fbpx

Sullivan Library – Data Mining Standard in Kyoto

Read Adrian Sullivan every week... at StarCityGames.com!
Tuesday, March 3rd – With the recent changes to the Pro Tour structure, deducing the best decks and strategies in Standard post-Kyoto is trickier than usual. In today’s Sullivan Library, Adrian takes us through his data-mined Top 8, factoring in a number of previously-unconsidered tangibles…

Now that the weekend is over, we have yet another Pro Tour in the books. Gabriel Nassif rose to the top, piloting a Five-Color Control deck that many top players and designers (StarCityGames.com own Patrick Chapin, among them) had a hand in making. But was his deck the “best” deck? Or was it, as some claim, Luis Scott-Vargas who actually piloted the best deck, and would have won it all were it not for the timely confluence of multiple Mulldrifters and Volcanic Fallout after a Head Games combined with a deeply unlucky hand in the fifth game.

In tournament Magic, things happen. It’s often pretty rare that the “best” deck takes the cake in a big Event. The variance is significantly able to make things go the wrong way, especially in metagames where a “best” deck might enjoy 60-65% wins in the common, competitive matchups. Ten years ago, Zvi Mowshowitz probably had the best deck of the Urza’s Block Pro Tour with “Zero Effect,” a Cycling-based combo deck, but while he made Top 8, he didn’t win the event.

Mixed events exacerbate the difficulty in figuring out who had the best deck, or even the best version of a deck. Time and time again, I know that I’ve seen people play card-for-card versions of the Burn deck that Jon Finkel played, drawn by Finkel’s star power, primarily, and not noticing that Finkel finished with only four wins in the lower parts of the brackets. Shockingly, Antii Malin, the World Champion, got almost no notice for his version of the list, despite its very impressive performance, not only simply within the format, but against strong resistance. Where a typical PT would have given us a good 14-or-so rounds to get at least a “best guess,” instead we only have eight rounds…

If we just take the “Standard Standings” from the event, as they are given to us by Wizards, we have a clear example of the problem: a three-way tie for first, an eight-way tie for fourth, a two way tie for 12th, and a 27-way tie for 14th.

Taking the “Top 13” (19 points and above), we get the following decks:

1 Vengeant Weenie
3 Boat Brew
1 Elves
2 Five-Color Control
2 Doran
1 Faeries
2 Black/White Tokens
1 White Weenie

This is gives a different taste of the event than what we see represented in the “Virtual Top 8” (Top 8, and tied for Top 8):

2 Black/White Tokens
1 Vengeant Weenie
3 Boat Brew
1 Doran
1 Five-Color Control
1 Faeries
1 Mono-Red
1 White Weenie

What’s even more aggravating is that neither of these means of measurement actually give us a real sense of what did well. Clearly the path of Cedric Phillips, favorite mage of many of us, was much harder than Takayuki Nagaoka. Cedric’s last rounds were played against people much higher in the standings, whereas Nagaoka’s resistance was clearly weaker in round 11-14.

Part of what is so difficult is that it is completely impossible to really get a sense of how things should have played out if this were “merely” a Standard tournament. The deep filtering that would happen at the top is gone. We merely have conjecture. The “Top” decks at the Wizards coverage of the event shows all of the decks with 15 points or more, i.e., 5-3 and better. If we break it down a little further, it becomes more interesting. Here is how things look, archetypically (with some similar archetypes joined for convenience), assuming coverage-master Bill Stark got his archetypes correct:

Decks represented by at least one 6-2 or better player:

Archetype: Number played / players at 15 or better / player at 18-20 / players at 21+

Boat Brew: 98 / 30 / 9 / 3
Doran: 3 / 2 / 0 / 2 ** (Wow!)
Five-Color variants: 56 / 12 / 4 / 1
White Weenie/Vengeant Weenie: 39 / 15 / 6 / 2
Faeries: 50 / 10 / 4 / 1
Elves: 4 / 3 / 1 / 1
Red/Blightning: 44 / 6 / 3 / 0
Black-White Tokens: 19 / 5 / 2 / 1

A look at this data indicates something quite interesting. Doran/X and Black-Green Elves players did incredibly. There weren’t very many of them, and nearly all of them shot right to the top. That’s incredibly impressive. The problem in paradise is that we have such a small sample to work with, it’s hard to know what our standard deviation would be. There is probably something to pay attention to here, though. Other than that, the decklists can be described in pretty simple term: Homogenous! On the whole, there seemed to be less risk-taking. With less time to prepare for the format (compared to focusing purely on it), there were few to no surprises (an effect I expect to see replicated in future events with the split format).

Of the decks that saw popular play, the percent that got 15 points or more is in the following order:

1st Place — Vengeant/White Weenie, at over 38%

Led by the inimitable Cedric Phillips, this archetype appears to have probably been the deck to play if you merely wanted a winning record.

2nd Place — Boat Brew, at around 31%

Brian Kowal (et al.) creation, “The Boat Brew” was dubbed the deck of the tournament by a bunch of people. Not only was it the most popular, but it also represented.

3rd Place — Black-White Tokens, at around 26%

Championed by current PT Master Luis Scott-Vargas, this deck got more than a quarter of its participants a winning record, performing just above…

…The rest, Red-based Aggro (14%), Five-Color (21%), Faeries (20%)

All of these performed under the Expected Value of attending the Pro Tour; fully 24% of the competitors received 5-3 records or better in Standard.

