MTGO Vintage Metagame: May  July

Introduction
Bah, it's four in the A.M. Here's some data.
Top 8 Results
Of note, the Paradoxical lists with the asterisks contained Monastery Mentor as their primary win conditions. Neither list was the standard 4X Mentor Outcomes, but rather eccentric lists from players such as @iamfishman and @brianpk80. These lists contained 2 Mentors each, with Brian running one in the SB "to come in against other Mentor decks". If counted as part of the Mentor archetype, that brings the total up to 28.4%.
Metagame Breakdown
This breakdown is based on archetypes and not tags. The actual percentage of Mentor in the metagame is slightly higher. Unfortunately, data collection has been inconsistent as Ryan and I have missed events and relied on others to help us. It's a significant amount of work and we are immensely grateful to @desolutionist and others for that help. However, we don't have tag breakdowns available for every event. The percentage of Paradoxical Mentor is typically low, normally 24 players on a given day. A reasonable approximation would put the percentage of Mentor at about 25% of the metagame, or on par with Shops. Again, these decks aren't necessarily focused on Mentor and the number of copies is pretty variable, but since Mentor is currently on the community's radar for restriction, it's best to keep the metagame saturation in mind.
Archetype vs Archetype MWP
The color codes are as follows. For win rates, green corresponds to >50%, red to <50%, and yellow to =50%. The problem with establishing a range is that the sample sizes (and therefore uncertainties) aren't consistent. We expect much more variance from the 3 matches Eldrazi had against Blue Control than the 171 matches Mentor had against Shops. The second table shows those match breakdowns, with green corresponding to >50, yellow 25 to 50, and red <25 matches. I'll leave it to the readers to discuss the implications of this.
Trends in Major Archetypes
Just to show how variable win rates and even metagame percentages can be on a weekly basis. Also, it looks that Shops is trending upwards in metagame percentage while Paradoxical is trending down.
Conclusions
It's 4:00 am. I'm going to bed.

Combined your tables, and added a corresponding column to see how deck are performing compared to their metagame presence...
Unsurprisingly, Shops and Mentor are far exceeding their metagame percentages, while everything else ranges from around 0 to massive negative.

Thank you for this great job.
Out of curiosity, do the number on MTGO and paper vintage correlate or are there signicative differences ?
Edit : i mean, there are difficulties and bias with all those numbers (for exemple, classification of decks can be difficult some times, or on MTGO one single player winning a lot with a particular deck, ...). It is not criticisme of course but just statistical problems linked to our favorite game. Those difficulties/bias are not the same on MTGO and paper vintage, so if the numbers still correlate it would be a strong hint that those bias are small enough to be neglected and would make the numbers even more strong. I am wondering if we are able to have such a precise analysis.

Great stuff thanks for putting this together!
Not sure if I'm reading the chart right, but does this imply that literally everything beats oath?

Interesting data. Has anyone published any trends from paper metagames? I would be very interested to see a comparison.

@cutlex said in MTGO Vintage Metagame: May  July:
Interesting data. Has anyone published any trends from paper metagames? I would be very interested to see a comparison.
Kevin and I do. We recorded our podcast Monday. Should be live soon.

For amusement, I calculated the Nash equilibrium metagame assuming all these win percentages are exactly correct. Obviously due to sample size that's not the case, but maybe it's interesting. I did this with excel solver instead of doing the linear algebra by hand, so there will be some noise in here compared to the true exact values.
All values are percentages
At equilibrium, shops is 4.8 with 59% win rate
Eldrazi is 7.8 with 50
Dredge is 9.7 with 50
Combo is 11.7 with 50
Oath is 3.5 with 31
Outcome is unplayed
Mentor is 18.9 with 50
Delver is 14.0 with 48
Big blue is 24.8 with 51
Blue control is 4.6 with 52.5
And other is 0.2 with 44So based on the data we have people should be playing a lot more slow blue control.
Shops is the only deck that has a much better than 50% win rate at equilibrium. Not entirely sure how that persists.

@walking.dude This is at once, very silly, very awesome, and pretty enlightening.