So Many Insane Plays - Grand Prix Madrid Nearly Killed Me, and I Wasn't Even There

This was the most insane Magic article-related project I’ve ever undertaken.

When the Wizards coverage team said that they wouldn’t be giving a metagame breakdown of the GP: Madrid field I shot an email to Aaron Forsythe, telling him that I’d do it. I’d done it for each of the 2009 SCG Legacy Opens, so I figured it wouldn’t be that big of a deal. Sure, it would take more time, but it would be worth it. I could data mine the largest Magic tournament in history.

It took a month for the decklists to arrive. The first thing I did was type up the Top 64 decklists (see Appendix). That only took a few hours. I was ambitious enough to think that I could create a spreadsheet labeling each archetype to each player’s name on the player list, just as I had done for the SCG Opens. I could then do a number of fancy things. Someone could then do a complete analysis of every matchup played in the tournament. I could also generate graphs showing the distribution of archetypes throughout the field. That proved to be very slow-going.

I decided to simply tally up the entire field by archetype with one simple goal: create a complete metagame breakdown, and then try to learn which archetypes outperformed the field and how they performed relative to each other. There has been a great deal of debate in the Legacy community about the relative strength of a number of archetypes, from Zoo, to Merfolk, to Ad Nauseam, to Dredge. Is Zoo a good deck? Or does it simply seem to perform well because it’s so popular? Conversely, is Dredge a good deck? Or does it simply underperform because the pilots are less skillful and less experienced? And so on.

The StarCityGames.com tournament data provides empirical answers to a number of these questions. For starters, there have been 10 SCG Opens, more than enough tournaments to give us a very clear idea of what archetypes tend to Top 8. For example, we know that in the first nine SCG opens, Dredge only made Top 8 twice. That’s 2 out of 72 possible Top 8 slots. When we compare this data to the metagame breakdowns of the field, we can get an even better idea of what performs and what doesn’t. For example, in the first 9 SCG Opens, there were a total of 81 pilots playing Dredge. Only 2 of those 81 made Top 8. That tells us something about the strength of Dredge, particularly when we compare it to other archetypes using similar metrics.

Starting last fall I began to extract complete matchup data (including all swiss pairings) by archetype, and Jared Sylva has carried this work even further than I had. This data, aggregated into large samples, is statistically significant. This data is even more revealing than Top 8 statistics because we begin to have a concrete sense of which archetypes tend to win which matchups. When you have 100 recorded matchups between Zoo and Merfolk in the last three months, and the results from each tournament are telling us very similar things, the data is very persuasive.

In spite of all of this, there appears to be an anti-empiricist wing in the Magic community. These players raise doubts about the strength of the SCG field and the quality of the pilots and other unknown and unknowable variables, which calls into question the basic reliability and predictability of this data. One line of thinking goes something like this: some decks are clearly more difficult to pilot than others, and therefore we can’t credit even large sample sizes. It’s probably not the case that most Zoo pilots are incompetent, so the Zoo matchup data is reliable, as a body of data. But the Dredge or ANT pilots are much worse, on average, and therefore the data is more likely to mislead.

There are a number of ways to rebut these claims. I can think of no good reason to think that the skill distribution (whatever that curve may look like) for Dredge pilots is somehow substantially different than the skill distribution for any other major archetype. Why would Dredge pilots, on average or in the main, be any worse than CounterTop pilots? Even if the archetype is more difficult to play, the distribution of skill among the archetype shouldn’t be that different. Perhaps more importantly, even if the skill distributions were very different, such that there were much fewer high-end players of certain archetypes, other data — not simply matchup data — would tell us this. For example, we would see that some percentage of an archetype — particularly when piloted by well-known and highly regarded players — would make Top 8. Yet, with Dredge’s tiny Top 8 percentage casts doubt on that. Can we really say that 79 of the 81 Dredge pilots were bad? I find that to be highly implausible.

But more deeply at issue is a much older debate, a centuries old dilemma. It’s an epistemological question, not a data question. Magic players know it by the old label “results-oriented.”

Results Oriented? Theory versus Data

The Scientific Revolution fused two competing schools of thought: Cartesian rationalism and Baconian empiricism. These schools of thought can be understood in many different ways. In our vernacular, it’s the theory versus results. Which do you trust? Which do you credit? Descartes put his faith in reason and theory. He believed that he could deductively understand the universe through pure reason. Francis Bacon and others believed that empiricism and data was the route to truth, an inductive process. He was fundamentally an experimentalist. Gather lots and lots of data, then pose hypothesis and test it with experiments. Newton fused both traditions, the inductive and deductive traditions, the rationalist and the experimentalist schools, into modern science.

