I’ve been working on original Magic decks for years now, and one of the most important things I’ve learned in building new decks is this: you have to know the decks that you expect to face.
You can have the most creative deck in the world. It means very little, though, if it doesn’t have a shot against the field that you’re hoping to play against. More than anything, you have to have prepared enough to have a deck that isn’t good against just some deck, but things that are actually like what you hope to be playing against.
My first real advantage in this came about a million years ago. I’d gotten a tip from Rob Hahn (then working within Wizards) that the big card for Exodus was going to be Oath of Druids. He imagined a deck with “Verdant Force/Archangel/Sliver Queen/Benthic Behemoth” as “Basically, it’ll be a U/W control deck possibly with Tutors to go get the
Oath…” He was way off by a lot, but it did inspire me when I built the Counter-Oath deck that blew the U.S. Open to bits that year. But, the big thing was this: what do I test it against?
From here came the huge tip. Back in the day, Regionals information was not a public resource. If someone wrote a tournament report on the Dojo or USENET, that would be the way that you’d be able to know about a deck. A friend of mine had gotten all of the Regionals qualifying decklists. Every one. There were so many decks to test against, but how could we know we were even doing at all reasonable testing? Which deck could we choose? The Madison crew didn’t have infinite time, and so we had to marshal our resources. From this, came my first experiments with amalgam decks.
This early amalgamating was not particular sophisticated. Take two decks (or more) of the same archetype, and look at them, judging for yourself what they did right or wrong, and put all of the right bits back together again. Sometimes, two decks of the same archetype are so wildly different, you really just have to accept that they are distinct and not try to smoosh them together. This helped narrow down all of the Sligh decks, and all of the Blue/White control, and all of the everything into much more manageable sizes. My Oath deck had a successful showing, and I got to meet Dave Williams and a number of my fellow Cabal Rogue members in the flesh finally. It was a great weekend, and a large part of the reason was the preparation that went into the final form of the deck, honing it against the amalgamated versions of the standard archetypes. I look at the final result, and I can’t help but think of all of the ways that the deck could have been so much better, but deck building was a rougher art back then…
In thinking about this, I can’t help but think back, way back to Pro Tour: Tokyo in 2001. Brian Kowal and I designed this deck:
3 Darigaaz’s Caldera
6 Mountain
3 Shivan Oasis
11 Swamp
1 Urborg Volcano
4 Blazing Specter
3 Crypt Angel
3 Flametongue Kavu
4 Nightscape Familiar
4 Ravenous Rats
3 Skizzik
4 Thunderscape Battlemage
3 Bog Down
3 Ghitu Fire
3 Terminate
2 Void
Dave Williams would take it to his first Pro Tour Top 8, and the O’Mahoney-Schwartz’s would get their best Constructed finish to date with it. The big change that they made (one that Dave would tell me was a huge error), was removing a Flametongue Kavu for a second Void. I had told Dave that 1 Void was the tried and true properly tested number, but it didn’t feel right to them, and they moved it around a little bit, eventually agreeing that it was a small mistake.
Dave would later update the deck and do quite well with it in Grand Prix, though once the format in this form had shifted, Brian Kowal and I realized that the big error in the deck had been not to include three or four Thunderscape Familiar. Yes, Thunderscape Familiar, the 1/1 for 1R with first strike, and an amazing ability to reduce the cost of the Black spells in the deck. The reduction of costs was great, especially when it came to Blazing Specter, but the additional, huge factor was that it was a 1/1 first striker. A quick look at all of the Red/Green decks with X/1 creatures will show just how big of a deal that this is.
When Brian and I made this deck, we were fortunate enough to be working with a lot of people who were throwing out their ideas of what various decks might look like. There were the regular gang of Cabal Rogue members and friends. There was our local crew, then including Bob Maher. Mike Hron couldn’t be convinced to play our deck, opting instead for a much more typical Black/Red deck, but what he did do was regularly throw bizarre concoctions our way that greatly impacted a lot of our decisions in how to build the deck.
It was a new Pro Tour format, without anything to really base our decisions on, so we largely based it on what little information that we could get on what the format might look like. We guessed a little here and there. We would weasel intel out of people on IRC and elsewhere. We did the best we could, and I think we did pretty well. Our Black/Red deck was the only one of many to make the Top 8, and we did it by both having reasonable intelligence on what might appear, and not getting blindsided by what were the “powerful” cards in the format — cards that didn’t serve the other four Madison folk who got swayed by Hron’s extra Voids and Pyre Zombies.
Zvi Mowshowitz went on a tear with his deck, The Solution, largely because he totally called the format for Red/Green. His deck is perhaps the best example of a deck put together to just K.O. the expected archetype.
