Affinity At Grand Prix Portland

Zvi Mowshowitz goes over the ins and outs of the Affinity deck that he played at the Modern Grand Prix in Portland a couple weeks ago, where he finished in 3rd place.

Two weeks of intense work got me 9-7 and no cash at Pro Tour Dragon’s Maze. A dedication to not caring about the tournament whatsoever got me into the semifinals of Grand Prix Portland.

Maybe I’m doing this preparation thing all wrong.

No preparation is, of course, a misnomer. Without the years of playing Magic I have under my belt, I wouldn’t have understood what my opponents were up to or how the cards in my deck worked, and things would have been hopeless. Instead, I knew exactly what was going on by turn 3 of every match and felt a bit rusty for it not being clear by turn 2, which led to me sequencing my lands wrong in one game in a way that could have been fatal.

I also wouldn’t have known where to turn to get the best list! What I did not do, however, was try to make my own mark on the decklist that I used, which I instead got directly from my teammate Paul Rietzl. I knew that he was providing me with a list that had been tuned within an inch of its life and that any modifications I made were more likely to subtract from the deck rather than add to it. When I was emailed this list, I hit the forward button on my phone to register it sight unseen:

The Ins and Outs of the Deck

This list impressed me throughout the tournament, with the only cards I would consider for the maindeck that weren’t there being the fourth Etched Champion, the third Galvanic Blast, or possibly the second Master of Etherium. You can’t add more than one additional three-mana card to the deck, but it’s possible that you can get away with five. If the field wants you to have a third Galvanic Blast, that is certainly reasonable. What is even harder is finding cards that can leave the maindeck. The second Memnite is the only card I feel one could reasonably cut unless the metagame lets you move the second Galvanic Blast to the sideboard (which right now it almost certainly doesn’t), and I’m reluctant to do even that. The deck is highly linear, so most of the card choices are completely forced.

Sideboarding is very dangerous for this deck because of how important it is to maintain deck flow. Every card you put in that isn’t part of the deck’s core engine weakens the engine and makes it less likely that you will have the ability to get out of the gate fast and follow it up with good action. Without a fast start in the first two turns, opponents can take you apart with their more expensive cards, and without a boosting effect of some kind to make your creatures bigger, they’re only one or even zero power and the opponent will have plenty of time to catch up. There are cards you can trim from the deck without posing too big a risk, but there aren’t many:

If protection from all colors isn’t relevant either because they have colorless cards, they have abilities that ignore protection, or weren’t trying to stop your guys in the first place, you can remove all three copies of Etched Champion.

If they have sweeper effects, you can cut two one-drops—usually Memnite. It’s possible that you can also cut one Signal Pest, but I suspect this is generally bad.

If life gain and evasion are both irrelevant, you can cut copies of Vault Skirge, but if this is true then you can’t cut Memnite.

Galvanic Blast can freely leave the deck if there’s nothing worth killing with it.

Finally, you can trim a copy or in extreme cases two copies of Thoughtcast if you are bringing in a bunch of colored cards and therefore don’t want to overtax your mana / you won’t be playing out as many artifacts in the first two turns.

Everything else is untouchable.

These cases are rarely going to come up together, and even if they did there is a cumulative effect to cutting artifacts that cost zero and one mana such that cutting more than four of them is unthinkable and even four requires that the cards you’re cutting be very poor in context and that the number of two casting cost artifacts be going up rather than down. Think of cutting Etched Champion or Galvanic Blast as a “free action” since those cards don’t fuel the deck’s engine, but if those cards don’t leave you’re not bringing in more than four cards.

It’s very easy when sideboarding or when building a sideboard to begin grabbing “good cards” but forget to think about what the decklist will look like after you bring all those cards in. You need a core deck engine that still works, you need the right curve, and you need to not overstress your color requirements. If you don’t do these things, your sideboarding will make the deck worse rather than better. This is very easy to do.

That’s why the sideboard is constructed this way; a third or fourth copy of a colored spell narrows your options and increases the number of things that can go wrong, so such choices must be accompanied by extreme caution. A third Spell Pierce or Ancient Grudge would be high enough impact and reliably useful enough to be considered, alongside the third copy of Galvanic Blast, and of course there is Paul’s single copy of Relic of Progenitus.

