Dr. LoR’s Meta Report – Patch 3.6 Week 1

Hey folks, Dr. LoR here to take a look at early Patch 3.6 data. Patch 3.6 brought some massive changes to Runeterra, with buffs to 24 cards (including 16 champions), nerfs to 5 important cards including Sun Disc, 3 new cards, and 28 more cards impacted by the play/cast rules change. Although there are a variety of new decks, many familiar faces are still around.

• The 379k matches of data are from April 27 to May 2, including 191k matches of Plat+ (Platinum, Diamond, Masters), courtesy of Mobalytics premium. Notably, Mobalytics ranks data are imperfect and come from optical recognition for Mobalytics deck tracker users. This is especially true at the start of each season, when ranks reset. There are some ranked games that are therefore uncategorized (~43% right now; please help fix that by using Mobalytics Deck Tracker!).
• Mobalytics archetypes are defined as a combination of champions + regions. This means, for example, that Bandle Tree Noxus is spread across multiple archetypes. I therefore combine some archetypes that are largely the same across different champion combinations. This mainly affects Targon’s Peak (Asol + Zoe/Braum/Tryndamere),  Aphelios Viktor (sometimes with 1x Vi, Zoe, Yuumi), Yordle Rally (Tristana Demacia w/ Fizz/Kennan/Teemo/Gnar/Yuumi), Swain BC (Gnar/Teemo/Swain), Bandle Tree Noxus (some combination of Fizz/Teemo/Kennen/Lulu/Ziggs/Gnar/Poppy), Kindred P&Z Sentinels (w/ Kindred/Vi/Elise/Senna), Rally Elusives (Zed/Lulu/Poppy) archetypes. In general, a 3-2-1 champ split is defined by the 3-2 champs, whereas the 1x is more of a powerful additional unit.
• I will be referring to Bayesian WR (bWR), which is win rate (WR) that has Bayesian smoothing toward 51.5%—the average WR in these data. To understand Bayesian smoothing, check out this primer…basically, we’re trying to avoid statistical flukes for archetypes with less data. A deck where one very strong player went 18-2 is probably not a “true” 90% WR deck. A deck with a lot of data (>1000 matches) is likely to have a WR closer to its “true” WR.
• I calculated meta diversity using the Herfindahl-Hirschman Index (HHI). TL;DR is that smaller HHI is more diverse and most of my meta reports have been in the 150-350 range, but as high as 500 in the Lurk and AzIrelia metas. (Sun Disc’s peak in week 1 of patch 3.4 saw HHI reach 456.)
• I provide 95% Bayesian credible intervals when talking about WRs (using a beta[35,35] prior since LoR deck WRs are essentially always between 35%-65%). 95% credible intervals provide a range of WRs around the average WR for a deck, and it means that we can be 95% sure that the ‘true’ WR for the deck lands somewhere between these two numbers. Of course, ‘true’ WR isn’t a static number since the meta is always evolving but the point of the 95% CI is to give a sense for the uncertainty in WR estimates.

Most Popular Archetypes

Here are the play rates (PRs) for the 20 most popular decks by rank (plus Plat+, the combination of Platinum, Diamond, and Masters). These decks have at least 1% play rate in Plat+ and together they account for about 50% of the Plat+ meta. None have more than ~5% playrate. New decks to the top 20 are underlined.

• Sun Disc (Azir Xerath, sometimes + 1x Zilean) is the biggest playrate loser of the patch. That said, it’s still the most played deck and seems to be holding onto the crown due to fewer aggro decks and more midrange and control decks.
• Caitlyn Ezreal gained a lot from the Ezreal buff (from the cast/play merge) as well as the increase of midrange Demacia and Freljord decks.
• Pirates (MF with Twisted Fate / Gangplank / Katarina) became the scariest deck on ladder for most of the previous meta, with versions running anywhere from 1-3x Twisted Fate becoming the predominant versions. It continues to prey on less streamlined decks in the early meta. For a little experimentation of its own, some decks even include 1x Katarina.
• FTR soared in popularity with Sun Disc’s decline and was a big winner of the patch. It also happens to do well into Pirates, Caitlyn Ezreal, and various smaller control decks (ft. Jayce and/or Lux). However, heavy SI Freljord control decks have known weaknesses to Rally decks and landmarks, so I expect those decks to balance it out on ladder.
• Akshan Udyr is a newer deck that takes advantage of the Udyr rework. Although very smooth when it works, it doesn’t seem like this happens as often as necessary for this deck to remain so popular.
• Jayce Lux has been floating around for ages (just ask Davebo) but has received a major buff from the cast/play merge. It’ll only become stronger when Jayce’s copied spell is fixed in a future patch (probably 3.8) to also trigger Lux.
• Yuumi Pantheon is another big winner this patch, with winrates soaring to pre-nerf rates. Part of this is because fast and slow spells now trigger fated before resolution, but also because the meta has shifted to be more advantageous for it.
• Garen Jarvan Elites is the most prominent of the Elite decks that take advantage of Garen’s buff and potential Rally weaknesses. The version featured here is mono-Demacia but there are other region splashes and different champion combos that may end up being better.
• Jayce Sentinels (mostly with 2x Elise, 1x Vi) is a collection of decks featuring the Sentinels early game and big spells, along with usually some number of Albus Ferros as an additional finisher rather than Ledros. With a proactive gameplan and control tools, it has long been a strong deck that has gotten a boost from the cast/play merge changes for Jayce.
• Targon’s Peak (Asol mostly with Zoe/Tryndamere/Trundle/Braum) contains to maintain its popularity and even holds a reasonable WR with the increase of slower decks on ladder.
• Ashe Noxus (w/ LeBlanc) makes a triumphant but perhaps short-lived return to popularity. Despite receiving buffs to both champs, it seems to still be poorly positioned in the meta, with bad matchups against all four the most played decks.
• Ziggs Taliyah BC continues to impress, with few bad matchups and some very good ones. Joining it is Taliyah Malphite, a bit more of a slower, midrange deck that takes advantage of Malphite’s buff and absolute domination against other board-centric midrange decks.
• Thresh Nasus, Draven Sion, and Reputation are old favorites from seasons past that have received buffs and have thus made their way back to the top 20 for the first time in a while, with mixed success.
• Joining them are Lurk,  Darkness, and Teemo Caitlyn, longtime crowd favorites whose fans just refuse to play new decks. In this case, it seems like Lurk is actually well-positioned, which is more than can be said for the others.
• Azir Irelia also maintained its narrow foothold in the meta, which surprisingly good results thanks to becoming better positioned in the meta (FTR, Sun Disc, and slower midrange decks are all positive matchups).
• Kicked off the top 20: Riven Viktor, Fizz Lulu, Spider Burn, Aphelios Viktor, Viktor Lee, Viktor Vi. Most of the old aggro decks and Viktor decks that people played to death last season have dropped in popularity, whether due to nerfs or just wanting to try out new decks.
• Meta diversity report: With an HHI of 139, the Plat+ meta is highly diversified. Looks like a great time to be playing Runeterra, with no decks that account for more than 1 in every 19 games!

