for publishing some of the most fundamental principles of Magic. Almost all strategical theory can be
traced back to the Dojo’s loyal writers, and any serious Magic player owes these old vets a debt of
business in 2000. In a last-ditch effort to save the four years of wisdom that had been collected there at
the time, the editor asked the community to archive the articles for future reference. The best of the
Dojo articles are reprinted here because they’re still vital to Magic today… StarCityGames.com merely
reprints them, adding links to clarify older cards that new players probably won’t have seen so that they
can understand some of the strategy. Many of the Dojo’s writers are still active in Magic and write for
other sites; give them a shout-out for helping the community grow.
Building Broken Decks Volume V: More Mail,”More than Just Me”
Previous editions of Building Broken Decks can be found on The Dojo following:
Lessons from Texas
The Mail Bag and More On Deck Selection
The Wakefield Error and”The Decks to Ignore”
If you don’t want to read your fellow readers’ comments, questions, and suggestions, skip ahead to the “More than Just Me” section… but you’ll miss my bi-weekly bashing of moneydrafter and grad student extraordinaire, Tim McKenna. Oh, and let me warn you, even if you do jump ahead, this week’s installment of Building Broken Decks has only two deck lists… One is awful and one is untuned. 🙁
First of all, let me apologize to anyone who I did not mail back personally over the last set of articles. I have been extremely busy with non-Building Broken Decks obligations (including, as you might expect, my own process of building decks). Please don’t be discouraged, however. I do read each and every one of the personal emails that you all send, and appreciate the feedback.
This time around, the vast majority of emails consisted of new deck lists or requests for my upcoming deck lists. I haven’t had much time to answer the former sort of emails, and have to apologize again. There have been some interesting ideas, but most of their creators seemed to want to save them for next month’s State Championships. As for the latter, whenever I intend to share a new deck list, I will do so in this column; check here in upcoming weeks for Invasion analysis.
There were also a couple of players who objected to my assertion in Volume III that you shouldn’t play”your own style,” whatever that means. These objections ranged from the completely understandable (“I don’t have the cards to have multiple competitive decks changing every week”) to the unreasonable (“Forbidian is more than a style, it is an actual deck, and more than just creatures and counterspells; player X’s success has to do with his heightened experience over time blah blah”). Now of course there is a great disparity between players who have vast collections of in-print cards (when I first started, they called these guys”Mr. Suitcase”) and those who have much smaller collections.
In case you were wondering, I tend to fall in the second column. I get almost all of my cards from PTQs and drafts. I habitually borrow cards from my trading-crazy friends, like Paul Jordan, Josh Ravitz, and Tuna. Just this past weekend, when I won the NY Grudge Match qualifier, I had to borrow Treachery from a nice player I had never met before named Hector Claveria (thanks man!). To make a long story short, I don’t think that small collections necessarily imply an inability to be competitive, but I can understand that line of reasoning, especially from new players. The other line of argumentation is less valid for many reasons. For one thing, any highly heralded deck, be it Forbidian or Sligh or whatever, fits into the matrix of the metagame matchups just like any other design; some decks are better, yes, and perform better. That is what my comments on deck selection were about… there is nothing mystical about these decks. They are composed of cards in such ways to generate either cards or time, or they fail. As for associating certain star players with certain iconic decks, that is less interesting for the average Dojo denizen, don’t you think?
On a related topic, steak-eating buddy Mark Schick pointed out a nice distinction I failed to make in my rather sweeping”play your own style” criticism. Mark pointed out the difference between not playing the”best deck” in a format, for example Trix during last year’s Extended PTQ season, if a player is not familiar with the deck, and playing a less optimal deck with which he is familiar. Mark used the successful Three-Deuce as his example in the email, and while no one argued that we should ban Mountain Valley, we can all recognize reasonableness of Three-Deuce as a deck choice if someone were not going to play Trix; Three-Deuce may have been one of the best decks in that format, after all. In this case, Trix is a very difficult and math-intensive deck, whereas Three-Deuce, while less optimal, still boasts a number of powerful cards, and is a comparatively straightforward creature/elimination deck. I think this is a perfectly defensible argument, but then again, it at no point suggests that someone make a deck choice on the basis of it being his”style,” but instead points out one of many reasons that someone might make an alternate deck choice. Considering the fact that no tournament that I have ever heard of has been 100% composed of a single archetype, we can only assume there are dozens of such reasons… I would just hope that the instances of”because it is my style” will go down in the future.
