Study · published September 21, 2026

Flawed hits: 30 Steam games that sold and still disappointed

A flawed hit is a Steam game with 500+ reviews and a 40-79% positive score. People bought it, so the hook worked. A good share of its reviewers then did not recommend it, so the execution did not.

That makes these games more useful to study than outright failures, which mostly show that nobody came. A flawed hit proves there was demand for the idea, and it documents the mistake that cost the game a better score, in the words of players who paid for it. For ten flawed hits in each of 3 genres, the 142 most helpful reviews were read in full and written up: what the game was bought as, where it let players down, and what doing it right would look like.

flawed hits, each written up as a post-mortem

genres, ten games in each

142

most helpful reviews read for every game

4,260

reviews read in full across the study

What failed, across genres

One table per genre. Each row is a failure pattern, the meter shows how many of that genre's ten games showed it, and the last column names them.

The n-of-10 figures are judgments, not counts. Each was tallied by hand from the ten post-mortems of a genre, and a post-mortem records an impression of what reviewers said, not a number of reviews.

Each report also names the combinations of failures that recur across its ten games. They are listed under each table, with a link to check your own game against them.

10 games · report dated September 21, 2026

Souls-likes

All ten were praised, even by people who did not recommend them, for art direction and atmosphere, and several for an interconnected world in the Dark Souls 1 mould. None of that saved them. What sank them is what happens between the bosses: combat that does not register hits or parries reliably, the trip back after each death (runbacks, scarce checkpoints, rationed fast travel, mazes with no landmarks and no map), a handful of enemy types thrown at the player in groups, thin weapon variety, and releases that were not finished. Players accept a hard game. They do not accept a game that wastes their time or does not read their inputs.

Failure patterns across the ten souls-likes, with how many of the ten showed each one
Failure pattern	Games (of 10)	Which
Combat that does not feel fair or reliableHitboxes that do not match animations, parries that do not register, enemies that track through dodges, no hit feedback	8 of 10	All except Dark Devotion and Salt and Sacrifice
The way back is a punishmentLong boss runbacks, scarce checkpoints, fast travel that is rationed or locked to one point, forced backtracking, slow ladders and elevators	7 of 10	Mortal Shell, Ashen, Hellpoint, Dark Devotion, Salt and Sacrifice, Bleak Faith, Nostalgaia
Thin weapon and build varietyWeapons that share movesets, few archetypes, one style forced on everyone	7 of 10	Mortal Shell, NECROPOLIS, Ashen, Dark Devotion, Salt and Sacrifice, Enotria, Crimson Moon
Bugs, soft-locks, crashes, or shipped unfinished	7 of 10	NECROPOLIS, Ashen, Hellpoint, Bleak Faith, Enotria, Nostalgaia, Crimson Moon
Few enemy types, used as ganks and mob spam	6 of 10	Mortal Shell, NECROPOLIS, Ashen, Salt and Sacrifice, Nostalgaia, Crimson Moon
Bosses that are unfair, broken or damage sponges	6 of 10	NECROPOLIS, Ashen, Hellpoint, Salt and Sacrifice, Bleak Faith, Enotria
Samey or maze-like levels with no landmarks and no map	5 of 10	Mortal Shell, NECROPOLIS, Hellpoint, Bleak Faith, Nostalgaia
Healing or stamina economy that frustratesScarce or non-refilling healing, slow stamina, enemies with no stamina limits	4 of 10	Mortal Shell, Hellpoint, Dark Devotion, Bleak Faith
A genre mash-up that fights itselfSouls-like plus roguelike, extraction or monster hunting	4 of 10	NECROPOLIS, Dark Devotion, Salt and Sacrifice, Crimson Moon
Broken promiseSequel that drops what the first game had, playtime claim, "1.0" that players call early access	4 of 10	NECROPOLIS, Salt and Sacrifice, Bleak Faith, Enotria
Too little content for the price	4 of 10	Mortal Shell, NECROPOLIS, Enotria, Crimson Moon
Controls, camera and missing basicsNo rebinding, broken camera or lock-on, floaty movement, no pause or options	3 of 10	Dark Devotion, Bleak Faith, Crimson Moon
Co-op that does not holdLost progress on disconnect, awkward matchmaking, shared saves	3 of 10	NECROPOLIS, Ashen, Crimson Moon
Platforming that kills more than the enemies do	3 of 10	Hellpoint, Dark Devotion, Salt and Sacrifice
Abandoned after launch	2 of 10	NECROPOLIS, Ashen

