When we build and position decks, my group uses a different set of paradigms than I think make up the typical vocabulary of deck choice viability.
Controlling How You Lose
My absolute biggest influence on deck choice (as opposed to, and in some ways more important than, deck design or development) is Zvi Mowshowitz. He's probably not the same Zvi as may have influenced you - a distant, if respected, writer - but the Zvi who is my friend was for a long stretch the consensus best deck designer in the world.
One of the basic elements of Magic, taken mathematically, is that deck X and deck Y may have very similar blended win percentages against the field. We speak in broad numbers that might be as much as 80% accurate. So for instance, given the Regionals 2002 Standard metagame, ZevAtog, Kibler's RUG, and Mono-Black [Torment] Control might be the decks that we identify as our short list.
ZevAtog is the top-down best deck of the three, and probably the format (which has not progressed, yet, to the point where it can be effectively fought by even better Psychatog decks, Deep Dog [which doesn't yet exist], or Trenches or Opposition variants). ZevAtog is extremely powerful, and home-grown (good!)... but known.
Mono-Black Control is a strange case. In 2002, we had not yet developed the effective anti-Compost techniques that would incentivize us to actually play Mono-Black in 2003, and lived in fear of that two-mana enchantment. Barring Compost, Mono-Black Control, home grown by the Evil Don Lim, was a superb choice; more than any other deck, it crushed ZevAtog (something no other deck in our gauntlet could claim). It crushed creatures. We just didn't want to lose to Compost.
That left Kibler's RUG. Kibler's RUG had a very similar Win/Loss breakdown to ZevAtog across the format. In addition, the RUG was not known, which was a huge incentive. The tiebreaker was that RUG was consistently 55-60% over ZevAtog in testing, whereas ZevAtog was consistently, you know, 50% against itself. We went with RUG main line with Don playing 'Tog.
Josh Ravitz played brilliantly, split Fact or Fiction with an expert eye, and hit his mana; he qualified. From our "Mike Flores's kitchen table" testing group, Don Lim also qualified (with ZevAtog). Paul Jordan went through most of the Swiss before taking his second loss in round seven or eight.
I was out after two rounds.
The most influential conversation ever to grace my Magic-related development occurred after this disappointing second loss. My first round I dropped in two to Josh (and I couldn't beg for a better person to lose to), but I lost the second round to an inferior U/G/R deck, though I think that I could have played a little bit tighter so as not to lose on my mana taps. Zvi made a very good point at that point:
I was playing a deck where I could lose to my mana much more often than other decks. I needed G on the first turn for Birds of Paradise, U on the second turn for Merfolk Looter, and required RRR for Violent Eruption, all in the same deck!
Let's be honest: More than 80% of losses in Magic start with imperfect mana development. No one is saying that the great players can't play out of slow mana draws, but flood and screw are colossal mental barriers for players of all levels. Manascrew is the reason you might be able to beat Jon Finkel over his own kitchen table (or the Pro Tour), and remains an integral and fundamentally important element of the game. You can't guarantee that you will hit your drops.
In this deck, even if you hit your lands, you might not hit your colors. If you hit your colors, you might not like that fact very much... I mean Kibler's RUG had a mana base that only Brian himself could love, with eleven Ice Age/Apocalypse "pain" lands and City of Brass! (I have always hated a City of Brass.) You want to lose to an inferior player? Run City of Brass. Better yet, draw the damn City of Brass.
My second round, I think I drew a bunch of Coasts and the City, and I would have had to be tight as the shell on a leatherback sea turtle to have eked out the win. In reality I mentally high-five myself when I make a play that's already on the table; I'm loose like the village bicycle.
Zvi said something that almost – almost - could have come out of Jamie Wakefield's mouth. He walked me, logically, through a series of arguments, and taught me that my deck decision was logically wrong.
He was fine with our short list (Black, 'Tog, and U/G/R), but not fine with our process from that point. Black was a reasonable choice given its dominating matchup over 'Tog, but playing a deck that packed to Compost with no plan against Compost was going to violate the whole "controlling how you lose" sequence once we got to that point. Now assume Kibler's RUG and ZevAtog really were similar on graded percentage (say 70% against the field)... What is the tiebreaker?
My argument against ZevAtog: I can't beat Black. I didn't want to play against the Black.
