Vintage Championships 2016 - Decklists and Metagame Analysis

@Topical_Island The boxes that the decklists are in are (I believe) all the same size. You could put them all in one sheet and then have a double filter, the first which checks if the "gush/workshop" card is in the deck (where "deck" is a range corresponding to the dimension of the decklist boxes), and the second checks if the "TKS/Basic Island" is in the deck and with what frequency.

This would be a fair bit of work, but it should be programmable. Don't feel obligated though.

last edited by diophan

@diophan I don't, but I'm very interested in knowing, even just for myself. In the mean time this stuff Swords is just fascinating. It is so widely played and there is just know association with winning at all... it has me kind of staring at the matrix over here, scratching my head and muttering, "what does it mean?"

I'm doing Jace, TMS right now. I'll get bored soon and look into using the filters you suggested. Thanks for the tip.

Eliot Davidoff has 6 Swords to Plowshares in his decklist... if that matters.

@Topical_Island Fixed. I used Copy/Paste for the templating and looks like I forgot to update the SB. Thank you for your help.

@ChubbyRain No problem... not to be picky either (I really hate when I am and when people are too) but there are some decks that lack the swiss points, so as I comb the data I'm just throwing those out on everything where I'm trying to tie cards to performance. But are those swiss points recorded somewhere?

@Topical_Island Look on the "Player Info" sheet. If it's not there it should be in the winrate file. I'll glance through the meta data file and update the people without points that I find. Although instead of using swiss points a more accurate metric would be wins/losses since points treat 1-1 drop the same as 1-8.

last edited by diophan

@diophan Sorry to be a dunce about this, but which page is win loss on?

@Topical_Island It's a separate file, Win Percentages by Archetype and SubArchetype (click the link in the OP). You can get the total wins and total losses of each player from there.

In retrospect it probably wasn't a good idea, but I did the winrates in a separate file so that it wouldn't clutter up the decklist file. Also the DCI has several people's names incorrect and if we corrected them then I'd have to correct them on the sheet for every single round, which I didn't want to do.

last edited by diophan

@diophan No worries. Thanks again for all the work.

Some more breakdown here. Decks that used Mindbreak Trap and Mental Misstep together, did much much better than ones that only used Misstep.

@Topical_Island does that stat mean anything? Say, just for example, that Mentor decks play both and Pyro decks play one or the other. In that case it could be Mentor that increases those decks placing, not the misstep/mindbreak pairing. Unless you're breaking down by archetype, I'm not sure you can trust that conclusion.

@Topical_Island said in Vintage Championships 2016 - Decklists and Metagame Analysis:

Some more breakdown here. Decks that used Mindbreak Trap and Mental Misstep together, did much much better than ones that only used Misstep.

I think that's spot on, considering that after the tournament even Montolio was playing Mindbreak in Shops. Wondering though, how did decks do without Misstep, but with Mindbreak and Force?

@Lesbimagical Yeah, I'm doing a lot of these, and very often I get to the end of computing it all out (Swords) and I'm just scratching my head going... is there meaning there?

I think it does mean something though. Basically you should test 3 Missteps and a Trap before you just play 4 Misstep without ever thinking about trap... that's my takeaway.

Although I appreciate the effort, I think that the sample size is much too small for such a high variance game that Magic is to tell if there is a correlation.

While that might be the case, one could technically provide statistical uncertainties along with percentage win rates. It would give a clearer picture since most people don't really have a good idea of how large the sample size needs to be before they'll feel satisfied with the conclusion.

Both statements I agree with. What I'm basically doing here is combing the data for hypotheses rather than conclusions. For example, I only said that people should test Trap. I think that Mindbreak Trap is a little underplayed right now. I thought that before, and this data supports that hypothesis in a mild sort of way. I'm not claiming it confirms the conclusion that you should certainly play Trap or anything. I'm only claiming that as a card, it outperformed other counterspells (not technically countermagic btw) in this tourney, which is just a fact. And since this forum is about sharing knowledge, I humbly suggest that people test it in place of possibly superfluous Missteps.

More crazy facts about champs.

There were 342 "real" decks. (Excluding the non-sense decks) And 342 x 4 = 1368, which is the maximum possible number of any single card being played in the field. (Excluding basic lands).There were 528 copies of Mental Misstep at champs, or 39% of the maximum possible number. That is the highest number of any card.

As staggering as that number might seem, it shouldn't at all be surprising. On mtgtop8, Misstep is the most common card in vintage decks top8ing for the last year, when you count copies in sideboards. It's in 61% of mainboards, at an average of 3.6 copies per mainboard. So that's an even higher rate of saturation among top 8 decks. 55% of the maximum number of Missteps that could appear in top 8 decks of the past year.

Yet in this data that just got compiled from champs, (thanks again Matt and Ryan) there is basically no association with actually winning games. (At least going by Swiss points.) And the main reason for that is probably (and I'm just hypothesizing here) that there are just so many Missteps floating around. As many Missteps are on the winning side of match ups as losing ones. They just saturate the format.

The clear exception here is Shops and Eldrazi style decks that just sidestep the issue by playing few or no cards at the 1CMC level. And those decks do enjoy a better than 50/50 win rate as we all know. Which might mean that Misstep as a card is both, the most commonly played card in Vintage, and loses more games than it wins. (Though pitch Dredge might balance out its win loss somewhat in regards to Shops decks, I really haven't done the work to check out those hypotheses yet, and that would take a long long time.)

@Topical_Island said in Vintage Championships 2016 - Decklists and Metagame Analysis:

Yet in this data that just got compiled from champs, (thanks again Matt and Ryan) there is basically no association with actually winning games. (At least going by Swiss points.) And the main reason for that is probably (and I'm just hypothesizing here) that there are just so many Missteps floating around. As many Missteps are on the winning side of match ups as losing ones. They just saturate the format.

What about blue non-misstep decks? Did those out perform blue misstep decks?

@spook That's a great question. Gimme and hour and I'll look it up... care to define a "blue deck" or should I do that for us?

This is amazing, thank you guys! I was going to ask if you were going to try to control for player skill, how would you do that? Use DCI ratings, assuming you can get them? That would be incredible.

  • 70
    Posts
  • 54170
    Views