Patch 2.17 State of the Meta
Hey folks, Dr. LoR here with an updated meta report. Patch 2.17 is the third patch since Bandle City dropped, and it’s been a while since we’ve published a meta report (since Patch 2.12!). This unintentional hiatus was due to the fact that Mobalytics had an issue where certain decks using multiregion champions or cards didn’t appear in the data and I didn’t want to publish a report that was missing 20% of the meta!
The lack of meta reports since then was quite unfortunate as Bandle City was the largest expansion since the game launched and it changed the meta quite a bit, especially after the emergency patch (2.14.1) before Worlds knocked Akshan Sivir and Azir Irelia down a couple pegs. The next balance patch is supposed to arrive on October 20th, which is also when the season resets, so let’s see if we can learn anything about the current meta for those of you looking to do last minute climbs.
- The 157k matches of data from October 3rd to 9th for the analysis below come from Mobalytics premium.
- Mobalytics archetypes are defined as a champions + regions combo. 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 Bandle Tree Noxus (Fizz/Poppy, Fizz/Teemo/Poppy, Fizz/Lulu/Poppy, etc.), Shellfolk P&Z (Ezreal/Vi/Viktor), Poke City (TF/Gangplank/MF), Yordle Burn (Ziggs/Poppy/Teemo/Darius), and Rally Elusives (Zed/Lulu/Poppy) archetypes.
- Mobalytics ranks data aren’t perfect and come from optical recognition for Mobalytics deck tracker users. There are many ranked games that are therefore uncategorized. I think this is part of why there are so few Gold users. Most Mobalytics users probably have reached Platinum or higher by now.
- I will be referring to Bayesian WR (bWR), which is win rate (WR) that has Bayesian smoothing toward 50.4% — 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.
- I calculated meta diversity using the Herfindahl-Hirschman Index (HHI). TL;DR is that smaller HHI is more diverse and most of my past meta reports have been in the 300-500 range.
- I include margins of error and 95% confidence intervals when talking about win rates. 95% confidence intervals take the win rate +/- the margin of error, and it roughly means that we can be 95% sure that the ‘true’ win rate for the deck lands somewhere in-between.
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). Most of these decks have at least 1% play rate in Plat+ and together they account for about 60% of the meta.
- Bandle City made a huge splash, with all 4 of the most played decks featuring champions from the expansion, and 13 of the top 20. That’s pretty impressive considering how many champs LoR already had. Some of this is from novelty, but some of it is definitely from overtuned cards in Bandle City.
- Draven Sion has been the top deck of the expansion, consistently one of the most played decks of the expansion and amongst the top decks in terms of win rate. It’s been particularly oppressive in Masters, with a near 10% playrate, although this isn’t that bad compared to historically popular decks.
- Yordle Burn (Ziggs Poppy Noxus, sometimes with Teemo or Darius in place of one of the champs), is a deck that soared to the top of the meta in WR and subsequently in PR over the last month. While its popularity has cooled a bit, it’s still a very big part of the meta and prevents essentially any other pure aggro deck from existing.
- Darkness was the most popular week 1 deck, fell off quite a bit with the rise of Zoe Nami, but has risen again due to Yamato’s performance at World’s. Coincidentally, this and Draven Sion were the most ‘prebuilt’ decks of the expansion.
- Zoe Nami (aka Starfish), has lost popularity recently as the ladder became less friendly. It was the most played deck of Patch 2.15 (around World’s) with some almost unbeatable draws, but perhaps players are tired of its rather repetitive play pattern.
- Bandle Tree Noxus (usually Fizz Poppy, or Fizz Lulu Poppy, but Teemo and Ziggs sometimes get in there, too) is a deck that was quite popular in the first month of the expansion, fell off a bit, but has recovered in popularity due to good matchups into Darkness and other control decks.
- The most popular “old” decks are Lurk, Plunder (Gangplank Sejuani), and Akshan Sivir Demacia, with Zoe Lee, Feel The Rush (Trundle Tryndamere), and EZ Karma a little behind. Zoe Lee in particular has grown in popularity recently and had a good showing at Seasonals, so I expect to see more of it on ladder over the next week.
- Rally Elusives could be considered an old deck since Lulu Zed was one of the boogeymen pre-expansion, but it’s been quietly hanging out as one of the hidden gems of post-expansion as well, with either Lulu or Zed being replaced with Poppy. The core of the deck has stayed largely the same, looking to win with a Rally (or two) before the opponent can stabilize.