Zvi just leered at me. I deserved it. Torment being new or not, this was kind of a miserable argument. How many Mono-Black Control decks was I really going to face in eight or nine rounds? One? Probably none. Even if I did play against one, would it be the Don Lim version that could beat ZevAtog? Would the opponent understand how to play against the Standstill and beat Draw-Go Magic (this was a year before Undead Gladiator)? With even Don playing 'Tog himself, the chances of this being an issue were around 0%. Table this and move to RUG...
Zvi's argument against RUG: You lose to your own mana.
Say the two decks will perform to similar limits on the numbers. In the 30% of matches you are "supposed" to lose from 'Tog, you probably have a lot more control over the games. Many of these are the games Dave Price wins with The Stare, or Jon wins with incredibly tight resource manipulation... The games (unheard of in 2002) where the lethal 'Tog comes over with no Upheaval support. Of the 30% or so of games Kibler's RUG is "supposed" to lose, a much greater percentage will be because of mana issues, either not being able to play its spells, or getting mauled by its painful mana base. You have to get a little lucky to win or qualify at any (well, "most") amateur-level events (PTQs and so on), but with ZevAtog you get a lot more margin, even with the same math.
They're consistently awful at Magic.
Pro Tour finalists are bad at Magic. A fair number of Pro Tour Champions make embarrassing on-table errors. Amateurs? I feel fairly certain the average PTQ/Regionals players make an average of three mistakes per turn.
We often test as a group, with primaries on each side of the table, and secondaries watching and commenting from the sidelines, learning and recording, making recommendations. Geoffry Siron once told me that every [serious tournament] player at essentially every skill level will find and make the tightest play on every stack... If you give them 24 hours. The best players, like Jon Finkel, make the best play effortlessly and immediately. Communal thinking gives us the best chance to make the best plays during testing.
For example, in the finals of the mock tournament we ran Tuesday night, Game Two of the finals saw Julian Levin swinging with a Loxodon Hierarch wearing Armadillo Cloak into Gaea's Might Get There piloted by Mike, one half of this year's Two-Headed Giant State Champions. Mike had no idea what to do – and let's be honest, this is a terrible scenario for Gaea's Might Get There. I looked up from my drubbing of Tenacious Tron and said, "Put all your guys front of that."
"Well you're on seven if you take that hit, and he has a Tribe Elder back, and going up to like thirty either way. If you take the hit, you're basically dead on board if he has... well, anything. You have two turns max and you probably won't get another chance to kill the Hierarch with what - a redundant Isamaru? - as your only card in hand. You might as well do it now."
Mike triple-blocked and kept a single one-drop, Julian flashed Cabal Therapy to strip the last Hound of Konda from Mike's hand, and they were in topdeck mode with Mike controlling one creature and Julian up a million life. Some turns later, thanks to the explosive power that gave Raph back-to-back GP titles, Gaea's Might got there and Mike won the game (though Julian's triple-Hierarch draw ended up winning the third).
Because we largely test with constant table talk, walking through everything that seems to matter, we tend to have better expectation numbers than other groups, even when they're wrong... But not all of this translates to real tournament play. It is even less likely for players who don't have solid strategic foundations to begin with.
30% might not equal 30% because on one side. The bad guys might be shipping as much as 15% back (and on the other side, only 5%) with their play. I mean almost anyone with a reasonable draw from a reasonable deck can take down even a Pro Tour Champion who keeps a two-lander when those lands are Yavimaya Coast and City of Brass. Our deck choice wasn't the worst (Josh qualified, after all), but it didn't give us – or at least me - the maximum margin or maximum value.
How does this translate to practical Extended deck positioning?
I wouldn't soon play either deck in the present climate, and would be even less likely to run Tenacious 'Tron or some similar. All these decks get a chunk of free wins... But they are all, in one way or another, painfully vulnerable to commonly played and cheap answer cards. I love Affinity. I test against Affinity a fair amount online. I've never lost a single game I've drawn an Ancient Grudge. Will none of my opponents think of maybe playing Ancient Grudge (i.e., the best card in the format)? I'm astonished Saito was able to make Top 8 of Singapore. I mean, he's awesome and everything, but like, people do nothing but scoop on the second turn in sideboarded games with Tenacious Tron. Sometimes they try to stretch it out until the fourth or fifth turn, but then the other guy will expose Ancient Grudge and... I mean... let's be honest... It's out of your hands.