Nonetheless, there remains a tension between the two schools, despite the Newtonian synthesis. At what point do you credit data? When data does not accord with theory, it is often easy to find a reason to discount the data.

A few years ago, Richard Feldman wrote a seminal defense of being Results Oriented. Mark Young took him to task in the forums, and many other since — particularly professional Magic players — have critiqued the idea of being â€˜results oriented.’ For the most part, the two camps have talked past each other. As Gerry Thompson says, “Just because you flipped a coin and it came up heads does not mean that the coin flip will always come up heads.” Yet, if we flip a coin 100,000 times, we probably have a good idea when it will come up heads — or at least, how often.

At core is this: Magic players, especially professional Magic players, have a highly developed sense of theory. They have a great deal of experience across formats and over time, and they have well developed frameworks for analyzing data and understanding metagames. Then, they test with other pros and develop further theories about the metagame and a format. When data — particularly results like an SCG Tour^® naments — does not accord with the theory, what is a Pro to do? You probably trust your skills; you trust your knowledge. And that’s exactly what the argument is: it’s an appeal to authority, for sure, but it’s built upon the notion that Pros â€˜know better’ because of their greater experience and skill. Shouldn’t we credit LSV’s opinion over the results of a few large tournaments?

Maybe. Maybe not.

A few years ago, I started applying systems theory principles to Magic. One of the key lessons of system theory is that human beings are notoriously bad at operating in complex systems. Professor Sterman at the MIT Sloan School of Management has done an excellent job showing why and how our cognitive processes fail us with various modeling.

In short, humility is required. Theory is always incomplete. The reason that central bankers, military strategists, Wizards R&D, and investment bankers often fail or come up short is the same. The system is dynamic, complex, and constantly evolving. Every Pro player’s theoretical framework is based upon data that they’ve analyzed. Yet, when every other player is trying to outsmart the field, the field/market/opponent isn’t going to sit still. There are numerous dynamic feedback processes.

Let me put it in Magical terms. Every player arrives at a tournament with a deck that they’ve tested and prepared with. In the ideal world, they’ve devised plans for each matchup and they’ve predicted the metagame. Yet, if each player does the same thing, not every player can be right, or else more people would win than lose. Yet, for every win there must be a loss. What happens is that the metagame evolves. Matchups change the moment of the tournament. If a Zoo pilot believes they are favorable against Merfolk, and go into the tournament with that expectation, they may be surprised to lose to a Merfolk pilot that devoted unusual sideboard space to the matchup. The theory — the paradigm — that the Magic player brings to the tournament is outdated unless it accurately accounts for all of the adjustments that every player will make, both in terms of deck choice and deck design. This is, of course, impossible. And this is, of course, why I have long believed that metagaming — anticipating or projecting the metagame — is the most important skill in Magic. It conditions all other decisions.

When Player X tests Dredge and finds that they win most matchups, I question their methodology. Dredge pilots may do great in testing, but poorly in tournament, and I believe there are a number of important reasons for this. First, Dredge pilots in testing know what their opponent is playing, and more importantly, what they are running against them. In actual tournament conditions, you often don’t know what your opponent is playing, nor exactly what sideboard solutions they are running. If you guess wrong and bring in the wrong answer to their hate, you could easily lose a match. Yet, in testing, these are givens. The key skill with dredge is defeating your opponent’s hate. Testing conditions do not and cannot accurately test this skill, which only matters for a 2/3 of game match in a tournament.

This is, again, why I credit tournament results, and why I will always trust them more than any individual’s opinion, no matter how skilled they are at Magic. Tournaments are like markets. They are complex systems. Even the smartest investors sometimes can’t outperform the market (Warren Buffet has a famous bet to that effect). Tournaments present conditions that cannot be fully tested in a lab or under experimental conditions (which is what playtesting is). Yet, I don’t fetishize tournament results.

I understand that all of the results that have occurred in the past in no way determine tournament results of the future. Past results cannot predict — with any precision — future results. However, they do structure and constrain future results. They set the conditions under which future results will occur, and they help us understand how future results may play out. They suggest avenues for which the future may unfold. That is the value of being results-oriented. It helps us discredit theories that have failed. And it helps us better understand metagame dynamics so that we may begin to understand how the metagame may evolve moving forward. For example, if the data so far tells us — as it does, that Ad Nauseam beats Zoo and that Zoo beats Merfolk, this becomes the initial conditions for which the metagame may evolve. Zoo and Merfolk pilots may attempt adjustments that change this basic matchup result, and they may be successful or not. We may not know exactly where the metagame is evolving, but we can at least foresee possible trajectories based upon current data. And we can use this to plan and prepare. And, if you are more accurate than your competitors in your planning, you will be a winner.

