A Lille More Data: Computing the Conversion Rates and Matchup Results for Each of Standard’s Most Popular Decks

The coverage team gets a lot of questions during a GP weekend. Onsite people mostly ask if Riley Knight can sign their Reliquary Tower. Online readers mostly ask for deck lists. Another standout: How does the Day 1 metagame look?

The interest is understandable. If we knew how many of each deck archetype entered the tournament at the onset, the data for Day 2 would hold a lot more meaning. It would show how successful each archetype was relative to the others. Without this critical piece of information, all of the Day 1 success stories remain merely anecdotal.

The question comes up so frequently because we rarely, barely ever, have an answer. At Grand Prix São Paulo earlier this year, Willy Edel gave one guy the job to do nothing else for two days straight but to sift through all the deck lists. If you wonder why it isn’t done more often, well, that’s why. No one’s got time for this.

But even back home, at the local Draft I go to every Wednesday to spend my Magic Online downtime, I was met with: “Hey, how did the Day 1 metagame in Lille look?”

This last time, I went prepared with an answer. That is, I suggested to people that they wait for this very article.

Day 1 to Day 2 Conversion Rates

I took a random sample of one-third of all Grand Prix Lille deck lists. Since the event was the largest tournament in Guilds of Ravnica Standard to date, this meant 446 deck lists in total. And since Day 2 players’ lists were already separated from those who failed to make the cut, I took extra care for my sample to contain about 20.7% of random Day 2 deck lists.

20.7% is the share of players who made Day 2 overall—277 of 1,339—so it acts as our baseline. If a deck put a larger share of its pilots into the second day, it did well. The larger the share the better. If a smaller share of pilots reached the Sunday rounds, their deck underperformed. The following is sorted accordingly.

Some explanation:

First column: The random sample of 446 Day 1 deck lists included eight copies of Turbo Fog. Eight is 1.8% of 446.

Second column: The 277 Day 2 deck lists included eight copies of Turbo Fog too, although not the same eight. Eight is 2.9% of 277.

Third column: We assume that the full Day 1 field of three times as many deck lists included three times as many copies of Turbo Fog, meaning 24. Thus, we assume that eight of 24 Turbo Fog players, or 33.3%, made it to the second day.

You may notice that the Day 2 percentages differ a bit from the ones published as part of the official coverage during the weekend. This is due to the fact that I drew slightly different lines when sorting the Day 1 decks into archetypes. Several Golgari Midrange lists featured a minor blue splash, most often for Muldrotha, the Gravetide and/or Hadana’s Climb. I decided to lump them in with the rest of the Golgari swarm. This also affected one Day 2 list, which we counted among the “others” previously. One Arclight Phoenix deck in Day 2 exhibited all of the archetype’s aggression, complete with Crackling Drake, Maximize Velocity, and Thud, but splashed for Deafening Clarion. We didn’t count it toward “Izzet Phoenix” initially—a mistake, I believe, and it is included here under the new label of “Arclight Phoenix.” Finally, I expanded one category to include all decks that were rooted firmly in Dimir territory with regard to either color or the surveil theme.

One notable finding is (the reasonable assumption) that Golgari Midrange increased its metagame share from one day to the next. Still, both Jeskai Control and Arclight Phoenix, the next two biggest players in the Day 2 field, outperformed the boogeyman. As did many others, although not what is listed as “others” above. Those decks did abysmally, which really is no wonder since the archetype determination happened after the fact.

Boros Midrange did poorly too, whereas aggressive decks full of white 1-drops performed best. At least, we assume as much, although low numbers such as theirs leave more room for doubt. For a better assessment of their merits, it would be nice to have more data.

Luckily, I do…

I Know What You Did Last Sunday

Including playoffs, Grand Prix Lille’s second day consisted of 918 individual matches. Only one match featured the unfair battle of player versus bye, and there appear to have been ten unintentional, as well as three intentional, draws. I disregarded all of the latter 14, then worked out the combined records for our archetypes.

If we use the previous table to draw comparisons, we get the impression that Mono-Red Aggro did a lot more winning on Day 2 than it did on Day 1. The reverse is true for White Weenie and Turbo Fog. Big Red failed horribly, and the six Steel Leaf players ended up far from being champions too. In fact, they earned the dubious distinction of being most likely to drop from the tournament. Just look at how many more matches the six Dimir players completed!

Let’s zoom in and see how Golgari’s 335-368 record came about. The bad news, right off the bat, is that there were 135 meaningless mirror matches. Against unnamed decks, Golgari won eight matches and lost eleven. The rest went as follows:

Going by these results, I consider Golgari’s matchups against Arclight Phoenix and against Mono-Blue Tempo to be clearly negative, and its matchup against White Weenie to be clearly positive. Steel Leaf Stompy likely is a favorable matchup, while Mono-Red Aggro and Esper Control are likely to be unfavorable.

All the other matchups are too close to call, a conclusion I arrived at through a vastly unscientific method I like to think of as the “test of faith.” Imagine, for instance, you ran into a Golgari player at an event who claims to have an even matchup against Boros Midrange. You ask them their record versus Boros Midrange so far, and they admit to only winning nine out of 22 encounters. But then they add, as Magic players do, that one time they mulliganed into oblivion and this other time they never found a third land, even after double explore, and also this one unfriendly opponent totally threw them off their game and, what’s worse, they then made this huge blunder that cost them the match right there and that was a total embarrassment, and the bottom line is two of their losses should have been victories. Like, really.

Would you believe them? Funnily enough, you should. It is painfully easy in Magic for dumb luck or a dumb mistake to turn a game and a match or, come to that, two matches. If someone reframes a 9-13 record as an 11-11, I’m willing to give them the benefit of the doubt. If, on the other hand, they stretch such contention to three matches—well, now they’re just making excuses.