The combinations that recur in souls-likes

The beautiful maze

Striking world, reliable praise for atmosphere, then samey corridors, no map, rationed fast travel and long runbacks.

Seen in

Mortal Shell, Hellpoint, Bleak Faith, Nostalgaia

Check your game against this pattern →

The hybrid that halves both genres

A souls-like bolted to a roguelike, an extraction loop or monster hunts, where each half undermines the other.

Seen in

NECROPOLIS, Dark Devotion, Salt and Sacrifice, Crimson Moon

Check your game against this pattern →

The ambitious 1.0 that was not done

A small team ships a large 3D souls-like with soft-locks, broken bosses and placeholder systems.

Seen in

Bleak Faith, Enotria, Crimson Moon

Check your game against this pattern →

Read the 10 souls-likes game by game ↓

10 games · report dated September 20, 2026

Indie action RPGs

Every one of these games sold a systems fantasy: endless builds, a loot chase, being an unstoppable mage, a boss-rush soulslike, old-school MMO grinding. Nearly all were praised, even in negative reviews, for art, first-hours combat and the idea behind their build system. What sank them was depth that ran out: content that repeats or ends early, build variety that collapses to one or two viable options, loot and levelling that do not feel like progress, a promise the game did not keep, and, for anything online, instability and heavy-handed operations. In this niche the hook and the look get a game bought; whether the systems still hold at hour twenty gets it recommended.

Failure patterns across the ten indie action RPGs, with how many of the ten showed each one
Failure pattern	Games (of 10)	Which
Content runs out or repeatsThin endgame, few enemy or mission types, same maps on a higher difficulty, campaign that feels like an expansion	9 of 10	All except Hero Siege
Bugs, crashes, lost progressCrashes, broken quests, save or character loss, rollbacks, wipes after updates	7 of 10	Wolcen, Hero Siege, SpiritVale, Dungeons of Sundaria, Dragonkin, Van Helsing III, Relic Hunters Legend
Broken promiseMarketing or franchise expectation the game does not meet (not the advertised genre, sequel smaller than its predecessor, redesign that replaces the game people bought, co-op that forces solo play)	7 of 10	Wolcen, Hero Siege, Eldest Souls, Lichdom, Van Helsing III, Zengeon, Relic Hunters Legend
Build variety is an illusionA handful of viable skills or builds, one overpowered class or character, items that do not change how you play	6 of 10	Wolcen, Dragonkin, Lichdom, Van Helsing III, Zengeon, Relic Hunters Legend
Loot and levelling that do not feel like progressStat-stick items with no uniques, enemies that scale with you, upgrades gated or effectively unobtainable	5 of 10	Wolcen, Hero Siege, Dungeons of Sundaria, Dragonkin, Zengeon
Unfair difficultyDamage-sponge enemies, one-shots, difficulty that only inflates numbers, sudden spikes	5 of 10	Eldest Souls, Dungeons of Sundaria, Dragonkin, Lichdom, Zengeon
Abandonment, slow support or hostile community management	5 of 10	Wolcen, Hero Siege, SpiritVale, Dungeons of Sundaria, Zengeon
Story that is weak, unfinished or in the way	5 of 10	Wolcen, Dragonkin, Lichdom, Van Helsing III, Relic Hunters Legend
Combat or controls that feel wrongNo hit feedback, slow or locked movement, oversized hitboxes, mobile-style controls	4 of 10	Eldest Souls, Dungeons of Sundaria, Lichdom, Zengeon
Online operationsUnstable servers, wrongful bans with no appeal, bots and a broken economy, trading switched off	3 of 10	Wolcen, Hero Siege, SpiritVale
Levels that waste timeHuge empty maps, hours-long corridors, no map, distant checkpoints	3 of 10	Dungeons of Sundaria, Dragonkin, Lichdom
Grind or time-gates used as contentCity or rank grinds, daily caps, hundreds of hours without an upgrade	3 of 10	Hero Siege, Dragonkin, Relic Hunters Legend
Paid extras inside a paid gameStash tabs, classes, subscriptions	1 of 10	Hero Siege
The core system is never taught	1 of 10	Lichdom