GP: Madrid is a unique case study both because of its size and because of the presence of a greater absolute number of professional Magic players. It offers an opportunity to look at a dataset that is totally unique and offers highly skilled players throughout the field. But the tallying took much longer than I expected. Hour after hour passed and I barely made a dent in the pile. I watched three seasons of Lost tallying up the GP: Madrid metagame. Today, I present the results.

Lessons in Line Drawing

I have probably seen more Legacy decklists than anyone on the planet, between having looked at most of the SCG Open lists and all of the GP: Madrid lists. And one of the issues that I continue to struggle with is the taxonomy of Magic, the way in which we classify and sort Magic archetypes. Tallying up 2200 decklists is more complicated than it may seem. First, I have to use or create archetype categories. Second, I have sort decks into these categories.

As I wrote a few weeks ago, human beings are meaning-making machines. We create or borrow categories that are given to us and relentlessly sort things into them. It’s a largely unconscious process. Almost every chair you see is different, yet we have no trouble labeling something as a chair. Magic decks also come in infinite variety. And when I have to sort each decklist into an exclusive category, the classification system and the sorting system become very important, in ways that deciding whether an object is a fork or a spoon does not.

There are an infinite number of ways to classify archetypes. Perhaps one of the most logical ways of labeling an archetype is by color combination. That is a clearly ascertainable, objective standard of classification. We don’t tend to do that because we don’t find knowing the number of U/G/W decks to be very useful if those decks are very different in terms of their strategic and tactical aims. A U/G/W deck could be very aggressive or very controlling. Thus, Magic players tend to use hybrid classification taxonomies, that use historical labels that are well known (i.e. Zoo), color combinations, salient card names (i.e. Reanimator or CounterTop) and strategic posture (i.e. Control, or Combo or Aggro-Control).

Let me explain how I classified and sorted the GP Madrid decklists.

Zoo — 225 copies in the field

The most popular archetype in the field. The vast majority of the Zoo decklists were R/G/W, as you might traditionally see in a Legacy Top 8. However, I also counted decks that were R/G that looked very similar to the R/G/W lists, but didn’t have the white cards. So, if a deck was mostly Burn and a just a few Green spells, such as Goyf, I wouldn’t label it Zoo. But if it had Kird Ape, and other creatures, I would be more likely to call it Zoo. There were also a number of R/G/W/b lists, a.k.a. Dark Zoo, and just R/G/B lists, without the White at all. It’s amazing to think about how many Tiagas and Tarmogoyfs were played in this tournament.

Merfolk — 159 copies in the field

The second most popular archetype in the field. Merfolk also comes in a number of varieties. I was looking for the core cards: Lord of Atlantis, Reejerey, Vial, Etc. This is a far more homogenous archetype, strategically, but it does come with numerous splashes, mostly in the White or Green splash variety.

ANT — 125 copies in the field

This was a far more fractured archetype, as there are lists with a red splash, lists with a green splash, and just U/B lists.

Goblins — 124 copies in the field

The fourth most popular archetype. Again, this wasn’t hard to sort. There were a variety of splashes: Blue, Black, Breen, and White, but the core of the deck remained the same.

Dredge — 102 copies in the field

Tied for the fifth most popular archetype in the field. This archetype, fortunately, was also very easy to spot. There are a number of variations to it, but the core remains the same.

CounterTop-Progenitus — 102 copies in the field

This is a far more fractured archetype. Initially, I separated out the Bant and Supreme Blue (U/G/w/r with Firespout) versions of the archetype, but then just aggregated them as I saw more and more variety. I saw U/G/W/B and U/G/W/r and U/G/r and a few other color combinations. But the core of this deck remains largely the same: Natural Order, Counterbalance, Noble Hierarch, Tarmogoyf, etc.

U/G/x CounterTop — 99 copies in the field

Speaking of classification difficulties, I consider CounterTop Progenitus to be a sub-archetype of this more general archetype. But Wizards and SCG now separate the two, so I kept them separate. If we combine them, it becomes the second most popular archetype. I typically have kept this archetype Bant Countertop, except that more than a few lists were U/G/r or U/G/b or U/G/W/(r/b). The one commonality is that they are all U/G. The vast majority have a White splash.