For an article about the same subject by someone much smarter than I am, check out “Learning from Losing” by Magic’s best analyst.
I got a random treat in the form of a short note of encouragement from my former roommate (and The Dojo’s former editor), the aforementioned Charles”Tuna” Hwa. Tuna is perhaps the least well-known editor of this great site. In my opinion, he had the most difficult time of it, logistically, but left the site a great deal better than he left it, and in the hands of our current, lovable, Christopher Robin Senhouse. I think everyone should give Tuna a little hand for what he did during the interim period between my departure and The House’s current efforts.
There were, of course, the random notes of encouragement and praise. I think that half the reason that most of the game’s writers put so much effort into these articles every week or two is due to the encouragement that we receive from everyday players.
In the tradition of Tim McKenna, Lagges@aol.com attempted to give my last article a bad rating. Unfortunately for him (and the aesthetics of the page, mind you), the page rating scripts appear to be broken. sigh.
Speaking of Tim, I got not one, but two emails from my favorite Penn student (well, not counting Tuna or this cute girl Jen I met who is a graduate student of urban planning or something). Here they go:
wow a lot of insights in this article. the most useful email you
received said”metagame against the correct field”. i think that was the
problem with your pt junk deck, you made the first constructed deck
metagamed against the mirage sealed deck field.
do you want to money draft in philly this weekend?
SEND ME YOUR TYPE II DECKS
TIM MCKENNA, PHD 2006
The highest compliment, of course, was hearing from Mike Aten! Mike Aten is currently a student at NYU, but never shows up at the Ground; I met Mike and his brother Tim a few years ago when I still lived in Northeast Ohio. We all gamed at Reality Recess, which was the proving ground for a number of the Pro Tour’s favorite sons, including John Hunka, Jason and John Marks, Ben”Manascrew” Murray, Jason Opalka, Mike Turian, and Worth Wollpert. Aten is also qualified for the upcoming PT Chicago (unlike your four-time manascrewedinthetop8 narrator). Mike still owes me $20 from when I bashed his PatrickJ.dec in the brain with a Faerie theme deck for ante last year.
great article on why the ‘play a deck that is your style’ philosophy is dumb.
too bad i wrote it three years ago, you great big hack.
Gotta love Mike.
On the subject of old Ohio buddies, WotC’s latest Pro-turned-minion, Worth Wollpert dropped me a line last week too. Not many people know this, but besides driving my ass to Mike Guptil’s insanely successful Ohio Valley Regionals 1999, Worth was also instrumental in my making the Top 8. Who knew that Cursed Scroll was better than Erg Raiders? Thanks buddy.
I misquoted Bill Macey last week. Unlike your narrator, he never says”mize.”
Last but not least, edt (Magic’s best analyst, in case you missed that) asked that I write an article about Invastion and Investment. Investment is a theory that I developed with Erik Lauer in 1997-1998. edt asked that I compare Brainstorm/Soothsaying, Teferi’s Moat/Wrath of God, and Teferi’s Response/Tsabo’s Web, lest he steal the topic from me. I think I will eventually work on this kind of an article, but I would also not hate seeing edt’s crack at it. I enjoy edt’s strategy writing more than anyone else’s, I think, but he has been largely absent from the Magic analysis scene for months. Go ahead and write your article, Eric. I can tell you right now that I will probably agree with what you have to say about the first comparison but not the others. In fact, I will probably have to re-read my own article to do any kind of justice to the topic.
If you would like to do so as well,”Investment” can be found on The Dojo here.
“More than Just Me”
We have worked so far on the theory of deck design as a largely theoretical process. Besides my first article (which only touches on the idea of playtesting), we have talked almost entirely about decks that exist already (templating, selection) or new environment card choices in a vacuum.
