I encourage you to read this.
Everyone does a “Lessons From…” piece on a tournament, especially when they don’t do all that well, and the tendency is for these to be quite sparse in the way of actual lessons. I aim to avoid that pitfall, but nothing I’m going to say will be succinct. I haven’t written the article yet, but I’m imagining that it’ll be quite long, quite drifting, and altogether broad in its scope. One has a quite understandable tendency to skim through the meandering bits, and as we authors love to ramble this tendency is more than justified. That said, I promise actual content somewhere along the way. No fingers crossed.
The reason for the somewhat apologetic and altogether uncharacteristic preface is that a lot has happened over the last two weeks – both Magically and otherwise – and I’m hurdling along at such a prime velocity that it’s hard to separate events into distinct, quantifiable entities. My memory, like the prose of a Joyce novel, carries more emotive than literal force. Everything sort of folds together, blends like smeared watercolor, and as I set Snow face down on the desk in front of me I find it hard to picture myself sitting mundane and hungry in front of a computer screen keying away at an article. I feel rather like a glass-window observer taking notes, watching a kid blink in the monitor-light and rub his eyes.
By the time you read this I’ll be at a Political Science conference on the mechanics of international law sponsored by Americans for Informed Democracy and held in Washington, D.C. just a short train ride away from Nationals. As I brush up on my multilateral protocols and international agreements, my mind can’t help but wander to the politics of this year’s National Tournament as I watched them manifest from the floor. The campaigns for Invitational votes, the webs of prize-split agreements, the conflicts of interest between mutual friends. What the official event coverage didn’t tell you, for instance, was that DaRosa and LSV only decided to draw after the outcome of Cedric’s match had been determined; because Cedric won, the two of them could afford to draw into the Top 8 secure in their higher tiebreakers. This set off a rage of debates on whether or not it was within the rules to base one’s decision to draw on the outcome of recently-resolved matches, because feature match results are often made public before the end of the round.
Interesting stuff.
My own performance completely surprised me. I was expecting to dominate the Constructed portion of the event due to my recent history with 60-card decks, and would requite only a modest Limited performance to secure a cash berth. Instead I completely punted Constructed, going 2-5, but rattled off six straight Limited wins against solid players before losing to a Marshalling Cry in the tenth round that I wasn’t playing around because I had passed the guy no Fortifies and forgot there were other creature enhancement effects in White. The result was a strictly mediocre 8-6 finish that, while unexciting, wasn’t entirely disappointing either.
Since I’ve got them sitting here beside me, here are the three decks I played.
Draft 1 (4-0):
2 Emberwilde Augur
1 Riptide Pilferer
1 Looter il-Kor
1 Infiltrator il-Kor
1 Stingscourger
1 Grapeshot
1 Empty the Warrens
1 Grinning Ignus
1 Flowstone Embrace
1 Dead/Gone
1 Lightning Axe
1 Brute Force
1 Erratic Mutation
2 Cryptic Annelid
1 Fathom Seer
1 Aquamorph Entity
1 Thunderblade Charge
1 Pyrohemia
1 Crookclaw Transmuter
1 Keldon Halberdier
2 Veiling Oddity
Relevant sideboard:
1 Dandan
1 Logic Knot
1 Dismal Failure
Draft 2 (2-1)
1 Bound in Silence
1 Judge Unworthy
1 Lightning Axe
1 Errant Ephemeron
1 Linessa, Zephyr Mage
1 Serra Sphinx
1 Aquamorph Entity
1 Unblinking Bleb
1 Shaper Parasite
1 Snapback
2 Tidewalker
1 Erratic Mutation
1 Dismal Failure
2 Careful Consideration
1 Infiltrator il-Kor
1 Bonded Fetch
1 Viscerid Deepwalker
1 Prismatic Lens
1 Spiketail Drakeling
1 Cryptic Annelid
2 Plains
1 Mountain
15 Island
Relevant Sideboard:
1 Dismal Failure
1 Ghost Tactician
Standard:
Creatures (9)
Lands (23)
Spells (28)
- 4 Wrath of God
- 2 Mind Stone
- 4 Compulsive Research
- 4 Lightning Helix
- 4 Remand
- 2
- 4 Coalition Relic
- 4 Edge of Autumn
Sideboard
I listed the list (that sounds awkward) only because, despite my poor performance, Ervin did considerably better with it and we truly were crushing MTGO when I went down to playtest with him in Alabama. Some of the numbers that look wrong are right (1 Mystic Enforcer; sometimes you just need a way to get through or hold off a Vampire or Angel of Despair, and other times you need a big monster, but you never ever want to draw more than one) and some of the numbers that look right are probably wrong (one Mind Stone needs to be cut for a Foresee, most likely). On the other hand, the strategy may just be fundamentally flawed, or barring that might just be strictly weaker than some of the other options. The very fact that I don’t know the answer to that question, though, leads me to Lesson One:
Don’t cut corners.
