I’ve long been a proponent of deck amalgamation for discovery of approximating proper builds for Magic decks. As I’ve written before, I first started working over common Decklists back in the wake of Pro Tour Tokyo, 2001, when Team ABU had four members in the Top 8 with Red/Green Beats(no, not “Stompy” — “Heat”, if one must…). Fairly recently, Frank Karsten received notice for playing “Average” Faeries, a composite list of all of the successful Faeries decks he liked.
This idea of averages is not completely unreasonable to sleeve up and play. Typically, though, it is most useful in trying to employ an element of Magic’s hive-mind and collective intelligence. Not all of us are fortunate enough to, for example, have Paulo Vitor Damo da Rosa and Luis Scot-Vargas sit down with us together and hash out the list. Their debate would likely bring them to a pretty close consensus on versions of Swans, albeit with a decent likelihood of divergence. What you can do is compare their lists, and see where they agree and where they do not.
The result of this is not necessarily going to be the correct deck to play. What it is, however, is a deck that will behave most like a random version of a list of an archetype. Comparing Joel Calafell’s list (1st place), for example to Luis Scot-Vargas’s list (40th place), it is easy to see that there are major differences. LSV, for example, was the only player featured in Wizard of the Coast’s coverage who was playing with the Deny Reality plan (a plan, I believe, developed by Rashad Miller, if some reports are correct). Trying to reconcile LSV’s list into a list of other decks as a question of “averages” would give you a list with 1 Deny Reality. While the list you have in your hands, then, is “average,” it isn’t a proper way to amalgamate your lists. Really, you need to decide, do you want to play against the LSV build or not? Further, though, you can decide to play with his exact list, or a list like his, informed by amalgamation, or other non-Deny Reality lists informed by amalgamation with LSV.
If you’re trying to craft a decklist, you want it to be just that: a deck list. Averages alone don’t capture what you need to do when crafting a good list, you need to look further than that. If, for example, we look at the people who were covered by Wizards, we’d have the following list:
Joel Calafell (JC)— 1st Place
Paulo Vitor Damo da Rosa (PV)— 3rd Place, playing the exact same spells/sideboard as Calafell (heavy praise for those spells!)
Hugo De Jon (H)— 5th place, 1st in the Swiss, losing only 2(!!) games until the Top 8
Rui Casas Novas (R)— 74th place, undefeated day 1
Luis Scot-Vargas (LSV) — 40th place, top (covered) player playing Deny Reality build
So, if we look solely at their spells, we’d have:
Spell — JC / PV / H / R / LSV /// Mean
Bloodbraid Elf — 4/4/4/3/4///3.8
Swans of Bryn Argol — 4/4/4/4/4///4.0
Ad Nauseam — 2/2/3/2/0///1.8
Bituminous Blast — 2/2/0/2/0///1.2
Captured Sunlight — 2/2/4/4/2///2.8
Deny Reality — 0/0/0/0/4///0.8
Primal Command — 1/1/1/0/2///1.0
Seismic Assault — 4/4/4/4/4///4.0
There is a fair amount of agreement here, but, some of the particular details of the card advantage engines can be broken down like this:
4CC Cascade — 6/6/8/7/6///6.6
5CC Cascade — 2/2/0/2/4///2.0
Total Cascade — 8/8/8/9/10///8.6
Total plus Ad Nauseam — 10/10/11/11/10///10.4
So, really, there is a remarkable amount of agreement, here, about what is the appropriate set of cards to be running. There should be 10 or 11 “card advantage” cards, to go with the clear core of 8 combo cards and 1 Primal Command. This leaves us our 41 or 40 land.
So, what should the amalgamated versions of these lists look like? Initially, something like this:
4 Bloodbraid Elf (sorry, Rui)
4 Swans of Bryn Argol
4 Seismic Assault
1 Primal Command
2 Ad Nauseam
2 Bituminous Blast
2 Captured Sunlight
41 Land
LSV’s build essentially drops the Ad Nauseam and Bituminous Blast for 4 Deny Reality, to go with a land-attack plan. Both of these builds are straight-up the lists of PV/JC and LSV. Really, the strength or weakness of this plan as compared to the more straightforward plans offered by the other decks is something that you’ll have to grind out to see who is right. An Ad Nauseam is likely to deal 1.2 damage per card drawn — not too shabby. If your hand was already particularly full of spells, it could do as little, on average, of .9 or so. A big part of the question is, what are you looking for? In many cases, the answer is fairly simple: you’re looking for the next spell (or two). With Cascade, you can expect to just have a Cascade spell push you into whatever you need next, though typically, it will be a Seismic Assault in the end. The other thing you could be looking for, since you can count on Assault, is just a ton of land. The other thing you can be looking for is a counterspell — from your opponent. Here, the Ad Nauseam is seemingly used as a Mana Short of sorts. Counterspells, though, are actually fairly difficult for the deck, often making the deck resort to the beatdown mode. Part of what makes the LSV plan so interesting is the way that it fights against counters by attacking their resources.