Now, obviously, this isn’t an indictment on the top performers’ decks. Perhaps everyone had a wrong build of, say, the Red decks, except for the very best build, played by only a single player. It is an indication, though, of how everything went. Worth checking out, then, is the carry through of a deck’s top performers. The best three to four percent of competitors went 7-1 or better…

1st Place — Black-White Tokens, at 5%

To my mind this is saying something very interesting. The deck only barely managed to outperform the average in the bulk of the tournament, but when looked at in the highest levels of play, it was the highest performer (by a tiny nose). This seems to suggest that the deck requires a certain degree of play skill to operate at top efficiency, and that only if you have a certain threshold of skill would it reward you for it. Watchers of the podcast will surely remember the ecstatic glee that Brian David-Marshal and Randy Buehler exclaimed as LSV took a huge amount of time to wait, wait, wait before dropping Head Games on eventual champion Nassif. Perhaps, like Storm in Extended, playing Black-White isn’t something for anyone but the very best to take up…

2nd Place — White Weenie/Vengeant Weenie, at 5%

Only a smidge of a smidge down was White Weenie. Also worth noting, though, was just how top heavy this archetype was. It wasn’t just that it nearly performed at the same level as B/W for 7-1 or better performances; it wildly outperformed every archetype at 6-2 or better, butting fully 20% of its players into that record. That’s actually kinda crazy.

3rd Place — Boat Brew, at 3%

Again, Boat Brew continues to represent. Where some have claimed that the deck was overhyped, the numbers don’t at all agree. The deck managed to maintain itself as a clearly potent deck for anyone who picked it up. It’s unsurprising that Brian Kowal was among the 7-1s, but he definitely wasn’t alone.

So, maybe we have a general idea from those numbers, but what can we look to if we want to find the “real” Top 8 or 16?

One place to start is to examine the actual opposition that each player in contention faced. This is a lot of work, but it is doable. By going through the matchups of rounds 1 through 4, and then 11 through 14, you can actually deduce the total resistance that any single player faced. If we add up the points that any player they played received in Standard, we should have a good sense of the resistance that each player faced. Much like the DCI Reporter, I don’t think we want to completely handicap someone for playing against a losing player, so we’ll make the “minimum” points that can be received by a player 11 points (just worse than 50/50). Here, checking just those players at 19 and above, we get the following:

1st – Cedric Phillips, 22 points, Opposition: 11+11+18+11+18+18+21+6=114
2nd – Akimasa Yamamoto, 22 points, Opposition: 11+11+11+18+11+15+16+6=99
3rd – Takayuki Nagaoka, 22 points, Opposition: 11+11+15+11+11+15+11+5=90
4th – Brian Robinson, 21 points, Opposition: 15+11+11+15+21+11+0+15=99
5th – Gabriel Nassif, 21 points, Opposition: 11+11+11+11+0+15+18+21=98
6th – Yuichi Kishimoto, 21 points, Opposition: 15+11+11+18+11+0+15+15=96
7th – Brian Kowal, 21 points, Opposition: 11+11+18+18+11+11+15+0=95
8th* – Gael Bailly Maitre, 21 points, Opposition: 18+11+0+15+11+11+15+11=92
9th* – Sam Black, 21 points, Opposition: 18+0+11+11+11+15+11+15=92
10th – Raphael Levy, 21 points, Opposition: 0+11+11+11+11+21+11+15=91
11th – Shuu Komuro, 21 points, Opposition: 11+16+11+15+11+0+11+15=90
12th – Antoine Ruel, 19 points, Opposition: 0+11+11+15+11+15+15+4=82
13th – Luis Scott-Vargas, 19 points, Opposition: 11+11+11+15+11+0+15+7=81

Working out a tie-breaker for this 8th, I’d give it to strongest opposition in the loss each player received. For Sam, this means his tiebreaker is Kazuya Mitamura, with 12 points, versus Gael’s Yushitaka Nakano, with 18 points.

Here, then, is your Actual Opposition Top 8 for Standard from Kyoto (or, at least, the best we can figure of it…):

1 Vengeant Weenie
3 Boat Brew
2 “Doran”
1 Elves
1 Five-Color Control

That’s not exactly what we expected, is it…

1st – Cedric Phillips


2nd – Akimasa Yamamoto


3rd – Takayuki Nagaoka


4th – Brian Robinson


5th – Gabriel Nassif


6th – Yuichi Kishimoto


7th – Brian Kowal


8th – Gael Bailly Maitre


While there are many other ways that the weighting of the various options could result in a different “Top 8,” I, for one, am fairly satisfied with my examination of what the “real” Standard looks like. There are a lot of small take aways. The only countries represented in the 19+ points in Standard were the USA, France, and Japan. Black/Green seems to be far more powerful than we might have otherwise realized. Red/White is a force to be reckoned with.

Personally, I have some other nice take aways. Fellow Madisonian Brian Kowal let me know that the card I’ve been championing for the mirror-match (and versus and deck that makes tokens), Galepowder Mage, was a complete beating for him in the sideboard, and he was really happy that he had had four. It was also cool to see Sam Black right there at the top of the Constructed side. In the finals, I wasn’t sure who to be rooting for; I knew that I wanted any deck Patrick Chapin had been working on to end up doing well, but at the same time, LSV is perhaps the nicest high-level player we’ve had in the game.

In the meantime, I’m looking forward to a single-format event, Grand Prix: Chicago, right outside my back door. Needless to say, the event largely snuck up on me, and I’m finding myself less prepared than I’d like, but still, I’m hopeful. So complications have just arrived in my personal life as I started my most recent round of prep for the Grand Prix, but they are good ones: I’ve been accepted into the Media and Cultural Studies Ph.D. program at the University of Wisconsin — Madison. Overall, despite the stress of everything, it’s been a good week.

See you in Chicago!
Adrian Sullivan