Survival — 82 copies in the field

Here’s how this archetype broke out: there were 44 Bant Survival decklists (i.e. Survival with Force) and 29 other Survival decks (mostly with a heavier Black splash) on day 1. And then 9 other Survival decks that made Day 2 (not sure which subarchetype they were).

Burn — 81 copies in the field

Most of these lists looked the same. However, there was tremendous variety. I counted as Burn some lists with a very light Green or White splash (for example, just Jotun Grunt or sideboard cards).

Canadian Threshold —73 copies in the field

The tenth most popular archetype in the tournament. This has been an archetype in decline, but it still came out in force. Most of this archetype looked the same. Keep in mind that I separated out the Jace deck that looks somewhat similar.

Enchantress — 62 copies

Stax — 55 copies

This was an archetype that was heavily aggregated. Of the Stax decks that did not make day 2, the vast majority were mono White or W/G (45 total). There were 4 mono brown lists, 3 Black versions, and 1 W/B version.

Elves — 53 copies

There were several variants of this archetype. One runs Survival. Another just combo kills with a storm or burn spell. And another runs Natural Order. Some were hybrids of two of these three.

Eva Green — 52 copies

Affinity — 51 copies

Reanimator — 50

I only saw one Protean Hulk deck in the entire tournament.

The Rock — 41

The vast majority had a White splash. The reason it wasn’t PT Junk is because it was more controlling. It has Deeds, etc.

Aggro Loam — 40

B/w Suicide Black — 30 (32 with Dark Depths)

There were two other lists with Dark Depths.

German Fae — 29

Dreadtill — 27

Painter — 27

This was really multiple archetypes, such as Mono Red Imperial Painter and UR Painter with CounterTop, etc.

Dream Halls – 25

Lands – 25

Landstill — 25

The Landstill decks had tremendous variety. Of the versions that didn’t make Day 2, there were 3 UWB, 4 UWG, 3 WUBG, 7 U/W, 3 UWR, and 1 U/W with Counterbalance.

PT Junk — 24

This deck is W/G/B and aggressive. It tends to have cards like Serra Avenger.

Dragon Stompy — 23

Mono Black Aggro-Control — 21

A.k.a. Suicide Black, and its variants.

Bant-Progenitus — 20

This archetype looks like CounterTop-Progenitus without Countertop.

B/G Depths — 20

This looks very similar to Eva Green except with Dark Depths combo.

Belcher — 19

Bant — 17

This is mostly an Aggro-Control deck, and looks like CounterTop without the Counterbalance and Top. Some versions use Stoneforge Mystic.

W/G — 17

This is a container category, but it’s mostly mid-range fatties and some White disruption.

White Weenie – 16

U/W/x Control — 13

This includes Planeswalker control.

Stompy — 12

This is a Mono Green beatdown deck, mostly with Berserks.

Team America — 12

This is a UGB archetype with Goyf and Tombstalker as the main win conditions.

Faeries — 12

There were 8 UR Faries, 3 Mono-U faeries, 1 UWG, 1 UBG, and 1 UBW faeries.

U/G/B Dark Depths — 11

The G/B version with the Blue splash.

Show and Tell/Natural Order — 11

This is an odd deck to see a lot of.

Thopter Combo — 10

Boros — 10

This is W/R, a lot of burn, but with a very heavy White component — i.e. Grunt and Steppe Lynx, etc.

High Tide Combo — 8

Counterslivers — 8

U/W Tempo —8

This archetype has Fathom Seer.

Jace Fae — 7

Look at the 9th placed decklist to see what this looks like.

Crabs — 6

At first I didn’t think this was an archetype, but I kept seeing more of it. It uses Hedron Crab and jank like Archive Trap, and many had Glimpse.

Scepter Chant — 4

Mono White Control — 3

Aluren – 3

Sea Stompy — 3

Only three in the entire tournament, but one made Day 2.

B/r Suicide Black – 3

The Red splash.

Life Combo — 2

False Cure Combo — 2

Cephalid Breakfast — 2

Singletons: Hypergenesis Combo, Cruel Bargain combo, Kobold-Glimpse, Sneak Attack, and Draco-Explosion combo

Other: 69

This category was a catchall for decks that weren’t real or were completely rogue. For example, I put Standard (or worse) decks here. Over half of these decks were clearly budget decks. Many were ported directly from Extended or Standard. Somewhere clearly casual homebrews.

That was my tally. It turns out that 48 decklists somehow did not make it in the tally. I counted up all of the Day 1 decklists and then used Wizards Day 2 tally, so somehow something either went missing or got lost.

