Magic is not an exact science. Playstyle, both players' shuffling techniques, deck composition, and mistakes are just a few variables that you'd need to control or eliminate to get a truly rigorous scientific analysis of deck performance and winning probability. Instead we tend to focus on results and how a deck will play over multiple games. This might not be the most "valid" method, but when you dig into results, there's still a lot of valuable information that can be analyzed.
I've spent the previous week trying to explain the odd disparity between my results from Magic Online and paper Magic with Death and Taxes. To do this I've been playing the deck to the exclusion of all others and I have not changed my list since last week to prevent additional variables from distorting my data. I still don't have that answer, but I have learned a lot about my DnT list and its place in the metagame. It's not been what I expected, and it's left me wondering what DnT's actual role in the metagame could be. I think that the deck is worth investigating, but I don't think it's ready for prime time. That doesn't mean that it won't eventually be, though.
The Issue of Speed
There is no getting around it---the deck was surprisingly slow. There was never a time when actually aggroing out the opponent was correct, and frequently it was impossible. For reference, the deck that I was testing:
As a preface to the rest of the discussion, the sideboard was definitely wrong, but I never changed it to keep my testing consistent. As I said, I was trying to answer questions that earlier testing raised and at that time I thought the sideboard was correct and if I made changes it would have invalidated my data. The Disenchants were there for Ensnaring Bridge and Ghostly Prison, but I failed to account for how powerful Flickerwisp actually is and so Disenchant was unnecessary. They should have been another Kataki and Sunlance. Mirran Crusader was a little underwhelming as well for reasons I will explain later. It really wants equipment to excel and could have been more Mindcensors or possibly Tectonic Edge.
This deck has a bad curve. Out of 38 total spells 16 cost three or more. Of the remaining spells Serra Avenger effectively costs four due to its playability clause and Path to Exile is reactive, leaving 16 cards that we can play before turn three. Compared to a typical burn deck with only one- or two-mana spells, or Jund which typically runs 12 three-or-higher-drops and 27 two- or one-drops, DnT is alarmingly top-heavy.
The Clue from Thraben Inspector helped alleviate some of the pressure on your mana, but far too often your curve was a single one-drop and hand full of threes that took forever to deploy. Your one- and two-drops are a bit too anemic to really be considered aggressive. This isn't a problem for GW Hatebears thanks to Noble Hierarch, but the best non-green decks can do is Aether Vial which doesn't help the problem until turn four at the earliest. This made the deck feel incredibly clunky. It was prone to falling behind and losing to good curve-outs from nearly every other deck.
The speed issue was rarely a problem. Yeah, I fell behind a lot but once I hit three mana all my spells were effectively two spells that gradually made up for the tempo I lost the previous turns. It wasn't a guaranteed thing, mind you, and it is very hard to fight a really excellent curve, but if you do or the opponent just has an average hand, you will claw your way back with the value creatures. My losses tended to come from my slow curve letting me get swamped or from the opponent going over my head. I rarely lost the grindy games.
This was especially surprising against Jund. In my initial testing online I was demolished when Jund kept answering my creatures cheaply and then deployed a threat that went unanswered. My disruption was too fragile to be meaningful and Jund simply powered through. In paper the games played out as grindfests and DnT would either end up with the last threat standing due to threat density or it would race Tarmogoyf with a combination of fliers andKitchen Finks. Online, Crusader frequently saved me since only Lightning Bolt could answer it on the board and it ended the game very quickly, but in paper it ended up being just another threat and wasn't necessary.
I'm at a loss to explain this disparity. The best I can do is speculate that variance works differently on MTGO thanks to the shuffling program than it does in paper (interesting idea for a statistical/programming study if anyone needs a paper topic). If the number of good-to-great hands were somehow lower in paper than online, that would necessitate more grindy games between average hands, and in that situation the deck with the most two-for-ones will prevail. Of course that would mean the shuffling algorithm was somehow "correcting" for bad hands. This seems unlikely, but it's not impossible if the algorithm inadvertently selects for a mix of spells and lands.
