Subnautica 2 opened the week with 66k reviews in five days at 92% positive, against the backdrop of a legal fight that nearly ate the studio. A million copies sold in the first hour is wild and the data underneath is even more interesting.

English reviewers loved it (93.8% across 39k reviews). German loved it (94.3%). French and Russian and Brazilian Portuguese and Spanish, all comfortably in the 90s. Then there's Korean at 64% and Chinese at 74.2%. That's a 19-to-30 point regional gap on the same game, and it's not random as both audiences are flagging the same specific things.

Chinese-language reviews are particularly precise, with hostile creature balance generating around 18% of negative sentiment. Players describe being trapped in their bases by Hammer Sharks with no real way to fight back.

Performance optimization comes next at 9.25%, with the now-familiar story of Unreal Engine's Lumen bringing high-end rigs to their knees. Then there were complaints about end-game resource hunting that ends with completely softlocked saves, appearing in over 7% of negative sentiment globally and is the most-cited bug across all languages.

These are typical mostly-launch-week issues and certainly aren't holding back the overall 92% overall score.

Subnautica 2's forums seem to be pretty ok! 1,508 threads, 9,732 unique posters, and 74.9% of those posters own the game. That's right in the middle of the healthy 50-to-80% band. People posting about Subnautica 2 are mostly people playing Subnautica 2. The team is at 0.8% response rate, which is low, but given the legal context (the studio just finished an active courtroom fight with its publisher) the silence is understandable.

One thing worth noting before next week: Subnautica 2 is in early access. The reviews we have now are from 5 days of post-launch play, in the most-engaged opening window the game will ever get. As patches land and the technical issues resolve, the question is whether the regional sentiment gap closes. If the complaints are about specific bugs (the end-game resource softlock, UE5's Lumen performance), patches will fix the score. If the complaints are about underlying design decisions (creature balance, content length), patches might not be enough. Gameplainer will have the data either way.

For developers reading this: the Korean and Chinese sentiment data isn't a "those audiences are harder to please" story but a "those audiences are running the game on hardware configurations or playing in ways that surface the bugs the English-speaking audience hasn't hit yet" story. The complaints translate just as the player patterns do. Pay attention to the tail of your language breakdown, that's where your QA gaps live.

It's wild that Forza Horizon 6 is the second most reviewed game this week and it hasn't even shipped yet. These 26k reviews are from people who paid $120 for the Premium edition to play 4 days early. Standard edition opens later this week at nearly half the price, with a wider audience and presumably a different sentiment curve. So treat everything below as a snapshot of one specific buying audience, not the game. I know I'm going to be obsessed with Forza Horizon 6 just like I was the previous titles - I just haven't had a minute to buy it let alone play it 😅

It's immediately interesting where the audience is: 40% English and 30% Chinese, and the 28-point sentiment gap between the two.

Chinese sentiment is specific and technical with 21% of reviews expressing frustration with specific issues, 13% complain about performance, and 15% complain about price.

One of the higher upvoted positive reviews complained about the price (500 yuan, or approximately $70 USD): "I spent 500 yuan and feel like an impotent husband"

In the Steam Forums, community manager CM_Artie is in the lead as the most active, personally responsible for 1.5% of all posts. I took last week off from writing my report to deep dive into Mixtape's forums which were awash with hostility - it was almost a breath of fresh air to see so much positivity here with Horizon 6.

There were no dev replies in the two non-English focused forums, which is possibly a good spot to start turning the point gap between English and Chinese reviews.

There are a couple of posts from users a little annoyed at specific features, but the community of players are controlling the threads to a great extent. The Forum owner-to-poster ratio is in the higher band of nearly 77%, curious to see how this changes as the price drops later in the week with the standard priced product dropping on Thursday.

Directive 8020 is the latest entry in The Dark Pictures Anthology, Supermassive Games' horror series that's been running since 2019. The previous games shipped with co-op multiplayer. Directive 8020 did not, although the team announced months ago that it would arrive post-launch. The announcement didn't survive contact with launch week, and the data shows what happened next.

The headline score is 62% positive across 1253 reviews, which lands roughly in the middle of the franchise's range. Directive 8020 is closer to the franchise' weaker titles than its stronger ones, but that's not actually the interesting part of the data. The interesting part is who is leaving the reviews.

English reviewers came in at 66.1%, Russian at 48.7% with a 17-point gap. Russian-language reviews flag the missing multiplayer at just under 11% of negative sentiment and missing Russian voice acting at 15%, with one of the top complaints translating to "thanks to the devs for not adding Russian voice acting to the game for 4k!" (referring to the 4,000 ruble price). Chinese sat at 53%, French at 53%, German at 67%. The English audience is the most forgiving cohort by a clear margin.

Eight reviews actually praised the co-op multiplayer that doesn't exist yet. I checked, and it looks like a mix of confusion with Steam Remote Play (which lets you share local screen sessions with friends) and players who didn't realise the difference

The Steam Forums are where the structural story gets interesting. 870 unique posters, 213 of whom own the game (24.5%) which is unusually low. For comparison, the previous four Dark Pictures Anthology games all sit between 82% and 90% owner share on their lifetime forum data. Even The Devil in Me, the worst-reviewed game in the franchise, kept its forum at 82.3% owner share.

Directive 8020 opens at 24.5%, one day in.

