Data Mining For Modern In Valencia

Find out what the Modern metagame might look like at Grand Prix Richmond next weekend as Adrian Sullivan analyzes the data from Pro Tour Born of the Gods.

Was the first Valencia really so long ago? When I qualified for that Pro Tour way back in 2007, it was in a Block Constructed PTQ; will we ever see that again? Back then not only were Pro Tours single format, but that event was Extended, a format that has basically bit the dust. At the event itself, I felt like I had a great deck, but unfortunately a literal flood hit the event, causing the first day to be cancelled.

Evan Erwin practically changed Magic media with The Magic Show, which covered the event. Patrick Chapin hit the nail on the head when he described that episode thusly:

"Absolutely unreal. By far one of the greatest pieces of Magic media I have ever seen. Evan, you literally blew my mind. I am a believer. Please keep doing what you are doing and keep improving your product. You actually are the Magic storyteller."

Just in case you haven’t seen it:

Since then, of course, things have changed a bit. StarCityGames.com is still at the forefront of creating media content; just check out SCGLive and you’ll see what the cutting edge looks like. Nathan Holt and Shawn Kornhauser of Walk the Planes fame have taken the mantle as Magic’s current video storytellers. Pro Tours are mixed formats (like Worlds used to be), and there’s no Worlds as we once knew it. Heck, Pro Tours are named after the set that most recently premiered now rather than the city they’re in. The Magic world has changed a lot in those six-plus years.

One thing that remains the same though is a real need for good data. If you look at the Wizards of the Coast coverage of Pro Tour Born of the Gods in Valencia, you can get the list of the top decks from the event, but you still really need to slog through it to get to the heart of the matter. What is that first question?

What Is Good In Modern?

This is the question that serious tournament players always ask themselves about whichever format they’re addressing. Sometimes people’s understanding of a format simply isn’t attached to reality because they are only getting a very narrow picture.

To understand what I mean, let’s take a quick glance at the highest-profile decks in the event: the eight decks that played in the Top 8.

1st: Shaun McLaren – U/W/R Control
2nd: Jacob Wilson – Melira Pod
3rd: Patrick Dickmann – Splinter Twin
4th: Anssi Alkio – Splinter Twin
5th: Lee Shi Tian – U/R Control
6th: Christian Seibold – Affinity
7th: Chris Fennell – Storm
8th: Tim Rivera – Splinter Twin

The problem? Well, if you isolate for just the Modern portion of the event, half of the players in that list would not even be included among the very top portion of the field. For example, Shaun McLaren sits at 31st place among competitors if one just counts Modern.

Of course, even isolating for any particular Constructed portion of an event is a difficult proposition. When you’re dealing with mixed format events, you stop actually having the equivalent of "cream rising to the top" when it comes to successful decks. Mixed format events are better at discovering who the best players are in an event but less successful at helping to unravel what the best decks are. A big part of the reason for this is that even the best performing deck might not have been playing against other decks that were performing just as well.

If you’re a whiz at Draft, you might end up 6-0 in Draft and find yourself paired in round 12 against an opponent who went 2-4 in Draft; you both may have a 7-4 record going into the round, but your opponent went 5-0 in Constructed, while you went 1-4. Even outside of that, if we compare two players with 7-2 records in Constructed after round 15, this topsy-turvy nature of one’s opposition means that you really can’t tell how strong their opposition was overall.

The best solution I have to this problem is to simply accept it, go through the hard work of figuring out the Constructed strength of the opposition, and use that to create the equivalent of a tiebreaker based on the performance of only Modern. This does mean that the data isn’t perfect, but it’s probably as good as you’re going to be able to do without simply ignoring all of the data and waiting for Grand Prix Richmond to sort it all out.

Well, I’m not sure about you, but I tend to prefer walking into those events with good data in hand rather than walking out of them waiting for the data after the fact.

Wizards of the Coast lists 113 competitors as the “Top Modern Decks,” including every deck that had a 6-4 record or better. Much like AJ Sacher, this just isn’t enough rarified air for me. If we raise the bar just the barest amount, we can see the picture a bit more clearly. Take for example the top of the top of the field. In taking only the Modern results, there were ten competitors who had an 8-1-1 record or better in Modern. If we give them the tiebreakers based on their opponent’s Modern strength, here is your “Modern Only” Top 8 (the accompanying brief descriptions will likely be old hat to experienced Modern players):

Bryan Gottlieb’s Ad Nauseam ranks eighth. This deck has a simple plan: combo together Ad Nauseam with an effect that will stop you from dying, either Angel’s Grace or Phyrexian Unlife. When these two happen together, draw your entire deck and kill your opponent with Lightning Storm (powered by Simian Spirit Guide), acquiring whatever Pact of Negation and Slaughter Pact combo you need to ensure that nothing will stop you from winning.

