Alright, no more distractions. It's time to find out if the companion rules change actually worked. When the change came down, I hypothesized that there would be a general drop-off in playability, but companions would remain a factor in Modern. Lurrus of the Dream-Den would take the biggest hit, while Obosh, the Preypiercer would be unaffected. Now it's time to see if I was clairvoyant.

Companions will go down as a great "what if?" in Magic. They became legal, rose, and then fell at a point when only online play was possible. The online metagame has always been a bit weird and distorted compared to paper, and thus it's natural to wonder if the complete takeover seen online would have played out in paper. Online has always been more fluid and prone to follow-the-trend groupthink thanks to a lower barrier to entry and deck switching, so it's possible that more players would have resisted companions rather than join them under normal circumstances. I suspect little would have changed, since companion's opportunity cost proved too low. But it could have!

Methodology

The purpose of this inquiry is to find out if companions were provably nerfed. To investigate, I'm taking all the posted results from non-League events, tabulating the overall data (because we might as well look at where the metagame is heading, too), and then separating out the companions. I had intended to then do some statistics on the data, but as you'll see, that isn't necessary. The answer proved to be far more obvious than I expected. I also kept track of which decks were still playing companions.

The data is from the two full weeks since the announcement. This is partially so there was enough time for players to experiment and test their decks; taking a sample right after implementing the change would reflect the confusion, chaos, and conversion of the transition, yielding muddied data. That is, muddied unless tracking the evolution from the old to the new version of a deck is the point. I wanted to see where players' heads were once they'd had time to adapt. I've lumped all the singleton decks together as Other.

The other reason is that I'm not sure when the rules change actually hit MTGO. When the announcement went up, I remember it saying that the rules change would be implemented June 4 for MTGO. Later, it was changed to June 3. However, in one matchup I played on June 4, my opponent was still able to play their companion as before, though that didn't happen June 5. Technology is wonderful, isn't it? Regardless of anything else, I'm going to sample after the point I'm sure the rules were successfully updated online.

Week of 6/7

As a reminder, right before the change went in, Prowess was on top of the metagame by a wide margin. GBx and Eldrazi Tron were second and third respectively. Lurrus was the most played companion by far, and was present in roughly 50% of all decks. Yorion, the Sky Nomad was in second, representing 13.6%. Attendance had also declined over the course of companion's domination. It is even lower now, with only 108 decks in my sample. However, that fact doesn't necessarily mean anything. My source only recorded three Preliminaries each week, and two premier events week 2.

That is a very through shaking up of the metagame. Prowess has utterly collapsed, while Bant Snow decks have surged. In many ways, they've switched places in the standings. Eldrazi has held its numbers, which in this sample let it move up a place to second. Toolbox is in the same boat. BGx has simply disappeared; I didn't record a single placement for this week. There's no definite reason for this drop-off, but I do have a theory.

There's an impulse to claim that the metagame just reverted to its pre-companion configuration. This isn't true. This metagame data now looks like what players thought the metagame was like before Ikoria, not what it was actually like. The narrative centered on Bant Snow, Urza decks, and Amulet Titan being the best decks in Modern. In reality, RG Ponza was on top of MTGO, and by a wide margin. Ponza had the right disruption and clock to knock all those durdly decks around, and was rewarded accordingly.

My guess as to this deviation is that players went back to lists they knew or believed were good in wake of the change. The various varieties of snow decks were easier to tweak, and powerful choices on their own. Thus, they were the default choice. Everyone else was behind the snow players in terms of the learning curve and didn't correctly anticipate the metagame.

The Companions

As for the real crux of this investigation, I think the companion data speaks for itself.

They're just gone! Fallen from ~76% to ~12% of the overall metagame. Lurrus remains the most popular, but only by one deck. Where once I saw six or more companions per week, now there are three, who also happened to be the most prevalent three from before. One set of data isn't definitive, but it very strongly suggests that the nerf was successful, and that companions are something to be earned rather than the default.

Week of 6/14

One data set is insufficient to draw meaningful conclusions. So now it's time for last week's data. This sample, despite being one event lighter than the previous week, was up to 113 decks. I can't see this as anything other than an uptick in participation. This further suggests that the companions were affecting event attendance negatively. In turn, this interpretation would corroborate all the Twitter chatter indicated about players getting burnt out. Hopefully, this population uptick continues.

To further drive home the point, this week looks nothing like the last. Bant Snow took a beating, while Ponza surged to the top. I suspect these changes are linked, since again, Ponza appears to prey on Snow decks. Sultai Snow has come out of nowhere on the back of recommendations from high level players. Eldrazi Tron is just slogging along, though that shouldn't be surprising. It's been putting in strong online showings for a long time, though that rarely translates into paper.

