While we were preparing for Grand Prix Phoenix, Mark Jacobson shared a blog post by Allen Wu. (You can find his blog here.) The post was about using Monte Carlo simulations to find the likelihood of certain opening hands for Eldrazi Tron. The conclusion he had drawn was that you should be willing to mulligan to 5 cards in search of a busted opening hand, because it happens more often than you think. The post is super interesting and I recommend reading it.
More importantly, he took the simulation framework he built and applied it to Affinity. Specifically, the likelihood of having 2 mana on turn 1. The conclusion he came to was that every zero-cost card you play increases the likelihood of having 2 mana on your first turn by around 2%. Furthermore, it doesn’t matter much if it’s Memnite or Welding Jar—as long as it costs zero, you’re getting a huge gain on your busted openers.
Some of Allen’s other posts have more thoughts about Affinity, and specifically the first two turns of the game. It’s easy to notice after a few matches with Affinity that if you have an unchecked Steel Overseer, your chance of winning goes way up. As a result, your opponents make getting rid of it an enormous priority, often at the cost of their first or second turn. In Modern, the majority of the action in most games takes place in the first few turns, so throwing your opponents off on their most important turns by forcing them to react to you is a strong strategy.
Consider the following sequence:
In this case, not only is the B/R player way ahead, but they get to untap with three mana, only one of which they need to remove the Overseer. They are not going very far out of their way to both further their game plan and stop Affinity’s.
Compare it to this one:
Should they ignore the Overseer and cast Goblin Lore, or should they spend a mana to Bolt it and hit for one with Adept? Option A lets them develop their board at the cost of letting Affinity’s board develop explosively, while option B prevents both players’ development while passing the initiative back to Affinity. Neither option seems particularly attractive. In one single turn of the game, Affinity has neutralized the disadvantage of going second while stealing the opponent’s most important turn.
The conclusions I came to after reading Allen’s posts were that I should be building my deck to optimize my first few turns more, I should be more willing to mulligan marginal seven-card hands in search of 2 mana on my first turn (especially on the draw), and I should consider sideboarding differently on the play and on the draw. I took all those conclusions and came up with this list.
Notable features of this list:
2 Memnite, 2 Welding Jar
12 total 0-drops (with Mox and Ornithopter). I never liked Memnite, but now I realize how important they are for maximizing the explosiveness of this deck. Welding Jar, on the other hand, is really awesome and does lots of things. On top of adding 2% to your busted-turn-1 equity, it also provides a layer of backup for when you do play Overseer on turn 1.
3 Etched Champion, 1 Master of Etherium
I found this split by looking at some of my old lists from 2012, when Jund with Bloodbraid and Deathrite Shaman was the most popular deck, if not the best deck. It makes sense to build your main deck for the opposition you expect to face, and with everyone thinking Jund is the best deck and everyone else wanting to try to make Jace decks good, I wanted to choose my threats appropriately.
I kept one Master because you would still like the option to be really fast in some matchups, like Storm or Ad Nauseam, and also because overloading on Etched Champion can be a huge liability if you draw some colorless matchups like the mirror, Tron, Eldrazi decks, Lantern, etc.
2 Thoughtcast and 2 Galvanic Blast
When Jund and other B/G/x midrange decks were on top, having a lot of Thoughtcasts to reload was really important. That fell to the wayside when Bloodbraid got banned and the popularity of Twin exploded, requiring you to have 4 Galvanic Blasts in your deck to prevent dying to the combo on turn 4. Even after Twin got banned, B/G/x decks didn’t magically get better, and creature combo decks kept popping up out of nowhere, meaning that you didn’t have time to draw cards and needed the utility of Galvanic Blast.
2 Ceremonious Rejection
Somewhere along the line, my Tron matchup went from pretty favorable to kind of unfavorable. I’m not sure why exactly, but I suspect it’s some mixture of Sanctum of Ugin, Ulamog, and Walking Ballista. Previously I was attacking the matchup with cards like Thoughtseize and Spell Pierce, but Pierce can’t always hit important cards like Karn or Oblivion Stone, and attacking early mana development is not a reliable strategy. Thoughtseize is a bit more effective in this matchup, but still underwhelming—if they have two threats and you’re equipped to deal with zero, you didn’t really accomplish much.
Rejection was the card I was looking for because it stops Oblivion Stone no questions asked, and also forces them to commit their entire turn to a threat. You are often getting a huge trade-up on mana with it. I think this card single-handedly turns this matchup back into one I don’t mind facing.
1 Ray of Revelation
This is kind of a pet card of mine. I almost always have one and I’m usually the only one. Back in 2012 I didn’t have Aether Grid as a reasonable plan against Stony Silence, and I also liked having a card against Bogles, which was at peak popularity (until its comeback in the past two months, anyway). Nowadays, Bogles is a premier strategy in Modern, so it makes sense to me to have a strong sideboard card for it. It’s also not hard to find other fringe uses, like killing Detention Spheres out of U/W Control or having a card against Phyrexian Unlife out of Ad Nauseam.
0 Thoughtseize effects
Like I was saying in the Ceremonious Rejection section, these days I am much more into having my opponents commit their mana and their key turns than I am trying to remove single threats with discard spells. For the time being, I don’t think Affinity is a good Thoughtseize deck.
As for the Grand Prix itself, these were my matchups:
R3: Humans – lose
R4: Burn – win
R5: Jund – win
R6: U/B Mill – win
R7: U/W Control – win
R8: Burn – win
R9: Madcap Moon – lose
R10: Bant Spirits – win
R11: Eldrazi Tron – win
R12: Storm – win
R13: Humans – lose
R14: B/R Hollow One – win
R15: Naya Zoo – win
12-3, good for 16th place, $1,000 and 3 Pro Points.
I lost to Humans twice. In the second match vs. Zan Syed on Day 2, I made a lot of bad mistakes in game 3. I just plain didn’t have enough reps against Humans, since it became popular during a period when I didn’t think Affinity was very good. I think that matchup is good for Affinity, but not by much, and it’s easy for that edge to disappear if you don’t know all the ways that the Humans cards can interact with each other.
A lot can change in a month, but I really like Affinity for now. I think it’s good against Jund and untuned Jace decks. I don’t feel like the deck is underpowered relative to the other Modern decks like I did a few months ago. Really strange that the catalyst for this was adding even more power to the other decks by unbanning two cards and not taking any away.
Props to Allen Wu for giving me a new way to look at Affinity—not easy to do for someone who has over a decade of experience with a deck.