The combinations that recur in indie action RPGs

The Diablo-like that runs out

Good look and campaign combat, then stat-stick loot, few viable builds and a repeating endgame.

Seen in

Wolcen, Dragonkin, Van Helsing III

Check your game against this pattern →

The live-service indie

A loved core loop undone by servers, bans, bots, wipes or monetization the team could not operate.

Seen in

SpiritVale, Hero Siege, Wolcen

Check your game against this pattern →

The mislabelled game

Tags or marketing promise a genre the game is not.

Seen in

Eldest Souls as a soulslike, Zengeon as a build-driven roguelite, Lichdom's "thousands of spells"

Check your game against this pattern →

Read the 10 indie action RPGs game by game ↓

10 games · report dated September 20, 2026

Indie psychological horror games

All ten games were praised, even by people who refunded them, for atmosphere, visuals, sound and voice acting. None of that saved them. What sank them was the layer underneath: horror with no real threat, stories that do not pay off, walking with nothing to do, padding, and a price the length could not carry. In this niche, presentation gets a game bought; systems, story payoff and respect for the player's time get it recommended.

Failure patterns across the ten indie psychological horror games, with how many of the ten showed each one
Failure pattern	Games (of 10)	Which
Story does not landIncoherent, buried in notes or metaphor, obvious twist, or an ending that explains nothing / arrives out of nowhere	9 of 10	All except Fear Therapy
Padding and tediumSlow forced walking, backtracking, fetch loops, chores, pixel hunting	8 of 10	Scratch Creek, Observation, Layers of Fear 2, The Park, Kholat, Remothered, Fear Therapy, Autopsy Simulator
Horror that does not scareToothless or telegraphed enemies, recycled or constant jumpscares, no stakes	7 of 10	Amnesia AMFP, Lust for Darkness, Layers of Fear 2, The Park, Kholat, Fear Therapy, Autopsy Simulator
No agencyOn-rails corridors, puzzles that solve themselves, hand-holding, "click where told"	6 of 10	Amnesia AMFP, Lust for Darkness, Layers of Fear 2, The Park, Kholat, Autopsy Simulator
Broken expectationSequel drops what the series was loved for, misleading label, co-op bolted onto a solo series, lore that needs another game	6 of 10	Amnesia AMFP, Layers of Fear 2, Scratch Creek, Autopsy Simulator, The Park, Kholat
Comfort and controlsMotion sickness, headache-inducing filters, painfully loud stingers, unrebindable keys, unreadable text, punishing stamina	6 of 10	Observation, Layers of Fear 2, Fear Therapy, Remothered, The Park, Kholat
Bugs, softlocks, crashes	6 of 10	Scratch Creek, Observation, Fear Therapy, Autopsy Simulator, Amnesia AMFP, Remothered
Too short for the priceAll priced $13-25 for 1.5-5 hours	5 of 10	Amnesia AMFP, Lust for Darkness, The Park, Kholat, Autopsy Simulator
Trial-and-error deaths with distant checkpoints, usually in chase sequences	4 of 10	Scratch Creek, Layers of Fear 2, Kholat, Remothered
Player does not know where to go or what to do	4 of 10	Scratch Creek, Observation, Kholat, Fear Therapy
Unreliable stealthEnemy detection and spatial audio the player cannot learn	1 of 10	Remothered
Abandoned in early access	1 of 10	Fear Therapy

The combinations that recur in indie psychological horror games

The premium walking sim

Gorgeous, linear, non-threatening, obscure story, 2-4 hours at $13-20.

