All genres

Souls-like game design checklist: what players punish in Steam reviews

Answer 24 questions about your own souls-like. Of the 57 checklist items, 19 block a release.

Built from 12,327 reviews of 30 Souls-like games released 2026-03-21 to 2026-09-21, plus post-mortems of 10 flawed hits (142 most helpful reviews each).

The 6 ways souls-likes become flawed hits

A flawed hit sold well and still ended up with a mixed review score. Post-mortems of 10 of them show the same combinations of failures recurring. If a game fits one of these, that is the thing to fix first.

The beautiful maze

Looks like Mortal Shell, Hellpoint, Bleak Faith and Nostalgaia.

  • E1 Players can find their way. Landmarks, distinct areas and sightlines first; a useful map if the world is large or looks alike.
  • D2 Checkpoints are frequent, and fast travel works between all discovered ones. No rare consumable to unlock it, no single active teleport point.
  • D1 The run back to any boss is short. A working guide from the complaints: under a minute, and without a gauntlet of respawned enemies.

The hybrid that halves both genres

Looks like NECROPOLIS, Dark Devotion, Salt and Sacrifice and Crimson Moon.

  • L2 A hybrid is designed as one game. Souls-like plus roguelike, extraction or monster hunting works only if each half serves the other.
  • H2 The structure you chose is funded. A roguelike or extraction loop needs several times the rooms, enemies, weapons and bosses of a linear game of the same length.
  • G1 Several weapon classes with clearly different movesets. Weapons inside a class may share one; classes may not.

The 1.0 that was not done

Looks like Bleak Faith, Enotria and Crimson Moon.

  • I1 No soft-locks, progress blockers or frequent crashes.
  • I4 It is finished at 1.0, or it is called early access.
  • F3 Every boss is fair and tested. Readable attacks, openings, no heal loops or AI that stops, health that fits the player's damage.

Hard, but not fair

Looks like Mortal Shell, Ashen, Enotria and Nostalgaia.

  • C1 Hitboxes match the animations, for the player's weapons and every enemy attack, including lingering hitboxes after a swing.
  • C2 Defence is reliable, and more than one defensive style works. A correctly timed parry always registers; dodge has real invulnerability; the game does not force everyone into one answer.
  • F2 Encounters are sized for the combat system. A moveset built for duels does not meet groups of six in a corridor; late areas do not swap design for enemy and projectile spam.
  • F3 Every boss is fair and tested. Readable attacks, openings, no heal loops or AI that stops, health that fits the player's damage.

Co-op first, broken alone or online

Looks like Ashen, Crimson Moon and NECROPOLIS.

  • K1 Sessions are stable, and a reconnect restores the run.
  • K2 A game balanced for two is still good alone.

The sequel that shrank

Looks like Salt and Sacrifice.

  • L3 A sequel keeps the freedom the first game gave.
  • G1 Several weapon classes with clearly different movesets. Weapons inside a class may share one; classes may not.

4 of 32 pages · select the page to enlarge

PDF · 32 pages · instant download

Full Souls-like genre report

  • What players complain about, ask for and love, ranked, with quotes from 12,327 reviews of 30 games
  • Post-mortems of 10 flawed hits: what each was bought for, what sank it, and what doing it right looks like
  • Why players quit in the first two hours versus what veterans criticise
  • The complete checklist with every item's evidence, ready to print

One-time payment, no account needed to buy. 30-day money-back guarantee: if the report is no use to you, reply to your receipt for a full refund, no questions asked.

It documents what players punished. It is not a forecast for your game.

What players hate in souls-likes, from 12,327 Steam reviews

These are the first 10 of the 25 complaint themes ranked in the full report, in the report's order. The order combines how many games a theme appears in, how many reviewers raise it and how severe the complaints are, so a row can sit above one with more reviewers.

From 12,327 reviews of 30 Souls-like games released 2026-03-21 to 2026-09-21, plus post-mortems of 10 flawed hits (142 most helpful reviews each). Counts are reviewers, not players.

Top 10 complaint themes in souls-likes, with the number of reviewers who raised each one and the number of games it appears in
#Complaint themeReviewersGames (of 30)
1Combat feels clunky and unsatisfying26126
2Combat is basic and lacks depth23626
3The camera is problematic during boss fights64% of these reviewers come from one game.26919
4Keybinds cannot be remapped on keyboard, mouse, or controller63% of these reviewers come from one game.24126
5Hitboxes are inconsistent and problematic23322
6Difficulty balance18625
7Enemy and encounter design18625
8The game has numerous bugs19823
9Boss design14727
10Combat feel17024

The 19 things that block release

These are the items the checklist marks Critical: they separate well-rated from poorly rated games, drive early quits or refunds, or sank several flawed hits.

