In Development – Diversity and Crushing the Pro Tour

What’s your goal when you play in a tournament?

I hope all your answers center on some variation of “having fun,” or I’d have to recommend you do something else. However, winning is also fun, so let’s take a look today at how deck choices play into absolutely crushing at major Constructed tournaments. While we’re at it, let’s also take a look at just how diverse recent Pro Tours have been.

Diversity at the Pro Tour

The general philosophy for Constructed formats is that diversity is good. This makes intuitive sense from the player’s perspective, since it means you have a lot of options for decks you can bring to a tournament with some reasonable hope of success. “Diversity” is a different proposition at modern Pro Tours than it is for many other events, of course.

They’re engineered this way, coming in soon after a new set is released in an effort to avoid having a format be “solved” before the event starts. On the other hand, coming that soon after a new set has been released can mean that there isn’t time to figure out diverse approaches to a new metagame.

The Intuition About Diversity

Our intuition about diversity might be that it would tend to go up as a format’s size increases. Certainly, you can expect to see a wide variety of decks at a Legacy event. Of course, the real question is not just whether a wide variety of archetypes is present at a tournament, but whether there’s a wide distribution of archetypes.

Or, to put it another way, if 100 players at a PTQ are all playing the same deck, but the other 20 people are each playing something different, that’s a wide variety of archetypes but a very narrow distribution.

So, in looking at recent PTs, we might expect Modern to be the most diverse—followed by Extended, then Standard, and finally Block.

The Actual Diversity Breakdown

We could approach the question of “is this PT diverse?” in a lot of ways, but the easiest approach is to ask what percent of the field played the top three archetypes in each PT. The top archetypes are the consensus “best decks,” as decided on by the largest portions of the players attending the event. If the percentage of players playing the top deck or the top three decks is relatively small, then we might conclude that the format was “diverse” in the sense that there was no consensus best choice.

For today’s article, I’m going to focus only on the Constructed portion of all the mixed PTs since they moved to 10 rounds of Constructed (more on that below).

Here are the diversity breakdowns for the four Block Constructed PTs we’ve had since the switch to mixed format:

The percentage of players on the most common archetype varies from 21-27%, and the percentage of players on one of the top three archetypes varies from 45-59%. The least color-differentiated format, Alara Block, is responsible for the biggest overall chunk of players choosing one of the top three archetypes.

Moving on to Standard, we see:

In the three Standard PTs we’re looking at here, from 22-27% of players went with the top deck choice, and from 45-63% went with one of the top three choices.

That second percentage is pretty astonishing, with the 63% honor going to the most recent Standard PT—PT Dark Ascension.

This is a good time to mention that archetypes are conceptual groupings defined by people. This means that the 20% of people player “Delver” at PT Dark Ascension were running variations such as Spirit Delver, straight U/W Delver, and so forth. It’s up to you whether variations in Delver at PT Dark Ascension, much like variations in Jund at PT San Diego, really should mean breaking the overall archetype down into smaller categories.

I’d argue for the “how people use it” approach in this case. How would a typical player describe their opponent’s deck after a round? If they’d say “Delver,” then that’s a good category name.

One practical consequence of the top-heavy, non-diverse nature of PT Dark Ascension was that you had about a two-thirds chance of sitting down against one of the top three archetypes in each Constructed round. If you’ve successfully built your deck to take them down, that’s awesome. Otherwise, it might be pretty boring.

The two Modern PTs so far have looked this like:

Each PT saw about half of the players run one of the top three archetypes. However, PT Return to Ravnica extended the tradition of PTs Dark Ascension and Avacyn Restored by having a big chunk of its players run the most popular archetype—31% came in on Jund, meaning that you had about a one in three chance of playing against Jund in each Constructed round.

…and as we saw, if you were Stanislav Cifka that was awesome. If you were a third of the field, it meant playing mirrors in approximately a third of your games.

Finally, we can look into the sort-of-wayback machine at the now-defunct “reduced Extended” format, which applied for two PTs:

Curiously enough, these were by far the least top-heavy PTs out of our sample pool, with at most 35% of the players choosing one of the top three archetypes, and a fairly modest 16% choosing the top deck.

