Lessons from a Bad Beat with Storm

Storm is one of my favorite decks to play in Legacy, and between the Gitaxian Probes and the Duress effects, you often work with perfect or nearly perfect information about the contents of your opponent's hand. The kills can often be accomplished with no variance whatsoever once you've identified what you need to play around, and for this reason, I find it extremely interesting to revisit interesting situations with the Storm deck. Developing the right heuristics for situations you may see again in the future is especially rewarding because it's important to identify the optimal line as quickly as possible in the context of a real game.

I recently played Storm again at the Card Kingdom weekly, the first time I've done so since a successful run at the recent Mirkwood event in which I split in the finals. I often like mixing up what I play at weekly events to expand my range, but as larger events approach (such as the upcoming Mox Boarding House dual land event), I often go back to the decks I feel most likely to play in order to get the necessary practice. I went a disappointing 2-2 at this weekly, but I wanted to discuss a game I played against Infect, where I took a good line that nonetheless the most statistically "bad beat" in my time playing Storm. But in retrospect, there were arguably even better lines, and I think this is an illustration that in Magic, even the smallest of margins (in this case, a fraction of a percent) is relevant, and that it's important not to "settle".

The situation is as follows: I'm up a game against Infect, and my opponent is on the play and has played a Tropical Island and a fetchland without taking any other action on his first two turns. I believe I Pondered on the first turn, then on turn 2, I led off two Gitaxian Probes (paying life).

• His hand: Invigorate, Berserk, Daze, Windswept Heath, Misty Rainforest, Crop Rotation
• His board: Tropical Island, Wooded Foothills
• His life: 20
• My hand: Lotus Petal, Dark Ritual, Cabal Ritual, Cabal Ritual, Infernal Tutor, Brainstorm, Cabal Therapy
• My board: Underground Sea, Misty Rainforest (both untapped), no land drop available for turn
• My graveyard: Misty Rainforest, Ponder, Gitaxian Probe, Gitaxian Probe
• My life: 15
• Storm count:2

Note: I don't remember the exact contents of my hand, but I remember the calculations I went through in the process, so I took some guesses above in order to reconstruct the game state to be consistent with what I remember.

 At this point, I make the following observations:

• I am dead next turn to Crop Rotation into Inkmoth Nexus on my end step if I do not take any actions.
• As I do not have a Lion's Eye Diamond in hand, I may need to use the Cabal Therapy in hand to help myself get hellbent, especially since I do not need to Therapy anything from his hand. The following options assume I use my Therapy to take away my Brainstorm.

1. I am one mana short of killing him by chaining Infernal Tutors into the natural Tendrils for 20 life through the Daze in his hand. Note that this assumes that at least one Cabal Ritual resolves with Threshold, which I can achieve, but which he can disrupt with Crop Rotation into Bojuka Bog if timed correctly, but it would have no effect, as that would add an extra spell to the storm count anyway and allow me to simply cast one fewer Tutor in the chain.*
2. Past in Flames would kill him if he doesn't have Bojuka Bog in his deck, but it would be a dead end if he does. I believe most versions of Infect do run it, and while some opponents may disagree, I think Bojuka Bog is a card that Infect should bring in for this matchup. I recognize my opponent as primarily a Modern player, and given that Modern Storm is even more dependent on the graveyard than ANT, I would expect him to be even more likely than most opponents to bring in Bog.
3. A hellbent Infernal Tutor would allow me to cast Ad Nauseam through Daze from 15 life (after cracking my fetchland) with 2 mana floating and no land drop if he doesn't Bog me to take one of my Cabal Ritual off threshold, or 0 mana if he does. Generating the necessary Storm if I find the cards to go off is trivial, so the extra storm from Crop Rotation is essentially irrelevant. Passing the turn is a possibility if he does Crop Rotate for a Bog, or if I'm able to make him discard a pump spell, as long as he doesn't rip that same pump spell on his draw step or Probe into that card. (Brainstorm or Ponder into the pump spell would not be lethal, as he wouldn't have mana to activate Nexus, swing, while being able to cantrip and play Berserk.)