What we have ignored is the process that edt mentions in the article I linked to above,”The reason Magic is different is that it is impossible to predict how much any particular deck will beat any other deck, other than by playing.”
The principle vehicle for successful deck design beyond sound theory is playtesting. Before we proceed further along this line, I want to point out that I think that sound theory trumps playtesting every time. Even when you get to the percentages between decks that playtesting gives you, you have to understand two things (and even the game’s best pros get this wrong).
“Testing Lies.” – Zvi Mowshowitz
The percentages that your playtesting generate do not — absolutely do not — suggest the expected value of your deck’s performance against your opponent’s givven archetype during an actual tournament.
Zvi’s first comment is something that I think of as the”common sense” of Magic preparation, and is a corollary to my argument that sound theory is more important to deck design than hours put into playing a deck with nothing on the line. No matter how much you test, who your test partners are, you will rarely receive an accurate picture of what is going to happen when you sit down against an actual human opponent in a tournament. No matter how much we try to lie to ourselves, we make playtest choices that hinge on our knowledge of a playtest partner’s deck configuration, we take back (or don’t take back) plays that could mean a significant”Win” or”Loss” in the ever-important percentage column, proxy cards, and also have to deal with an appreciable difference between playing a partner and playing a real opponent in a tournament setting. All of these things together foul pure percentages, and corrupt the real value of playtesting.
You may have heard of the PT Junk deck that Adrian Sullivan and I played at last year’s PT Chicago; Tim McKenna mentioned the deck in an email I quoted above.
The deck was just awful.
4 Cursed Scroll
4 Mox Diamond
3 Demonic Consultation
4 Funeral Charm
3 Hunted Wumpus
2 Land Grant
4 River Boa
3 Simian Grunts
3 Aura of Silence
4 Swords to Plowshares
4 Scrubland[/author]“][author name="Scrubland"]Scrubland[/author]
3 Treetop Village
It tested incredibly well, though.
I remember looking at the deck list and knowing — just knowing — that the deck would be awful against Survival of the Fittest decks. It had all spot removal, almost no card advantage, and would have the wrong spot removal in tons and tons of matchups. I have no problem accepting the blame for this choice, though. To be honest, I was a little lackadaisical in my Chicago preparation until about the last week or two before the PT. By that point, the deck had tested extremely well, and it was in the final honing stages for our playtest group.
I decided to test it a few dozen times against Rec/Sur myself.
I kept winning. Out of nowhere I would randomly double-Funeral Charm a Hunted Wumpus for the last 10 points. I would randomly Duress a hand that my opponent only kept because he had Survival of the Fittest, and then win because his hand afterward consisted of overcosted 2/2 creatures. I would randomly score from 20 with River Boa when my opponent got a Tropical Island and I got a hand full of Demonic Consultations.
There were lots of problems with the testing process that had nothing to do with the W/L column. I was winning and winning and winning, but I never at any time felt like I should have been winning. I said the wins were random, and they were. They had to do with suboptimal designs across the table, high standard deviation draws that couldn’t be counted on in repeated performance, and the fact that I was a much better player than my opponent. I don’t think it matters what a W/L record says, if it equates to the deck with Hunted Wumpus but no real clocks and no real card advantage beating the deck that draws 2-3 cards per turn and plays with Tradewind Rider and Survival of the Fittest.
At the actual Pro Tour, I managed to make Day 2, but a loss threw me into the”Rec/Sur bracket”… and my initial ideas about the deck came all too true. I got Dawnstrider-locked by a deck with no black sources in play. I had to have the right answers against diverse threats over and over again. I was literally forced to play Hunted Wumpus (more than once) when there was both a Survival of the Fittest and a Tradewind Rider in play on the opponent’s side.
In even the non-Rec/Sur matchups, PT Junk performed a lot less excitingly than we had predicted. I was doing poorly against Jon Finkel mono-u deck (but then again, I was facing Jon Finkel, and we never tested against Dust Bowl Blue… check out the PT Junk land count). I lost to Chris Senhouse in the last round of the PT, even though I was”supposed” to beat Slivers. I did a lot of losing on Day 2.