This also made the deck very weak to Ancestral Vision. It's bad for a deck that grinds based on value creatures to face counterspells in the first place but counterspells backed by cheap card advantage is a nightmare. Worse, most Vision decks focus on big impact spells like Supreme Verdict and Elspeth, Sun's Champion which are nightmares for this style of deck.
What Does This Mean?
What these results suggest is that DnT needs to maximize the direct value creatures it plays and that synergies aren't particularly important. This means that the Eldrazi Displacers are out of place and should have been additional Finks and Blade Splicers. A deck like Eldrazi Taxes can make greater use of his ability thanks to Eldrazi Temple but any other deck should stay away. It also means that four Restoration Angels are a must to maximize the grinding potential and add more fliers.
I suspect that future mono-white DnT builds will be midrange-slanted to maximize the power of white value creatures and outgrind GBx decks. Whether or not they will be successful is up in the air, and an answer needs to be found for Ancestral Vision. Eldrazi processors might be the answer, but Jim Davis didn't have much success with that so we'll have to wait and see.
The Issue of Disruption
Death and Taxes is a force in Legacy thanks to the power of its mana denial package. Legacy decks tend to be light on actual, mana-producing lands to maximize their fetchland count because of the synergy with Brainstorm. (This has long made me wonder if Leonin Arbiter has a place over Phyrexian Revoker.) The ridiculous power of Wasteland and Rishadan Port, coupled with the tendency of the cantrip-heavy decks to play 18 lands total, means DnT can easily soft-lock other decks.
This is impossible in Modern. Ghost Quarter (even when it impersonates Strip Mine) is much weaker, there's no Port equivalent, and decks run more lands. The weird thing about land destruction is that while it really hurts land-light draws, it turns flood into an advantage and I lost a lot of games where my opponent answered each Strip Mine or Tec Edge with another land. We don't have Ponder and Brainstorm so you can't cheat on lands, and for a deck that wants to constrict its opponents' mana this is very bad news. Thalia is at her best against decks that like to play a lot of non-creature spells every turn and those decks are far less common in Modern than Legacy. The value of your taxing effects is far lower and consequently Modern DnT is far less disruptive than Legacy DnT.
I might not have been taxing my opponent consistently or for more than a turn or so, but the numerous little hiccups DnT induces combined together to make my opponent's play non-optimal enough for the midrange threats to catch me up and win the game. Thalia and Arbiter were frequently killed at sorcery speed so that I couldn't take advantage of them on my turn, which severely crimped my opponents' ability to develop their own gameplan. Aggro decks were often forced to use an entire turn's worth of mana to get past Thalia just so they could attack, and Arbiter proved to be far more threatening to those decks than he was against midrange or combo.
Even the value creatures were disrupting since it took far more to answer them and it could only be done cleanly with Anger of the Gods (the Kor Firewalkers should have been Burrenton Forge-Tenders for that reason). Rather than directly attacking my opponents' ability to play spells, my deck was constricting what spells could be played and that was enough. An early Thalia or Arbiter required an answer, Kitchen Finks needed two, Flickerwisp and Resto blanked a removal spell, and eventually the opponent had spent all their cards and mana trying to answer mine and something finally won me the game.
Of course this really didn't work against decks that just drop big bombs that invalidate incremental advantage. Against them the direct disruption was necessary, and unfortunately when Ghost Quarter wasn't good enough I was just dead. UW Control was a nightmare for this reason, though Jeskai was still good since it relies more heavily on incremental advantage and efficiency. Some decks can directly be answered with DnT creatures, like Aven Mindcensor blanking Scapeshift, but on the whole the lack of direct, universal interaction like Thoughtseize really hurts.
What Does This Mean?
Legacy Death and Taxes players need to accept that the Modern version is going to be weaker against combo and tempo, but in exchange you'll be better against fair midrange decks. This suggests that a complete role reassessment is necessary, which might in turn dictate a large-scale rebuild. I don't think that the other disruptive options will fix the problem I found, but I'll discuss that in greater detail next. I think the long-standing criticism that DnT is bad Merfolk or Zoo is unfounded because it's more directly fighting for space with Jund and Junk. What exactly this means for the deck's future I don't know at this time.