Prowess continues to languish at the bottom of the heap, but notably, Amulet Titan has also bottomed out. There was only one Amulet deck in the sample. I've always thought the deck was overrated, but still reasonable, so this disappearance is something of a mystery. This is especially true in light of it holding its ground during the companion era and being a solid deck beforehand. Given how strong Blood Moon is against the deck, it's tempting to lay the blame on Ponza's rise, but remember that Titan was solid when Ponza was really dominating over a month ago. There's another variable I'm not seeing here.

The Companions

On the subject of confirming results, let's move onto the companion data. And again, I think it speaks for itself.

Companion participation on Modern has fallen again, both in numerical and percentage terms. Now down from 13 companions and ~12% down to 11 and 9.72%, it is evident that the companion era is well and truly over. Once mighty Lurrus is now tied with Yorion. Obosh is the lone other companion. I was skeptical that Jegantha would keep seeing play due to its cost, and it looks like I was right. That there's only one Obosh left is significant.

Taking Stock

I don't think there can be any doubt, and therefore no need for any statistical analysis: the companion rules change successfully nerfed the companions. It would be cliché and blithe to add "a little too well," but this is the outcome that was intended, and I think it's justified. The way that Wizards talked about companions indicated that they were meant to be a fun but rare treat for constructed, not the omnipresent force they became. Having plummeted in metagame saturation and continuing to fall, the data clearly shows that companions are now Just Another Thing in Magic.

As for the overall metagame, it is far too soon to tell. Again, it's obvious just by looking at the standings that the metagame is far out of equilibrium and is trying to sort itself out. The only clear trends so far are GBx and Prowess disappearing: I have exactly one Jund deck in either sample.

I'm not sure how to explain the fall of BGx. Some blame a poor Snow matchup, but Jund was a fine deck before Ikoria, and Snow was far more present there. If Snow is the explanation, it must be a self-fulfilling prophecy, in which players inordinately fear the matchup enough to just not play the deck. Another suggestion, which I have no way to verify but does make sense, is that a lot of online Jund players sold part of their deck to pay for the Lurrus package. They need time to reassemble their decks. I know players that do this, so it's plausible, but again cannot be verified.

Prowess disappearing is more intuitive. It's effectively been exploiting the metagame this year, and now the hole has closed. Early on, it took advantage of Amulet Titan's non-interactivity. As Titan fell off, so did Prowess. Prowess made the best use of Lurrus, and was rewarded with the top spot in that world. Now the party's over, and interaction is more common.

My Predictions

Since I mentioned my predictions at the start of the article, Checkov's Gun demands that I reexamine them.

Lurrus

I noted that Lurrus was the most impacted, and in terms of its decline, that is definitely true. Prowess, as predicted, had to abandon Lurrus. Of the five Prowess decks I saw, only one had the nightmare cat. The rest resembled older versions. GBx took a much bigger hit than I expected. While I can't directly attribute that to companions falling off, the only Jund list neglected to run Lurrus. With both the card itself and the decks that fully embraced it suffering greatly, I'm claiming success on this prediction.

Obosh

Oh boy, was I ever wrong here. I thought that Obosh Ponza would just keep on keeping on. Instead, only one deck kept the companion. I don't need to guess; I know why I was wrong. I thought the only card lost to Obosh was Chandra, Torch of Defiance, and that's not too burdensome. I forgot that Ponza ran Bloodbraid Elf before companions. Elf's a good card, yeah? And far better than a clunky companion these days.

Yorion

I'm not sure how to call this one. Which is appropriate, considering that I thought Yorion would be ambiguous. The same types of decks still run Yorion, but it's more infrequent. I think I got the effect on Yorion correct, but the decks are still up in the air.

Reaching Into the Kit

As a note, the only archetype that still heavily relies on companions is Toolbox. Of the 10 Toolbox decks I recorded, 7 ran companions (6 Lurrus, 1 Yorion). This makes some sense, as cutting the three-mana creatures still leaves lots of combos available. And with infinite mana, tutoring for Lurrus is no burden. Another interesting note is that three of 16 Burn decks still have an incidental Lurrus. No Mishra's Baubles anymore, just the Lurrus. Because they can, without trouble.

Where Next?

The companion nerf has worked exactly as prescribed. This has left the metagame to readjust. Which it will continue to do for some time thanks to another set incoming. I think that Core 2021's impact will be relatively muted, but there will still be churn added to the existing readjustment. We'll have to wait and see what emerges.