Most Successful Archetypes

Below, you can see the WRs for the 20 most popular decks at Plat+, again grouped by rank. The error bars represent 95% credible intervals for the win rates. Just for fun, I also grouped them by tier. This is purely based on WR (plus top 20 PR) and is NOT an evaluation of raw deck strength. The right pilot can probably play most of these decks to a Tier 1 win rate and most can be part of a cohesive tournament lineup. There are also decks that would be Tier 1 or 2 that don’t show up here because they aren’t played enough for me to be confident about their win rates.

• Tier 1 features a healthy mix of midrange, aggro, and control. I expect this to shift quite a bit as players reduce experimentation and start exploiting top decks to climb for seasonals.
• Tier 2 is also a nice mix of decks, and it’s good to see Sun Disc down here instead of in T1. While the deck isn’t dead by any stretch, slowly it down by a turn or two (or forcing players to run Clockwork Curator) makes the deck a little bit more manageable.
• Meme Tier: Akshan Udyr is shockingly bad, and many of the other decks with 50% or lower winrates all suffer from having a poor matchup spread.
• Balance Watch: Week 1 is never time to worry, with decks like Yuumi Pantheon, Pirates, Azir Irelia, and FTR having well-known counters. Landmarks are looking generally powerful though and I could definitely see some tuning coming up in a future patch.

Here’s a longer list of the top 45 decks at Plat+, sorted by WR. The super-columns are split by rank, including Plat+ (platinum, diamond, and masters combined), Diamond+ (diamond and masters), and Silver- (silver, bronze, iron). Under each supercolumn, you’ll find how many matches were recorded with that deck, the playrate (PR) percentage, the raw winrate (WR), the Bayesian WR (bWR), and a 95% CI (credible intervals), which corresponds to a range of values we’re 95% sure the “true” win rate lies within. More data à tighter CIs. The last column is the change in Plat+ playrate from last report.

Underplayed and Overplayed Archetypes

In the chart below I’ve plotted win rates versus play rates of 20 decks to demonstrate which archetypes can be seen as the ‘kings’ in the current meta, and which decks can be labeled as ‘overplayed’ or ‘underplayed’.

In the graph, you can see dots and vertical error bars (‘whiskers’) for each archetype. These ‘whiskers’ represent the 95% credible intervals for win rates (the true win rate is 95% likely to be in this range). The X-axis (play rate) is placed on a logarithmic scale to prevent too much lumping on the left.

I also included the ‘hidden gems’ – 10 additional decks represented as red dots without ‘whiskers’. These decks have lower play rates, so they have no error bars – they would’ve been would be huge and distracting.

• Meta Kings: Yuumi Pantheon, Pirates, FTR, Caitlyn Ezreal.
• Overplayed: Targon’s Peak, Elites, Ashe Noxus, Reputation, Thresh Nasus, Teemo Cailtyn, EZ Kennen… and maybe Sun Disc?
• Underplayed: Ziggs Taliyah, Scouts, Gnar Darius, Azir Irelia, Lurk, Jayce Sentinels.
• Hidden Gems: Jayce Heimer SI, Lux Heimer, Akshan Sivir, Fizz Aphelios BC, Spider Burn, Viego Shurima. Pairing Heimer with other spells matter champs seems to be a good strategy this meta, with these champs all benefiting from the cast/play merge. Akshan Sivir (and Scouts, which finished at 21th place in PR) are both great at bullying FTR. Viego Shurima is fairly well positioned now that Sun Disc is less popular. Spider Burn is always good in the early patch meta. Fizz Aphelios BC, otherwise known as Afaelios, is somewhat hard to play deck featuring the Fizz Papercraft Dragon combo and a sizeable Fae package to go with Aphelios’s flexibility in the midgame.

Summary

The meta is wide open and features successful decks of all flavors. No single deck is dominating the meta in terms of playrate and there is still lots of experimenting going on to figure out the best versions of decks featuring the buffed champs. Truly a great time to be playing Runeterra!

Follow me on Twitter for more updates and deck optimizations, or join me on Discord to talk more about all things LoR data.