Seen in

Amnesia AMFP, Lust for Darkness, The Park, Kholat, Layers of Fear 2

Check your game against this pattern →

The good idea with broken systems

A strong, fresh hook undone by controls, AI, bugs or guidance.

Seen in

Observation, Remothered, Scratch Creek, Autopsy Simulator

Check your game against this pattern →

Read the 10 indie psychological horror games game by game ↓

What the hooks were

Every hook in these tables sold. Each game passed 500 reviews on the strength of it, so a developer making a similar promise can treat the demand as shown and spend the worry on delivery.

Souls-likes

The hook each of the ten souls-likes sold on
Hook	Example
One new defensive mechanic on top of the Dark Souls formula, from a small team	Mortal Shell (hardening, shells)
Souls combat in a procedurally generated roguelike, with 4-player co-op	NECROPOLIS
An art-first, gentler souls-like built around co-op and a village that grows	Ashen
"Dark Souls in space", with couch co-op	Hellpoint
"2D Dark Souls" in pixel art	Dark Devotion
Sequel to a loved 2D souls-like, now with hunts and online co-op	Salt and Sacrifice
A three-person team's megastructure world in the spirit of Dark Souls 1	Bleak Faith: Forsaken
A sunlit souls-like rooted in a specific culture (Italian folklore, masks)	Enotria: The Last Song
A parody souls-like about video game nostalgia	The Last Hero of Nostalgaia
Co-op souls-like roguelike with a gothic, metal identity at $20	Crimson Moon

Several hooks are proven and, going by these reviews, still not delivered well: the co-op souls-like that works solo too, the souls-like roguelike with enough content, the culturally specific setting, and the small interconnected world that respects the player's time.

Indie action RPGs

The hook each of the ten indie action RPGs sold on
Hook	Example
"The next Diablo / Path of Exile, but prettier and classless"	Wolcen
Cheap pixel-art loot ARPG with many classes and co-op	Hero Siege
Nostalgia for a specific old MMO, with no pay-to-win	SpiritVale (Ragnarok Online)
Boss-rush soulslike in pixel art	Eldest Souls
Budget co-op dungeon crawler with hand-made dungeons	Dungeons of Sundaria
A new build-crafting gimmick (hex-grid skill puzzle) plus a city to grow	Dragonkin: The Banished
A class power fantasy: "the mage is a total badass", no mana, no cooldowns	Lichdom: Battlemage
Finale of a liked trilogy, import your character	Van Helsing III
Anime art plus roguelite runs with friends	Zengeon
Sequel to a liked indie, as a co-op looter-shooter	Relic Hunters Legend

Several hooks are proven and, going by these reviews, still not delivered well: the classless Diablo-like whose builds stay diverse at endgame, the mage-only power fantasy, the no-pay-to-win nostalgia ARPG without a shared economy to police, and the honest boss rush.

Indie psychological horror games

The hook each of the ten indie psychological horror games sold on
Hook	Example
Sequel to a loved horror game	Amnesia AMFP, Layers of Fear 2, Fears to Fathom: Scratch Creek
Real unsolved mystery + known narrator	Kholat (Dyatlov Pass, Sean Bean)
Unusual point of view	Observation (you are the station's AI)
Taboo theme	Lust for Darkness (erotic Lovecraftian cult)
Distinctive place	The Park (abandoned amusement park)
Homage to a classic formula	Remothered (Clock Tower-style stalker horror)
A profession as horror	Autopsy Simulator
Cheap co-op scream session with friends	Fear Therapy

Several of these hooks are proven and, going by these reviews, still not delivered well: the AI point of view, the profession-as-horror simulator with real deduction, stalker horror with trustworthy audio, the real-mystery expedition.

