As much as I appreciate the huge deal it should be to gather all info and keep track of a LOT of decks to finally present us the charts, I also enjoy how you explain the changes and make some good points. Keep it up man.

Taking a couple steps back and seeing the data of the last months (with MS Excel help), I’m starting to clearly see a lot of trends that I supposed were a thing, but couldn’t mesure feasibly. In fact, I feel like giving some of my own September predicitons:

1) Burn will keep rising, maybe to April levels (9.2%). Merfolk and Infect will go on too!
2) Seems like Jund hit its top in July. It will fall to 3rd place.
3) Affinity falls along Jund, but not as hard. I don’t see Robots with less than 6% though.
4) Grixis Control won’t be able to keep up with Merfolk, which is sad being a grixis player myself.
5) Infect is back! Grixis Twin was the first step to Tier 1.

Also, on the Twin topic. As Trevor stated, grouping them could be tricky. One can argue about gameplan and whatnot, but in the long run, metagame wise, having them broken down into 3 or 4 (I’ve seen some jeskai twin) different decks is the wise choice. A combined 8% may be hiding some weird 6%/1.5%/1% spread (again, exaggerated, but you get the point).

Sure! It’s weird to look at the Tier 1 tables and not seeing the premier combo deck of modern, which somehow defines the Turn 4 rule (I know…), but that’s how things are. I used to think the UR Twin decline was just player migrating to the Grixis variant, but seeing both decks on the decline means something else. Maybe Grixis Control? I don’t know. But be sure it’ll come back. My bet is on the October metagame breakdown. Cheers!

A minor point: this discussion about grouping variations on archetypes comes up often, and there are good arguments for both sides. While it’s true that it can be dangerous to lump Temur/Grixis/UR Twin together under the same umbrella (as they have similar strategies but certain matchups can exaggerate the differences in deck composition) I have always been of the opinion that for the purposes of metagame information, they should be grouped. While it is undeniably true that no individual Twin archetype experienced enough success to place it in the Tier 1 bracket, combining the results shows us that Twin Variants account for 9.5% of the overall metagame, more than any other macro-archetype in the format.

This is important because to the casual reader, it can paint a different picture of “Twin on the downswing”. I am aware that you anticipated this danger as you talked at length about the possible reasons for Twin’s perceived decline, but the fact remains that there is a danger to this kind of thinking, as it can lead us down a path of: “Twin is under-represented, it’s safe to play something that loses to Twin, run headfirst into Twin variants that are almost 10% of the field”.

On the other hand, the danger of grouping Twin variants together is that it can be harder to perceive small shifts in representation. Combined, the Grixis/UR/Temur variants have a 4.2/3.5/1.8% representation, but this could theoretically shift to 1.6/0.8/7.1% in favor of Temur and those of us just paying attention to “Tier 1” numbers only wouldn’t catch the shift. This is an exaggerated point, of course, but with smaller numbers and shifts that don’t elevate a particular Variant to Tier 1, we could (and have in the past) see shifts where Temur Twin surpasses U/R in popularity for a short time and takes the field by surprise because they were caught looking at Tier 1 Combined statistics only.

Again, there are benefits and downsides to both strategies, I have just found that the risks involved with splitting variants can often be greater than the risks involved with grouping them. Regardless, excellent article and great points!

There wasn’t much to say about Abzan in this update. It’s at around 3% of the metagame, which is roughly where it was last time. There are signs that Abzan might make a resurgence in September, but those signs are mostly from September events, not from those in August. The only promising datapoint for Abzan mages is the 4.2% prevalence in Day 2 of SCG Charlotte, but unfortunately the deck’s paper and MTGO shares were a lot lower over the month.

That’s completely reasonable, and the article do go in great lenghts to explain what’s going on. I still don’t think that UR twin plays that differently from the other versions though, but this might because of bias from the decks I usually play (where I won’t change anything in deck building wheter they have tarmogoyf, tasigurs or not)

What you lose by not grouping them together is that someone who glances over at the Top Decks table on the front page might think that ” Twin” suddenly became unpopular. Meanwhile, someone who read all the way through and paid attention could see that still is important to prepare for any versions of Twin on the aggregate.

