Affinity is one of the best archetypes in competitive Modern tournaments. It consistently performs at the highest levels of tournament play, including a top 8 finish at the last Modern Pro Tour, Born of the Gods in Valencia. Results after the Pro Tour include a 2nd place and 3rd place finish at Grand Prix Richmond, the largest Modern event ever. In Magic Online swiss-format tournaments, Affinity is the deck most frequently winning packs, surpassing Splinter Twin combo and Birthing Pod decks.

The highly aggressive and proactive nature of Affinity means it has game against everybody. Affinity has the tools to outrace combo decks, outmaneuver the aggressive decks, and outlast the control decks. With a price point well below that of the other top-tier decks in the format, Affinity also provides some of the best value of any archetype in Modern.

Affinity can be a tricky deck to play, but with some experience the deck can become second nature. Even for the most experienced players it always provides plenty of decisions and lends itself to fun, interactive gameplay.

The skeleton of Affinity decks is relatively static, but there are many options for the last few slots. There are different ways to take the deck, but the list I advocate right now is this build, which finished third at Grand Prix Richmond, along with an updated sideboard:

This list favors consistency and solid opening hands. Note the extra land and reduced colored spell count compared to the 2nd-place list from GP Richmond. I tend to view Affinity as a combo deck where the choice to mulligan is the single most important decision you will make all game.  Thus, value anything that makes the opening hands more consistently explosive.

When it comes to sideboarding, historically, Affinity tends to be conservative. Affinity operates at peak performance through synergy, and it’s important that the pieces come together properly. Removing these pieces can ruin the deck on a fundamental level. Much like combo decks or Dredge decks, Affinity is driven by an engine of core cards that are essential to its proper functioning. While certain cards may not seem essential, removing them can be detrimental to the proper and consistent functioning of the deck as a whole.

Springleaf Drum and Mox Opal are essentially lands, and they are necessary if the deck is to function properly. They are also quite important post-sideboard as they enable the use of other colors of cards. Lands also join this category, and should not be boarded out in any normal circumstances.

Arcbound Ravager and Cranial Plating are the power cards of the deck, and they make everything else worthwhile. These cards allow the underpowered, rag-tag team of artifacts to compete with all of the other overpowered things going on in Modern. The rest of the deck exists to enable these cards, so planning on cutting them is simply bad strategy.

Make sure you always have some combination of at least 8-12 of these cards at all times.  Aim for 12 when at all possible.

Ornithopter and Vault Skirge are the basic cog creatures of the deck that help enable all of the other cards. Signal Pest and Memnite fall in the same category, but generally they are less useful and liable to be cut. The artifact mana needs other artifacts to turn it on, Thoughtcast has the Affinity ability, Cranial Plating needs a body, etc. While these cards may seem extremely weak in the face of sideboard hate, they actually can be useful for beating hate cards. A rush of threats is the best way to beat Stony Silence or Ancient Grudge, and these cheap cards contribute to the critical mass of aggression.

With all of the mentioned cards off of the table, that leaves only 19 potential cards to ever take out when sideboarding. For this list, that leaves us with the following cards as options:

The colored spells, Galvanic Blast and Thoughtcast. The former will sometimes be ineffective, the latter too slow. Sometimes they are cut when colored spells are brought in to balance the mana requirements.

Etched Champion and Master of Etherium, the 2-2 split being a maindeck hedge against the format, one tends to come out when the other stays.



Steel Overseer, which is weak in the face of the most hateful cards.

Signal Pest, which comes out against fliers like Lingering Souls. The card is also weak on its own and weak to sweepers, which is the reason to sometimes cut Memnite.

Finally, Welding Jar is non-essential to the deck and the least consistently-useful cog.

Here’s how I’d plan to board against the most popular decks in the field:

Affinity

Spellskite is great for redirecting removal spells, and it can even target a Modular trigger from Arcbound Ravager to effectively counter it. Dismember can really catch them by surprise, so try to save it for a blowout if possible. Keep in mind that the opponent is likely to have Ancient Grudge of their own, and to use your own very judiciously.

Splinter Twin

Against RUG Twin with Tarmogoyf, I would bring in both Relic of Progenitus and leave Spell Pierce and Torpor Orb on the sidelines.

The opponent will be less likely to win using the combo, so be prepared to play against a control deck post-sideboard. Cutting the weakest threats for more resilient cards and a disruption suite is effective against their control strategy. Steel Overseer goes because it plays poorly against hate and Anger of the Gods.

Birthing Pod

You must respect them as a combo deck, but they can also grind you out. Cut the weakest cards for a ton of disruption in an attempt to shut down their engines.

Rock/Jund

Relic of Progenitus shuts off their biggest threat, Tarmogoyf, which without their offense can become anemic. Spellskite does work blocking and controls their removal spells like Lightning Bolt.

UWR Control

Stony Silence is the biggest concern post sideboard, so bringing in a disruption suite gives you some counterplay. Steel Overseer comes out because it is weak against their hate and against their board sweepers. Relic of Progenitus contains their Snapcaster Mage at little cost to you.

Turn to the comments if you have any other questions!

-Adam