Data Analysis: The Top Modern Decks at End of January 2022

Aliquanto

The ban announcement may list "no changes" for Modern. But we know better. Every format changes, all the time, even without bannings. We gathered the data, we did the analysis, and now we can report the findings in our monthly update. Here is how Modern did change over the course of the past few weeks!


lurrus is fine

It is here! The long awaited banned & restricted announcement. There was so much hype for it after the deployment of the Pauper Format Panel and the bans in Pauper last Thursday. Tons of talks took place everywhere (Reddit, Discord, Facebook, Twitter, Twitch, YouTube, WhatsApp, forums, and I must forget a few platforms) in the run-up. Most content creators gravitated toward Lurrus of the Dream-Den as the most likely candidate for a ban, or companions in general. Ragavan, Nimble Pilferer, Urza's Saga, some removal from Modern Horizons 2 maybe? Possible unbans were called for too, without agreeing on one in particular.

While still speculating on different possible changes if any were to happen, a lot of people also called the "no change" in Modern. And that is indeed what we got. The latest announcement did not affect Modern. Nevertheless we received some feedback from Aaron Forsythe, Vice President of Design at Wizards of the Coast, saying that the format looked fine according to "data and sentiment."



He also made some comments on cards to unban that people bring up regularly, specifically Splinter Twin and Birthing Pod. Forsythe explained that that the latter is unlikely to ever come back given how threatening it would be now. One should note too that the trend of Wizards taking into account online feedback and referencing it is not over, as the tweet mentioned "sentiment," and the announcement used the words "community" and "feedback" five times.

Anyway, even without any action taken in Modern, I still need to thank them because it afforded me the perfect topic for this introduction, no matter the outcome! Now let us move on and see how Modern archetypes fared in the published tournament results from Magic Online over the last four weeks. Here are the events we are covering today:

Modern Challenge 2022-01-01
Modern Challenge 2022-01-02
Modern Challenge 2022-01-08
Modern Challenge 2022-01-09
Modern Challenge 2022-01-15
Modern Challenge 2022-01-16
Modern Super Qualifier 2022-01-16
Modern Showcase Challenge 2022-01-22
Modern Challenge 2022-01-23
Modern Preliminary 2021-12-27
Modern Preliminary 2021-12-28
Modern Preliminary 2021-12-29
Modern Preliminary 2021-12-30
Modern Preliminary 2021-12-31
Modern Preliminary 2022-01-03
Modern Preliminary 2022-01-04
Modern Preliminary 2022-01-05
Modern Preliminary 2022-01-06
Modern Preliminary 2022-01-07
Modern Preliminary 2022-01-10
Modern Preliminary 2022-01-11
Modern Preliminary 2022-01-12
Modern Preliminary 2022-01-13
Modern Preliminary 2022-01-14
Modern Preliminary 2022-01-17
Modern Preliminary 2022-01-18
Modern Preliminary 2022-01-19
Modern Preliminary 2022-01-20
Modern Preliminary 2022-01-21

Here are some quick statistics:

Number of decks in the data: 438
Number of different players in the data: 306
Number of different cards in the data: 759
Number of exact archetypes in the data: 69
Number of rounds played in the data (with Top 8s): 2,998
Average number of Swiss rounds in the data: 6.56
Minimum number of Swiss rounds in the data: 4
Maximum number of Swiss rounds in the data: 9
Number of events in the data: 29

Archetype Presence

We should be used to the process by now. As always, we start with a look at the most represented archetypes among the top finishes—Challenge/Showcase/Premier/PTQ Top 32s plus Preliminary 3-1 and 4-0 results.

They are weighted by the number of matches to make appearances in larger events more meaningful, which is relevant to try to determine the best decks later. This means, a deck that reached the finals of a nine-round event, adding three matches in the playoffs, has a weight of 9+3=12, whereas a Preliminary list comes in at a weight of 4. Interestingly, the ranking remains pretty much the same when you only take the number of appearances without weighing them.


Proportion of archetypes, out of 2,998 matches (threshold for Other: 60 matches)
Proportion of archetypes, out of 2,998 matches (threshold for Other: 60 matches)

Since last month, Izzet Murktide collapsed, now showing less than half of its previous results share. It still sits at a comfortable fourth place though, whereas Grixis Shadow multiplied its share by 1.5, moving up to first place. WURG Blink also saw reasonable growth, up to third place . In between, Hammer Time remains extremely stable. Mill is happy with those changes, the Shadow matchup much easier than Murktide, and gets back over the 2%. Less Blood Moon and more Drown in the Loch is also good news for Green Tron, appearing again, just like Blue Living End (though this one was barely under the 2% threshold anyway last time). Temur Footfalls and Burn see a rise as well, now clocking in right behind the biggest decks of the format.