One last statistic: Out of the 1942 Day 1 decks I tallied, 120 were clearly budget decks, or 6.17% of the field. Identifying budget decks was particularly challenging for a number of reasons. First, a lot of the times if someone is short a dual land, rather than substitute a shock land, they’ll skimp or otherwise alter the manabase. Other times, they’ll just play a different deck altogether, one that they can build optimally under budget constraints. So, when tallying up â€˜budget’ lists, I looked for a few key things: Does the decklist run shock lands instead of revised dual lands? If it does, then I labeled it a budget deck. Does the decklist which would normally run Wastelands not have them? Then I labeled it a budget deck. For example, if a Merfolk deck didn’t have Wastelands, I’d look to see why not. If I didn’t see an answer, I labeled it a budget deck. Or, if a deck that normally runs 4 Goyfs had fewer for no apparent reason, I labeled it a budget deck. And so on. There were also decks that were clearly budget or casual, someone’s kitchen table homebrew or someone’s Standard Jund deck.

In short, about everything you could possibly imagine in Legacy showed up. Now, the question is: how did these archetypes perform?

Take a look at this table, with a complete tally:

Archetype	Total	Day 2 Cut	Top 64	Top 32	Top 16	Top 8
Zoo	225	42	8	4	3	3
ANT	125	22	8	5	4	2
Merfolk	159	17	6	4	2	0
Dredge	102	13	4	2	0	0
CounterTop Progenitus	102	11	3	3	1	1
U/G/X CounterTop	99	12	4	1	1	0
Canadian Threshold	73	10	1	1	0	0
Goblins	124	12	3	1	0	0
Survival (mostly bant)	82	9	4	1	0	0
Reanimator	50	9	3	1	1	1
PT Junk	24	6	1	0	0	0
Aggro Loam	40	6	2	1	0	0
German (Urb) Fae	29	5	0	0	0	0
Burn	81	5	1	1	1	0
Lands	25	6	2	1	0	0
Landstill	25	4	0	0	0	0
Enchantress	62	4	0	0	0	0
Eva Green (B/G Suey)	52	3	0	0	0	0
Bant	17	3	1	1	1	0
Elves	53	3	0	0	0	0
Painter	27	6	2	0	0	0
The Rock	41	3	2	0	0	0
Faeries	12	2	1	0	0	0
Stax	55	2	0	0	0	0
Dreadtill	27	2	2	0	0	0
Thopter Combo	10	2	1	0	0	0
Belcher	19	1	1	1	0	0
Dream Halls	25	1	0	0	0	0
U/W/x Control	13	3	0	0	0	0
Dragon Stompy	23	1	0	0	0	0
Sea Stompy	3	1	0	0	0	0
Affinity	51	1	0	0	0	0
Suicide Black	21	1	0	0	0	0
B/G Depths	28	1	1	0	0	0
Team America	12	0	0	0	0	0
Bant-Progenitus	20	1	1	1	1	1
W/B Suey	30	0	0	0	0	0
W/B Depths	2	0	0	0	0	0
R/B Suey	3	0	0	0	0	0
White Weenie	16	0	0	0	0	0
W/G	17	0	0	0	0	0
Mono White Control	3	0	0	0	0	0
Scepter Chant	4	0	0	0	0	0
Jace Faeries	7	1	1	1	1	0
High Tide	8	0	0	0	0	0
Counterslivers	8	0	0	0	0	0
U/B/(G) Dark Depths	11	0	0	0	0	0
Stompy	12	0	0	0	0	0
U/W Tempo	8	0	0	0	0	0
Crabs	6	0	0	0	0	0
Boros	10	0	0	0	0	0
Natural Order/Show and Tell	11	0	0	0	0	0
Aluren	3	0	0	0	0	0
Life Combo	2	0	0	0	0	0
False Cure Combo	2	0	0	0	0	0
Other	80	6	1	0	0	0

Now, let’s put this table into percentages:

Archetype	Total	Day 2 Cut	Top 64	Top 32	Top 16	Top 8
Zoo	10.33%	17.72%	12.50%	12.50%	18.75%	37.50%
ANT	5.74%	9.28%	12.50%	15.63%	25.00%	25.00%
Merfolk	7.30%	7.17%	9.38%	12.50%	12.50%	0.00%
Dredge	4.68%	5.49%	6.25%	6.25%	0.00%	0.00%
CounterTop Progenitus	4.68%	4.64%	4.69%	9.38%	6.25%	12.50%
U/G/X CounterTop	4.54%	5.06%	6.25%	3.13%	6.25%	0.00%
Canadian Threshold	3.35%	4.22%	1.56%	3.13%	0.00%	0.00%
Goblins	5.69%	5.06%	4.69%	3.13%	0.00%	0.00%
Survival (mostly Bant)	3.76%	3.80%	6.25%	3.13%	0.00%	0.00%
Reanimator	2.29%	3.80%	4.69%	3.13%	6.25%	12.50%
PT Junk	1.10%	2.53%	1.56%	0.00%	0.00%	0.00%
Aggro Loam	1.84%	2.53%	3.13%	3.13%	0.00%	0.00%
German (Urb) Fae	1.33%	2.11%	0.00%	0.00%	0.00%	0.00%
Burn	3.72%	2.11%	1.56%	3.13%	6.25%	0.00%
Lands	1.15%	2.53%	3.13%	3.13%	0.00%	0.00%
Landstill	1.15%	1.69%	0.00%	0.00%	0.00%	0.00%
Enchantress	2.85%	1.69%	0.00%	0.00%	0.00%	0.00%
Eva Green (B/G Suey)	2.39%	1.27%	0.00%	0.00%	0.00%	0.00%
Bant	0.78%	1.27%	1.56%	3.13%	6.25%	0.00%
Elves	2.43%	1.27%	0.00%	0.00%	0.00%	0.00%
Painter	1.24%	2.53%	3.13%	0.00%	0.00%	0.00%
The Rock	1.88%	1.27%	3.13%	0.00%	0.00%	0.00%
Other Faeries	0.55%	0.84%	1.56%	0.00%	0.00%	0.00%
Stax	2.52%	0.84%	0.00%	0.00%	0.00%	0.00%
Dreadtill	1.24%	0.84%	3.13%	0.00%	0.00%	0.00%
Thopter Combo	0.46%	0.84%	1.56%	0.00%	0.00%	0.00%
Belcher	0.87%	0.42%	1.56%	3.13%	0.00%	0.00%
Dream Halls	1.15%	0.42%	0.00%	0.00%	0.00%	0.00%
U/W/x Control	0.60%	1.27%	0.00%	0.00%	0.00%	0.00%
Dragon Stompy	1.06%	0.42%	0.00%	0.00%	0.00%	0.00%
Sea Stompy	0.14%	0.42%	0.00%	0.00%	0.00%	0.00%
Affinity	2.34%	0.42%	0.00%	0.00%	0.00%	0.00%
Suicide Black	0.96%	0.42%	0.00%	0.00%	0.00%	0.00%
B/G Depths	1.28%	0.42%	1.56%	0.00%	0.00%	0.00%
Team America	0.55%	0.00%	0.00%	0.00%	0.00%	0.00%
Bant-Progenitus	0.92%	0.42%	1.56%	3.13%	6.25%	12.50%
W/B Suey	1.38%	0.00%	0.00%	0.00%	0.00%	0.00%
W/B Depths	0.09%	0.00%	0.00%	0.00%	0.00%	0.00%
R/B Suey	0.14%	0.00%	0.00%	0.00%	0.00%	0.00%
White Weenie	0.73%	0.00%	0.00%	0.00%	0.00%	0.00%
W/G	0.78%	0.00%	0.00%	0.00%	0.00%	0.00%
Mono White Control	0.14%	0.00%	0.00%	0.00%	0.00%	0.00%
Scepter Chant	0.18%	0.00%	0.00%	0.00%	0.00%	0.00%
Jace Faeries	0.32%	0.42%	1.56%	3.13%	6.25%	0.00%
High Tide	0.37%	0.00%	0.00%	0.00%	0.00%	0.00%
Counterslivers	0.37%	0.00%	0.00%	0.00%	0.00%	0.00%
U/B/(G) Dark Depths	0.50%	0.00%	0.00%	0.00%	0.00%	0.00%
Stompy	0.55%	0.00%	0.00%	0.00%	0.00%	0.00%
U/W Tempo	0.37%	0.00%	0.00%	0.00%	0.00%	0.00%
Crabs	0.28%	0.00%	0.00%	0.00%	0.00%	0.00%
Boros	0.46%	0.00%	0.00%	0.00%	0.00%	0.00%
Natural Order/Show and Tell	0.50%	0.00%	0.00%	0.00%	0.00%	0.00%
Aluren	0.14%	0.00%	0.00%	0.00%	0.00%	0.00%
Life Combo	0.09%	0.00%	0.00%	0.00%	0.00%	0.00%
False Cure Combo	0.09%	0.00%	0.00%	0.00%	0.00%	0.00%
Other	3.67%	2.53%	1.56%	0.00%	0.00%	0.00%

This table shows you each archetype as a percentage of each part of the field. So, Zoo, at the very top, was 10.33% of the Day 1 field, but 17.72% of the Day 2 field, and so on. This table gives you a sense of how well each archetype performed. For example, Zoo and ANT consistently outperformed their proportion of the field as a whole as they ascended the tournament.