The third copy of such spells is much worse than the second copy since drawing two can be absurdly awkward, but in the right context Spell Pierce and Ancient Grudge are so valuable that it’s still the best thing you can do. Thoughtseize can be excellent, but the sequencing on it can be very bad and where you want it you’ll usually also want Spell Pierce, so a third copy would be a trap.

When choosing what to bring in, remember this principle of keeping your touch light. There are four copies of Spellskite for two reasons: 1) where you want it even the second copy can sometimes be invaluable since it’s soaking up every answer they have and 2) because as an artifact and a creature it does far less damage to the engine to bring in Spellskite than to bring in your other sideboard cards.

Thoughtseize is a great card—one of the best in Modern—but it should only come in when it is extraordinarily good, which means opponents that are trying to beat you with one key card or a combination of cards for which they are unlikely to have a backup. Thoughtseize needs to be backbreaking. Spell Pierce is similar. If the entire game isn’t going to come down to whether one particular spell resolves, leave Spell Pierce in the sideboard.

Ancient Grudge doesn’t go in against everyone with artifacts; they have to have a ton of artifacts. If it’s only brought in for the mirror, it’s provided most of its value. Bringing in Blood Moon should be done even more reluctantly. If that card doesn’t outright end the game, don’t bother. Shutting down most of your opponent’s lands may sound like it’s worth doing, but for that much mana it’s not. Kill them instead. Only bring in Blood Moon when they have no way to handle it whatsoever or are completely disrespecting it and will be caught.

When playing Affinity, every game is a math puzzle. What’s the right sequence of plays? What can I afford to play around? The majority of the time there will be a clear correct answer, but often there are multiple reasonable options and you must choose. The best heuristic when choosing is to decide how good a hand you can beat. Playing to beat a hand of nothing as efficiently as possible doesn’t make sense unless your hand is incapable of beating anything, but their best draws will usually be unbeatable so it’s not worth trying.

Your goal is to do enough to reliably win if unanswered while preserving a solid game if they can answer. While leading with Arcbound Ravager and a place to put the Ravager’s counters greatly decreases the number of bad things that can happen to you and puts them in much greater danger of either falling behind or dying, actually moving in on Ravager for more than one or two is usually either an act of desperation or a case where the cards being sacrificed no longer matter. That doesn’t mean doing so is wrong, however, since games that go long rapidly become desperate.

Once you decide how to proceed, it is on your opponent to figure out a way to beat what you have. This was a familiar position for me to be in, as many of my Modern deck designs are about reliably posing difficult questions and daring my opponents to have quality responses. Our strategy here requires us to play a handful of key cards that represent the bulk of our pressure and to expose three or more cards to mass removal in order to properly exert pressure. If they answer it properly, either directly or with something more powerful, they win, and doing more than the deck can afford to do in order to try to stop them will only make things worse.

The Grand Prix Itself

Affinity turned out to be a great choice for Grand Prix Portland. The ban on Eggs caused people to greatly reduce the amount of artifact removal and Stony Silence that they had in their sideboards, with the Birthing Pod lists in the Top 8 going as far as to give up sparing one sideboard slot of Kataki, War’s Wage. That overadjustment won’t last in the wake of Portland, as Affinity is a tier 1 deck worthy of serious sideboard hate and the hate is very good. The Birthing Pod matchup plays completely differently if you are forced to always worry about Kataki, and if a Stony Silence hits the table, your deck becomes quite poor. Neither of these things happened in Portland.

What did happen in Portland was that my opponents took many mulligans and I drew many copies of Cranial Plating. The luck of the draw smiled upon me all weekend, and I only lost one game in the Swiss, conceding in round 9 to Paul Rietzl in the 74-card mirror to thank him for building such a great deck and to help a teammate on his quest for Platinum—and so we could go talk Block for an hour since when preparing for a Pro Tour time is short and every minute counts. I made some mistakes, but overall I am very happy with the quality of my play. I expected myself to spew far more value than I did given my level of experience with the format. I also had a ton of fun the whole time. Here are my observations on the various matchups I faced during the tournament:

Living End

I faced Living End twice. This deck is very good at having it—if by it we mean Living End on turn 3 or 4—but not so great at making sure what they get back is decisive. It’s important to put them to the test quickly so they don’t have time to build up a graveyard, which means trying for the fourth-turn kill or close enough to it that they risk dying to Galvanic Blast if you can. Sequencing your lands carefully is very important because you’ll want to preserve the Nexus that matters for after the first Living End to steal wins.