These decks are all awesome for sure, but in terms of personal preference, I don't like it when I have absolutely no say in whether or not I lose, in games dictated by cards that cost 0-3 mana, with "3" being effectively "2" in a deck that also packs to a 2.
One of the basic concepts in Magic that as far as I know has never gotten any significant press is Trump. Trump is a term related to power, ultimately - pardon the term – trumping the Fundamental Turn given the right conditions. Resource, offensive, and defensive speed obviously remain operational… But Trump, when it comes into play, is the most important element in any game that has not actually been decided.
Think of a game of Spades where the top cards are Big Joker, Little Joker, and Deuce of Spades, followed by Ace of Spades, King, Queen, and so on, then Ace, King, Queen, and so on of individual non-Spades suits.
In fourth position, you can take a book either with the seven of Spades or the Ace of Spades. There will be sandbag conditions where you will use the Ace so that you will fail to win a book later, but generally, you are going to want to use the seven. Using the seven instead of the Ace will probably pay off a book later. Hands go by, and both Jokers fall, but not the Deuce. One hand left. You're praying your teammate has the Deuce, because your Ace is no longer good. Your Ace is "good" in the abstract, maybe, but it's not good enough to win a book any longer; Deuce is Trump.
I actually learned about Trump in Masques Block Constructed when we figured out what you wanted to do in White-on-White was to be the guy controlling Mageta the Lion. We played all kinds of garbage - Afterlife main, five Disenchants to kill their Parallax Wave, Predator motherloving Flagship - to not lose to the other guy's Mageta. At the very least we played a million lands and four Lions to make sure we would draw and cast it more often. Mageta was trump over all the White creature strategies.
Think about Extended from a few years ago. Loop Junktion (Life) was a popular deck because it was relatively fast at establishing its nigh-infinite life combo and was pretty good against media darling Red Deck Wins. Life was utterly useless against decks like Mind's Desire and Aluren. Mind's Desire could care less about an opponent with 1,000,000,000 life, and Aluren would gain 1,000,000,002 and deal 1,000,000,001 at instant speed.
The speed/Fundamental Turn of Loop Junktion was not relevant because the opposing strategies essentially ignored what the Life deck was trying to do.
Their kills were Trump.
Trump is generally fluid based on matchup. Mind's Desire's "Millstone" kill was a tough sell against Cephalid Brunch (which just wanted to do the same thing to itself). Many matchups at Pro Tour: Charleston and even 2006 summer Standard came down to Simic Sky Swallower advantage. One of the reasons we switched from U/R Wafo-Tapa and 'Vore to U/W Wafo-Tapa for the NAC Qualifiers and NAC was that a resolved Simic Sky Swallower was no longer trump against a White deck (we had Wrath of God and permission).
In the present Extended, there are certain cards that basically read "I win" that you can play on turn 2 or thereabouts. These cars include Counterbalance, Trinisphere, and, most commonly, Destructive Flow; again, these are matchup-specific threats that don't win every matchup, but absolutely crush many common enemies. Trump usually refers to single cards or strategies beating whole decks – for example, Mageta the Lion lording over the entire strategy of playing a lot of White creatures, or gaining infinite plus one life and dealing infinite plus one life loss spitting all over merely gaining infinite life and winning in some arbitrary way. In some wise, the existence of trump presupposes the deploy of a successful strategy, in that strategy simply ends up being unsuccessful or irrelevant.
Consider decks that really really want to get Sensei's Divining Top and Counterbalance in play. They play special cards like Trinket Mage or even Muddle the Mixture to find one half or the other, and run cards that are often big liabilities like Dark Confidant due to in-theme positive interactions. Consider middle-of-the-road Destructive Flow decks... They tune towards Birds of Paradise and sub-optimal acceleration choices like Elves of Deep Shadow to get the proper mana for a fast Destructive Flow. Both of these strategies win blowout games against some of the best decks in Extended; both strategies require sacrifices on card power in order to find room and increase consistency for their key [planned] hosers.
It then follows that when these cards fail to generate any card advantage that more is lost than one or two topdecks to virtual card advantage. These decks can be like houses of cards, falling like dominoes, all their incentives gone with the diminished efficacy of a central theme.