However, the problem with this table is that it is weighted by the more popular decks. So, the 7% differential from Day 1 to Day 2 for Zoo is actually proportionately equivalent to the 1.5% differential for Reanimator from Day 2 to Day 2. How can we get a better measure of how archetypes performed given their proportion of the field?

Take a look at this chart. This chart graphs Top X penetration:

Archetype	Total	Day 2 Cut	Top 64	Top 32	Top 16	Top 8
The Field (Baseline)	100%	10.64%	2.87%	1.44%	0.72%	0.36%
Zoo	10.33%	18.67%	3.56%	1.78%	1.33%	1.33%
ANT	5.74%	17.60%	6.40%	4.00%	3.20%	1.60%
Merfolk	7.30%	10.69%	3.77%	2.52%	1.26%	0.00%
Dredge	4.68%	12.75%	3.92%	1.96%	0.00%	0.00%
CounterTop Progenitus	4.68%	10.78%	2.94%	2.94%	0.98%	0.98%
U/G/X CounterTop	4.54%	12.12%	4.04%	1.01%	1.01%	0.00%
Canadian Threshold	3.35%	13.70%	1.37%	1.37%	0.00%	0.00%
Goblins	5.69%	9.68%	2.42%	0.81%	0.00%	0.00%
Survival (mostly bant)	3.76%	10.98%	4.88%	1.22%	0.00%	0.00%
Reanimator	2.29%	18.00%	6.00%	2.00%	2.00%	2.00%
PT Junk	1.10%	25.00%	4.17%	0.00%	0.00%	0.00%
Aggro Loam	1.84%	15.00%	5.00%	2.50%	0.00%	0.00%
German (Urb) Fae	1.33%	17.24%	0.00%	0.00%	0.00%	0.00%
Burn	3.72%	6.17%	1.23%	1.23%	1.23%	0.00%
Lands	1.15%	24.00%	8.00%	4.00%	0.00%	0.00%
Landstill	1.15%	16.00%	0.00%	0.00%	0.00%	0.00%
Enchantress	2.85%	6.45%	0.00%	0.00%	0.00%	0.00%
Eva Green (B/G Suey)	2.39%	5.77%	0.00%	0.00%	0.00%	0.00%
Bant	0.78%	17.65%	5.88%	5.88%	5.88%	0.00%
Elves	2.43%	5.66%	0.00%	0.00%	0.00%	0.00%
Painter	1.24%	22.22%	7.41%	0.00%	0.00%	0.00%
The Rock	1.88%	7.32%	4.88%	0.00%	0.00%	0.00%
Other Faeries	0.55%	16.67%	8.33%	0.00%	0.00%	0.00%
Stax	2.52%	3.64%	0.00%	0.00%	0.00%	0.00%
Dreadtill	1.24%	7.41%	7.41%	0.00%	0.00%	0.00%
Thopter Combo	0.46%	20.00%	10.00%	0.00%	0.00%	0.00%
Belcher	0.87%	5.26%	5.26%	5.26%	0.00%	0.00%
Dream Halls	1.15%	4.00%	0.00%	0.00%	0.00%	0.00%
U/W/x Control	0.60%	23.08%	0.00%	0.00%	0.00%	0.00%
Dragon Stompy	1.06%	4.35%	0.00%	0.00%	0.00%	0.00%
Sea Stompy	0.14%	33.33%	0.00%	0.00%	0.00%	0.00%
Affinity	2.34%	1.96%	0.00%	0.00%	0.00%	0.00%
Suicide Black	0.96%	4.76%	0.00%	0.00%	0.00%	0.00%
B/G Depths	1.28%	3.57%	3.57%	0.00%	0.00%	0.00%
Team America	0.55%	0.00%	0.00%	0.00%	0.00%	0.00%
Bant-Progenitus	0.92%	5.00%	5.00%	5.00%	5.00%	5.00%
W/B Suey	1.38%	0.00%	0.00%	0.00%	0.00%	0.00%
W/B Depths	0.09%	0.00%	0.00%	0.00%	0.00%	0.00%
R/B Suey	0.14%	0.00%	0.00%	0.00%	0.00%	0.00%
White Weenie	0.73%	0.00%	0.00%	0.00%	0.00%	0.00%
W/G	0.78%	0.00%	0.00%	0.00%	0.00%	0.00%
Mono White Control	0.14%	0.00%	0.00%	0.00%	0.00%	0.00%
Scepter Chant	0.18%	0.00%	0.00%	0.00%	0.00%	0.00%
Jace Faeries	0.32%	14.29%	14.29%	14.29%	14.29%	0.00%
High Tide	0.37%	0.00%	0.00%	0.00%	0.00%	0.00%
Counterslivers	0.37%	0.00%	0.00%	0.00%	0.00%	0.00%
U/B/(G) Dark Depths	0.50%	0.00%	0.00%	0.00%	0.00%	0.00%
Stompy	0.55%	0.00%	0.00%	0.00%	0.00%	0.00%
U/W Tempo	0.37%	0.00%	0.00%	0.00%	0.00%	0.00%
Crabs	0.28%	0.00%	0.00%	0.00%	0.00%	0.00%
Boros	0.46%	0.00%	0.00%	0.00%	0.00%	0.00%
Natural Order/Show and Tell	0.50%	0.00%	0.00%	0.00%	0.00%	0.00%
Aluren	0.14%	0.00%	0.00%	0.00%	0.00%	0.00%
Life Combo	0.09%	0.00%	0.00%	0.00%	0.00%	0.00%
False Cure Combo	0.09%	0.00%	0.00%	0.00%	0.00%	0.00%
Other	3.67%	7.50%	1.25%	0.00%	0.00%	0.00%