Figures come from 12,327 reviews of 30 Souls-like games released 2026-03-21 to 2026-09-21, plus post-mortems of 10 flawed hits (142 most helpful reviews each). "n/10" figures are judgments from the post-mortems, not counts.

  1. The core promise

    A1

    Fighting feels good, moment to moment.

    Combat feel is praised by 3,575 reviewers and complained about by 2,182, across 30 and 29 of 30 games

  2. The core promise

    A4

    Bosses worth the effort.

    Boss design has the highest praise severity among the twelve most praised categories (3.6/5; 1,165 reviewers)

    when i tell you i am a good soulslike player and this fight is impossible, i mean it

  3. The first two hours (the refund window)

    B1

    Controls and camera are right on the first try, on keyboard-and-mouse and on controller.

    "The camera is problematic during boss fights" comes from 269 reviewers across 19 games (64% from one game)

    Abysmal control scheme that can't be remapped on controllers

  4. The first two hours (the refund window)

    B2

    The first fights already feel good, with the starting weapon and no upgrades.

    Combat feel is a larger share of complaints from players with under two hours played than from players past ten hours

  5. The first two hours (the refund window)

    B3

    It runs and does not break in the first session on ordinary hardware.

    Performance is a larger share of complaints from players with under two hours played than from players past ten hours

    Got stuck inside the first boss on four or five different occasions, forcing me to restart the game

  6. Combat craft

    C1

    Hitboxes match the animations, for the player's weapons and every enemy attack, including lingering hitboxes after a swing.

    Inconsistent hitboxes: 233 reviewers across 22 games

    marred by iffy, erratic and sluggish combat, janky hitboxes, mostly terrible AI companions

  7. Combat craft

    C2

    Defence is reliable, and more than one defensive style works. A correctly timed parry always registers; dodge has real invulnerability; the game does not force everyone into one answer.

    "Parry timing feels inconsistent and unreliable": 193 reviewers across 15 games

    combat and build variety falls apart as the game forces you into parrying everything 10x or more

  8. Combat craft

    C3

    The character answers inputs. No input delay, no floaty or sliding movement, recoveries short enough to react.

    "Combat feels clunky and unsatisfying" is the biggest combat theme: 261 reviewers, 26 of 30 games, no game above 13%

    Too floaty, especially with a game with traversal mechanics and precise combat.

  9. Death, checkpoints and the way back

    D1

    The run back to any boss is short. A working guide from the complaints: under a minute, and without a gauntlet of respawned enemies.

    "Boss runbacks are excessively long and tedious" across 8 games

    atrocious boss runs, what an awful and frustrating decision

  10. Death, checkpoints and the way back

    D2

    Checkpoints are frequent, and fast travel works between all discovered ones. No rare consumable to unlock it, no single active teleport point.

    Too few checkpoints across 7 games

    you cannot freely fast travel between Breaches

  11. World and level design

    E1

    Players can find their way. Landmarks, distinct areas and sightlines first; a useful map if the world is large or looks alike.

    An uninformative map: 166 reviewers across 20 games

    Map areas are huge mazes that branch off constantly, and all the corridors look the same.

  12. Enemies, bosses and difficulty

    F1

    Enough enemy types for the length, with different behaviours.

    "Enemy variety is very limited" is the biggest enemy theme: 191 reviewers, 21 of 30 games, no game above 20%

    the same handful of enemies that have limited movesets

  13. Enemies, bosses and difficulty

    F3

    Every boss is fair and tested. Readable attacks, openings, no heal loops or AI that stops, health that fits the player's damage.

    Boss complaints from 8.0% of reviewers in 30 of 30 games

    The final boss consisted of 25 minutes of hitting and rezing.

  14. Builds and progression

    G1

    Several weapon classes with clearly different movesets. Weapons inside a class may share one; classes may not.

    "Weapon variety is very limited": 86 reviewers across 18 games (48% from one game)

    every weapon feels almost identical

  15. Technical quality

    I1

    No soft-locks, progress blockers or frequent crashes.

    Bugs are raised by 12.7% of reviewers in 30 of 30 games, and are 8.5% of weak games' complaints against 5.2% in strong ones

    It's pretty easy to get soft-locked in this game one way or another.

  16. Technical quality

    I2

    Frame rate holds, including in the largest area. In a timing-based genre a frame drop is a gameplay bug.

    Performance is the clearest single-player separator in this sample: 7.1% of weak games' complaints against 2.8% in strong ones

  17. Co-op and online play

    K1

    Sessions are stable, and a reconnect restores the run.

