MTGO February 2017 Power 9 Challenge

As a signifcant improvement over last month, we had 63 players register for this month's Power 9 Challenge. Congratulations to ecobaronen for taking down the tournament!

Top 8:
ecobaronen - Jeskai Mentor Silence
ReneRandrup - Jeskai Mentor
Call1Me1Dragon - Leovold BUG Fish w/ Painful Truths
Blululululu - Jeskai Mentor Silence
kleuter - Jeskai Mentor Silence
jacksad - Omni Oath
thehello32 - Grixis Tezz
Becks84 - Foundry Ravager Shops

Top 16 Decklists:

Dishonorable 17th place mention: Myself

alt text

alt text

alt text

alt text

I think the most interesting dynamic of this tournament was the relative positioning of Ravager Shops, Outcome, and Mentor. Walking Ballista has given Ravager Shops a significant advantage over Mentor. However, if Shops pushes too hard in this direction they can fall prey to a traditionally strong matchup, combo. Shops may have spheres, but Outcome has an expansive manabase and Hurkyl's (which notably does not work against Stony Silence). Although we can always complain about sample sizes, this is supported by the archetype vs. archetype winrate percentages.

Note: Other includes 2 Leovold BUG Fish decks, 1 Null Rod Merfolk deck, and 1 goblins deck. The two BUG Fish players did fairly well at this P9. Big Blue was completely encompassed by Outcome (both Paradoxical Mentor and Paradoxical Storm). Paradoxical Oath was listed separately under Oath. If you want to play around with the data on various combinations of tags, save a copy of googledoc to your account and change the Include/Exclude tags fields (B24-B25).

Link to googledoc:

EDIT: Updated a couple tags based on the published decklists. Will update screenshots when I get home.

last edited by diophan

Better restrict something from shops

This may be a dumb question, but where is the best place to get a long-range set of data on archetype matchups percentages? Is it just something that you occasionally find for larger tournaments, or is someone tracking matchup data somewhere?

@jhport12 We did this with the first batch of data sets we collected. The problem is that decks and archetypes tend to be very dynamic...The Gush decks winning now are typically Jeskai Mentor but easily could become Grixis Pyromancer as an adjustment to both Outcome and other Gush decks. Even within specific deck types, people will tweak their sideboards to combat metagames, so a matchup (like Dredge) that is heavily influenced by sideboards is prone to varying based on metagame prevalence (more Dredge = more hate). It makes data collected over a span of months of questionable value and we didn't continue through with it.

I suppose that makes a lot of sense. Because even if an archetype's engine could arguably be measured, singular card choices could easily skew the results massively.

@Bibendum I assume this is a joke? Gush had the highest % of the meta and the highest win % against the meta.

@diophan Are the top 8 lists posted somewhere? I like the idea of silence mentor, and would love to see the list.

last edited by garbageaggro

@garbageaggro I'll post a link once Top 16 lists are available on WOTC's website. It should look something like this though:

@jhport12 That said, if you're interested in doing something like this we have/should have all the data publicly available.

last edited by diophan

@garbageaggro sarcasm doesn't translate well in text

@garbageaggro : don't take the list we played under "kleuter", that list is flatout bad 🙂

As the data points to an objective problem with Gush Mentor, I also want to add that subjectively I am really tired of seeing that deck. I have a friend who's a long time tournament mainstay in the Northeast that stopped playing Vintage specifically because of it; he just hated it. I suspect the feeling is widespread. Even as someone that's won a lot with it, I think it's time to put that blight out of its misery.

last edited by brianpk80

@Smmenen Walking Ballista was just released on MTGO and many players (Rich and Brian, for example) opted to play Shops because there simply wasn't time to test against Ballista. Using that event as a evidence of anything regarding the long-term metagame or health of the format is just foolish...

Note: I'm not saying that the 40 person event before it means much, either...Still, I'm always concerned when the consensus "best deck" entering an event actually emerges as the empirical best deck.

last edited by Guest

@Smmenen I'd argue that the fact that the circumstances in January had people try different decks only for things to swing back heavily in Gush's favor the following month supports the narrative, not the opposite. People had no time to really test so they decided to try out the "new hotness" from Aether Revolt, but Gush was back in force once people had time to actually assess things and adjust accordingly.

I think people would gain more insight from this if they didn't think in terms of binaries (shops is too good, restrict gush, etc.). Performance needs to be understood in the context of the expected and actual metagame.