The Issue of Build
As I mentioned last week there are a lot of different possible builds of DnT out there and each has it strengths. BW decks are very disruptive and have much lower curves. The Eldrazi builds have more impressive threats that can be accelerated out. GW decks have general acceleration, impressive creatures, and Gavony Township. My build is fairly middle-of-the-road compared to the other versions. While it seems well suited for grinding it's not that great against more extreme decks, and there is still room for it to improve on currently good matchups. Trying to find the optimal build appears to be a Herculean task.
I don't think that matters. Seriously. Let me explain: All the versions currently agree on the same eight disruptive creatures, Thalia and Arbiter. Everything else is just a complement to those cards. White has the best non-Blood Moon hate in Modern---the question is simply which piece is good when. I'm not just talking about Rest in Peace and Stony Silence. I mean that you can maindeck more direct hate in white than in other colors (no, Thoughtseize doesn't count).
For example, Torpor Orb is very good against a number of decks. There are many more times that the Orb is utterly dead, which is why it doesn't see maindeck play. If you really need that kind of effect (Kiki-Chord is rampant in your local meta, say) you can get the same effect without it being completely dead by playing Hushwing Gryff. Sphinx's Revelation getting you down? Spirit of the Labyrinth laughs at that card. If Abzan Company is running riot, Samurai of the Pale Curtain trades or defeats every non-Wall of Roots card in the deck and breaks up the combo. If you need to hate on a specific deck/card, white has an answer attached to a maindeckable creature.
What Does This Mean?
I think that DnT is a deck that can beat any deck individually, but can't be built to beat everything simultaneously.
I'm seeing nods of understanding and confused, frightened stares in equal measure. Let me explain. I'm not certain that it is possible, or even preferable, for a stock version of Death and Taxes to exist in Modern. In Legacy the ubiquity of Brainstorm means that decks have enough in common for the same kinds of hate to be equally effective against a wide swath of the metagame. Modern is too diverse for that to be true, so instead you need to focus your deck to beat an expected metagame and then adjust week to week.
It's a tuner's deck, pure and simple. If you think that Jund is going to be highly played one week then Loxodon Smiter is where you want to be. Heavy Burn necessitates mono-white. Lots of combo calls for BW. You need to be on top of metagame shifts and adjust accordingly.
A Few Other Thoughts
- Inspector Clouseau (Not my nickname for Thraben Inspector, but I like it) is very good. Unless the metagame gets less fair it is the best one-drop available. A 1/2 for one is better than you think on the defense and the Clue is very important when you have no two-drop or as you head to the lategame.
- Fliers are surprisingly underrepresented in Modern. The more you can justify playing, the better.
- Flickerwisp is surprisingly hard to play optimally. There have been several occasions where post-match analysis has shown that non-intuitive chains involving Flickerwisp and Restoration Angel would have won games against Affinity and Goblins that I lost. There are times when maximizing the number of your own creatures flickered has been correct and sometimes you just want to take the opponent's lands. If you're going to play this card, you have to learn to look at all the angles and really work to extract maximum value from the card.
- If your own Leonin Arbiter gets Path to Exiled and there's no other Arbiter on the field, you get to search without having to pay.
- Avoid Strip-Mining early, except against Tron and Scapeshift. I know it's the big appeal of the deck but it is much harder to mana screw your opponent in Modern than in Legacy, and even if it sets your opponent back it stunts your mana as well. In a deck so full of three-drops that can be fatal.
There is real potential to be found in Death and Taxes in Modern, and I am not abandoning the deck by any means, but I am putting it on the shelf for a while. When the format is still in considerable flux it's not the time to play such a metagame-dependent deck. Once things get a little more settled and it is possible to really target your deck then it will be worthwhile to try again with specialized builds. If this proves successful it might be an interesting choice once the PPTQ season rolls around.
As always, if you have your own take on the deck or experiences different from mine, I'm eager to hear from you in the comments and our Contribute page is always open.