It’s hard to say exactly what to make of that. Was that Extended format intrinsically more diverse, or was it simply so unpopular that players didn’t develop or test for it unless they absolutely had to? An unpopular format may be diverse by default, since no one will play it extensively enough to figure out a “best deck.”

Crushing Formats via Deck Choice

So what do we do with all this knowledge about diversity?

As I touched on in the introduction, winning is fun. Now, it may be fun for its own sake, or because you get to exercise your skill as player, or because you executed on a metagame-based plan, or any combination of those and other factors. Winning requires solid play, practice, and all those other elements that we regularly write and talk about—and for Constructed play, the journey to a win begins with deck choice.

So what do you play?

Do you play the “best” deck?

Do you metagame against the consensus best deck?

Or do you play something off the beaten path entirely?

Crushing the Format

Since moving to the 10 rounds Constructed/6 rounds Limited format for PTs, Wizards has been reporting on all Constructed lists from players earning 18 points or better in the Constructed rounds. That’s everyone who did better than winning half their Constructed games, which is nice and all, but doesn’t necessarily help us think about what the best deck choice might have been for that tournament.

We could instead look at just those lists run by players who ran the tables in the Constructed portions. My cutoff for today’s article was pretty strict—X-2 or better (24+ points). This limits the selection to somewhere between 10 and 20 deck lists per PT.

An X-2 record at the PT really is crushing the format. It’s the kind of record that, if we mapped it onto an SCG Open or PTQ, would let us Top 8 the event. Given that making that top eight is a necessary stepping stone on the path to winning, it’s reasonable to examine the deck lists that get us there.

Obviously, this entirely sidesteps player skill as a consideration. Luis or Paulo could ace the Constructed rounds of a PT with a ham sandwich, so the deck alone does not determine the winner. However, what we can say about deck lists whose pilots went X-2 or better is that it didn’t keep them from winning, which is a solid qualitative starting point on the way to a win.

Risk Versus Reward—Play the Best Deck, or No?

There are obvious standout performances for rare deck choices, such as Cifka’s win with Second Sunrise. We all know—and can see from the metagame breakdowns at each PT—that many, many people try to run their special deck choice and fail. So which case are you? Perhaps more to the point, do you need to pick the consensus “best” deck, do you need to metagame against it, or something else entirely?

One basic way to look at this question is to ask about “enrichment” or “depletion” of certain archetypes in the 24+ point category. In other words, if all decks were basically the same and every match were a coin flip, then we’d expect to see decks earning 24+ points in approximately the same proportions as we saw in the starting field in general.

When we instead see a disproportionately high percentage of the 24+ club going to an archetype, we’d say that archetype is “enriched” in the 24+ set. Similarly, we’d say it’s “depleted” if a smaller percentage made it into that 24+ club. Converting that into a quantitative metric, a “enrichment/depletion” number of less than 1 means that deck underperformed in getting its pilot into the X-2 or better category, and a number higher than 1 means that the deck did very well in getting its pilot into the X-2 or better category.

To be very clear—decks scoring over 1 on this metric do better than average in crushing the field.

So using this metric, what happened with each of the top choice archetypes for the 11 PTs we’re looking at today?

There’s a lot going on in this chart, but you can ignore much of it. The two points to look at are the column marked with the green arrow and the one marked with the blue arrow. The green column shows the enrichment (or depletion) score for the most popular archetype at each PT. Again, a value over 1 means that this archetype did better than chance in terms of taking its pilots to a 24+ point record. The blue column shows the enrichment (or depletion) score for all other archetypes at the PT, combined.

Our first take home is that in Block Constructed, you take your future success into your own hands when you choose to go with something other than the most popular archetype. This perhaps fits better into our intuitive understanding of Block Constructed as a “smaller,” less diverse format. It’s small enough that even on short notice, an obvious consensus best deck will emerge—and then you don’t have a lot of additional time to figure out how to actually defeat that deck while remaining solid against the rest of the field. On average, any deck that was not the most popular archetype performed worse than chance in terms of hitting the 24+ point mark at Block Constructed PTs.