The 30 games, one by one

Each card gives what the game was bought as, where reviewers say it failed (the emphasis is the source report's), up to three verbatim review quotes, and what doing it right would look like. Quotes are anonymous and unedited, typos included.

Souls-likes

There are 76 flawed hits on the 382-game watchlist for this genre. These ten were picked as small- and mid-team games an indie developer can compare with: $19.99-39.99, a clear hook, spread across 2016-2026, each with 400+ usable negative reviews. Eldest Souls was left out because it is already in the indie action RPG set.

Mortal Shell

2021 on Steam
$29.99
67% positive
10,707 reviews

The hook

Bought as a gorgeous Dark Souls-like from a small team with one new idea: harden mid-swing and inhabit different "shells". Hardening, weighty combat, art and sound are praised.

Where it failed

Failed on a hub and areas that look the same with no map and almost no fast travel, scarce healing, an inconsistent parry, the same few enemies placed in groups, and a forced trek back out after every boss.

From the reviews

“Very few landmarks to know where you're going.”

“everything looks the same, so you will get lost a lot”

“the same handful of enemies that have limited movesets”

Done right

Landmarks and sightlines before a map is even needed; a healing source that refills at checkpoints; no more than a couple of enemies per encounter in a combat built for duels; a shortcut out after each boss instead of a second gauntlet.

NECROPOLIS: BRUTAL EDITION

2016
$29.99
63% positive
3,696 reviews

The hook

Bought as souls-style combat in a roguelike dungeon from a known studio, with four-player co-op. The low-poly look, stamina combat and co-op are praised.

Where it failed

Failed because it runs out in an afternoon: a handful of samey levels, difficulty by enemy count, sluggish uncancellable attacks, one sponge boss, co-op saves that fail, and development that stopped.

From the reviews

“Final run time: 1:34:49, and that is hours:minutes:seconds.”

“The visuals in the first seven levels are the same with slightly different colors.”

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

Done right

A roguelike needs far more rooms, enemies and bosses than a linear game, so cut the procedural promise or fund the content; a boss every few floors; enemies with different behaviours instead of more of them.

Ashen

2019 on Steam
$39.99
67% positive
3,280 reviews

The hook

Bought for its art, a gentler take on the formula, a growing village and seamless co-op. Art, music, the village and climbing are praised.

Where it failed

Failed on weapons that all play alike with no stats or magic, late dungeons with a single checkpoint and long boss runs, an unfair final boss, an AI partner that falls off cliffs, awkward "passive" co-op, and loot that is mostly crafting junk.

From the reviews

“The weapon variety isn't interesting--essentially an assortment of clubs and axes”

“every weapon feels almost identical”

“atrocious boss runs, what an awful and frustrating decision”

Done right

A few weapon classes with truly different movesets beat many reskins; a checkpoint before every boss; if the game is balanced for two, the AI partner is a core feature, not a fallback; normal co-op lobbies.

Hellpoint

2020
$34.99
76% positive
2,582 reviews

The hook

Bought as "Dark Souls in space" with couch co-op. The cosmic-horror station, interconnected layout, secrets and split-screen are praised.

Where it failed

Failed on platforming deaths in a game whose jump is clumsy, huge look-alike mazes with no map, fast travel that needs a rare consumable, healing that does not refill at checkpoints, and technical jank.

From the reviews

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

“you cannot freely fast travel between Breaches”

“Jumping puzzles that will ruin your day due to the clunky movement”

Done right

Keep platforming optional unless movement is built for it; free fast travel between discovered checkpoints; checkpoints restore healing; distinct colour and architecture per zone.

Dark Devotion

2019
$19.99
74% positive
1,925 reviews

The hook

Bought as a 2D Dark Souls in atmospheric pixel art. Art, bosses and the feel of combat once learned are praised.