    "Multiplayer does not function properly" across 6 games

    we get disconnected frequently in the middle of a run

  18. Honesty of the promise

    L1

    If you call it a souls-like, it delivers the core (deliberate stamina combat, exploration, a death that costs something), or the store page says what it is instead.

    Complaints that a game does not live up to the label are few (2.6% of reviewers) but among the most severe (3.7/5) and appear in 28 of 30 games

  19. Honesty of the promise

    L2

    A hybrid is designed as one game. Souls-like plus roguelike, extraction or monster hunting works only if each half serves the other.

    A hybrid that fights itself in 4/10 flawed hits

    Necropolis isn't a roguelike. It's not a good roguelike atleast.

What players ask for

Not requirements. These are the recurring wishes, mostly written inside positive reviews, and they are not part of the check.

  • A map, or a better one (the top wish: 84 reviewers, 14 games; minimap 12 games).
  • More weapon types and classes (11 and 6 games).
  • More content: maps, areas, bosses (9 and 8 games).
  • Cosmetic character customisation (12 games).
  • A parry where there is none, and deeper combat mechanics (12 and 9 games).
  • Difficulty options (13 games), New Game Plus (7 games), fast travel (10 games).
  • Co-op (9 games).

Questions about the souls-like checklist

What do players complain about most in souls-likes?

In 12,327 Steam reviews of 30 recent souls-likes, the three highest-ranked complaint themes are: Combat feels clunky and unsatisfying (261 reviewers across 26 games); Combat is basic and lacks depth (236 reviewers across 26 games); The camera is problematic during boss fights (269 reviewers across 19 games). For "The camera is problematic during boss fights", 64% of these reviewers come from one game. The ranking combines how many games a theme appears in, how many reviewers raise it and how severe the complaints are, so it is not a sort by reviewer count alone. Counts are reviewers, not players. The table on this page lists the first 10 of 25 ranked themes.

What makes a souls-like a flawed hit?

A flawed hit sold well and still ended up with a mixed review score. Post-mortems of 10 of them show 6 combinations of failures that keep recurring: "The beautiful maze" (Mortal Shell, Hellpoint, Bleak Faith and Nostalgaia); "The hybrid that halves both genres" (NECROPOLIS, Dark Devotion, Salt and Sacrifice and Crimson Moon); "The 1.0 that was not done" (Bleak Faith, Enotria and Crimson Moon); "Hard, but not fair" (Mortal Shell, Ashen, Enotria and Nostalgaia); "Co-op first, broken alone or online" (Ashen, Crimson Moon and NECROPOLIS); "The sequel that shrank" (Salt and Sacrifice). The games in brackets are the ones each pattern describes. The patterns come from reading the most helpful reviews of each game, so they are judgments, not counts.

How many questions are in the souls-like check?

The quick check is 24 questions and takes about five minutes. It covers all 19 release blockers on the checklist, plus the few other items the flawed-hit patterns depend on. Before the questions it asks "Will the game ship with co-op or any kind of online play?", and a "no" skips the ones that would not apply. The full checklist has 57 items, and the deep check asks the rest. Every question is about something you observed in a playtest or a build, and "Not tested yet" is a valid answer.

Is the result a prediction of my Steam reviews?

No. The checklist says itself that its scoring thresholds have not been calibrated against real review scores, so the check never turns your answers into a predicted rating or a sales figure. It lists the release blockers you failed, the ones you have not tested yet, and any of the 6 flawed-hit patterns your answers match. Every game in the sample found an audience, so a clean result does not get a game noticed either.

What this checklist cannot tell you

It is not a forecast. Its own scoring thresholds have not been calibrated against real review scores, so this page never turns your answers into a predicted rating or a sales number. It tells you whether you are about to repeat a documented mistake, not whether the game is good.

  • Built from 30 recent games plus 10 older flawed hits. Only part of the 382-game watchlist has been crawled and none of the older games is LLM-analysed; percentages will move when they are.
  • The tag is broad: 3D souls-likes, 2D "soulsvanias", boss rushes, roguelite hybrids and one extraction PvPvE game share the sample. Mark items N/A rather than forcing a score (platforming for a game without jumps, maps for a boss rush).
  • Big-budget games were left out on purpose, so expectations set by FromSoftware-scale productions are only visible through what reviewers compare small games to.
  • Launch-window reviews over-represent controls, performance and bugs, and under-represent runbacks, late-game balance and endings. The flawed hits cover that gap, with judgments, not counts.
  • Reviewers are not all players: they skew engaged and opinionated. Ranks are more trustworthy than absolute percentages.
  • Survivorship: every game here reached an audience. A perfect score does not get a game noticed.

Checklist version 1, 2026-09-21.

Other genres

A game that avoids every mistake above still has to be found. Check the page that has to sell it.