By the same token, I wouldn’t put much faith in a passionate Boros Midrange mage who submits a 13-9 record as evidence of their excellent Golgari matchup. And it is indeed the same token, because our null hypothesis has to be an even matchup. Only strong evidence can support another claim.

With that in mind, let’s move on:

Or maybe, let’s revisit. The battle between Jeskai Control and Mono-Blue Tempo may be a case in point for what I just mentioned. On the one hand, I can imagine how the presence of a 3-mana sweeper that catches everything except for Tempest Djinn warps the matchup. But aren’t fish-style decks—and Mono-Blue Tempo qualifies—supposed to beat control? Others will know and I hope they’ll leave a comment below. Either way, 75% for Jeskai definitely raised an eyebrow this side of the internet, and a 6-2 doesn’t pass my test of faith.

Meanwhile, I’m willing to accept that Jeskai Control is an easy matchup for Big Red players. As in, they might as well concede and do something more productive with their time. Wash cotton candy or something.

We’re only down to the third most played deck and already we find ourselves reduced to this. Except for Phoenix beating Golgari, we can’t ascertain any trend with any kind of certainty anywhere. My apologies to Nate, who asked in the comments to my previous article what Arclight Phoenix’s good matchups are supposed to be. I would have liked to answer the question with data, but the data didn’t cooperate. Even 24 Phoenix players generated too little of it.

Individually, none of Mono-Red Aggro’s numbers serve as evidence of strongly favorable matchups, and they are way too low to make a statement almost all across the board. In the aggregate, though, this paints an impressive picture. With a negative record in but a single matchup, one could say it’s no wonder Mono-Red Aggro won the tournament. Or, one might argue, this merely reflects what was an uncharacteristically hot run.

In the absence of relevant data to add, I’d like to share a somewhat pertinent anecdote:

In Round 10, we decided to feature on camera one of these new breed of Turbo Fog decks splashing, of all things, Ritual of Soot. Frank Karsten gave his best Vaevictis Asmadi impression (see artwork from Legends) when he heard of the double-black splash in Bant, and the wisdom of the inclusion continued to look questionable.

While the card would have been great against virtually any White Weenie deck from the history of the game, against Theau Mery’s version with four each of Dauntless Bodyguard, Hunted Witness, and Adanto Vanguard it proved quite embarrassing. The duel was over so quickly that we managed to show two additional matches that round, more than in any other round on Sunday.

Meanwhile, Simon Görtzen proved why he sat in the expert seat, explaining, expertly, that Turbo Fog always had problems surviving a very fast creature rush. White Weenie’s complete lack of reach or ways around Root Snare appears to put Turbo Fog at an advantage in the matchup, but it’s a deceptive advantage. In truth, the battle is decided, more often than not, before Turbo Fog even gets to the point where it can chain fogs.

It’s good to know that, despite what we saw in the finals, Mono-Blue evidently can beat Mono-Red.

None of Selesnya’s numbers are large enough to give any clear indication. I can see the deck win a match against Mono-Red Aggro, and I can definitely imagine how it might lose to Esper.

For metagame breakdown purposes, Frank and I decided not to make a distinction between Boros versions that ran Angels and those that used other expensive threats such as Siege-Gang Commander. Some evidence suggests that Angels fared better.

The morning fog blew over quickly in Lille. With only barely positive results in two matchups and negative records in all the others, Turbo Fog came closest to a reverse image of Mono-Red Aggro.

I leave the the remaining matchup tables here for the sake of completeness. The additional results may not provide much information by themselves. But if you have some results of your own, you may be able to corroborate your findings or to find them challenged. After all, it can never hurt to have more data. White noise this is not.

Big Red, finally, is a sad case. When Goblin Chainwhirler ruled the day, it was advisable to go bigger than other red mages. The same may even be true in the red mirror today. Sure enough, though, the one deck Big Red didn’t get to face all day had to be Mono-Red Aggro.


Some strategies probably don’t deserve further attention. In particular, Steel Leaf Stompy and Big Red didn’t do well either on Day 1 or on Day 2. Meanwhile, I wouldn’t worry too much about how unsuccessful Mono-Red Aggro appeared to be on Day 1. It’s not that I don’t trust my sample size in this regard. I do. Rather, I’ve seen the same effect before, over and over and with the same archetype. Mono-Red often is one of the cheapest decks in a format and the most popular choice specifically with the weakest players in the field. It’s been a cliché for a long time and for good reason.

I don’t expect Golgari Midrange to vacate the position of most played deck any time soon, although its metagame share could and possibly should decrease. Its biggest problem in Lille was Arclight Phoenix and it will be interesting to see whether or not the archetype can adapt. Arclight Phoenix itself showed no major weaknesses in Lille at all, and that’s notable in its own right. Then again, it was more or less the new kid on the block and a field previously ill-equipped to deal with it may yet turn hostile. It remains unclear if Standard even has enough tools for everybody to defend against recurring Phoenixes as well as against a Crackling Drake high on Maximize Velocity. Simply dialing up the numbers of Vraska’s Contempt can’t be the answer.

The puzzle of the format seems far from solved, and I’m excited to see what players come up with at the Pro Tour!

Until next time, I hope you meet lots of your good matchups, and I hope I could help with that.

1 thought on “A Lille More Data: Computing the Conversion Rates and Matchup Results for Each of Standard’s Most Popular Decks”

  1. Pingback: Standard at Pro Tour Guilds of Ravnica in Numbers

Comments are closed.

Scroll to Top