By any metric, this was a strong launch week for Slay The Spire 2.

The English-speaking audience is sold. Graphics, multiplayer, new characters - the praise is consistent and the negatives are minor. Balance complaints sit at 11.8%, bugs at 2.3%. For an early access title, that's a mighty clean picture.

The more interesting story is what non-English players think. Russian reviewers flag reused content as their top negative at 27%. Korean players put "too similar to the first game" at 15.7% - their single biggest complaint. Chinese players note balance issues and translation gaps. Japanese players are mostly positive but flag a specific ask: insect-type enemies are considerably more detailed than in STS1, and some players want an option to tone them down

The "is this a sequel or a DLC?" debate is almost entirely a non-English phenomenon. English players are 7.2% bothered by it. Everyone else is considerably more vocal. If MegaCrit is reading their own forums, they're reading English threads - the one place this complaint barely registers.

The original Slay the Spire's all-time reviews tell you who's complaining. Chinese players make up 37.1% of STS1's total review base. Korean players are 4.9%. These aren't new players with inflated expectations - they're the core audience. The people most familiar with the original are the ones least convinced this is a sequel.

Speaking of reading the forums - only one developer posted to the Slay The Spire 2 forums this week - a Valve moderator who locked a thread after it "devolved into non-productive argument." MegaCrit was nowhere to be seen.

Bungie's Marathon has landed with a positive 89.03% overall. But that number papers over a 30-point regional split that deserves attention.

English: 91.3% positive. Chinese: 61.0%. That is not a localization issue. Every other major language group - Russian at 83.6%, Korean at 78.2%, French at 92.3% - sits within a normal range. Chinese players are specifically angry about optimization (30.3% of negatives), UI complexity (25.1%), and task guidance (22.7%). These are onboarding and performance problems, not cultural preference gaps.

There's a second anomaly: Russian is the second-largest language group at 6.8% of all reviews, but the top actionable from Russian players is that the game doesn't work in Russia. They reviewed a game they cannot fully access. What the score would look like with working regional access is an open question.

The top upvoted review on the entire page is a Linux incompatibility complaint with 5,194 votes. The second most upvoted is a monetization complaint . Neither is a recommendation.

On the Steam Forums, the developer response rate is logged at 0.3% across 1,486 threads. But Bungie posted zero times. The four "developer" responses logged in the data were all Valve moderators closing threads for "non -productive argument ." 8,256 unique posters, 1,486 threads, and the only people showing up were Valve's own staff doing housekeeping.

This isn't a scaling problem - Bungie has the headcount to manage community at scale across bungie.net, Discord, and Reddit simultaneously. Steam forums are a choice, and they've chosen not to be here.

Discord is opt-in and self-selecting; Steam forums are public, indexed, and open to anyone with a complaint. For a studio with a historically careful relationship with public criticism, not responding may be deliberate. The absence isn't neglect. It's a position.

Poker Night at the Inventory is a remaster of a 2010 game and the community received it warmly. The TF2 item drops are doing a lot of conversion work. A search through all reviews finds over 13% of reviewers mention Team Fortress (or TF2) items as their reason for buying.

The mechanical complaints are consistent across both languages with enough data (English and Russian): AI behaviour is the top negative in English (13%), save data wiping is present in both, and item drop rates are called out as frustratingly slow.

Russian players are 14.7% of all reviews - the largest non-English audience by a wide margin. Their top complaint is that the game has no Russian language support. They bought it anyway. They reviewed it anyway. The game still doesn't speak their language.

The dev team posted to 1.4% of forum threads - low by any standard, but significantly more than Marathon or STS2 managed between them.

Esoteric Ebb launched this week to a 96.38% positive rating. Players are drawing direct comparisons to Disco Elysium - praising its writing, humor, and CRPG mechanics in the same breath. Writing is the #1 thing players love at 20%. Writing is also the #2 complaint at 8%, with the specific criticism being that it doesn't reach Disco Elysium's bar. The comparison is a compliment and a challenge at the same time.

58.4% of all Steam Forum threads received a team reply. In General Discussions specifically, 71.4% of threads got a response. The top two posters are both listed as developers: Snail with 199 posts, and Paige from Raw Fury with 57.

Forum users who reviewed the game left 92.7% positive reviews. That conversion rate sits below Over The Top's 96.7% (coming up next) but well above Marathon's 88.0% and Poker Night's 89.7%.

English dominates at 84.6% of reviews and no other language had enough volume for a full sentiment report. For a text-heavy CRPG launching without full localisation, that's expected. The audience that found it is clearly the right one.

Over The Top: WWI launched this week with a 91.87% positive rating with 788 reviews. The smallest game in this week's report, and the one with the most visible developer presence.

The team responded to over 76.3% of the total 304 Steam Forum threads. Three developer accounts appear in the top posters: COPOT (184 posts), Frederick_William_of_Prussia (119 posts), Olafson (111 posts).

Forum users who reviewed left 96.7% positive reviews - the best conversion rate in the report. The correlation is not subtle. Small team, active forums, highest forum-to-positive-review rate of the week.

The game's issues are acknowledged and specific: optimization ("runs like potato on high spec rig"), AI that doesn't build trenches, no iron sights, steamrolling balance. These are solvable problems and the community appears to believe they'll be addressed, which probably explains the score gap between the raw gameplay frustrations and the overall 91.87%.

French and Chinese scores sit lower (76.5% and 66.7% respectively) but the sample sizes are too small - 17 and 9 reviews - to really draw conclusions.