Several of my friends, guys who haven’t been playing much lately, have asked me what they should play in the grinders for Nationals. This question is tricky, because I genuinely think most of the commonly-played Standard decks are perfectly good choices. This is because Wizards has decided (correctly, I think) that the way to make Standard interesting is to offer a wide variety of powerful cards that support different strategies, so that each deck is fun and impressive. I don’t think there’s a clear best deck, because they’re all so strong.
Clearly, that’s not acceptable. “Just play whatever you want, they’re all equally viable” isn’t a responsible or accurate conclusion. The decks all have strengths and weaknesses, and some definitely outweigh others.
An ideal mathematical approach would be to make a chart listing every deck you’d consider playing on the left side, and every deck you’d expect to play against across the top. Add percentages based on how frequently you expect to play each deck next to the name at the top, and how frequently you expect to beat each deck in the corresponding box on this chart. For example:
| Deck A 25% | Deck B 18% | Deck D 15% | Deck E 15% | Deck F 15% | Other 12% | |
| Deck A | 55% | 50% | 70% | 20% | 60% | 70% |
| Deck B | 50% | 50% | 55% | 55% | 30% | 60% |
| Deck C | 60% | 48% | 45% | 50% | 70% | 50% |
In this chart, “Deck A” is the name for the most popular deck in the format. Your Deck A is 55% in the mirror because you think you have a pretty good understanding of the matchup and some pretty solid tech. This is one of the reasons it’s one of the primary decks you’re considering. Deck B is the second most popular deck, and Deck C is a homebrew that you built primarily to beat Deck A, but that incidentally gave it a pretty good matchup against Deck F. It’s not represented in the columns at the top because you don’t expect other people to be playing it.
This chart makes it reasonably easy to figure out which deck will win more matches on average. Just multiply the percentage that you’re winning against the percentage that the deck is being played and add them up.
Deck A: .55 x .25 + .5 x.18 + .7 x.15 + .2 x .15 + .6 x .15 + .7 x .12 = 53.65%
Deck B: .5 x .25 + .5 x .18 + .55 x .15 + .55 x .15 + .3 x .15 + .6 x .12 = 49.7%
Deck C: .6 x .25 + .48 x .18 + .45 x .15 + .5 x .15 + .7 x .15 + .5 x .12 = 54.39%
In this case, it would be correct to play Deck C.
Note that the percentages next to expected decks is not a direct metagame breakdown. It’s very close, and realistically, a direct metagame breakdown is as close as you’ll be able to come to the number you should be using, but the number you should be using is actually the frequency you’ll play against each deck. This needs to consider both how many of each deck enters the tournament and how long it stays in the tournament (essentially, how much it will actually win). If using a direct metagame breakdown, give slightly less credit to decks that you expect to perform poorly.
The problem with this model is that you never actually have all of this information. It’s easy to do in hindsight, particularly using Jared’s Too Much Information column to find all the percentages, but the best that it can do is tell you what the metagame for a past tournament was, and how a composite person playing a composite list of the decks you are considering would do. It doesn’t tell you how you actually perform with your version of the deck against a field you’ll face in the future.
A long time ago, I wrote an article about how I chose to play my deck after Grand Prix: LA last year by playing a ton of MTGO Constructed queues with several different decks, trying to ascertain my win percentage against different decks, and to discover how frequently I was playing against each of them on MTGO. Using that, I tried to predict the upcoming GP metagame. That led to me playing Mono Blue Faeries, with which I was very happy. In general, I found the process highly effective for learning the format and choosing my deck, but it’s extremely labor-intensive and still imperfect.
The fact of the matter is that we never have enough information to really treat Magic as a science to this degree, at least if we want reliable data. Late in the season, we might have enough idea of what decks there are, and what the matchups are like, so we can make a chart like this based on guesswork rather than solid data and it will be reasonably accurate. Right now, if I were to try to make that chart, it would be something like:
| Jund 18% | UW 18% | Ramp 13% | Naya 10% | RU Combo 15% | Mythic 8% | Mono Color Aggro 12% | Other 6% | |
| Jund | 50% | 40% | 40% | 60% | 60% | 60% | 45% | 60% |
| UW | 60% | 50% | 55% | 50% | 35% | 40% | 60% | 60% |
| Ramp | 60% | 45% | 50% | 60% | 40% | 40% | 50% | 60% |
| Naya | 40% | 50% | 40% | 50% | 65% | 55% | 45% | 60% |
| RU Combo | 40% | 65% | 60% | 35% | 50% | 30% | 55% | 70% |
| Mythic | 40% | 55% | 60% | 45% | 60% | 50% | 45% | 60% |
| Overrun | 45% | 40% | 55% | 50% | 55% | 50% | 50% | 60% |
These numbers are off the top of my head, based on my general assumptions about how these things go. Some of them could easily be completely wrong. Also, this is the first time I’ve done anything like this, and I don’t know what deck it will tell me to play. That said, overall, these numbers look fairly reasonable to me. If you disagree with some of them, it’s extremely easy to change them around, but I think this is a reasonable starting point. So, let’s see what it says:
Jund: .5 x .18 + .4 x .18 + .4 x .13 + .6 x .1 + .6 x .15 + .6 x .08 + .45 x .12 + .6 x .06 = 50.2%
UW: .6 x .18 + .5 x .18 + .55 x .13 + .5 x .1 + .35 x .15 + .4 x .08 + .6 x .12 + .6 x .06 = 51.2%
Ramp: .6 x .18 + .45 x .18 + .5 x .13 + .6 x .1 + .4 x .15 + .4 x .08 + .5 x .12 + .6 x .06 = 50.2%
Naya: .4 x .18 + .5 x .18 + .4 x .13 + .5 x .1 + .65 x .15 + .55 x .08 + .45 x .12 + .6 x .06 = 45.95%
RU Combo: .4 x .18 + .65 x .18 + .6 x .13 + .35 x .1 + .5 x .15 + .3 x .08 + .55 x .12 + .7 x .06 = 50.9%
Mythic: .4 x .18 + .55 x .18 + .6 x .13 + .45 x .1 + .6 x .15 + .5 x .08 + .45 x .12 + .6 x .06=51.4%
Overrun: .45 x .18 + .4 x .18 + .55 x .13 + .5 x .1 + .55 x .15 + .5 x .08 + .5 x .12 + .6 x .06=49.3%
The first thing I notice about these numbers is that they’re strikingly similar, and that all the differences are well within the margin of error of me making up numbers. Also, I feel like there’s a good chance these numbers overall are just confirming personal bias. That said, I’ve been pretty happy with my deck selection for the last year or two, so feel free to take these numbers as you see fit.
