of time lends a lot of leverage.

What are the sources of information around predicting opposing decks?

Remember, this is far from an exact science and Magic is a game of small n but having the additional tools of "getting lucky on matchups more often than the average player" is a hell of a superpower.

Just as a note, these techniques tend to be most effective when beatdown, combo, and certain types of control are good (or at least popular) and least effective when flexible/midrange and over-the-top type decks are good.

● Last week's Top 8s - We are almost overwhelmed with not only this information from sites like this one, but commentary on what each Top 8 might mean. Any time you have the Impulse to copy a new/different deck that just appeared in a Top 8 remember that there are thousands of others who had the exact same Impulse. Copying last week's tech is not one step ahead; it's only one step ahead of last week.
● Obvious strategies that are enabled by new cards or changes in a format - Dragons of Tarkir was begging players to shift from Planeswalkers to Dragons. Silumgar's Scorn is that good. Personally, I bought a play set of Draconic Roars without any idea of where they might see play. Dragonlord Ojutai is kind of super obvious even if Dragonlord Dromoka and Dragonlord Kolaghan aren't. Just from Magic Origins it is pretty obvious we will see some new/different Planeswalker strategies, a resurgence in burn, and interest in both Elves and Goblins. Stormbreath Dragon will be weaker than pre-Magic Origins as white gets some tools to fight it, whereas the Theros Block gods (and other interesting enchantments and enchantment-enablers) may spawn new decks.
● Not-obvious cards and strategies should not be considered, though. Dromoka's Command is a card that was definitely going to impact Standard; Crucible of the Spirit Dragon was not. It turns out Crucible of the Spirit Dragon is very good, but preparing for it (relative to Dromoka's Command) would be a waste of time. That is, unless you have a small enough tournament and can take advantage of:
● Talking to players from a certain area to see what they are thinking - Just ask! This works at every level from LGS to the Pro Tour by the way.
● Time! We are about to hit a rotation; players are much less likely to have access to new cards the first couple of weekends, so even if there are awesome new cards in a new set, it is less important to consider them, near-term.


* Coin flips on the other hand...