Anyways, I love those articles and please keep them coming. : )

I think Abzan is definitely poised for a comeback. Will it surpass Jund in September? I don’t think so: there are still enough matchups where Jund is going to be a better option. Will Abzan get back to tier 1? That’s much more likely, and I imagine we’ll see it by the end of October (if not earlier). Abzan is a great choice against a Jund and Grixis heavy metagame, which is probably what many of the pro players are expecting. Then again, the last time the metagame polarized heavily towards Abzan was February 2015, and that saw a huge movement from the unfair decks like Infect. We are already seeing Infect rising up, and I think this metagame knowledge will keep players ahead of a possible Abzan spike. The deck is bound to go up, but I think it won’t be as successful as many are predicting.

I definitely agree that Grixis Control’s rise is at play in Grixis Twin’s drop. From the article: “A more plausible explanation is the parallel rise of Grixis Control, the “better” Grixis deck. Grixis Control saw an uptick in the past month, and some part of that was likely Grixis Twin converts. ” I’m not sure what percentage of Grixis Control mages are in this category, but as you suspect, I don’t think they are insignificant.

I don’t like combining multiple decks into one supertype because you lose a lot between the different versions. UR Twin plays very differently from Grixis and Temur Twin (those latter two are more similar), so it’s not quite fair to combine them. As I said in the article, “the metagame is open to different Twin variants, so you’ll need to consider all of them in your own testing.” Going into September, I would keep testing and expectations open around all three Twin versions. You might face any of them because, as you suggest, the collective URx Twin share is definitely high enough to guarantee the matchup. It’s up to you to figure out which specific versions you prepare for.

This is the first time where no single Twin deck qualified for tier 1, which created this interesting issue around the deck classification. Normally, you have at least one Twin deck up there, so players can be ready for that main version and then be mindful of the secondary ones. If this issue persists, we can maybe revisit how Twin is classified, but for the most part I prefer the granularity gained by keeping them distinct. It’s definitely helped us in the past to observe trends, and we lose more than we gain if we eliminate that detail.

As Rackhamm said, it’s easy to underestimate how powerful Bolt is against Infect. The Hierarchs alone make the spell very powerful in the matchup, as does the prospect of a turn 1 Glistener Elf. That’s not a card you want to Path: transitioning from turn 1 G-Elf into turn 2 Agent with Vines backup is not what I want to help an Infect player into. I wouldn’t want to rely on Bolt alone, but Bolt in tandem with the usual suspects of Snapcaster (in URx decks), Path (in UWR decks), and Terminate/Decay (in BGx decks) is way stronger and scarier than just those other removal spells without Bolt behind them.

Obviously you don’t side bolt out or anything, but it isn’t ideal. Nothing is surefire (like I stated in my original comment) but there are several other removal spells that are harder for Infect to play around. I just don’t think it should be looked at like the pinnacle of Infect hate.

Absolutely agree that separating the twin variants seems wrong. Does adding tarmogoyf honestly change the entire deck? no. How about replacing roast with terminate and adding tasigur? nope.

As an opponent you play in fear of the exact same thing: Eot exarch, untap, twin. That is what defines these decks, not whether they play tasigur, tarmogoyf, or electrolyze (or whatever card we consider to be the defining difference for straight UR).

I agree with this point – I’ve faced UR Twin, Grixis Twin, and Temur Twin, and I can’t say that the strategy was very different. Grixis has more of a control Plan B while Temur favors midrange, but at its heart, it’s still an interactive deck that has a combo win-con in its back pocket. I think it’s probably time to revisit whether they should all be considered separate decks or logged under “Splinter Twin”.

As for the decline… can at least some of it be attributed to the rise of Grixis Control? It seems to me that the natural crossover for a disillusioned (or bored) Grixis Twin player is to take the combo pieces out and just play Grixis Control instead.

Its true that they can pump in response, but bolt is still very good against infect. Bolt turn 1 glistener or noble hierarch? Yes please. Bolt your guy eot when you play him or on my turn? Yes go ahead and pump him while he has summining sickness please. You can also bolt in response to pump spells. Plus bolt decks tend to run more removal alongside lightning bolt. Even path and terminate aren’t surefire removal when they can get nullified by vines, spellskite, pierce, or blessing.