Hillforts

Introduction

Hillforts, released 22 July 2026 by Five Head Games, sets a card-based autobattler in a medieval setting: draft cards to build an army, research supporting technologies, and lay siege to an enemy castle across matches designed to be short and replayable rather than long campaigns. It sits Positive with English-speaking players — 86.5% across 37 reviews — and much of that goodwill traces directly back to a well-regarded pre-release demo, with several reviewers reporting dozens of hours spent there before ever buying the full game. Reading through the sample turns up something more specific and interesting than simple continuity, though: a genuine, detailed disagreement between reviewers with real stakes in the genre about whether the changes made for the full release improved or damaged what the demo had already gotten right.

A demo's goodwill mostly carrying forward

The dominant pattern among positive reviews is continuity. Multiple reviewers who spent significant time with the free demo — one citing roughly ninety hours — report the full release delivering on or exceeding that experience, crediting new civilizations, technologies and a "Conquest" mode for building meaningfully on the original concept rather than replacing it. Art direction and music draw specific, repeated praise for staying clean and enjoyable without becoming repetitive, and the short, quick-paced match structure is repeatedly cited as a genuine strength — losing a run doesn't feel devastating when the next one is minutes away. Steam Deck performance comes up unprompted more than once, with reviewers specifically valuing low power draw for travel sessions.

A detailed, unresolved dispute about civilization balance

The most substantive material in this review's research is a direct disagreement between two reviewers, both citing real experience with the genre, over the same specific game feature. One detailed review — from a reviewer describing thousands of hours across other autobattler games — argues that new, strong civilization bonuses introduced for the full release broke the demo's more careful balance: bonuses like a siege-focused nation making an aggressive siege-heavy opening overwhelmingly strong, in that reviewer's view, remove the natural counterplay that made the demo feel adaptable and instead reward whoever best exploits their assigned civilization's bonus. A separate, positive review directly and explicitly rebuts this, citing nearly the same civilization examples as evidence that the bonuses instead deliver exactly the kind of build variety a good autobattler should have. Both reviewers are describing the identical mechanic; they simply reach opposite conclusions about whether it's good design. This review has not played enough matches across enough civilizations to adjudicate that dispute independently and reports it here as genuinely open, but flags it clearly for readers who specifically valued the demo's balance and are deciding whether the full release still delivers that.

A launch bug, and how much is actually new

Beyond the balance dispute, that same detailed review raises two more concrete points. It reports that profile reset and unlock-all buttons did not register clicks promptly at launch, an issue the developer is reported to have addressed in a public reply. It also compares the game's content compendium between demo and full release, estimating that only roughly 10 to 20% of units and technologies are genuinely new, given that the demo already contained an estimated 80 to 90% of the full game's content — a specific, checkable claim rather than a vague impression, though this review has not independently verified that percentage. Separately, even satisfied reviewers describe the overall game as short, with one saying it can be fully "finished" within a few hours of dedicated play, and several requesting expanded late-game modes and more starting-difficulty options for repeat playthroughs.

A small, specific privacy detail

One further technical point is worth reporting plainly rather than folding into general complaints: standard Unity engine telemetry files are present in the game's save directory with their default settings enabled, and there is no in-game option to disable this. The same detailed reviewer who identified this notes the files can be edited manually to opt out, but that requires a player to know to look for them — it isn't surfaced as a setting within the game itself.

Randomness that frustrates before it clicks

A smaller, more specific thread concerns the game's combat randomness rather than its civilization bonuses. One reviewer describes an early stretch where consistent improvement felt impossible, attributing the frustration to random targeting among units in combat, before the underlying logic eventually "clicked" and the experience became substantially more enjoyable. That's a meaningfully different complaint from the balance dispute above — it's about combat legibility during the learning curve rather than civilization design — and it's worth flagging on its own, since a player bouncing off in the first hour or two for this specific reason might be missing a game that becomes considerably more satisfying once its systems are better understood. Separately, and less charitably, two other reviewers independently describe the game's randomness as excessive even after significant playtime, with one comparing it unfavourably to Mordheim and another calling the experience after launch "about as fun as solitaire" — a reminder that not every player who bounces off the game's RNG does so purely from an unfamiliarity that later resolves.

Final Verdict

Hillforts succeeds at the thing its demo already proved it could do: a tight, replayable autobattler loop with clean presentation and matches short enough to make losing feel low-stakes rather than punishing. What keeps this from a more confident score is a genuinely unresolved question at the heart of its design — whether the civilization bonuses added for the full release are the source of real strategic variety or the thing that broke a more careful balance the demo had already found — argued in detail and good faith on both sides by reviewers with real genre experience. Add a minor launch bug, a real question about how much content is actually new, and an undisclosed telemetry detail, and this reads as a well-executed, worthwhile purchase for genre fans rather than an easy, unqualified recommendation.