4 Coastal Tower
10 Island
10 Plains
4 Crimson Acolyte
4 Galina’s Knight
4 Meddling Mage
4 Stormscape Apprentice
4 Voice of All
4 Absorb
4 Exclude
4 Fact or Fiction
4 Repulse
But, this event also had another big story. Alpha Beta Unlimited. They would accomplish a feat that I’m not sure has ever been accomplished since: they put four Red/Green decks into the Top 8. That’s right. Half of the Top 8 was their Red/Green deck.
Only one problem. They weren’t card for card copies of each other. In fact, they were all quite different.
It was at this event that I did my first real amalgamating.
The lists:
| 10 Forest | 9 Forest | 10 Forest | 9 Forest |
| 1 Keldon Necropolis | 2 Keldon Necropolis | 1 Keldon Necropolis | 2 Keldon Necropolis |
| 9 Mountain | 9 Mountain | 10 Mountain | 9 Mountain |
| 4 Shivan Oasis | 4 Shivan Oasis | 4 Shivan Oasis | 4 Shivan Oasis |
| 4 Blurred Mongoose | 4 Blurred Mongoose | 4 Blurred Mongoose | 4 Blurred Mongoose |
| 2 Flametongue Kavu | 2 Flametongue Kavu | 2 Flametongue Kavu | 2 Flametongue Kavu |
| 4 Kavu Titan | 3 Kavu Runner | 4 Kavu Runner | 2 Kavu Runner |
| 4 Raging Kavu | 4 Kavu Titan | 4 Kavu Titan | 4 Kavu Titan |
| 4 Skizzik | 4 Raging Kavu | 4 Raging Kavu | 4 Raging Kavu |
| 4 Thornscape Battlemage | 3 Skizzik | 3 Skizzik | 4 Skizzik |
| 4 Thornscape Familiar | 4 Thornscape Battlemage | 3 Thornscape Battlemage | 4 Thornscape Battlemage |
| 2 Yavimaya Barbarian | 4 Thornscape Familiar | 4 Thornscape Familiar | 4 Thornscape Familiar |
| 4 Ghitu Fire | 4 Ghitu Fire | 3 Ghitu Fire | 4 Ghitu Fire |
| 4 Urza’s Rage | 4 Urza’s Rage | 4 Urza’s Rage | 4 Urza’s Rage |
| Ryan Fuller | Phillip Freneau | Lucas Hagar | Chris Benafel |
It’s clear that these decks were all designed together. What isn’t clear, though, after the dust settled, is what you should test against. Should you test against Ryan Fuller, who had the number one seed in the swiss? Should you test against Lucas Hagar, who ended in third place?
It seemed clear to me that you should probably test against a version of this deck that was as “average” as could be. Here is the “average” (mean) deck that you get out of these decks…
9.75 Forest
1.5 Keldon Necropolis
9.5 Mountain
4 Shivan Oasis
4 Blurred Mongoose
2 Flametongue Kavu
2.25 Kavu Runner (the inclusion of this card seems to be a true point of contention)
4 Kavu Titan
4 Raging Kavu
3.5 Skizzik
3.75 Thornscape Battlemage
4 Thornscape Familiar
.5 Yavimaya Barbarian
3.75 Ghitu Fire
4 Urza’s Rage
A quick glance at the clear decisions of the deck gives us this mana:
10 Forest
1 Keldon Necropolis (with a potential +1)
9 Mountain (with a potential +1)
4 Shivan Oasis
24 Land
Because of the expected 24 mana, this would seem to give us a manabase of 10 Forest, 1 Keldon Necropolis, 9 Mountain, and 4 Shivan Oasis. Continuing to the spells (36 of them), we get:
4 Blurred Mongoose
2 Flametongue Kavu
4 Kavu Titan
4 Raging Kavu
3 Skizzik (and maybe one more)
4 Thornscape Battlemage
4 Thornscape Familiar
4 Ghitu Fire
4 Urza’s Rage
33 spells. This leaves us our question marks.
Only Fuller chose to run Yavimaya Barbarian (2), and not run Kavu Runner at all. Further, the number of Runners is erratic for the rest of them, counting at 2, 3, and 4. Since the deck seems like it might want to run Runner, we could try simply averaging the three that did run it, and we’d fill out our 60 cards with 3 Kavu Runner. There is still that nagging question of maybe running a 4th Skizzik. The real question becomes, then, 3/3 on Skizzik/Runner, or 4/2? It becomes worth noting that each of these three players had access to a full set of 4 and 4 for both cards after boarding. Relooking at these three players, for their averages on these two cards, I find that it is 3 on Runner and 3.33 on Skizzik. I’m fine with that split, especially from the question of curve consideration. That gives us this final list.
Amalgam Red/Green Invasion Block
4 Blurred Mongoose
2 Flametongue Kavu
4 Kavu Titan
4 Raging Kavu
3 Kavu Runner
3 Skizzik
4 Thornscape Battlemage
4 Thornscape Familiar
4 Ghitu Fire
4 Urza’s Rage
10 Forest
1 Keldon Necropolis
9 Mountain
4 Shivan Oasis
It becomes worth nothing that this list is only one card off of Freneau’s list (exchanging a Forest for a Necropolis).
