Thursday, March 11, 2010

the 95% rule

Quick thought on general combo deck construction.

When building or tuning any combo deck, especially engine-based ones where you draw your entire library, all you need to do is make sure that when you fully combo off, whether it's with Mind's Desire, Ad Nauseam, Glimpse of Nature, or anything else, you have a 95% chance of winning the game against a random opponent. This means that you shouldn't play otherwise useless cards that get your 95% to 99% or 100% in an effort to be super sure.

Last season, GerryT wrote an article on that incarnation of Elves that included one Mirror Entity as the kill, with two Thoughtseize main, and nothing such as Grapeshot, Eternal Witness, or the like. The reasoning is that if all your creatures came into play that turn so you can't kill them with Mirror Entity, or they killed it earlier or something, you can just put your deck on the table and Thoughtseize them a couple times to get rid of any way of them possibly killing you. Sure, it's theoretically possible that they have the perfect combination of cards to kill you when you have 50 power on the table after losing the two best cards in their hand, but it's incredibly unlikely, and you'll lose more games to playing do-nothings like Grapeshot.

Sometimes, this'll lead you to realizing that you don't even need the combo at all. When people first started playing Reveillark, it was with an infinite combo that did a bunch of crazy stuff with Body Double and Mirror Entity. When the deck started getting more popular, they realized that they would get that 95% win just by having good cards in play, not through an infinite combo. It was superfluous, so it got cut.

The same applies to this season's version: some people insisted on Grapeshot just to be super-duper-sure their opponent actually died when you combo'd, instead of playing a card with relevant text outside the combo, like Primal Command. This is bad. Don't do it.

By throwing in bad one-ofs to try to go from a 95% chance of killing to 100%, you're giving in to The Fear. You'll lose more to drawing dead cards than to whatever crazy situation you think of.

Wednesday, March 3, 2010

extended banned list

Current Extended banned list consists of:

Sensei's Divining Top
Aether Vial
Disciple of the Vault

SDT and Skullclamp obviously need to remain banned; the former simply for time reasons, the latter because have you read the card seriously just look at it. The latter two, though, seem like they don't need to be banned any more. First, let's look at the reasons for banning them in the first place:

"Last season's Extended metagame was pretty amazing. There were lots and lots of decks—all of them very powerful and very fast. Of course, with the format rotating in October, we lose a lot of that diversity.

The fact that nothing approximating a normal control deck was viable in the format was not a huge problem considering the wide range of options. But with half the decks rotating out, we had to take a serious look at starting the “new” format off in a place more conducive to a traditional “rock-paper-scissors” metagame.

What we didn't want to do was take good decks and blow them to smithereens like we did with Affinity in Standard three months ago. Instead we opted for the “surgical strike” option on some of the more resilient aggressive decks.

Aether Vial is a pretty ridiculous card. It costs but one mana, and it makes all of your creatures essentially uncounterable free instants. It messes up permission, combat, you name it. Basically it's like a super Dark Ritual that gives you 17 free mana over the course of a game. Somehow it has survived getting the axe in other formats, but knowing how much of a head start it gives Goblins and Affinity in this format—and how much it cripples control—we can't justify leaving it around.

Disciple of the Vault was a less obvious candidate, but in general we really dislike the “Oops, I win” draws that Ravager Affinity is capable of, especially since there isn't much anyone can do to stop them once they start going for the Disciple combo-kill. We certainly don't have a policy of banning cards preemptively—Disciple has proven itself again and again. Affinity took the crown at the last Extended Pro Tour using a card pool that was significantly bigger than what will be available in LA this fall; it probably won't even need Disciple to remain competitive.

Will there still be a good Goblin deck? Absolutely. Affinity deck? Yes, hopefully with some new twists. Pro Tour LA should be very interesting—and it needs to be. It has large shoes to fill based on the popularity of the last Extended season."