Most of the problems with our PT Junk deck, including actually showing up to a Pro Tour with that pile of crap, could have been dealt with during the testing process. As I said before, I was a lazy playtest partner, and didn’t give Adrian enough support during a good portion of the available months. Some of the problems I can identify via 20/20 hindsight are:
The Lan D. Ho problem: We didn’t give our opponents any credit. Lan D. Ho tries not to make this mistake, as I mentioned last week. Our deck designs were really boring”last year” decks with very little attention to the new Mercadian Masques options. I am not suggesting that all deck designers are equal, but assuming that your opponents in whatever format you meet them will make no significant improvements to an archetype, and will then show up for a tournament, is ridiculous.
The Zvi Mowshowitz problem: At the PT, we would be playing against opponents at least as good as we were. We didn’t test like this was our attitude. Zvi, on the other hand, tries to test against deck design builds and levels of player skill more-or-less commensurate to the tournament in which he is entering. To paraphrase him,”I’m not interested in beating the best Aggo Waters, or the most boring Aggro Waters, but the Aggro Waters that I might actually play in this tournament.”
Between these two mistakes alone, you can basically reduce months of playtesting to zero useful information, but we had more problems on top of that…
Bad work ethic: Me again.
A complete refusal to play the best decks, even though we had them: We could have played some version of Necro-Donate (our Cabal-teammate Shane Neville developed the deck well before Chicago), Necro-Pebbles (Adrian learned about the deck during a trip to Europe just before the PT), free spell Necro (our friend Nate Heiss sent us the Midwest version right before the PT), or good old Maher Oath (for the love of God, I wrote a rant about how it was the best deck… and it turned out that it was), but we were bullheaded. The mere existence of these sorts of archetypes kind of refutes the idea that other playtest groups don’t have (or at least share) innovation beyond existing non-templated archetypes. If you don’t prepare for them, you are asking to lose.
These kinds of mistakes are the sort of thing you should think about but probably not dwell on. Following is a new 2-tier approach to playtesting that I don’t think any group practices. I developed it quite by accident for Building Broken Decks. Most playtest groups are fabulous at one side and awful at the other side. The described method is not perfect by any means; I suspect that it will be better and better with longer possible playtest periods.
Testing for the Decks
Testing for the Players
Testing for the Decks is the first step in this new method. During this step, the player objective should be to learn about the interaction between decks. It is during this stage that you can work on deck percentages, predict what the common archetypes will be, and so on. I think that keeping percentage track of deck performances is a good idea during this stage, because it lets you know what decks seem to be performing well against predictable opposition. As I mentioned before, you should not take the percentages that you glean from Testing for the Decks testing as sacrosanct, or even accurate, when you approach a tournament setting.
The wins and losses of this stage merely contribute to some possible percentage that exists out of real time called”the matchup.” Say you win something like 7-3 against a friend, and decide that”the matchup” is now 70% in your favor. You proceed to lose”the matchup” 0-2 in the first round of your store’s weekly Type II. Is that an anomaly? Were you supposed to win? Was n too small? I don’t think it matters that you won 7-3 in testing, because that does not imply that the matchup is 70% in your favor; that is to say that you do not look at any given trial (say, a tournament match) and compare it to this fourth dimensional concept. What in actuality happens is not that you are now contributing two more trials to”the matchup.””The matchup,” after your round 1 loss is now 7-5.
Even when n is very large, say 700-3 rather than 7-3, it is constantly in flux. Every trial contributes, rather than being compared against,”the matchup.” Nor am I arguing that you cannot at all predict a deck’s performance in any matchup, just that some players seem to approach interaction percentages as something to hide behind rather than something to which they are contributing.
I may not have explained that as well as I should have, but I want to move on to additional dynamics of the Testing for the Decks stage. During this stage, there are a couple of different ways to look at which player should play which deck. In the very early stages, the more familiar, or founding, player should probably play one deck over another. As time passes, players should probably switch off decks in groups of 10, in order to minimize the impact of play skill on test results (remember, we are looking to understand the decks, not improve the players, at this stage).