It’s no secret: I have a terrible record at mixed-format tournaments. I am not a “natural gamer” by any stretch of the imagination. When somebody introduces a new game to me, I hate playing it until I have time to sit down and think through its determinate factors, because I don’t have that intuitive feel for a winning strategy without some academic study of the process. So what usually happens is that I concentrate on what I enjoy — Constructed — only to get battered about the head and neck by people who realize that you have to do something besides twiddle your thumbs for the other seven rounds.
Coming into Nationals, I resolved to do two things. The first was obviously to tell the story of how John Moore’s sister revealed to us that she had a penchant for chugging rubbing alcohol*, but the second was to work on my Limited game so that I wouldn’t make the same mistake I had been making for the last seven or eight years. The problem was that, in doing so, I forgot about what made me a good Constructed player in the first place: namely, rigorous, systematic, intensive playtesting. Ervin and I talked extensively about how to draft, what constitutes the most valuable pick in any given set of cards, how to tell exactly what your neighbors are drafting, how to reap the rewards from later packs, etc., and consequently I was confident that my 40-card game was better than it ever had been before. At the same time, the unjustified, unqualified assumption that our deck was where it needed to be based on its MTGO success rate and on-paper power level probably cost me more wins than all of that extensive Limited testing made up for.
It’s not just a matter of tuning the list. Though we’re probably a few cards off from the optimized list of the archetype, what mattered most is that I truthfully didn’t really have much of an idea how to play it. The core strategy came from Neon and Fob’s Relic deck of about three weeks ago, but our addition of Lightning Helix allowed the deck to be much more interactive than it was otherwise. We understood the format enough to capitalize on that. At the same time, although our core strategy was the same against both aggressive and control decks (kill all their guys and blow up all their lands) I wasn’t terribly familiar with how to go about implementing it.
The way I usually playtest goes as follows: I sit down across from Richard Feldman with a gauntlet of decks. It’s about nine in the morning, so we may or may not get a cup of coffee to perk us up depending on whether or not we’re playing on top of his website’s** beta server. We then proceed to play ten-game pre-and-post-board sets, five on the play and five on the draw, with whatever deck we feel like we’ll end up running. For each game, we sit and talk about what specifically won each game, what could have gone different, what might have worked better if the deck was composed differently, what might have worked worse, etc. It’s usually evident very early on in the process whether or not the deck is total garbage, so around game twenty or so we start to really tweak the numbers and figure out what is “essential” and what is “nice.” After two or three days of this, we emerge with a well-oiled machine.
This type of testing does more than just refine the list, though. It allows you to become intricately familiar with what the deck can and cannot do, what exactly your plan is in every matchup, how to eke the most value out of every single one of your cards. This is why I often wind up playing either interactive combo decks (StarWarsKid Heartbeat, High-Tide, CAL, Hermit Druid/Ashen Ghoul in old Extended) or long-game control decks (Korlash, Tenacious Tron, Trenches, Psychatog). It’s not that these decks are harder to play than mainstream aggressive decks; to the contrary, I think Sadin’s R/g list is one of the most difficult-to-play decks to ever grace Standard. But these kind of decks do reward intimate familiarity with the list and the matchup, because they look at so many cards and give you all kinds of options.
The Relic deck rewarded these same types of things, but I wasn’t equipped to take advantage of them. While we analyzed the composition of the sideboard very carefully, and consequently believe it has all the tools the deck needs to win games 2 and 3 against anything, I kept some marginal hands and sideboarded out some incorrect cards based on improper knowledge of how a matchup is supposed to pan out. Being in Boston for two weeks took my finger firmly off the pulse of the evolving metagame, and as a consequence I wasn’t able to adapt to everything I needed to adapt to in order to play out a game optimally. As anyone who was watching Ervin in the feature match area could tell you, simply not making mistakes with the deck is an absolute nightmare. Moving from that point to actively winning games without the requisite systemic knowledge to do so is borderline impossible.
The final thing I want to emphasize before moving on is that simply playing a lot of games with a deck doesn’t really constitute playtesting, because it has no ultimate goal in mind other than “win as many games as possible.” I’m of the opinion that in order to truly optimize a list, in order to really figure out how a matchup works, you have to have communication between both players about what they’re scared of, what they thought was in each other’s hand (after the match, of course), what could have happened differently, etc. Also, if something – a number of a particular card, or the card itself, or even the absence of another card – “feels off,” it’s important to play some number of games with that particular issue in mind. Otherwise, you’ll have some numbers about how many games you win and lose, you’ll have some vague “feeling” about a matchup, but you won’t have much actual data on which to base your decisions.
