Core Keeper: amount of content appears 2.2 times as often in negative Japanese reviews
Players writing reviews in Japanese mention the amount of content far more often than usual. The dataset contains 1,512 Japanese-language reviews collected between March 2022 and November 2026.
Own research by 648 News. Every figure in this article was counted by us from 1,512 Steam reviews written in Japanese, collected between 2022-03-08 and 2026-11-11.
The collected sample matches the platform total
Players writing reviews in Japanese recommend Core Keeper at a rate of 91.1 percent. The collected sample shows 91.1 percent positive reviews. This matches Steam's all-time figure for the game, which also reports 91.1 percent positive reviews across 1,514 total reviews as of November 2026. Both counts reflect a Very Positive reception. The confidence interval for the collected sample ranges from 89.5 percent to 92.4 percent. This confirms the high approval rate remains consistent. Every entry was sorted into specific topics by the analysis model. The model successfully processed all 1,512 entries in the dataset.
The numbers
Content length stands out against typical releases
Reviewers talk about the amount of content, length, and replayability 39.0 percent of the time. That is 1.5 times more often than the typical game. They praise the extensive crafting systems, the structured progression tied to mining new ores, and the sheer volume of bosses and biomes. Some note that the game feels complete despite being a sandbox. They also appreciate the variety of biomes and the detailed pixel art. Players writing reviews in Japanese mention performance, optimization, bugs, and crashes only 5.1 percent of the time. That figure sits well below the standard 14.5 percent. They also discuss the story, scenario, and writing at a rate of 4.5 percent. This falls short of the usual 10.9 percent baseline. Players writing reviews in Japanese mention UI, controls, and key bindings at 3.1 percent. They discuss music, sound, and voice acting at 2.7 percent. Both figures remain low compared to standard baselines.
序盤はまっとうに面白い。Stardew Valley視点でTerrariaをプレイするとでもいった感覚。
The beginning is genuinely fun. It feels like playing Terraria from a Stardew Valley perspective.
Complaints focus on progression and difficulty
The amount of content, length, and replayability splits opinion sharply. It appears in 47.4 percent of reviews that do not recommend the game, which is 2.2 times the usual rate for negative reviews. Reviewers cite a narrow progression path, heavy inventory management, and a reliance on specific tiered ores to unlock new areas. Difficulty, balance, and performance also divide players. Difficulty appears in 30.4 percent of negative reviews, compared to just 8.5 percent of positive ones. Performance problems show up in 20.7 percent of negative reviews versus 3.6 percent of positive ones. Gameplay, combat, and core mechanics appear in 71.1 percent of negative reviews, compared to 55.3 percent of positive ones.
[b]知名度が上がってからだいぶ時間を置いてプレイしたつもりだけど、まだ待ったほうがいい。[/b]
[b]I tried playing it a while after its popularity rose, but I think you should still wait.[/b]
Long sessions correlate with higher approval rates
Players with the least play time recommend the game at a rate of 83.1 percent. They spent up to 24.2 hours before writing their reviews. The least-logged group talks most about gameplay, combat, and core mechanics at 51.3 percent. They also mention the amount of content, length, and replayability at 31.7 percent. Players with the most play time recommend it at a rate of 96.0 percent. They had played at least 139.3 hours when they posted their feedback. The most-logged group discusses gameplay and combat at 64.3 percent. They talk about content length at 42.9 percent. The difference between the two groups is 12.9 percentage points, with a p-value of p < 0.001.
Methodology and sources
Data
Steam user reviews, retrieved with Steam's public review listing between 2022-03-08 and 2026-11-11: 1,512 written in Japanese. Steam's own all-time totals for Core Keeper were read on 2026-11-12.
How the reviews were classified
Every review was read by a language model, which rated how positive it is, how much concrete detail it gives, and which of 15 fixed topics it talks about. The counts and percentages in this article were then computed from those labels in ordinary code, not by the model. Automatic classification makes mistakes, so single percentages should be read as approximate. Reviews of fewer than 15 characters were not sent to the model; they count as positive or negative from the reviewer's own recommendation and carry no topic.
Limits
People who write a Steam review in Japanese are not a representative sample of everyone who plays in Japan, and the language of a review does not establish where its author lives. Steam's default filter for off-topic review activity was left on, so periods Steam marked as review bombing are excluded from the counts in this article. Reviews that arrived after the retrieval date are not included.
Links
Data from Steam. Steam and the Steam logo are trademarks and/or registered trademarks of Valve Corporation.