Note that the worst performing deck in my head is Naya, and the best, by a very small margin, is Mythic. These decks have a very similar core, and I’d like to take a moment for a brief aside on why I think Mythic is a better deck outside of made up numbers. But I’ll try to do it in a way that ties in more generally to my discussion of deck selection, so let’s step aside for a transition to explain how this fits into a larger picture, and then come back to my tangent. I’ve probably lost you at this point. That’s okay. It’ll all become clear.
I think too many people list their decklist as a collection of around 36 spells, and then say something like, “and the 24 lands to make them work.” The habit of doing this rather than taking the mana in a deck seriously has led a lot of people to some misevaluations of decks. I think comparing Naya to Mythic is where this becomes particularly egregious. It’s easy to look at both decks, and understand that you’re using some acceleration creatures and Knight of the Reliquary to play some sweet spells, and then compare the Red cards to the Blue cards. Yes, Cunning Sparkmage and Bloodbraid Elf are better than Mana Leak; Rhox War Monk; Jace, the Mind Sculptor; Sovereigns of Lost Alara; or whatever other Blue card Mythic is playing. Yes, Bloodbraid Elf really is better than the Blue options. However, those Blue cards aren’t bad; they’re actually pretty sweet in their own right. I’d guess that they’re good enough that some people even play Mythic because they think they want to cast the Blue spells more than they want to cast the Red spells.
For me, that’s all missing the point. I play Mythic over Naya for the lands. Yes, Raging Ravine and Celestial Colonnade are both awesome lands, and no, Celestial Colonnade is not so much more powerful that it makes me want to play the Blue cards. The issue is that Celestial Colonnade taps for the two colors that aren’t Green, while Misty Rainforest gives me “an untapped Green dual.” This means that I can satisfy my non-Green requirements with tapped lands, which allow me to play more untapped Green lands so that I can play Birds of Paradise or Noble Hierarch on turn 1. Most Naya decks have between eight and ten Green sources that enter the battlefield untapped, and they have to work for them. Mythic has closer to thirteen while being better at enabling its other colors. That’s a real difference. Aside from that, Noble Hierarch’s ability to tap for Blue mana makes the splash even easier. If you’re just playing G/W, adding Blue is almost completely free. Adding Red takes serious effort. All of this just makes the cards that both decks are playing subtly more consistent and powerful in the Mythic deck, which to me makes a huge difference.
Generalizing, I think consistency is something people can often fail to properly appreciate if they’re just looking at new lists and don’t have a lot of experience playing the decks. Sometimes a deck looks awesome on paper, and then you pick it up, and it just feels much clunkier than you possibly could have expected.
Another point on deck selection: One often hears Magic players who don’t play as much complaining that competitive Magic is all Rock-Paper-Scissors. You choose a deck and just hope you don’t play against the decks it can’t beat. In my experience, in almost every format I’ve played a lot, I eventually find a deck that I feel I can pilot to consistently beat almost everything. I try not to spend too much time on decks that have some really terrible matchups unless I think those matchups will almost never come up. Otherwise, I like to work with decks that have close matchups, and try to figure out good plans for the matchups that are traditionally harder. For this reason, I’d try to stay away from decks on my chart that have numbers below 40 somewhere, even if their overall numbers are pretty good. This may not be rational, but I like to feel like I’m never dead when I sit down.
On a related note, once you’ve chosen a deck, it’s reasonable to use a chart like the one above to figure out how much attention you want to devote to any given deck in your sideboard. You could assign a “Needs Work” index to each of your bad matchups, a function of the percentage below fifty that you expect the matchup to be multiplied by the percentage of the field you expect the deck to be. So Jund would have a “Needs Work” index of .1 (10% below 50% match win) + .18 = .018 against UW, and an index of .05 x .12 = .006 against Mono Color Aggro decks. This doesn’t directly tell you how many slots to dedicate to anything in your sideboard, since the exact increase in value that any given card gives you multiplied by the frequency that you bring it in to get that percentage is actually the most important metric for sideboard cards, but it might be correct to weight that primary metric in some way by the Needs Work index.
This is where I should conclude by saying what I’ll be recommending to my friends for the grinders and why, but I’ve really already said all that with my chart. At this point, if you don’t know what to play this weekend, my recommendation is to take a serious look at the chart, question both the matchup percentages and the metagame percentages, adjust them where you feel confident in information you have that disagrees with my guesses, and let the numbers guide you when you’re uncertain.
Thanks for reading.