* Note: I later noticed that it is possible for my opponent to take me off Threshold for both Cabal Rituals if he casts Crop Rotation for Bojuka Bog sufficient early, but it is exceedingly unlikely he would commit to such a play, as that takes away his clock and his best piece of disruption, unless I show him a Past in Flames, or at the very least, a Cabal Ritual. By casting Dark Ritual, followed by Lotus Petal, then cracking both the Petal and the fetchland in play, I will have achieved Threshold without having shown him any cards that are dependent on the graveyard, while having up to 5 mana available. This will allow me to put two Cabal Rituals on the stack if necessary (Cabal Ritual, opponent responds with Crop Rotation, I respond with another Cabal Ritual), while being able to play around Daze the whole line. I will assume going forward that the opponent would not take such an aggressive line, since it would trade his Crop Rotation (which represents his clock) and Bojuka Bog for my Dark Ritual and Lotus Petal, leaving me still with five cards in hand (unknown from his perspective, but two Cabal Rituals, Cabal Therapy, Brainstorm, and Infernal Tutor).

Of these options, the Ad Nauseam plan seems the most attractive, as it likely allows me to win or safely pass the turn, although it's not completely risk free. Since some of my friends were curious, I would guess I'm somewhere between 70-90% to win if I cast Ad Nauseam, depending on whether he takes me off threshold for Cabal Ritual; this is an estimate based with no calculations performed whatsoever, and based only on "gut-feeling" from my history of flipping cards to Ad Nauseam.

I decide on a better line (option 4). As I am one mana short of killing him with a Tutor chain through Daze, I can simply use one mana to cast the Brainstorm in hand to get an extra storm, Therapying anything left stranded in my hand if I can't cast it. If I manage to hit a ritual effect, an artifact, a discard spell, or a Probe, I can even further increase storm to allow for Tendrils for 22-24, thereby play around anything as obscure as Crop Rotation into Radiant Fountain.

Unfortunately, my Brainstorm hit three lands, leaving me without a way to get my Tutor hellbent. I cast Cabal Therapy on him instead, and he cast Crop Rotation for Nexus, so I took his Invigorate. I was able to draw a non-land spell on my next turn, and I attempted to go off, but he drew the Force of Will as his unknown card from his draw step, and I was dead shortly thereafter.

How likely was this to happen? I boarded out a land, so I had 14 lands in my deck, of which three were already either in play or in the graveyard, so 11 copies left in deck. There were at the time 49 cards in my library, so the probability of my Brainstorm hitting three lands (the only possibility that leaves me without being able to Therapy anything) is C(11, 3)/C(47, 3) was 1.02%. In other words, I was about 99% to kill him on my turn 2 and probably still likely to kill him on my turn 3 if he didn't kill me first or find a Force of Will.

I am still convinced I was correct to Brainstorm instead of taking one of the three lines above, but after thinking about the situation in retrospect, there are some things I could have done differently:

• First of all, after my Brainstorm failed, my Therapy targeting him should have taken Berserk instead of Invigorate. He likely has three more copies of Invigorate in his deck, whereas he is unlikely to have more than one copy of Berserk. No combination of pump other than Berserk + Invigorate (besides the exceedingly unlikely Brainstorm into a second and third Invigorate) would kill me on the next turn, and there was almost no way I wasn't going to go for it right away if I got another turn.
• I could have fetched before casting Brainstorm (option 5) to leave me with only C(10, 3)/C(48, 3), or a 0.79% chance of whiffing, which is about 22.3% less chance of whiffing than if I had Brainstormed without fetching first. However, if I took this line and still hit three lands, but I'd also be unable to get hellbent on my next turn, which would leave him with two draw steps to find the kill or a Force. Note that this means that an Invigorate or a Vines is also a kill in this situation, in addition to Brainstorm into one of the above, regardless of whether I took his Invigorate or Berserk. By taking option 4 (Brainstorming then fetching), as long as I avoid drawing land on my draw step (9/45, or 20%), I have the ability to try again next turn if I get one.
• There may have been an option 6 available to me: since the Brainstorm I cast was only there to generate storm and would actually have been better if it didn't resolve, I could have played it after casting Probe but before playing my land for turn to hopefully induce Daze. If he does so, I easily win with a Tutor chain into Tendrils (basically option 1), even beating Radiant Fountain if that's somehow in his deck. The only snag is I don't remember the exact contents of my hand after drawing for turn 2 (in other words, I'm not sure what my hand was before I cast Gitaxian Probe). It may have been important for me to resolve the Probe through Spell Pierce, which means that I might have already made my land drop before I cast Probe to see the contents of his hand.

Hopefully, this was interesting to you. The detail into which I went may border on neurotic overanalysis, but as a Magic player, I think it's important to maximize even the smallest margins because the extreme corner cases do come up. This is especially so for a Storm player, since we frequently work with perfect information, know exactly what combination of cards are necessary when going off, and need to make such calculations relatively quickly within the context of a tournament setting.