On the other hand, Azorius Control has a way harder Grixis Shadow matchup than it ever had (historically preying on it) and might need to rethink its structure or wait for a meta change to come back. It also was stronger against Murktide than the current iteration of Shadow. The following table summarizes all of these changes for all of the decks that were above 2% in presence.


meta changes

Among those decks, I think I need to mention that the difference between WURG Control and WURG Blink keeps getting harder to identify automatically. Indeed, in recent events, kanister and a few other players developed a version cutting the "bad cards," taking both Top 2 spots at the same event and placing another one in the Top 32. Looks like replacing Ephemerate with Ragavan, Nimble Pilferer to make the deck even more expensive is a path to try for next month.


However, many people raise strong arguments against the use of pie charts in representing such data. Even though they are widely used in that context in Magic, bar charts are instead recommended. Besides, while I set a bar at 2% of presence for readability on the pie chart (and to be fair I should even just stop after ten or twelve decks anyway), the tier list is based on decks that are above the average presence. And this time, the average presence is much lower than 2%, sitting at 1.45%. So that is why I present to you another graph for the presence in the results today:


Proportion of archetypes above the average presence of 1.45%
Proportion of archetypes above the average presence of 1.45%

With that, three more decks appear: Creativity Combo, Dredge, and Gruul Midrange. I don't remember the last time we showcased a Gruul list (also known as Ponza), so here it is!


It is also worth showing a Creativity list again. The latest versions changed win conditions, going from Emrakul, the Aeons Torn to Archon of Cruelty.


Archetype Win Rates


95% confidence interval on the win rate of the most represented archetypes
95% confidence interval on the win rate of the most represented archetypes (click images to view at full resolution)

Do you think that so many win rates this much above 50% is abnormal? Don't worry, it was all planned. Or at least, it is to be expected. After all, as mentioned earlier, we only have data on the best-performing decks in Challenges/Preliminaries. So rather than the actual win rate, it is rather a metric of whether the deck tends to win tournaments or is at the lower end of the Top 32.

Belcher still looks very well positioned, and the most played decks also seem to have high win rates. Among the outsiders, Creativity also looks quite interesting, though its lower presence (meaning less matches) means that its win rate is even less reliable than the others, as you can see with the large interval. Besides, it might just have surprised people with the different build.

Displaying Win Rate and Presence Together

So now we have the data for both the presence and win rate. Let's visualize it.


Win rate by presence for the most represented archetypes
Win rate by presence for the most represented archetypes

It is interesting to note that there seems to be a stronger correlation between the win rate and presence than we saw previously. The big exception is Belcher, which is famous for being underplayed. (It is not a playstyle a lot of people enjoy, and it also has to be underplayed, in a way, because, when it is played a lot, it receives more hate and immediately becomes much worse.) We should also be glad to note that more decks appear on the graph than previously, which suggests that last month's trend of increased concentration on a smaller number of decks was just a temporary thing.

Once again, do not worry. I know that you like to see what happens outside of the top tiers too, so here is the complete picture (with tiers based on presence only, the only metric most other tier lists would use):


Win rate by presence for all archetypes (circle diameters based on number of different players)
Win rate by presence for all archetypes (circle diameters based on number of different players)

Are you wondering what that lonely dot in the top left side corner is? You can guess that it comes from a Preliminary given its minimal presence. Here is the corresponding list (which actually looks closer to a midrange than an aggro deck now that I check it manually, so this could need an update):


From Data to Tier List

Now that we have the data on both metrics for the most represented decks (above the average presence), we can use the same normalization formula as before on both of them. Summing up then gives us a value we can use to rank the decks.


Sum of the normalized win rate and presence for the most represented archetypes
Sum of the normalized win rate and presence for the most represented archetypes

Note that I recently changed the formula for the points a little. Here are the old and the new formula:


Previous equation for the sum of the normalized metrics
Previous equation for the sum of the normalized metrics

Updated equation for the sum of the normalized metrics
Updated equation for the sum of the normalized metrics

Basically, as someone mentioned months ago in the comments (yeah, it took me a while to get to it)—the values don't go up to 1 when you divide as you do in the first formula, because you already reduced the values on the numerator by subtracting the minimum value, so the maximum of those values divided by the maximum of the initial values is not 1. Overall, it doesn't change much, but it might re-equilibrate a little the impact of both variables in the presence, as the mark for the maximum value of presence will be the same as for the maximum value of the win rate.

The tier list is then generated by setting the average of those values as the threshold for tier 1.5, and then adding or subtracting standard deviations to change tiers.


tier list

If you want to toy with the template, you can find it here on Tiermaker.

Conclusion

This is it for the month! If you want some more data, you can also check out the win rates of decks with or without companions as well as some of the most powerful MH2 cards on Twitter. I had prepared these in anticipation of the banning announcement.

See you in about a month for the next metagame update! (Or maybe earlier for another type of article? Who knows …?)


Opinions expressed in this article are those of the author and not necessarily Cardmarket.



0 Comments

To leave your comment please log into your Cardmarket account or create a new account.

Mentioned Cards

cardPreview