The other important lesson, and I knew this, was
MTGO ISN’T REAL.
I knew this. I always have known this, and that’s why I never ever ever EV-er playtest extensively on it, or base any of my metagame estimates on what the Magic Online climate looks like. In fact, as of Tuesday, I had a really good read on what the Nats field would be: a bunch of Rakdos decks, a bunch of Blink decks, and some midrange strategies trying to capitalize on the makeup of that particular field. The problem was that we’d still tested mostly online, and so hadn’t had the time to figure out how to trim the fat for our expected field. Moreover, although we knew that Tenth Edition would have more of an impact because of what rotated out of the format as opposed to what rotated in, not having played a single game against cards like Mogg Fanatic or Head Games, or even Arcanis the Omnipotent, came back to bite us in the ass.
Also, what is good on Magic Online is not necessarily good for a given tournament. I had this conversation with a lot of people on the weekend and many of them didn’t quite understand what I was getting at, so I’ll try to be as clear as possible.
Whenever you’re playtesting a given format in real life, what you’re really doing is playtesting a given format for a particular tournament. Ervin and I were talking about how to break a format when Antonino walked up and said something along the lines of “nobody ever breaks formats, and you shouldn’t try to do so.” I adjusted my language somewhat – and he agreed with my subtle change – by saying that when you talk about breaking a format, you really mean you’re trying to break a particular tournament. Ichorid wasn’t even all that big a force by the time last year’s Extended season rolled around, but it sure was a tremendous beating at GP: Charlotte. It doesn’t actually matter how objectively good any deck is if it wins more games, for whatever reason, than anything else people are playing on July 27-29, 2007.
What this means is that you usually want to get some idea of the expected field, then tune all of your existing decks to have good matchups against the three or four most common decks in that field. Often it’s easier to come up with something entirely new, but the point of these last couple paragraphs is that you’ve always got a target that you can hit.
On Magic Online, this is generally never the case.
Obviously you’ve got Premier Events that sort of act as milestones by which you can measure the evolution of a format, but because information is transmitted so fast the environment is more or less in a state of flux. More importantly, though, the pool of people who take the time to participate in MTGO Constructed tournaments is not necessarily representative of the people you’re going to be paired against at a particular event. The average skill level may be higher or lower, access to online card pools may be more or less difficult (I wouldn’t dream of altering any relevant Constructed deck because I couldn’t get cards in time, for example, but online people will often just throw cards into decks to see how they work) and archetype preferences may vary widely online from what people want to play in real life. Perhaps most importantly, all big real-life tournaments (Nationals, Grand Prix, Pro Tours) are heavily influenced by what is being chattered about in the event hall immediately before the event. This year’s U.S. Nationals, with Gerry Thompson’s G/W/U Blink deck at the forefront, is a prime example of that. Magic Online is simply not going to reflect these changes in time for you to do anything useful with them because such a small portion of its players are attending the relevant “IRL” event.
Now, there are some exceptions to this principle – kind of. Fob and Cheon do a ton of their testing online, of course, and Ervin and I even got our deck’s skeleton from them a couple of weeks ago. At the same time, they live with one another and can systemize what they are doing constantly. You don’t randomly happenstance upon the idea of putting Arcanis the Omnipotent into a deck, for example. It’s all part of a guided process.
I guess the reason the tournament was so disappointing for me was that in a sense I did a lot of things right, and three of my losses were due to some incredibly bizarre circumstances (no Wrath in 30 cards, no Fetters/Enforcer/Wrath/Take Possession for a Lightning Angel in about that same number, and a game loss for power-sliding into my seat all of 13 seconds late despite scrambling around the play area for ~5 minutes because I couldn’t physically find table 18 like a giant idiot). I also did plenty of things wrong, though, and they cost me. I’ll admit that. A lot of people do extremely well at Magic without putting in much practice at all, but I just don’t have that much raw talent. My gut instinct is that a lot of people stumble into these same pitfalls – thinking they’re ahead of where they really are – and cost themselves a lot of matches as a result.
Yeah, don’t do that.
Next week? Back to block.
Zac
* Just ask me sometime. Although several people asked me if I was, I am not in fact a nominee for the “storyteller” invitational vote. At the same time, I can’t ever resist spinning a yarn about someone else’s embarrassing escapades.
** Isthatonegood.com, and it’s really awesome. You should check it out if you haven’t already.
**BONUS SECTION**
Ervin and I also designed the R/g/b deck that won the first Nats grinder. While it has a nightmarishly difficult time beating a Mogg War Marshal, its matchups against basically everything else are so good that I’d have to recommend it as the default aggressive deck to play if you don’t want to sling Gargadons:
Creatures (18)
Lands (22)