Where it failed

Failed on controls that cannot be rebound, a platform-filled world with no jump and one-way doors, gear lost on every death with a single teleport point, enemies that ignore the stamina rules the player lives by, and missing basics such as pause and options.

From the reviews

“there is no jump in this game.”

“Not rebindable in any way, shape, or form.”

“Abysmal control scheme that can't be remapped on controllers”

Done right

Full rebinding on day one; decide whether the game is a roguelike or an exploration game and design death around that choice; enemies obey visible rules; several active teleport points.

Salt and Sacrifice

2023 on Steam
$19.99
62% positive
1,738 reviews

The hook

Bought as the sequel to Salt and Sanctuary, now with mage hunts and online co-op. Animation, art, the skill tree and crafting are praised.

Where it failed

Failed because the hunt loop makes every boss a chase across the map through respawning mobs, enemies juggle and knock the player off ledges, weapon freedom shrank compared with the first game, and levelling means leaving the level.

From the reviews

“who thought that was FUN? why do I have to chase bosses around through the entire game while dealing with trash mobs”

“chasing them from room to room and back again, like a gaggle of kids chasing a soccer ball”

“constant ganks or instant reactions from enemies that can juggle you for days without much you can really do about it”

Done right

A sequel keeps the freedom the first game gave; a borrowed loop has to suit 2D platforms and knockback before it is built around them; two or three relocations per hunt, not seven.

Bleak Faith: Forsaken

2023
$29.99
75% positive
2,839 reviews

The hook

Bought for a three-person team's vast megastructure and the promise of Dark Souls 1-style interconnection. World scale, art and music are praised.

Where it failed

Failed on floaty, delayed movement and weak controller support, hitboxes and stamina that make combat unreliable, minutes-long ladders and elevators across an empty world, and a release players called unfinished.

From the reviews

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

“I'd wager at least 20% of my playtime was spent climbing up or down ladders.”

“THERE ARE TWO 6 MINUTE ELEVATORS where you have to wait 3 minutes for the elevator to come down”

Done right

Character control first, world size last; a world a small team can fill; traversal measured in seconds; call it early access if it is.

Enotria: The Last Song

2024
$39.99
69% positive
2,184 reviews

The hook

Bought as a bright, Italian-folklore souls-like with a mask loadout system. Level design, art and a generous parry are praised.

Where it failed

Failed because it shipped with soft-locks and broken boss behaviour, the combat funnels everyone into parrying, weapons in a class share one moveset, and it was shorter than advertised.

From the reviews

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

“I even got soft-locked from starting the 3rd region”

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

Done right

Dodge, block and parry all have to be viable if builds are the pitch; test every boss and every quest state before 1.0; advertise the playtime testers actually recorded.

The Last Hero of Nostalgaia

2022
$24.99
78% positive
2,038 reviews

The hook

Bought as a witty parody souls-like about game nostalgia. World design, the item-memory system, the look and the narrator are praised.

Where it failed

Failed on fast travel limited to one chosen point in a game that demands revisiting places, hit detection that fails on slopes and at range, late-game gank fights, soft-locks, and NPC quests that fail without warning.

From the reviews

“you can only choose 1 shrine that you can teleport back at a time”

“Level design, bosses, hit detection, a lot of it just started to unravel for me”

“wonky combat and enemy gank are incredibly frustrating”

Done right

If a system asks players to return to old places, fast travel to all of them is part of that system; NPC quests that cannot be failed by simply progressing; keep the last third at the standard of the first.

Crimson Moon

2026
$19.99
59% positive
2,792 reviews

The hook

Bought as a co-op souls-like roguelike with a gothic, metal identity at $20. Look, music, bosses and weighty attacks are praised.

Where it failed

Failed on a camera that fights the player, an unreliable parry, co-op that loses gear on disconnect, no rebinding at launch, a tiny inventory and a loadout only reachable after picking a mission, and two weapon types for the whole game.