Playing a deck that does not fall prey to such strategies can be considered a kind of passive Trump. Eternal Dragon is a good example of a mislaid Trump card. People don't realize how troublesome Eternal Dragon has always been. He took a lot of real estate away from Akroma, and has always spit on Exalted Angel. In the context of this discussion, he's basically trump. Eternal Dragon is a seven. No one plays sevens. Ergo, Counterbalance doesn't stop Eternal Dragon. If the rest of the opponent's deck is a sick 'Tog deck, the slow Eternal Dragon deck might still be in trouble… But if it's a creature deck, probably Eternal Dragon is going to beat everybody up and then beat the opponent up. It is a 5/5 flyer, after all.
Another good Eternal Dragon example might be decks with Exalted Angel for life defense and Eternal Dragon for draw versus decks with Pulse of the Fields for life defense and Eternal Dragon for draw. The Exalted Angel deck might be better at closing a game against Goblin Bidding before the Bidding opponent "goes off," but the Exalted Angel is in fact a major liability fighting the opposing White deck.
Imagine you actually hit someone with it.
Wow, would that be bad.
You go to twenty-four, he goes to sixteen, you promptly lose the game 30 turns later.
Why? Easy. The first Angel hit just turns on the opponent's Pulse of the Fields, which will now be infinite depending on time. I guess you could start mana burning or something. Unless you can win in a single strike (say, with a massive Decree of Justice overload) you can't really win on damage, or at least not with incremental packets of damage. So basically, it's going to come down to who has more lands in his deck, Eternal Dragon advantage, and timing. Recognizing how bad your Exalted Angel is against the opponent's trump of Pulse of the Fields is the first step into not, you know, attacking with the Angel.
This is an example of one player's ability to mitigate the power of the opponent's cards in a bad matchup... First he has to realize not just that, but why, he is the dog.
An important - maybe more important once you've got the fundamentals down - and related concept is the false, or lost, trump. For much of the season, a Chalice of the Void with two counters on it was essentially good game against Loam. That is no longer the case, with Engineered Explosives now being featured main deck on zero, Shattering Spree coming out of the sideboard on one mana, and Putrefy being essentially ever present on three mana. While it's probably still good in the abstract to try to play a Chalice of the Void for two, which should buy you a turn or turns against the Devastating Dreams if nothing else - putting all your eggs in that particular basket is folly given the more informed Loam players in the format.
Similarly, if you stick Troll Ascetic + Worship against Boros in a sideboarded game and the opponent, an honorable friend, doesn't scoop, you should look as far into the future as possible and get the gears rolling. Why didn't he scoop? The answer is probably "He has Ray of Revelation [somewhere in his deck]." At that point, you must, must, must realize that your little Worship "lock" is not good. Worship might buy you some time, but it won't win you the game. At that point, you have to flick on your strategy button and try to figure out how you are supposed to win this stupid game.
Will one more Worship win it? No. Ray of Revelation has flashback. You need three Worships and you need him to not draw two Rays in order to win via a Worship lock. Your other option is going to be killing the opponent as quickly as possible, leaning on Worship for a couple of extra turns but not relying on it for victory.
Playing for this plan does not presuppose the exclusion of finding more Worships, mind you. Remember: If all he has is Disenchant, you still want more Worships; you just need fewer to lock the game out. Playing to win quickly never ceases to be important because no number of Worships will beat two Rays.
Some players want to play the most powerful or flashy deck, while others want to win with a pet deck that showcases a lot of individual effort and tuning. Others see a format inefficiency and want to hammer a particular popular deck into the ground no matter what, to never let it pass, let alone qualify.
Some players think the best deck is the one that plays the most powerful cards. Others believe that the best deck is the one that breaks the most powerful cards in the most heinous and ugly way. Still others assign decks based on matchup percentages, picking the deck in their gauntlet that has the highest percentage against the predicted field, irrespective of tactics or nuance.
Each of the various approaches has some merit, but I try not to be slave to any of them. What my group tries to do is pay the least amount possible for wins while at the same time making losses as expensive as possible. Sometimes in my forums you'll see a comment like "Your deck is really good against beatdown... Don't you think you could give up 10% in the beatdown matchup in order for 20% in this minor matchup?"
