Toward the end of this week, I’ll be submitting a research proposal to the NIH. The research my collaborators and I are proposing will include a tremendous amount of systematic analysis aimed at better understanding the nature of fundamental microbial biology.

I mention this because the release of a new set affords us the opportunity to apply some systematic analysis before we decide on our first deck designs in the new Standard and Extended formats. In fact, this is the time when it serves us best to be analytical, because it’s an especially easy time to fall victim to narrative and assumption.

Narrative and assumption undergird a lot of our poor decisions generally, and our misunderstanding of card value and deck design in particular. ‘Narrative’ is our tendency to fix on an explanation that sounds good without ever bothering to test it. Stock market reports are rife with narrative. “The market went down today as traders responded to poor reports from the job sector.” Do you suppose the news agency carried out a representative poll of stock traders and asked them what drove their behavior that day?

Probably not. Instead, the reporter found an explanation that sounded fine to him, and ran with it.

In the same sense, I’ve seen a lot of interesting conversation around the new manlands following the release of Worldwake. I tend to think they’re undervalued right now, as lands that simultaneously fix your mana and beat face are simply tremendous. However, the conversation that really caught my attention was not just about whether animatable dual lands are good or not, but the assertion that they’d be “terrible in Jund.”

On the face of it, this is a reasonable narrative. Jund decks have tricky mana, and you already have a number of lands that either always or conditionally enter the battlefield tapped. A deck that can be completely shut down by a handful of janky blue enchantments certainly doesn’t seem like a viable candidate for adding even more “slow” lands.

It’s a reasonable narrative. But is it true?

The Ravine assay

When we’re trying to test a specific trait, we want to have the fewest possible number of moving parts. In scientific terms, we want to have just one variable to keep track of. This is a perpetual problem in Magic deck design and testing, especially in the immediate wake of a new set. We have 145 new cards to account for, some of which may render entirely new archetypes viable. This means, in general, we go with our fuzzy impression of the worth of certain cards as well as our broad knowledge of core deck building concepts such as the mana curve.

That said, I found myself with a very specific question on my mind last week.

Are the new creature lands “terrible in Jund?”

Based on my experience playing with and against Jund decks, I wouldn’t have been happy putting old-style lands such as Treetop Village into a Jund deck. The speed lost to them entering tapped would not be balanced by their abilities. In contrast, with the ability to swap in the new lands one-for-one with the M10 duals that already live in a Jund deck, I imagined that the potential speed loss would not be as severe – after all, those M10 duals frequently come in tapped in a Jund deck anyway. Given that, it was my hypothesis that the new Worldwake duals would yield more benefit than cost for a Jund deck.

I decided to test this hypothesis by evaluating in a system where I’d only be changing one variable – the M10/Worldwake dual mix. To make sure I was focusing on just that change, I decided to test a Jund mirror where that would be the only difference.

Here I’ll digress back to science again for a bit. Sometimes, when we change something, the effects are obvious. If I delete the gene that lets you metabolize sugar, you die. At other times, the effects are subtle. In those circumstances, we like to apply a “stressor” of some kind to accentuate any subtle effects. For example, if I delete the gene that lets you digest lactose, you may not notice it unless I stress you by having you drink a nice, tall glass of milk.

I wouldn’t want to stick around for the outcome on that one.

In this case, I figured that the appropriate stressor, given that I was testing a change in the mana base, was land destruction. With this in mind, I decided to test what was effectively a “pre-boarded” Jund mirror, using the following frame:

Ravine Test Jund (not recommended)

The “TEST SLOT” position was made up of some mix of Rootbound Crags and Raging Ravines to give a total of four R/G duals. For example, the Jund-3 test deck has one Rootbound Crag and three copies of Raging Ravine.

I then went ahead and tested near-mirror matches between the Jund-0 deck (containing 4 copies of Rootbound Crag) and each of the Jund test decks from 1 through 4. For clarity’s sake, that means:

Jund-0 (4 Crags, no Ravines) versus Jund-1 (3 Crags, 1 Ravine)
Jund-0 (4 Crags, no Ravines) versus Jund-2 (2 Crags, 2 Ravines)
Jund-0 (4 Crags, no Ravines) versus Jund-3 (1 Crag, 3 Ravines)
Jund-0 (4 Crags, no Ravines) versus Jund-4 (no Crags, 4 Ravines)

For each of these near-mirrors, I ran a set of six matches, with each deck on the play half the time. I used my general playtest mulligan rule, which boils down to “mulligan to six, and then keep mulliganing and stay at six cards.” I’ll explain that more in the future, but it’s a good time-saver if you have already done the basic testing on your mana and you’re now testing other aspects of the deck.

I’m reasonably sure this is not a statistically powerful evaluation of these matches, but I am just as sure that I couldn’t take that many Jund mirrors over the course of a weekend.

So what did I learn?

First, it’s kind of depressing to cascade right past one of your Ravines.

Second, there’s very little cost to having even up to four Raging Ravines in the deck in the place of Rootbound Crags. In my testing, there were a couple instances of a Raging Ravine cutting off one potential play on a single turn, but none of them were game-losing. Again, even at twenty-four games across four matchups, the sample size is small. However, consider that in the Jund-0 list that runs four copies of Rootbound Crag, the likelihood that you’ll be able to play an untapped Crag by turn three on the play is about 60%. Slightly more than half the time, you get that land untapped. Converting this to “never untapped” with the Ravine thus isn’t even the equivalent to swapping a normal land with a tapland – you’re getting an animatable land at about “half price” in this case.