What we have for variance in the lists largely comes from a question of their card advantage resources. Hugo De Jong, for example, up the Ad Nauseam count to three, likely because he wants to be able to reduces the reliance on the Seismic/Swans combo, and increase the possibilities of a Seismic/Ad.N. combo (draw 8 land, getcha). Rui Casas Novas, on the other hand, takes a strange middle road where he eschews the potential Bloodbraid plus manland draw for a full set of Captured Sunlight. Of the lists, Casas Novas was the real outlier, running no Primals and only three Bloodbraid. He’s also the only one of the 5 to not make it in the money, although 74th out of almost fifteen hundred is certainly no mean feat (top 5%).
The real place where there is massive disagreement is in the land, and this is no surprise, given the divergences in spells. Let’s examine the mana requirement of the various builds:
JC/PV/H: RRRBBWGG(W/U)
R: RRRBBWG(W/U) — one less G requirement
LSV: RRRBGGWU — one less B requirement, one more specific U requirement
These are all, still, very similar. So what makes up the breakdown of their mana? It almost sounds like an incredibly boring question, but when you consider that mana is two-third or more of the deck, this is actually a very critical question.
Land — JC/PV/H/R/LSV///Mean
Forge[/author]“]Battlefield [author name="Forge"]Forge[/author] — 2/2/0/0/0///0.8
Cascade Bluffs — 2/2/0/4/0///1.6
Fire-Lit Thicket — 4/4/4/4/4////4.0
Ghitu Encampment — 2/2/3/4/4///3.0
Graven Cairns — 4/4/3/4/4///3.8
Mountain — 1/1/2/1/1///1.2
Mutavault — 0/0/2/1/0///0.6
Reflecting Pool — 4/4/4/4/4///4.0
Rugged Prairie — 0/4/4/4/3///3.0
Spinerock Knoll — 4/4/0/3/0///2.2
Sulfurous Springs — 0/0/0/0/1///1.0
Sunken Ruins — 0/0/0/0/2///0.4
Treetop Village — 4/4/4/4/4///4.0
Vivid Crag — 4/4/4/4/4///4.0
Vivid Creek — 0/0/0/0/1///0.2
Vivid Grove — 4/4/4/0/4///3.2
Vivid Marsh — 2/1/0/4/4///2.2
Vivid Meadow — 4/1/4/0/0///1.8
Wooded Bastion — 0/0/2/0/0///0.4
Total Land — 41/41/40/41/40///40.6
Clear agreement exists on some of the cards. At large, we can mostly (though not entirely) agree on:
4 Fire-Lit Thicket
4 Graven Cairns
4 Reflecting Pool
4 Treetop Village
4 Vivid Crag
From there, though, there are other considerations to begin looking at.
Amount of RR land — 10/14/11/16/11///12.4
Amount of WW land — 0/4/6/4/3///3.4
Amount of UU land — 2/2/0/4/2///2.0
Amount of BB land — 4/4/3/4/6///4.2
Amount of GG land — 4/4/6/4/4///4.4
Amount of total filters — 10/14/13/16/11///12.8
Amount of “dedicated” R land — 23/27/20/28/23///24.2
Amount of “dedicated” W land — 6/7/10/8/6///7.4
Amount of “dedicated” U land — 2/2/0/4/3///2.2
Amount of “dedicated” B land — 6/5/3/8/11///6.6
Amount of “dedicated” G land — 12/12/14/8/12///11.6
Amount of Man-Lands — 6/6/9/9/8///7.6
Amount of Vivid/Pool — 18/14/16/12/17///15.4
Amount of CIP-tapped — 20/19/19/19/21///19.6
Okay, that’s a lot to think about. Are there any conclusions that we can really run from this, though?