Sometimes the answer is yes, but usually the answer is no. I try to pick decks that have lots of good matchups and not very many bad matchups. I don't really want to get percentage in the bad matchups, which I am assuming on the baseline are less common, if I don't get to win them. I can be wrong, of course, but if I'm not wrong about the kinds of decks I will be fighting, I try to win as many of my expected matchups as possible even if I am left with difficult esoteric ones. For example, if I think I will play three beatdown decks and I have an eight out of ten shot at winning against beatdown, I will not trade down to a seven out of ten in exchange for a 10 to 30 bump in my Tooth and Nail matchup. I'm still going to lose the damn Tooth and Nail matchup, and now I have a much more real shot of losing one of the beatdown matchups.
One of the things I try to do the most is to stay nimble and change constantly, which is fine because I absolutely love playtesting and trying out new cards and ideas. Certain cards like Fortune Thief, to a lesser extent Grand Arbiter Augustin IV and Gnarled Mass, and certainly Gilded Light lose trump value over time. For example, if Gilded Light becomes known, people start playing around it. Fortune Thief is actually Godawful useless if an opponent has the most tiny inkling that you might be playing it, even when it is an automatic win if they don't see it coming. Same on Grand Arbiter: Dragonstorm players start running Dreadship Reefs and they can break your Arbiter lock.
Winning with quirky-viable individual card choices rather than top-down obvious cards like Birds of Paradise, Life from the Loam, and Umezawa's Jitte presupposes an information imbalance. With some cards, like Gnarled Mass, time equals adoption, not simply reaction, meaning that the same deck can have a short run negative delta of as much as 20% against the field as your limit approaches 50%.
I play about two more basic lands in most decks than most people playing similar decks in Extended. It's so easy to hit your mana with Onslaught duals... You don't have to run a million Ravnica duals too. Sometimes you don't want to run the maximum number of Onslaught duals, and six can be right in a two-color deck (or maybe just the four), because you want to preserve even that one life point against beatdown. Don't you want to thin? Don't you want to hit colors? Of course, and of course. I am also very limited in my scope, being old and having played by Brian Weissman's rules since when Brian was still playing Magic at a high Q-rating.
The incremental draw quality of an Onslaught dual thinning your deck is less than 2%, and sometimes you are actually going 2% the wrong way; the value of a basic land versus a self-inflicted dual land wound is 5-15% against a beatdown deck. Fifteen percent? Have you ever thought of that against a deck like Levy's, or God forbid, Tsuyoshi's new Firecat Boros? I think of the beautiful array of available mana fixers in Extended as a necessary evil that can help produce an advantage, not a baseline given that should be pushed to the furthest possible extreme in every deck. Just for point of reference, I have a Rock deck that I have been testing that has eight Onslaught duals, three Ravnica Block duals, and twelve basics (along with four Birds, four Sakura-Tribe Elders, and four Solemn Simulacrums... I can probably give up a Forest). This would be an example of a three-color deck that consistently hits its colors without hurting itself more than is necessary.
The big picture element that I always try to keep in mind when picking a deck or building a new deck, more important than anything else, is that my deck be fast. I don't mean I need third-turn kills - just that I have relevant cards to play in the first three turns, consistently. Ideally, I want to play a card every turn, and I want whatever I play to be slightly better than whatever my opponent plays, but this is obviously not possible 100% of the time.
In general, I like my cards to have a direct effect on the board (or failing that, game state) immediately, so I tend to like cards like Destructive Flow, Flametongue Kavu, Loxodon Hierarch, Keiga the Tide Star, etc. I find that these kinds of cards, positioned oddly enough, force my opponents to act and react, and hopefully make short-term mistakes.
There is definitely not only one way to approach deck design and positioning. I am completely useless at combo decks, and pretty bad at any top-down design initiation, even though I am very good at anticipation and attack...
To make a long story short, the unique elements I use to build decks don't work everywhere. Today's sections outline some of the techniques that I, personally, try to use in order to create efficient decks. For about the last three years, my decks have tended to approach tournaments with greater share value than archetype decks, even when they play weird cards or combinations of cards, and that value comes from attacking the format from different angles than might be expected, utilizing forgotten linears, and positioning in such a way that the opponents' expected cards and strategies are least effective, wherever possible.
Before you ask, I have no idea what I am playing tomorrow, but it will probably be some mid-range polychromatic Green deck with 4/4 for four that somehow allows me to gain four. Just a guess.