As I watched Worlds over the weekend, there was a clear storyline going into the top eight—ChannelFireball had taken four of eight slots and guaranteed that the player of the year would be one of four (!) of their team members. Impressive, right?
By the time the semifinals started, there was one CFBer left. He was out by the finals. Was this a tale of pairing variance swinging away from Americans, or can we learn something more from this? First, let’s look at the Standard portion records of the four ChannelFireball players who made top eight, all of whom played Tempered Steel:
Conley Woods: 6-0
Luis Scott-Vargas: 5-1
Josh Utter-Leyton: 5-1
Not bad, right? All of them looked pretty good in the early Swiss rounds. So why did they all seem to do worse in the top eight? Patrick Chapin had a fairly quick (if overly simple) answer:
On its face, this is entirely true. Worlds Top 8 is best of five, so sideboarded games matter more. Premier Event top eights are run with full information, meaning that your opponent will get to see your decklist before you shuffle up. But what does this mean?
If you examine this a little more, you start asking better questions and coming up with more interesting answers. After all, what Patrick is telling us—albeit not in as many words—is that there are decks that get worse in a top eight and that Tempered Steel is one of them. So let’s start asking better questions:
Why does Tempered Steel get worse in known information situations and/or situations where sideboarded games matter more?
Is this a theoretically based circumstance? That is, are there attributes of this situation that can be applied elsewhere and still be true and useful, or is this a totally unique situation?
If it is a theoretically based question, how can we apply that knowledge? Are there other circumstances that closely resemble this one in which we’re more likely to participate?
And now, let’s answer our questions:
Tempered Steel is a very strong game one deck. It has a powerful, linear strategy (cheap and/or evasive artifact creatures plus power amplification), so it will perform at a consistent level of power in game ones. It expects to win a lot of game ones.
It gets worse in known-information situations for a bunch of reasons. First of all, in the Top 8, opponents know exactly what’s going on. They aren’t going to get a bad scout or make a guess about what deck their opponent is playing, keep a slow hand, and get crushed. They’re going to mulligan the hands that can’t beat Tempered Steel. This hurts Steel to a degree, but not as much as you might think.
What hurts it more is that a deck with as large a core component as Tempered Steel is not going to have as much room to improve as other decks. What do I mean by “core component?” Well, let’s look at the deck:
Creatures (21)
Lands (10)
Spells (29)
In his article yesterday, Luis stated that the core of the deck—the cards that were never under consideration as possible cuts—were:
4 Memnite
4 Signal Pest
4 Vault Skirge
4 Glint Hawk
4 Glint Hawk Idol
4 Tempered Steel
4 Dispatch
2 Origin Spellbomb
4 Mox Opal
19 lands
So Channel locked in all but seven spells and fifteen sideboard cards at the beginning of their process. The upshot is that if you have 30 spells that you always, always want in your deck at the beginning of your first game, how much can your deck improve for games two and three (and four and five)? Far less than the deck that only locks in the following:
4 Primeval Titan
4 Rampant Growth
2 Green Sun’s Zenith
1 Birds of Paradise
1 Thrun, the Last Troll
Tempered Steel doesn’t have a ton of flexible slots. It can’t just cut a bunch of its beats and bring in some great anti-Titan strategy that totally sidesteps sixes and sweepers. It can’t not get wrecked by Ancient Grudge and Manic Vandal. As a powerful linear strategy, there will occasionally be cards that are very good against exactly what it’s try to do. Ancient Grudge and Arc Trail and Manic Vandal are very good at stopping metalcraft from happening, for instance. In a best-of-five match, your best game—game one—is worth a third of a match win, not half. You’re going to have to beat Manic Vandal, Ancient Grudge, Naturalize, Arc Trail, and whatever else people have for at least two games, if not the full three.
Adapting to those cards isn’t impossible, though. Channel had Heroes and Spellskites to create a way around a ton of Ancient Grudges or Gut Shots or what-have-you. That, though, is where the “known information” half of the problem comes in. If people know that you have a Plan B sideboard, it gets a lot worse. Hero of Bladehold is one such example, although Legacy takes it a step further with full transforms, typically involving either Tarmogoyfs out of a combo sideboard or Show and Tell + Emrakul, the Aeons Torn out of Painter or Hive Mind, both of which are combo decks that play Intuition.
The reason why Plan B sideboards are good in the Swiss is easy to understand: they have Arc Trail and Incinerate versus your Hero of Bladehold and Spellskite instead of your Memnite and Glint Hawk, so you get to crack with Hero. Once that happens, they’re probably going to die. Hurray for sideboards, right? Well, sort of.