This table measures Top X penetration. What does that mean? It shows the percentage of the archetype in the field as a whole that made it to Day 2, Top 64, Top 32 and so on. This is a more direct indicator of archetype strength in the field. The top line shows the baseline, i.e. the field as a whole, for you to compare to each archetype.

Notice that Zoo, ANT, CounterTop-Progenitus, and Reanimator were the only four major archetypes to consistently outperform the field.

In terms of making Day 2, Lands, Zoo, Reanimator, and ANT were the top 4 performing decks.
– Of the archetypes with at least 25 copies in the field, Lands was the best performing deck in terms of making day 2. A quarter of its representatives made the Day 2 cut.
– Contrary to popular perception, Zoo was an extraordinarily well performing deck. Over 18% of the Zoo decks made Day 2, second only to Land in the archetypes with at least 25 copies in the field.
– Reanimator was, in terms of day 2 penetration, the third best performing archetype, with 18% making Day 2.
– Very close behind was ANT, with 17.6% of the ANT copies making Day 2.

When you look at Top 64 penetration, things shift a bit. Land remains the best performing archetype in terms of making Top 64, followed by ANT, Reanimator, and then Aggro Loam. Zoo falls substantially. Some people will misread this table. Zoo did very well in terms of making day 2. It was a top performing, disproportionate to its presence in the field. The problem for Zoo was that it was only an average performer for most of Day 2. It fell into the middle of the pack. That takes nothing away from its high performance in making day 2.

When you look at Top 32 penetration, Lands and ANT are tied as the best performing archetypes with 4% of each archetype making the top 32. Here, CounterTop Progenitus and Merfolk put their foot to the pedal as the third and fourth best performing archetype, and Reanimator falls to 5th best performing.

Overall, the best performing decks in the whole tournament were Lands, Reanimator, Zoo, and ANT. But which of these was the best overall? Despite being the best performing archetype in terms of Top X penetration for making Day 2, Top 64, and Top 32, Land failed to make it into the Top 16. Land dominated this tournament up until the top 16, where it disappeared. For that reason, I can’t label Lands the best performing deck in the tournament. That honor has to go to Reanimator, which held its ground with the highest archetype penetration into the Top and the Top 8, despite falling to 5th in the Top 32. ANT is a close second, with the best Top 16 penetration of any archetype, and second only to Reanimator in terms of Top 8 penetration. Then, Zoo and Lands are clearly the third or fourth best performing archetype, depending on which stats you credit. Zoo had the second highest Day 2 penetration and the third highest Top 16 and Top 8 penetration compared to Lands number one performance up until the Top 16. You choose.

In sum, Reanimator and ANT were the top 2 performing decks respectively, followed by Land and Zoo, which were also outstanding. Land was the top performing deck in the entire tournament up until the top 16. And Zoo far outperformed its presence in the field, despite being the most popular deck.

And now, decklists!

Reanimator

Andreas Muller

1st Place at Grand Prix on 02-28-2010

Legacy