Martin Muller’s Melira Pod would be seventh. If you haven’ seen Melira Pod, you probably haven’t been playing much in the way of Magic in quite a while. This deck patches together Melira, Sylvok Outcast with a sacrifice outlet to either gain infinite life or deal infinite damage with a Kitchen Finks or Murderous Redcap respectively. In addition, it includes the infinite life/infinite power combo of Archangel of Thune + Spike Feeder. All of this is fed by Birthing Pod and Chord of Calling, and even if the deck doesn’t combo off, it still has a large toolbox of powerful cards that can shut down the opponent.

Petr Sochurek’s Merfolk ends up sixth. This isn’t Legacy Merfolk; unlike that aggro-control staple, Petr’s Modern Merfolk runs no true counters and instead has only Vapor Snag (and arguably Tidebinder Mage and Cursecatcher in some matchups) as a means to try to control the flow of time in the game. This deck has shifted into a largely aggro space, with the slightest touch of disruption backing up a deck with more lords than Game of Thrones.

Here we have the first of the decks in our "Actual Modern Top 8" that appeared in the Pro Tour Top 8. Phillip Dickmann’s RUG Twin ranks fifth. Splinter Twin (aka “Infinity Faeries! Infinity Faeries!”) combines a copying effect with a creature that can untap a permanent when it comes into play. Together an army of infinite creatures is created, which kills the opponent. Dickmann’s list vastly diminishes the single-minded nature of the typical Splinter Twin deck, instead including powerful creatures and removal so that it can also just function like a RUG Midrange deck.

Chris Fennell’s Storm would be fourth. Another competitor who made the Pro Tour Top 8, Fennell’s Storm deck uses Goblin Electromancer and Pyromancer Ascension to increase the power of the cheap library manipulation and "ritual" effects, creating an explosive turn where a Grapeshot will kill the opponent with a huge storm count. A fairly singular deck, its "alternate" kills can include sideboarded Empty the Warrens as well as the occasional Lightning Bolt (copied tons of times).

Lee Shi Tian’s U/R Control ends up third. In some ways this is the "breakout" deck of the tournament. The combination of Blood Moon, Spreading Seas, and Vedalken Shackles along with a healthy collection of counterspells makes this U/R Control deck the best performing deck planning on saying "no" to the opponent’s game plan. Its game plan? Stay alive while shutting down the opponent’s outs and then kill them with something or other.

Mattia Fornacini’s Tribal Zoo ranks second and is the first pure aggro deck on our list. It runs a huge number of cheap one-mana beaters, complementing them with Tarmogoyf and Geist of Saint Traft. Ghor-Clan Rampager masquerades as a spell here, creating huge bursts of damage supplemented by efficient removal spells as well as the namesake Tribal Flames, which is capable of dealing five damage bursts because of the collection of Ravnica dual lands.

Our sole archetype repeat in this Modern Top 8, Anssi Alkio’s Splinter Twin ends up in first. Unlike Dickmann, Alkio is far more dedicated to the combo, with many more cards intended to protect its infinite army. Where Dickmann’s list could be said to shift a little toward midrange, interested in developing the board and interacting with the opponent, Alkio’s deck is a little more controlling, with Lightning Bolt, Electrolyze, and Snapcaster Mage helping to just play clean up on anything problematic the opponent may do.

A Bigger Picture

This is a diverse collection of decks, particularly when you note that the "ninth place" deck was the unusual Auras deck piloted by Seamus Kelahan. If we gather all of the decks that have a 7-3 record or better, you can get a good sense of the top portion of the field:

Splinter Twin: 28, 25, 25, 24, 24, 24, 21, 21 ,21, 21
Tribal Zoo: 27, 21
Auras: 25, 24, 22, 21, 21
Storm: 25, 24, 22, 21
Melira Pod: 25, 23, 22
Ad Nauseam: 25, 22
U/R Control: 25, 21
Merfolk: 25
BG/x Midrange: 24, 22, 22, 21
Living End: 24, 21
Kiki Pod: 24
Tribal Bant: 24
U/W/R Midrange: 22, 22, 21, 21, 21
Big Zoo: 22, 22
Burn: 22, 22
Affinity: 22, 21, 21, 21
U/W/x Control: 22, 21, 21
Gifts Rites: 22
Naya Zoo: 21, 21, 21
Tron: 21
Scapeshift: 21
KCI: 21
Faeries: 21

Obviously Splinter Twin was the big winner of the day, with fully ten copies of the deck in the top portion of the field. But this can be misleading; during the reign of Trix (Illusions/Donate) after the banning of Dark Ritual in Extended, there were a ton of Trix decks that were doing well at large events, but if you looked at the numbers, they were vastly underperforming given just how many people were playing them. It was actually better to swap decks with a random competitor in the event than play Trix in that era!