From the reviews

“Probably the worst camera experience I've had in a souls-like game.”

“a full 180 degree snap rotation”

“we get disconnected frequently in the middle of a run”

Done right

Camera and lock-on tested against the largest enemy in the tightest room; reconnect restores the run; rebinding and a hub loadout screen at launch; in a roguelike, weapon types are the replay value.

Check your own game against the souls-likes checklist →

Indie action RPGs

There are 54 flawed hits on the 239-game watchlist for this genre. These ten were picked as clearly action RPGs, indie-priced ($7.99-24.99), with a clear hook, spread across 2014-2026, each with 300+ usable negative reviews.

Wolcen: Lords of Mayhem

2020
$19.99
55% positive
61,837 reviews

The hook

Bought as the good-looking, classless Diablo-like. Graphics, weighty campaign combat, the story's first acts and the dual-resource idea are praised even by detractors.

Where it failed

Failed on bugs that outlived launch, a handful of viable builds, stat-stick loot, a campaign that stopped on a cliffhanger, a repetitive endgame, and online play that lagged, wiped characters on big updates and was eventually shut down.

From the reviews

“every item in the game is a BORING stat-stick”

“they're just nerfing the the few viable builds on endgame”

“It ends on a cliffhanger.”

Done right

Ship fewer skills but make most of them endgame-viable, and buff the weak before nerfing the strong; give items effects that enable builds, with deterministic crafting; finish the story you start; build the endgame loop before launch, not after; do not take on online infrastructure you cannot keep running.

Hero Siege

2014
$7.99
67% positive
36,859 reviews

The hook

Bought as a cheap, addictive pixel-art loot game with many classes and co-op; the loot loop, class roster and co-op are still praised.

Where it failed

Failed because a full redesign replaced the game long-time owners had bought, because nearly everything beyond the base game costs extra, because season launches overload the servers, and because of bans and a hostile relationship with its own players.

From the reviews

“EVERYTHING needs to be bought. You need more inventory space? Fork up 4 dollars.”

“Stash tabs cost money, inventory tabs cost money, in-game housing requires a subscription.”

“This game in it's early days used to be awesome.”

Done right

Evolve a live game without deleting the version people paid for (keep a legacy mode); charge once or sell cosmetics, never storage and classes in a paid game; size seasons to what the servers can take; keep moderation and ban appeals away from whoever is arguing with players.

SpiritVale

2026
$14.99
48% positive
4,763 reviews

The hook

Bought by Ragnarok Online veterans as a modern, no-pay-to-win grind MMO-ARPG from a tiny team. The grind loop, free respecs and itemization are praised.

Where it failed

Failed almost entirely on operations: bans that players say hit legitimate accounts with no real appeal, bots that wrecked the economy, servers that went down daily, and trading and the auction house being switched off.

From the reviews

“his account was suddenly banned for allegedly using a bot”

“Player trading has been disabled with no ETA”

“Servers are down multiple times per day for multiple hours”

Done right

An online economy is an operations commitment: server-side bot detection before launch, a human review and a real appeal for every permanent ban, load testing at playtest scale, and never removing trading as a fix. If the team cannot staff that, make the game self-found or co-op without a shared economy.

Eldest Souls

2021
$19.99
76% positive
3,125 reviews

The hook

Bought as a pixel-art soulslike. Boss designs, art, music and the three skill trees with free respec are praised.

Where it failed

Failed because it is a boss rush in an empty world sold under a soulslike label, with oversized boss collision that pins the player, slow movement, bosses with huge health that kill in two hits, and a short run for $20.

From the reviews

“This is a boss rush game, where you travel through an empty world from one boss fight to the next”

“Not being able to dodge through bosses, having dodges on cooldown and unpredictable attack patterns”

“Player movement feels slow and sluggish.”

Done right

Label it a boss rush; hitboxes match the sprite and the dodge passes through bosses; shorter fights where each hit matters; let players choose boss order so one wall does not end the game at boss two.