During the Testing for the Decks stage, proxy cards are not only acceptable, but encouraged. It is stupid to build every test deck card-for-card, especially when a non-performer can be discarded at any time.
I think that take-backs are also a good idea at this stage. It horrified me a few days ago when I saw one of my favorite deck designers make a material error during an alpha-stage playtest game, realize he was going to lose, and say”Things I have to learn…” That is horrible in this stage. You will never learn anything useful if you refuse to allow a viable deck to win because you fuck up during early stage testing. It is incredibly easy to make mistakes that you would never make a month later when running a new deck; this is only compounded when playing with proxy cards.
I think that one of the important things to note during this stage is not only which deck beats which other deck how often, but the mechanism by which it does so. Is it a stomp? Are the games close? Did it come down to life? Answers? Decking? The closer the games, the more varied the possible win conditions, the less reliable the percentage test results will be.
Testing for the Players is the near-opposite of Testing for the Decks. I have found that a lot of quality players test using a very stiff form of Testing for the Decks at every stage of preparation. I’m not sure if this is better or worse than not testing at all.
During Testing for the Players, the focus is on making the players perform better when playing decks with which they are already familiar. At this stage, play skill should be a factor. Players in tournaments have wildly different results for any number of reasons. Because not every player has equal skill or an equally boring deck, Testing for the Players should mold around these differences.
Testing for the Players can be accomplished in set/tabulation-based sessions like Testing for the Decks, but I have also heard of much more interesting methods. Both the YMG pros and the Neutral Ground pros hold mock tournaments to see how various decks perform in short (best of 3) game sets, against varied opponents. One of the game’s most celebrated designers routinely masquerades as someone else on IRC in order to see what”everyone else is playing” during pickup games of Apprentice.
Minimal proxy card use should be allowed during this stage. Jon Finkel has decried”poker deck” use at all, as it removed much of the intuition a player could have with his deck. Chris Pikula once said that the Deadguys would never test Survival of the Fittest decks on Apprentice, or with proxy cards if they could help it…”Part of that deck is picking it up and looking through it for the card you need; a lot of the time you think you know what you need, but there might be a better play… you won’t know if you can’t see the cards.”
Take-backs at this stage should be kept to a minimum. Players will never improve if they are allowed to take back every error they make in testing. Some players are particularly ruthless about this. Don’t worry if it seems they are being jerks. If you are losing it usually means you shouldn’t be playing your deck in a real tournament.
Decks should constantly be tweaked to improve performances in various matchups. Testing for the Players tries to simulate”real life” deck interaction, with real players and real consequences for mistakes. That also implies that real players can make significant improvements using information they have gathered from previous tests and trials. I remember Alex Shvartsman beaing me 7 games in a row in Trinity Green v. Napster testing the week before Nationals. Alex went home and I changed 1 card in my main deck, and then played Jon Finkel in the same matchup; I went 13-0. The change I made helped me beat Alex 2-0 when we met for real, on Day 2 of US Nationals.
The most important aspect of testing is to never become attatched to your results, your wins, your losses; never complain. If you are manascrewed and lose, you probably deserved to lose. Are you expecting to never get manascrewed in a real tournament? That manascrew loss is an accurate result and should be taken into account for purposes of”the matchup.” Some players become romantically involved with their decks. Please don’t do so during the testing stage.
Up Next: Ancient History
Stupid Bonus Invasion Deck Section
This is a version one of many decks I’ve thought up for post-Invasion Type II. I only played one playtest game with it, but I won. No one would play me afterward. 🙁
MikeyP got quite excited when he heard I was working on a U/G deck… but then got a lot less excited when I told him what it did. No, I don’t think this deck would be very good againt”Counterspell.”
2 Moss Diamond
4 Sky Diamond
4 Fact or Fiction
4 Power Sink
4 Rhystic Deluge
2 Natural Affinity
4 Skyshroud Claim