We can look for similar information by checking out the breakdown of the Valencia data:


The reddened data shows when a deck outperforms the expected finish for any random deck in the tournament. With 252 players of 393 making day 2, we would expect about 64% of players to make that bar. Decks in red exceeded that measure. At Honolulu, the first time I did this kind of analysis, only 34% of players actually made day 2, for example, but the threshold has been lowered, resulting in fully 64% of players making day 2.

One of the big surprises is that Splinter Twin, the deck most represented in the top portion of the field, really struggled to make day 2. After that, though, those who did make it generally greatly outperformed the field. While there are many possible explanations, one that comes to mind for me is that a lot of people probably had substandard Splinter Twin decks or were bad pilots of the deck; once they were out of the running, the deck performed admirably since it was narrowed down to the best lists and pilots.

The other heavily represented deck was Zoo, which had practically the opposite result with a good conversion to day 2 rate but poor results thereafter. The most obvious explanation is that Zoo was able to beat up on the decks that weren’t the top of the field but was in turn beaten up by the most prepared decks. Zoo, despite its immense popularity, just didn’t perform all that well.

But what this also highlights is the difference between a day 2 conversion and an actual winning record. Remember, this is a mixed format event, and you can achieve a day 2 largely on the back of a 3-0 draft; there are some number of 1-4 Modern decks that made day 2.

Take one of the most dramatic examples of this, Amulet Combo, which placed 80% (!) of its pilots in day 2. (Scapeshift performed similarly, with very poor results for its many pilots despite so many of them making day 2). Looking at the results, though, no one managed to get a 7-3 or better record with Amulet of Vigor. My expectation is that the players who played Amulet Combo were among the better players but that the deck itself is truly unspectacular, though with only five pilots we could easily be looking at a sample size issue since even one successful competitor would have put it in the red.

Still, when we compare those results to those of U/W/x Control, the results are striking. U/W/x Control also only had five pilots, but it took three people to seven wins or better. I expect that these controlling builds, whether they’re U/W or U/W/R, are likely very good in the current metagame.

Some decks have another very interesting characteristic: the best results with the deck float to the highest portion of the field but not elsewhere. Take Melira Pod, which greatly underperformed in achieving day 2 and getting into 7-3+ but slightly overperformed at the level of Top 32. These decks might actually just require a very talented player to get results out of at all. You could see similar returns from various B/G Midrange lists and Burn.

Anyone can potentially have a good day, so the decks which only had a single competitor put them over the expected results can be looked at with a little bit more skepticism. In this case, that would be Merfolk and Kiki Pod. These decks could well be very good but really need more examination.

The decks that you can in confidence call good are these:

Splinter Twin
U/R Control
U/W/x Control

The decks that might require very good skill to achieve good results:

Melira Pod
B/G Midrange

The decks that look like they just underperformed:

Living End
Kiki Pod

As for the rest, they either just weren’t spectacularly positive or negative or didn’t have enough of a population playing them for there to be results that can be easily gleaned from the corrupted data that a mixed event provides. For example, it could be that Merfolk is actually The Truth, but there simply isn’t enough information to make a great prediction based on the results from Valencia.

As a deckbuilder and lover of Constructed, all of this is a part of why I so greatly preferred the days of the single format Pro Tour. I wanted to know that U/W Control was definitively the best deck in the room and that it was better than the U/G Combo deck that was brewed up by Team X. I wanted to know that Big Zoo was better than Counter-Cat Zoo. Unfortunately, there isn’t actually a way to unravel this information the way that the Pro Tour works now. I understand the reasons why Wizards of the Coast has chosen to go with mixed format Pro Tours, but the purist in me just wants to know.

Right now I’m in the midst of brewing for Grand Prix Richmond. I’m still not sure if I’m attending or not, but if I do, I can tell you this: none of the things I’m contemplating fit into either the "Actual Modern Top 8" or the Top 5 performing archetypes above. I’ve got my Burn deck, I’ve got my Rock deck, I’ve got my U/W Midrange deck, I’ve got my Zoo deck (perhaps a bad call?), and I’ve got a new spin on an old deck that wasn’t so fond of Deathrite Shaman.

The next step is to bash them all against the Modern lists above and to pay special attention to the five archetypes that performed the best in Valencia. I recommend you do the same no matter what you’re planning to play in Richmond.