If your opponent has a [semi-]transformative sideboard that would blow someone out, it’s very effective in a best-of-three, hidden information match. In a match where hidden information doesn’t exist, it’s less effective. Even if it wins one game, you still aren’t out of the woods, since you have to win at least one more sideboarded game. This would never be a problem in the Swiss!
So, to recap: Tempered Steel is a powerful linear deck—as distinct from “aggressive deck”—that doesn’t gain as much from sideboarding as other decks; it wants to win a lot of game ones and can bank on the occasional loose game-one keep from an unsuspecting opponent in the Swiss; and its Hero of Bladehold/Spellskite plan is much more valuable when they don’t see it coming and when they don’t have another game to build their deck correctly.
How is this useful to Legacy players? Well, there are a ton of powerful linears in Legacy. Having trouble thinking of any? Here’s a reminder:
- Artifacts (Affinity, Painter/Grindstone, Ancient Tomb/City of Traitors)
- Tribal synergy (Goblins, Merfolk, Elves)
- Graveyard (Dredge, Reanimator)
- Storm (Five-color storm, Ari Lax U/B Tendrils, disregarding me)
- Nugs (c’mon)
There are more, but that’s off the top of my head. But how does this apply to Legacy? Well, as it turns out, there’s a rather commonplace known-information situation that has come up more Sunday evenings than not in this year: the top eight of an SCG Legacy Open. In the elimination rounds of each of these Legacy tournaments, players get three minutes to look at their opponent’s decklist. Matches are still best-of-three, but if you’re facing down a turn-one Putrid Imp, you’ll know. If you’re going to have to evaluate your opening hand based on whether it can beat a Goblin Lackey or an Aether Vial, you’ll know.
Linear decks aren’t going to do a whole lot with that information. You keep your Aether Vial; you keep your Putrid Imp; you keep your Painter + Intuition regardless of opponent. The thing is, you aren’t going to get a free win from playing your Imp on turn one and watching your opponent frown, slump in their chair, and play a Volcanic Island and pass, tipping their Spell Snare + Stifle hand all too well. Every deck that makes decisions is going to have more information with which to make those decisions, and (presumably) that means that they’ll make better decisions. This means that if you’re on the decision-light end of the matchup, you’re less favored than you otherwise would be. And what’s the most decision-intensive card in all of Legacy? Brainstorm, of course.
Do you realize how sick it is to play an answer-oriented deck and never make a speculative keep? No wonder the Brainstorm decks do well in top eights: they get perfect information! Making keeps with full knowledge of an opponent’s game one capacity is unbelievable. Brainstorm is the best card in the format, for sure, but the way that SCG Open top eights are organized, Brainstorm decks are given further advantages.
It doesn’t end with Brainstorm decks, though. A Knight of the Reliquary deck playing against Reanimator knows that it needs a Knight to get Karakas, so it can afford to mulligan to its key card. The Reanimator deck can look over a Zoo list, check for the presence of Karakas, and instead of getting Jin-Gitaxias with its turn one Entomb, it can get Sphinx of the Steel Wind.
The problem is actually furthered in sideboarding. A lot of anti-hate sideboarding from combo decks doesn’t work as well in a top eight. If you’re playing Painter against Zoo and plan on boarding in Show and Tell + Emrakul, the Aeons Torn, don’t be as surprised when they just happen have Red Elemental Blast or Knight of the Reliquary into Karakas for your backup plan.
So now that we’ve established that linear decks are at a disadvantage in a context with perfect information, let’s ask a few more questions.
Can linear decks change the way they build their decks to compensate for this lost advantage in elimination rounds?
If so, what does that look like? What are a few current deckbuilding models that are inefficient, why are they inefficient, and how can they be improved?
Don’t mistake this article as me saying “don’t play powerful linears.” It is true that the current configuration of SCG elimination rounds favors flexible, midrange strategies that can gain value from perfect information in decision-making, but that’s not a death knell for linears; it just means that linear decks have to polarize their strategies.