Your key card is Arcbound Ravager since Ravager allows you to sacrifice all your creatures and get them back after the Living End, which will often allow you to keep the advantage, especially if you are working Cranial Plating or they didn’t draw enough reach. Their deck does one thing and does it well, so you want to mulligan aggressively to find a hand that can beat what they are going to do.

Sideboarding means bringing in Thoughtseize and Spell Pierce since taking away their ability to Living End is game over. Grafdigger’s Cage doesn’t work on Living End, so don’t use that, but do use Relic of Progenitus if you have it.

Melira Pod

I also faced Pod several times, including Matt Nass twice; I almost punted a game to him in the Swiss by miscounting and letting him go to 1, but luckily he couldn’t take advantage. Always look to see how close they are to winning outright and whether they are threatening Kataki, War’s Wage after sideboarding since both require keeping Galvanic Blast available unless you are far enough behind schedule that trying to hold up the burn means you lose anyway. Abrupt Decay has become common for them, so if they’re representing it don’t walk Cranial Plating into its crosshairs.

Evasion is the key to beating them. They can stall the ground easily, so Memnite is rather useless, but the skies are hard for them to contest, so your plan is to use that to force them to win before they’re ready to do so or to spend their cards on removal instead of assembling their combination, which usually leads to victory if your draw is good. Holding up Galvanic Blast is wise whenever you can, and Whipflare is great but you shouldn’t get too greedy with it.

Splinter Twin

Your maindeck has few ways to interact with them, so if they have it (the combination with backup and some amount of interference to bridge to it) you’re going to lose. The good news is that your kill is far more reliable than theirs so often they won’t have it on time, and it’s vital that you not give them extra time to find it.

Sideboarding here allows you to bring in Spellskite, Spell Pierce, and Thoughtseize along with a third Galvanic Blast, which allows you to interact with them, and you can buy this largely by taking out Vault Skirge but also need to reduce the high end and trim Thoughtcast to avoid wrecking the deck. The bad news is that they will often get Ancient Grudge in exchange, which is very bad for you, and Engineered Explosives, so be careful about overcommitting cards that cost two mana to the board, especially if multiple Spellskites are involved.


There is more than one Scapeshift deck, as evidenced by the lists in the Top 8. The traditional build plays classic control cards like Cryptic Command and will try to use removal to disrupt you, which makes bringing in some copies of Spellskite attractive as a way to protect Cranial Plating—your best card by far.

Against the version I lost to in the semifinals, partly due to what on a lot of reflection was taking the wrong percentage play (trying to deny him business spells rather than trying to deny him mana despite his mana being crazy good and his business spell count being quite low), you’re facing only mass removal but no pinpoint removal, so there’s nothing to protect with Spellskite, but you’re facing a faster clock and spewing on to the board walks right into his Pyroclasms. Again, it’s important not to get too greedy. If they have it, they will beat you, so make having it as difficult as possible for them rather than trying to win unwinnable scenarios.

Jund & B/G/W

They have lots of powerful individual cards that make you cry but have trouble dealing with your best stuff. Long term they will grind you out, so you must apply early pressure. This is one place where aggressively giving away cards to do damage but balancing that with preserving your ability to play in future turns is most important, as is keeping in mind when you will die to their army. Making a Vault Skirge large to gain a lot of life can often be the difference.

Both of my matches against them involved lots of complicated lines for both sides, which led to my unintentional draw when we both spent a long time trying to solve a game that played like a chess problem. Evasion is the key to success, as you can then move in with Arcbound Ravager or Cranial Plating, and Lingering Souls is your worst nightmare. Without Lingering Souls, it’s very hard to lose provided you have a good draw, but if they do get out Lingering Souls, you either need to stick a Steel Overseer to overpower the blockers or go around them with Etched Champion. Whipflare can also clear a path.

Going Forward

I would happily play an essentially unchanged list if I got to play Portland again. Going forward, more hate will make the deck not quite as good, but it would be a mistake to oversideboard or change what you’re doing too much in order to combat a few problematic cards. There’s no good answer available for Ancient Grudge other than to battle through it. Stony Silence is an argument for bringing in Thoughtseize more aggressively or holding up Spell Pierce early, but if you need to do those things in the dark against unknown white opponents, it’s probably time to choose a different deck.