The hostility pattern on the forum confirms what the owner-share suggests. The temperature chart shows essentially zero hostile-tagged posts in the weeks leading up to launch, a sharp spike on launch day (May 12), and then dropping back back down through May 16th. At least one of the hostility leaderboard's top accounts has been banned by moderators, and the drive-by traffic appears to have moved on. Likely both effects together.

The Supermassive_CM Steam account, which has been the top poster on every Dark Pictures forum since 2019, isn't in the top 5 on this game's forum yet. Worth keeping an eye on whether they show up in the coming weeks as the launch noise dies down.

For developers reading this: there are at least two distinct flavours of forum brigading visible in the data right now. One is the sustained capture pattern, where non-owner posting share inverts and stays inverted for weeks. The other is the drive-by spike, where the hostility lands on launch day and self-resolves within days as the brigaders lose interest. Directive 8020 is the second kind. The owner-share will probably climb back toward the franchise norm over the next month as actual buyers settle in and the launch-day visitors drift off. Gameplainer will have that data.

Outbound is Square Glade Games' cozy vehicle-based exploration game, released May 11. I covered the demo back during Steam Next Fest , where it sat at 83.9% positive across 1,351 reviews with German players in the 90s, English players in the 80s, and Chinese and Russian players in the low 70s. Three different audiences with three different problems and the same score.

Eight days post-launch of the main game, the score has dropped to 72.87% across 1,158 reviews. Every language rating got worse but not evenly.

German players moved 5 points, English moved 10, Russian 22, and Chinese 25. The Traditional Chinese audience, which barely registered in the demo data, came in at 36.8% positive across 19 launch reviews. That's the angriest cohort across the entire week's report.

Among the complaints, the repetitive core gameplay loop sits at 15.4% of negative sentiment globally, with players describing the loop as "Collect resources → Rest → Collect resources → Rest" (though this quote was tagged as a negative inside a review that was ultimately a recommendation!). Multiplayer crashes are tagged "straight up broken" by 4% of reviews. Physics bugs show up at 8.4% of negative sentiment, with one of the better quotes being "we literally had a flying car with no upgrades installed" .

Pricing complaints sit at 14.3% of negative feedback, many of which were about a $2.99 Day 1 DLC vehicle that lands as predatory (side-note - the bus DLC is currently 100% positive rated, so the folk that bought it liked it!). Other folks complained about Kickstarter keys arriving late.

So far, this is a normal launch-week story for a small studio shipping an ambitious game. Lots to fix, none of it impossible. The story gets harder when you look at the storefront page, as the top "Not Recommended" review on Outbound's Steam page right now isn't about any of those problems but about how the developers responded to negative reviews. The review reads "Negative reviews exist for a reason. It's not a good look when developers ask players to change or remove their negative reviews", and it currently sits at 1472 votes. It's pinned at the top of the page because it's the most-upvoted review, which means it's the first thing every prospective buyer reads.

The Square Glade Games team is by far the most engaged developer in this week's report. The team's main account has made 131 forum posts this week, and a second team account has another 39. That's 170 posts from the studio across a single week, against a 32.6% team response rate, which is up from 28.8% during the demo period. The engagement has been increasing the entire arc from Steam Next Fest through launch. The team is trying!

The forum's 38.7% owner share looks low at first glance, but Outbound is a demo-having game and the demo audience lingers in the forum population without converting to owners. Last week's analysis on demo-having games suggested that pattern, and Outbound fits it. The owner share will climb over the coming weeks as actual buyers settle in. Not a brigading signal, just a demo-having game still in early conversion.

Two weeks ago I covered the pattern across thirteen established games : demo-having titles cluster around 40-80% owner share, against the 80-95% band for games that launched without a demo. Outbound's demo audience lingers in the forum population without converting to owners, at least not yet. The owner share will climb over the coming weeks as actual buyers settle in.

For developers reading this: Square Glade Games is doing two different things at once. Forum engagement is high (32.6% team response, 170 posts from studio accounts) and isn't directly the subject of any complaint. Review-reply engagement is much smaller in volume (about 3% of reviews have a developer response) but it's concentrated on negative reviews and it's mixed in result. At least one player has flipped their review from Not Recommended to Recommended after the dev replied directly with an acknowledgement of the bug they reported. Meanwhile, the most-upvoted review on the game's Steam page right now (1472 votes) accuses the developers of asking players to change or remove their negative reviews.

Replying to reviews is producing both outcomes. When it works, it's invisible. When it backfires, it ends up pinned to the top of your storefront page.

Here's what an invisible win looks like:

Psychic Kung Fu Master is a tactical RPG that launched mostly into a Chinese-speaking audience, sitting at 62.6% positive across just over 1000 reviews.

The complaints are technical and specific (missing save slot overwrite, a broken teammate cultivation system, character art described as "kindergarten-level crude"), though the dev has been patching them one at a time.

The forum is in a normal state. 70% owner share, no brigading signal, devs at 2.8% response rate which is low but acceptable for a small Chinese RPG launch.

Worth flagging: the dev shipped a save slot overwrite fix within days of launch (the most-upvoted negative review's update logs it), but the fix landed after the initial review wave. The lesson for developers: QoL features that look obvious to a 2026 audience deserve to be in the launch build, because the score that lands in the first 72 hours is hard to shift.

Three reviews have flipped from Not Recommended to Recommended on Psychic Kung Fu Master since launch. Two of those flips followed direct developer responses on specific bugs (one player came back after a fishing minigame pointer fix, with the dev's reply acknowledging the issue and an update landing two days later).