That “on average” is an important qualifier, however—more on that in a moment.

In the three Standard PTs covered in this analysis, the top archetype did somewhere between “okay” and “meh.” At both PT Dark Ascension and PT Paris, it was on average marginally better to be playing something other than the most popular deck choice.

In both Modern PTs it was at least marginally better, on average, to play something other than the top choice.

Extended was simply all over the place. At Austin, it was a great idea to be on the top choice of Zoo, whereas at Amsterdam it was catastrophic to be on the top choice of Scapeshift—at least in terms of crushing the Constructed rounds.

The Real “Best Deck”

There will always be some kind of consensus “best deck.” However, as Cifka’s and similar performances have shown us, there will often be a “true best” deck that correctly preys on the field and can power out ridiculous performances (such as not dropping even a game through most of the Constructed rounds, which is what Cifka managed).

Even among the most popular archetypes at each PT, we see a rock-paper-scissors metagame in action (or perhaps frozen in time). As a choice of consensus “best deck” coalesces, a significant portion of the field will want to play something else, both to avoid being the most common target for metagaming and to try and leverage their own metagame understanding into free wins.

In other words, the second and third most popular archetypes are often a response to the most popular archetype. That raises the question of whether our true “best deck” is likely to arise from that pool.

Here’s a look at how the top three archetypes in each PT performed.

In this chart, each row has a PT and its top three archetypes, in order. Archetypes marked in green were enriched by 2-fold or better, which is our best-case scenario for enrichment. So at PT Paris, for example, Stoneblade was an excellent choice out of the top three archetypes. Archetypes marked in yellow were enriched by less than 2-fold, which is good, but less exciting. Archetypes marked in light red (pink?) were depleted by up to 2-fold, which means they did poorly but not horribly. Archetypes marked in a bright, rich red were depleted by more than 2-fold, which meant they were a pretty poor choice if one wanted to run the tables in Constructed.

The first thing that stands out here is that the second choice is terrible in Block Constructed. There are no good results here for the second most popular archetype, and in most cases it’s just plain old bad. Going as deep as the third choice started to turn things around in Zendikar and Innistrad Blocks, but remained pretty poor in Alara and Scars of Mirrodin Block Constructed.

In contrast, Standard, Modern, and Extended are more mixed bags from which I can’t really derive good conclusions. Make of them what you will.

Or Maybe the Really for Real “Best Deck”

…but that was a look at how the three most popular archetypes performed. What if we instead ask which archetypes did much, much better than expected in terms of hitting the 24+ point mark?

To avoid some probabilistic skew, I’ve cut out of consideration any archetype that saw fewer than 5 players running it. The way the enrichment calculation is done, if 1 player runs a rogue build and does quite well, their pet deck will appear massively enriched, which is misleading.

Here’s the same table we just saw, with two added columns:

In only two cases, at PT Paris and PT Avacyn Restored, did one of the top performing Constructed decks come from the three most popular choices. In general, decks in this category saw player tallies coming in at 2-5% of the field.

This means that even in those cases where playing something other than the top-choice deck is on average a poor choice—such as every Block Constructed PT—some select archetypes massively outperformed the most popular deck.

That returns us to a risk-reward question that has a little bit of nuance, depending on the format we’re playing in. In Block Constructed, the risk is clear—deviating from the most popular choice carries an average risk of not leading to a clear Constructed winner. In other formats, it’s less clear, but obviously, there’s a lot of room for uncommon deck choices to crash and burn.

A Little Homework on Metagaming and Success

So, the take home from part one was that format diversity may not run quite the way we expect, and that recent PTs have seen what may be an increasing concentration of players in the top one to three archetypes.

The take home from part two is that the most popular archetype is only the average best bet in Block Constructed, and that the true performers in terms of crushing the format tend to live outside the top three choices.

Let’s close today with a bit of conceptual homework about metagaming at the PT, taking another look at that last chart:

magic (at) alexandershearer.com
parakkum on twitter


Scroll to Top