IMO the evolution of the metagame and the results from these snapshots should be thought of more like a stochastic process/Markov chain.

If you look at the daily results leading into the P9, there was a larger than usual proportion of paradoxical and oath decks. Assuming people aren't operating on anything higher than the 0th or 1st level, this rewards people playing something like Mentor Silence and disincentivizes shops. To be clear, Mentor Silence can play Stony Silence in the maindeck, which does huge damage to a large portion of the metagame, can FOW/MBT a turn 1 play, and has access to powerful white sideboard cards for Oath and Dredge. Also note that gush decks were near an all time low, so they were less incentivized to cannibalize themselves for the mirror. Shops was put in an awkward position where they have a hard time dealing with the mirror and token generators (ballista), being able to run null rod to deal with paradoxical, and beating Saheeli Oath.

Contrast this to the last P9. Coming into the tournament there was a ton of Mentor Silence. Paradoxical hadn't been doing well because it was being oppressed by this deck. This scares people away from the achetype, and the field was such that shops was allowed to crush with ballista. Even though Mentor Silence ran two null rod effects, they aren't Leyline of the Null and it's in no way guaranteed they can resolve with such a small manabase while being taxed.

last edited by diophan


I think this is a good point. I only ever play Saheeli Oath, but I adjusted by sideboard and maindeck to have 1 Stony Silence each because I had seen an uptick in Paradoxical Outcome decks (including some PO Oath decks) during the week's run up to the P9 Challenge.

@Hydra said in MTGO February 2017 Power 9 Challenge:

@Smmenen I'd argue that the fact that the circumstances in January had people try different decks only for things to swing back heavily in Gush's favor the following month supports the narrative, not the opposite. People had no time to really test so they decided to try out the "new hotness" from Aether Revolt, but Gush was back in force once people had time to actually assess things and adjust accordingly.

Except that's not actually true.

The aggregate daily data of February so far shows that Gush decks are far from dominating, and that Gush Mentor in particular is far from the "consensus" or clear best deck, at least empirically.

There have been 100 reported decks so far in February, which makes the numbers easy to calculate. But there were 27 Gush decks, 28 "Thorn" decks, and 11 Paradoxical decks (and 8 Dredge and 6 DPS decks).

In terms of individual archetypes or decks, Ravager Shops is by far the best performing deck in February, at least according to aggregate daily decks, with 21 appearances. In contrast, there were only 15 Gush Mentor decks, and that's combining Esper Mentor, Sylvan Mentor, Jeskai Mentor, and UW Mentor.

I'll post the full February (and January) metagame breakdown at the end of the month, so folks can see how all of the various decks are performing.

As I wrote in my "Notes on the State of Vintage" article last month, there are micro-trends (for example, PO decks pretty much put up all their results in February in the last two weeks), and then month to month trends as well.

But one thing is clear: the aggregate daily results from January and February establish that Gush Mentor is far from dominant. It's not even 20% of the metagame, and it's way below it's performance in 2016.

last edited by Smmenen

Looking strictly at daily event results is a poor way to conclude anything about a meta game, especially when compare to a 6-7 round event with a Top 8. 4 rounds of Swiss can contribute many factors to why a deck may appear to be more common, the first of which is how many appearances one deck manifests itself while being played by the same person. Luck, variance, and actual number of deck registrations (i.e. What decks were register and not cashing) is far more vital to determining what decks cash in a 4 round Swiss tournament that occurs at a (near) daily rate. With a card pool that recently expanded and virtually only added anything on consequence to a single deck, we can also expect to see more people gravitate to testing that deck in preparation for local tournaments and/or the Power 9. I think it speaks volumes that despite being everywhere for the last month in daily event cashings, Shops were a bit player in this weekend's meta game. When players decided which deck to use at a large and significant tournament, the deck most often chosen was Gush.

@Smmenen I could with your conclusion only if I were lining up to play a Daily Event or other similar 4 round event. Most paper events are not, so this carries limited value outside of trends of play, rather than what is actually competitive.

@Smmenen As others have stated, is this data going to be including all known points of data from larger events as well as in person events? Judging based on data from dailies I don't think would mean much, just because they are small events that don't pay out much so people will use them to try out new things. If that larger event data is included then I will be very interested to see what is reflected in it.

That's way more Leyline of the Void than I'm comfortable with.

  • 21
  • 16265