On of the first things that I find really interesting is that LSV’s list, despite actually having a significantly greater need for blue does not run any significantly larger amount of blue mana. Hugo De Jong is a huge outlier here, running no dedicated blue mana. At the other end, we have Rui Casas Novas with his four Cascade Bluffs.
Another thing that sticks out in my mind after some observation is just how close PV is to the averages of this list. He’s slightly different, in some places, but is primarily on board with the average numbers. In fact, if we go through the mana, correcting for outliers (like Joel Calafell running 0 Rugged Prairie), we get the following:
4 Fire-Lit Thicket
4 Graven Cairns
2 Cascade Bluffs
4 Rugged Prairie
1 Mountain
4 Reflecting Pool
3 Ghitu Encampment
2 Spinerock Knoll
4 Treetop Village
4 Vivid Crag
3 Vivid Grove
2 Vivid Marsh
2 Vivid Meadow
2 ???
This is only a slight nudge on numbers from where they were, and certainly seems to dodge uber-excesses of filtering and egregious use of comes-into-play-tapped. We need 2 more land, yet, and they should generally be neither of those types, if we want to pay attention to the successful lists.
Two lands? Our candidates from the lists, above, are as follows:
Forge[/author]“]Battlefield [author name="Forge"]Forge[/author]
Mountain
Mutavault
Sulfurous Springs
Seems pretty grim, honestly. The amount of pain lands, in total, average at 1, so maybe we can do something with that Forge[/author]“]Battlefield [author name="Forge"]Forge[/author]. The final slot? At this point, we’re really down to preference. For my part, I’d like to get that man-land count up, and throw in the Mutavault or slip into comes-into-play-tapped territory.
With that said, here is one version of an amalgamated Swans list. Your mileage may vary…
4 Bloodbraid Elf
4 Swans of Bryn Argol
4 Seismic Assault
1 Primal Command
2 Ad Nauseam
2 Bituminous Blast
2 Captured Sunlight
4 Fire-Lit Thicket
4 Graven Cairns
2 Cascade Bluffs
4 Rugged Prairie
1 Mountain
4 Reflecting Pool
1 Forge[/author]“]Battlefield [author name="Forge"]Forge[/author]
4 Ghitu Encampment
2 Spinerock Knoll
4 Treetop Village
4 Vivid Crag
3 Vivid Grove
2 Vivid Marsh
2 Vivid Meadow
Essentially, this is the PV list, with a slight mana change. Dropping 1 Forge[/author]“]Battlefield [author name="Forge"]Forge[/author], 2 Spinerock Knoll, and a Vivid Grove for 2 additional Ghitu Encampments, and an extra Vivid Marsh and Vivid Meadow. Alternately, you can think of it as Joel Calafell’s list, with —1 Forge[/author]“]Battlefield [author name="Forge"]Forge[/author], -1 Vivid Grove, -2 Vivid Meadow, and —2 Spinerock Knoll for +2 Ghitu Encampment, +4 Rugged Prairie, depending on how you choose to think about it.
The thing about a decklist like Swans, there really is a lot of room for play with regards to your spells, but we can actively see what was generally done, and try to make a list that will perform as optimally as possible for testing against. We want to have a deck, when it comes to our testing, that will act in ways that is closest to “a Swans deck” rather than a particular one. Optimally, we would test against all of the versions, but few among us really have the time.
The subtleties of these mana differences are real. If, for example, you’ve decided that your way to attack a particular list is to go after their primary mana producers and leave the filters, you might find this very effective versus the Rui Casas Novas list, but underwhelming versus Joel Calafell. Other strategies, for a controlling deck, might be to go heavily into counters and reduce the resistance to creatures. This might be a fair strategy versus Joel Calafell or Paulo Vitor Damo da Rosa, but could easily be a mistake versus Hugo De Jong or Rui Casas Novas. The amalgamated deck provides something that is at least close to each of these lists so that you can approximate them all, provided they are a “radical” departure, like LSV’s.
The next thing that is going to happen in the evolution of Swans, I’m sure, will be ways to make the mirror match stronger. Perhaps this could be an increase of sideboard cards like Aura of Silence and Pithing Needle, or perhaps it will be an increasing reliance on some other element. In testing for this mirror, I think it would be fairly critical to have a reliable “baseline” version to test against so you don’t over-inbreed your testing.
Good luck to everyone with an upcoming Standard event!