Goblins obviously isn't a concern since Onslaught rotated out, so let's just talk about Affinity. First, if you look at Affinity in Legacy, where both Disciple and Vial are legal, in an environment with a lot of countermagic (making Vial a lot better), absolutely no one plays Aether Vial. No one. In fact, it lists the top twenty cards most often played in it, and Vial doesn't even make that list, while Relic of Progenitus does. This says to me, rather clearly, that if it were unbanned, Affinity wouldn't even want it. To be blunt, Springleaf Drum is just a better card at doing the same thing, since you don't have to wait a turn. Zoo, same thing: the Legacy versions play about 90% Extended-legal spells, and they don't play Vial either, ever, so it's pretty easy to assume they wouldn't want it. Who would? Well, the most common deck that uses it in Legacy is Merfolk... a deck known for how it gets crushed by the format's Zoo. I'm also unsure how synergystic blue men could possibly compete with indestructible 20/20s on turn two... Unsummon, I guess? Doesn't seem like a solid plan. Even if it does enable a Merfolk deck or some better version of White Weenie, is anyone really not okay with this? Aggro decks that aren't Zoo have a very tough time in the format currently, and Vial would make them better, but not absurd, especially since almost no one plays real countermagic (one of the main reasons Vial is a useful card in Legacy).

Now on to Disciple of the Vault. To put it simply: Affinity sucks in Extended right now. It's not like last season where you needed some Kataki or Ancient Grudge to beat it; no, it just loses to most decks even with zero dedicated hate cards. Again, synergistic creatures have a pretty hard time dealing with 20/20s, turn one 3/3s, or turn three kills. Disciple, at its best, gives your Ravager the additional ability of making them lose one life; this would make the deck better, but certainly not great, in the format.

Affinity: "turn one Frogmite, turn two Ravager and Disciple! Oh what now!!! You thought it was banned BUT IT ISN'T"
Dark Depths/Elves/Hypergenesis/Living End: "oh I remember that card. Cool. Kill you."
Affinity: "oh. Too bad my deck still plays no relevant disruption and does nothing other than play undercosted or free creatures."

Worst-case, if I'm wrong about how good it is, it'll be another deck and people will have to go back to actually sideboarding against the deck. Doesn't seem like a bad result.

Monday, March 1, 2010

one game

In the spirit of one of the all-time best Magic articles, One Game by Richard Feldman, I'm going to post a fairly detailed walkthrough of a single game with my latest version of the Elves deck in the third game of the third round of a Magic Online daily event. My opponent is playing a fairly interesting UW Thopter list, deviating most noticeably from most Thopter lists in playing Tron backed up by various signets and talismans. Talismen? Notable cards seen before include Path to Exile, Repeal, Wrath of God, Day of Judgment, Hallowed Burial (?!), Mindslaver, Engineered Explosives, Academy Ruins, and Chalice of the Void. I did not see what his actual win condition was, since I scooped to Mindslaver recursion in the first game. In the second, the game went long, so I had to actually recur Primal Commands for once to keep him off Wrath mana, leaving me fairly little time for the third game. Worse yet, losing the first game means I'm on the draw for the decider... never a good thing for this deck; it shifts almost every matchup to be much worse than when on the play. For sideboarding, I brought all three Shamans- two Viridan, one Loaming- as well as a single Ranger of Eos, removing Cloudstone Curios and a Boreal Druid. The thinking here is that Academy Ruins recursion plus shitloads of mana is something that I just can't beat, so a 2G 3/2 that shuffles their library is pretty good. I'm unsure whether or not I want the third Ranger of Eos (I cut down to two sideboard, one main), but decide against it, since with two Primal Command and Regal Force in the deck, it could lead to some clunky draws.

Opening seven:

Five land, ouch. Glimpse plus Nettle is certainly tempting, but an easy way to think of lands past the third in a hand is that they're mulligans already, making this hand a "double-mulligan" and a toss. Yes, it has your two most critical cards (Nettle for early pressure and Glimpse to combo if they tap out), but the supporting cast isn't there.

Well hello there. Pact, when backed up with Elves in the opening hand, is only slightly worse than having Glimpse. Fist-pump keep.