What do I mean by “polarize their strategies?” Well, if you’re playing a linear deck, you’re going to occasionally crush people who aren’t prepared for you, right? So if you’re okay with having a few auto-win matchups, you have to accept the existence of an auto-loss or two. “Narrowly beating everything” is the job of midrange decks. After all, while they can theoretically win any game, they can also lose any game. As it turns out, people can make a string of bad judgment calls and throw away a game where they have all the answers. You shouldn’t be looking to build your linear deck in a way that lets you throw away sideboard games by mis-sideboarding or over-sideboarding. I see a lot of linear decks’ sideboards that look like this:
Creatures (23)
- 1 Kira, Great Glass-Spinner
- 4 Lord of Atlantis
- 4 Merrow Reejerey
- 4 Silvergill Adept
- 4 Cursecatcher
- 4 Coralhelm Commander
- 2 Phantasmal Image
Lands (9)
Spells (28)
I have nothing against Andres or his deck, but I disagree with the design philosophy of his sideboard. He’s not playing Brainstorm; he can’t draw extra cards except with Silvergill Adept; and yet he’s playing a sideboard jammed with one- and two-ofs. I get that a lot of his sideboard cards have similar but subtly different utility—Threads of Disloyalty and Sower of Temptation come in against similar decks, but you want Sower against Knight decks more and Threads against Delver and Tarmogoyf more.
Still, is the one Flusterstorm and the one Spell Pierce in the maindeck even good? Andres knew that he would have to beat Red Elemental Blasts, but he only has one Kira, Great Glass-Spinner. Sure, the Flusterstorm is a great story to tell when you blindside someone with it, but what about when you draw it against U/W Stoneforge and they play a Stoneforge Mystic on turn two and it’s not that Daze that would make this a real game? Why not just play the best possible aggressive deck you can?
Why not play four Kira? I get that you have “more options” if you play a 322221111 sideboard, but at some point, don’t you know what you have to beat? This is nothing against Andres, but much more about the school of thought that teaches us to always give ourselves options, no matter what deck we’re playing. The Legacy metagame is a fairly easy read. Merfolk knows that it’s going to see Lavamancer and Red Elemental Blast in every red deck it plays against. Why not just play a bunch of Relics and Kiras and be done with it? Instead of stealing their Tarmogoyf or Knight of the Reliquary, just always threaten to kill it. Why get cute when you could get them dead instead? Why make your Merfolk deck a worse midrange deck than the one you’re fighting?
I see this sort of thinking in a lot in Legacy decklists that should be focusing on redundancy. If you have tutors or a bunch of card selection, sure, play all the one-ofs you want. If you’re going to see 25 cards a game, then sure you can play that random one-of sideboard card that’s a total blowout in two specific matchups. But if you’re a deck with no manipulation? Why would you want to make your deck more inconsistent? They’ll see your cute one-of coming in the top eight, that’s for sure. You aren’t going to fool anyone with your sideboard when they already looked at your decklist.
If you want to play in a tournament with the goal of “doing well,” you can still play anything and sideboard anything. That’s the beauty of Legacy—anything can do well. That trope still holds true, even in an era where Brainstorm looks better than ever. But in terms of winning? Aggressive deckbuilders need to decide what they want to beat.
You can’t decide you want to have “some game” against Affinity and Painter as Merfolk and jam a few Energy Fluxes and Pithing Needles into your deck. If you want to really beat graveyard decks, play all four Relics. If they draw their Ancient Grudge/Nature’s Claim and their perfect hand, so be it. But if you never play against Painter or Affinity and you play against Dredge twice and Reanimator once, you’ll be real glad you cut those Energy Fluxes for the extra Relics.
If you want to beat Lavamancer and REB, why not dedicate yourself to it? Play four Kira. Just always have it and always have it in a relevant time frame. Don’t play cards like Flusterstorm in your maindeck when you aren’t playing your eighth free counter that is going to stop a Stoneforge Mystic from resolving. If you want to beat Batterskull, really care about it. Play something like Stifle, Steel Sabotage, Annul, Silent Departure, or Carry Away. Decide what you want to beat, beat it, and accept the possibility that you might battle Affinity and that, when you do, you’re not favored to win the match.
If you do make the top eight, though, you’ll have a deck that takes a lot of decision value away from an opponent because you will have out-decked your opponent. They’ll board in their Red Elemental Blasts for you like always, but you’ll have a ruthlessly efficient line of attack. They won’t have any decisions to make—either they’ll have it or they won’t. At the end of the day, when you’re playing a linear deck, isn’t that the question you should be asking?
Not “Should I board in my second Umezawa’s Jitte?”, not “I hope I draw one of my two Echoing Truths,” but “You’re dead on board—do you have it, or do I win?”
Good luck, think everything through, and I look forward to your questions.
Until next week,
@drew_levin on Twitter