Third, the Ravines rarely made a positive difference, either. Most of the time a Jund deck has some kind of viable play in hand, so your mana is constantly in use. Unlike Treetop Village and friends, the Worldwake lands require a minimum of three mana to activate. This means you’re typically unable to do something useful and animate one in the same turn. Note that I was testing the costlier Raging Ravine (four mana to activate) and a Jund deck could also profitably run Lavaclaw Reaches, which I did not test.

However, in the rare cases where a Raging Ravine managed to activate and not eat a Lightning Bolt, it was a real beast. A progressively growing 4/4 in the late game is tremendously hard to deal with, especially in an environment where the bulk of the playable creatures don’t deal more than four damage and decks depend on Lightning Bolt for removal.

The third discovery had nothing to do with my hypothesis going in, and that pushed me in a different testing direction altogether.

Testing the edge condition

One of the givens of scientific research is that you’ll often find that the most interesting result that comes out of your work doesn’t really relate to why you started the work in the first place. For example, you might have started studying extreme thermophiles simply because you were curious about how bacteria could live in boiling water. Later on, another researcher might have invented an amazing DNA amplification method that just needs some high-temperature enzymes to work. Combine the two, and you have PCR, the fundamental basis of much modern biotechnology and criminal forensic science.

If you were to look at my notes from the first round of Raging Ravine testing, you’d see “KRB” written in many of the game notes. That stands for “kicked Ruinblaster,” and it was often the deciding factor in many of the games.

This suggested that perhaps I ought to consider a completely different land from Worldwake. After all, if Ruinblaster is the deciding factor in so many Jund mirrors, what will Tectonic Edge do?

In looking at Tectonic Edge using my “Jund mirror” system, I was not going to simply swap Edges in one-for-one with color-producing lands. The mana base almost certainly wouldn’t stand up well to that kind of insult. Instead, I decided to start swapping in Edges in place of spells, swapping out some of the higher-cost action on the logic that although your Edges keep your opponent off mana and sometimes off a color as well, they also necessarily slow down your own mana development. The swap-out schedule looked like this:

1 Edge – Remove 1 Broodmate Dragon
2 Edges – Remove 1 Broodmate Dragon and 1 Garruk Wildspeaker
3 Edges – Remove 2 Broodmate Dragons and 1 Garruk Wildspeaker
4 Edges – Remove 2 Broodmate Dragons and 2 Garruk Wildspeakers

Then I went and ran another series of test matches as before between Jund-0 and Jund-TE1, 2, 3, and 4. The number is the number of Tectonic Edge swapped in, naturally.

By the way, it’s at Jund mirror number thirty-six or so where you really start to develop a headache.

So what did I learn this time around?

First, I literally never once found myself with a Tectonic Edge in hand, wishing it were one of the cards I’d sided out. There were no test games that turned on the absence of a Broodmate or a Garruk.

Second, Edge is significantly less effective than Ruinblaster. This comes down to the fact that you can hit a kicked Ruinblaster on your turn four on the play and take an opponent down to two lands. In contrast, a turn four Edge on the play does nothing in particular. In addition, Ruinblaster gives you a body, which means you can eat one of their lands and present a blocker if they already managed to get something onto the board.

Third, Tectonic Edge is tremendously effective at keeping someone off of their big mana spells. If your deck is engineered to work with about four mana, you can happily sacrifice your Edges to keep your opponent off of their five- or six-mana spells all day long. This came up in the test matches from time to time, but would be much more evident if your opponent chose to run a higher-end Jund deck featuring Siege-Gang Commander and other expensive cards.

General conclusions on Edges and Ravines

Although I was specifically asking about Tectonic Edge and the Worldwake duals in the context of Jund, I am more generally interested in their application in the upcoming Standard, and I think we can generalize from my observations here.

First, the Worldwake duals are legitimately amazing. They are quite playable in a deck with fairly picky mana requirements, which means we can expect them to be eminently playable in decks that only run two colors, or for which their three-color requirements are less stringent. Pretty much any G/W deck is going to want to run three or four copies of Stirring Wildwood, for example. Remember that we were already playing lands that enter the battlefield tapped just to fix two colors and perhaps gain a slight bonus of life. Getting that same color fixing combined with something to do with your mana from the midgame onward is tremendous.

Second, Tectonic Edge stands to really wreck decks that depend on getting to higher mana counts. It certainly appeared to be worth it to side in extra land over “action” spells in the test Jund mirrors. The ability to prevent big spells – or, indeed, to keep your opponent off a color – is also tremendous. Tectonic Edge is a “fixed” Wasteland in the sense that it’s not quite as much of a kick in the gut as Wasteland, but it is still quite good.

I’d convert these observations into the following actionable statements:

I recommend running Worldwake duals over M10 duals, unless extensive testing shows that it does not work in your specific build.

I think every deck will need to have four copies of Tectonic Edge in the main deck or the sideboard.

Both of these statements are meant for Standard, of course. I do think the Worldwake duals have a place in Extended as well, but I’ll need to do some more testing before I can say anything decisive about them.

I’ll close today by reiterating that I’m a big fan of just testing things out. It’s important that we critically evaluate our narrative thoughts, whether they’re “these lands are terrible in Jund” or “these lands will be amazing.” More to the point, it’s fun testing these options instead of just ruling them out ahead of time. As with any experiment, it’s exciting seeing what works, what doesn’t, and what unexpected results will lead you off in an entirely new direction.