This is not an exact science, but it is a lot better than deciding that you like some particular list out of that hat on a bunch of similar lists.
Often, deck tweaks are made because of specific concerns that a deckbuilder might have. Fuller’s list seems to clear have decided that he cared about Blue, for whatever reason. His list might be just the list to test against if you are trying to give your Blue deck as hard a time as possible, or it might be the list to test against if you know that the field you’re expecting will obviously include more Blue, and so a rational player might be more prone to a list like Fuller’s.
This is in many ways comparable to looking at various different versions of Goblins. Often they are trying to accomplish a certain thing for a certain meta, and thus their decisions are informed by that. Unfortunately, if they aren’t right there in front of you, you can’t ask them their intentions.
In those cases, it is simply better to amalgam the deck to get a test-deck rather than just pick a particular version out of a hat.
Let’s see what we can do with amalgamating when we apply it to something more current, like Dredge from Grand Prix: Vienna.
Dredge is actually a perfect deck to amalgam for playtesting purposes. The decks can actually perform wildly different depending on the makeup of the deck. Testing against just one version of the deck can be deceptive. How well your deck performs against it could be a function of whether or not they run Tireless Tribes, how many dedicated reanimation targets are run, and what they are, and any number of other factors. The best bet would be to test against them all, but sometimes, you simply cannot do that. The other best bet would be to test against the version of the deck that you know you will see because some particularly skilled player in your area will be running it. Unfortunately, you can’t always get that particular list to play. Rather than pick a list out of a hat, amalgamating is a method that will get you a reasonable version of a list, and have it be reasonably close to various versions that you might play against, thus giving you closer estimations to actual lists if you don’t have the time to test against each version.
So, let’s check out the lists!
The land is really interesting. Winkelmann is the only person choosing to go with the pain land approach (as a means to support Tireless Tribe). We can’t really smush together the mana of both, so while I’ll average his total number of land, to get the average number (15), I won’t use his land numbers to figure out land.
Averaging out the number of search lands of the remaining three gives us 5.667, an easy round to 6. This easily works out to running Saitou’s exact land:
4 Cephalid Coliseum
3 Flooded Strand
1 Island
3 Polluted Delta
4 Watery Grave
15 Land
The spells shows a remarkable degree of agreement:
1 Akroma, Angel of Wrath
1 Cephalid Sage
1 Flame-Kin Zealot
4 Golgari Grave-Troll
3 Ichorid
4 Narcomoeba
4 Putrid Imp
4 Stinkweed Imp
4 Breakthrough
4 Bridge from Below
4 Careful Study
That’s a whopping 34 spells that these decks are in complete agreement about. That is pretty remarkable. This leaves us 11 slots to figure out.
Here are our averages:
3.25 Cabal Therapy
1.25 Golgari Thug
1.5 Darkblast
2.75 Dread Return
1.5 Tolarian Winds
Winkelmann’s Tireless Tribes are partly to blame here, making him the lone dissenter on 2 Darkblast and 2 Tolarian Winds. Taking the majority opinions, we have these numbers:
3 Cabal Therapy
1 Golgari Thug
2 Darkblast
2 Tolarian Winds
3 Dread Returns
Our 11 slots exactly.
And so we have Amalgam Dredge:
3 Cabal Therapy
1 Golgari Thug
2 Darkblast
2 Tolarian Winds
3 Dread Returns
4 Cephalid Coliseum
3 Flooded Strand
1 Island
3 Polluted Delta
4 Watery Grave
1 Akroma, Angel of Wrath
1 Cephalid Sage
1 Flame-Kin Zealot
4 Golgari Grave-Troll
3 Ichorid
4 Narcomoeba
4 Putrid Imp
4 Stinkweed Imp
4 Breakthrough
4 Bridge from Below
4 Careful Study
In this case, we have a rare situation. The Amalgam Dredge list is exactly Tomoharu Saitou’s list. I have to say, I’m not surprised…
The thing to remember about amalgamating is that it is a shortcut. If you had the time and manpower, you really should test against various versions of all of the different decks that you can get your hands on. An amalgam list is good for exactly one thing: approximating what a deck with numerous variations can look like. Testing against an amalgamated list is going to give you the best approximation of what your results might be against numerous lists.
The basic process is looking at the average (mean) to get a sense of what the deck will generally look like and using the average (median) to hone out the weird numbers, throwing out outliers to fit in more general concerns. If a deck is clearly using analogs (Polluted Delta and Flooded Strand), recognize that your card counts should sometimes be 6 or 7 or more, depending. You want to end up with a deck that isn’t a total mishmash of differently conceived approaches, so remember to throw out certain elements of decks that just don’t fit the mold of the amalgamated archetype.
Sometimes we don’t have all the time resources we’d like to allocate to testing. Amalgamating is a great way to make more time, and get the most use out of a limited time budget.
Until next week!