More fist-pumping occurs. The only decision here is whether to lead with the Forest or the fetch, and whether to get a basic or a Temple Garden. Since his only spot removal seems to be Path and Repeal, there's not much advantage to try anything fancy with tricking him at the expense of life, so I just go with the Forest, in case I draw one of the two Rangers eventually.

Well that's not what I wanted to see, but it's not a huge deal. I have the Archdruid, and can Pact up Shaman when it's necessary to do so. I decide to fetch up a basic again, since the odds of drawing a Ranger before another white source are rather low; however, life total barely matters at all here, so it was a misplay. I should have gone for the Garden.

Boo land. Pacting for Shaman is briefly tempting, but it's possible he's slow-rolling the second white source (or that he'll draw it) in order to Wrath me for an easy game win. Instead, I begin bringing beats. He takes the four and Thirsts at end of turn, ditching another Chalice, then completes the Tron, plays a signet, and Repeals Archdruid. This isn't looking good.

Two lines of play here: either Shaman his Chalice and lay the Nettle, threatening eight damage with the Archdruid I play next turn, or play Archdruid this turn and intend to go crazy on the next. The latter seems much better, since he can just get blindsided by losing his chalice and seeing my board explode out of nowhere, not letting him untap with the knowledge that I've dealt with Chalice.

Speaking of exploding out of nowhere, he plays Triskelion (?!?!?!) and the turn I was setting up disappears. Thankfully, I've practiced enough with the deck to draw another Archdruid.

If I don't commit in a major way to the board, Triskelion recursion with the new Foundry is going to knock me out of the game in a hurry. For that reason, it seems best to play out everything I can and hope to overwhelm him.

I play Heritage Druid as the second elf rather than Nettle Sentinel, since I'd rather he Path the 1/1 than the 2/2 Nettle. Interestingly, he opts to Path Viridian Shaman rather than Heritage Druid, I assume for a similar reason. Here is where I've started rushing my decisions due to the limited time on my clock, so I can't fully calculate what he'll do with Triskelion turn after turn. I don't even think I saw the Foundry-Ruins interaction here that allows him to do that at all; I was overly focused on my side of the board, and it cost me.

Pacting up Shaman doesn't do anything here, especially with Chalice in the graveyard ready to go on top of his library. I could get Loaming Shaman to stop that from happening, but that eats up two turns and doesn't do anything about the Triskelion. The best way out of this seems to be to Ranger up Warcaller to potentially get enough guys outside of Triskelion range so that I can kill him soon, after Pacting for Shaman to deal with the Chalice he'll play.

As expected, there's the xx, and much like the band, it's pissing me off a bit, but can get tuned out temporarily before it gets brought back up. The first Pact obviously has to get VShaman, then I can play out my hand... but what does the second get? I reason that Archdruid is the best call, but this is wrong. If I had ended this turn by fetching up Loaming Shaman instead, I can stop his crazy Triskelion recursion and make him have one of a very small number of removal spells either as the last card in his hand or on top of the deck to win the game. If he does, he wins on the spot, but it still seems superior. I should have realized that, by playing Archdruid, he can just kill it with Trisk, so if I attack with Nettle Sentinel into his four untapped Thopters, then I can only tap three Elves plus four land and lose. Fortunately...

Well, that's a relief.

I attack with everyone, and make a huge mistake by holding on to the Llanowar in my hand until next turn for... some reason.

Note his five life, my four power worth of unblocked creatures, and summoning-sick Llanowar.

The deck attempts to apologize by handing me a win in the form of Visionary into Pact, but again I politely refuse, this time by searching for Archdruid. Trying to play too fast was getting to me.

...and I lose and drop from the tournament.

Short version: "he had turn two Chalice, obviously drew the Urzatron and Academy with no tutoring, and started recurring Triskelion, and I don't get a single Glimpse the entire game, so obviously I lose. Nothing I could do about that."

Next time I do this, it'll be a game I played well.