- Shellfolk P&Z (Ezreal Vi although sometimes only one and sometimes with Viktor instead) was popularized by ImpetuousPanda before Worlds but it really took off when Alanzq included it in his winning lineup (winning the only game it wasn’t banned). The deck is full of choices so lacks the win rate to maintain its popularity but it definitely feels very busted when you do manage to stabilize with Curious Shellfolk in play.
- Meta diversity is very healthy, with an HHI of 233 in Plat+, the lowest it’s been in any of my full meta reports. Even in Masters, it’s only 277.
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% confidence intervals for the win rates.
- The decks are sorted by Plat+ WR. Just for fun, I grouped them by tier, but I’d put Zoe Nami in Tier 1 based on its Masters WR. Note that the tiers aren’t literal power rankings, just groupings by WR.
- The strongest decks have fairly aggressive game plans. Yordle Burn is the most aggressive, while Rally Elusives is more combo-oriented, Draven Sion and Yordle Rally (Lulu Poppy BC Demacia) are more midrange, and Bandle Tree is midrange with a combo finish. Indeed, looking at the top 20 decks, very few are hoping the game lasts more than 7 or 8 turns. Again, Rally Elusives is the only ‘old’ deck amongst these, although the most popular variety is Zed/Poppy.
- Many “old” decks are still holding up in the middle up the pack, such as Plunder, Turbo Thralls, Akshan Sivir Demacia, Lurk, and Zoe Lee. They each have strengths into some of the popular decks but don’t have as much raw power as some of the newcomers.
- Many decks follow the common pattern of lower WRs at higher ranks, particularly aggro and midrange decks. The major exceptions are (unsurprisingly) Zoe Nami and Zoe Lee. I would’ve thought Shellfolk would show that pattern as well, but I guess it’s so hard even most Masters players don’t do that well with it.
- Popular decks with <50% WRs tend to have fun play patterns when they win, and that certainly holds true for EZ Karma, FTR, Shellfolk, Caitlyn Draven, Teemo Caitlyn, and Zilean Xerath. Someone’s gotta play the bad decks to prop up the good decks’ winrates…thanks for your time!
Here’s a longer list of the top 45 decks at Plat+, sorted by WR. The super-columns are split by rank, including a Plat+ (platinum, diamond, and masters combined). Under each supercolumn, you’ll find how many matches were recorded with that deck, the playrate (PR) percentage, the raw winrate (WR), and the Bayesian WR (bWR). Margin of error is a measure of sampling error: It roughly means that the deck’s true WR is 95% likely to be within +/- the margin of error. More data -> smaller margin of error.
Let’s plot the most popular 25 decks’ win rates versus play rates (for Platinum+) to see if any decks are being over or underplayed. A few important notes:
- Each deck is shown as a blue dot plus vertical error bars (‘whiskers’). These error bars represent 95% confidence intervals for win rates. More data à smaller error bars.
- The x-axis (play rate) is on a logarithmic scale to make more room on the left.
- Yordle Burn and Draven Sion deserve their high playrates as the two of the most powerful decks in the meta. Rally Elusives, Bandle Tree, and Yordle Rally in the same ballpark but are underplayed, perhaps because Mobalytics didn’t show these decks until last week.
- Darkness, Shellfolk, Teemo Caitlyn, Zilean Xerath, Deep, and Targon’s Peak are overplayed relative to their low win rates. Darkness is the most defensible due to its raw power but it’s quite hard to pilot well.
- The red dots on the left are 5 ‘bonus’ decks that I wanted to highlight as having few matches but good win rates. Fizz Nami BC is the OG Nami deck and still quite strong if you’re tired of playing the exact same game every match with Zoe Nami. TF Swain is still around and pretty strong. Elise Senna P&Z is a control deck usually with Ledros as finisher. MF Poppy is Scouts but without Quinn (Poppy is just a super strong champion). Finally, Tristana Sivir is a midrange deck with a combo finish reminiscent of Zed Sivir (it goes wider on board but has fewer ways to protect champions)—it’s a personal favorite that I highly recommend (created by FaintHD).
The release of Bandle City and nerfs to the most powerful decks of the previous meta (except Rally Elusives) has radically changed the meta. Diversity is high although the strongest decks haven’t changed that much since the emergency patch (2.14.1) right before World’s qualifiers. The next patch is October 20th and the developers suggested that there would be a big balance patch. I for one am hopeful that it’ll shake up the rather stale meta even without any new cards for the start of the new season.
Follow me on Twitter for updates! My next full meta report will be 1 week after Patch 2.18, but I will be posting smaller meta updates and deck optimizations in the meantime!