Dungeons of Sundaria

2023
$9.99
74% positive
3,061 reviews

The hook

Bought as a $10 old-school co-op dungeon crawler. Large hand-made dungeons, distinct classes and easy co-op with split-screen are praised.

Where it failed

Failed on level scaling that makes levelling up feel like getting weaker, combat with no feedback, progress-destroying bugs, hours-long dungeons with no map, and development that stopped.

From the reviews

“level-scaling system that actually makes you WEAKER”

“hit detection is non existent”

“the combat is clunky and floaty”

Done right

Never scale enemies past the player's gains; hit reactions and i-frames before content; dungeons in 30-minute chapters with a map and landmarks; fix save and quest bugs before adding anything.

Dragonkin: The Banished

2026
$24.99
72% positive
2,240 reviews

The hook

Bought for its hex-grid skill-building puzzle, pet dragons and city building, by players tired of seasonal ARPGs. The grid, snappy combat, visuals and quality-of-life options are praised.

Where it failed

Failed on a city grind that reads as padding, enormous empty maps, difficulty that only inflates health, loot with no uniques and rarity locked behind level, a few dominant skills, and a campaign full of escort and defend objectives.

From the reviews

“Upgrading your city to the maximum level requires an absurd number of hours spent repeating the same activities.”

“The maps are absolutely, mindbogglingly large”

“Higher difficulty means enemies take longer to kill”

Done right

Maps a third of the size with purposeful side dungeons; build-defining uniques from named bosses; higher difficulties add mechanics, not health; meta-progression that unlocks options instead of gating power behind hours; no escort quests.

Lichdom: Battlemage

2014
$9.99
57% positive
2,088 reviews

The hook

Bought on the promise that "the mage is a total badass": first-person spellcasting with no mana and deep spell crafting. Spell effects, visuals and voice acting are praised.

Where it failed

Failed because the crafting collapses into one rotation, the same few enemy types repeat for 15+ hours of linear corridor levels with distant checkpoints, and the central system is barely explained.

From the reviews

“an amazing concept and good production value is ruined by awful writing, a lack of direction, bad combat balancing”

“walk through long corridor levels”

“spell creation system is barely explained and extremely convoluted”

Done right

A few spell archetypes that really play differently, available early; enemies that demand different answers; short arenas with frequent checkpoints; teach the crafting system in play with sample builds.

The Incredible Adventures of Van Helsing III

2015
$14.99
64% positive
1,211 reviews

The hook

Bought by fans to finish the trilogy with their character. Katarina, the setting, music and the core combat are still praised.

Where it failed

Failed because it was smaller than its predecessor in every way players cared about: no character import, level cap cut to 30, skill trees cut down, unbalanced new classes, a short campaign, and frequent crashes.

From the reviews

“Between VH1 and 2 you could import characters. They removed that from VH3.”

“Level cap reduced from VH2 down to 30”

“The skill trees have been significantly dumbed down”

Done right

A sequel keeps what the series let players build, or says loudly that it does not; never ship a finale with less depth and length than the previous entry; test class balance before adding classes.

Zengeon

2019
$19.99
66% positive
1,229 reviews

The hook

Bought for anime art, five characters and roguelite runs "with hundreds of combinations", solo or co-op. Art, animations and co-op are praised.

Where it failed

Failed because items do not change how a character plays, three fixed maps and a few enemy types run out in hours, the controls feel like a mobile port, and updates stopped.

From the reviews

“The controls are VERY awkward and ignoring the weird laggy/stutter effect and yanky controls, it makes the game difficult.”

“The gameplay gets very repetitive really quick”

“Devs clearly abandoned the game”

Done right

In a roguelite, items must create archetypes, not percentages; real map and encounter variety before launch; native keyboard-and-mouse and controller schemes; a healing system that works when the player is in trouble.

Relic Hunters Legend

2025
$19.99
65% positive
1,101 reviews

The hook

Bought as the co-op looter-shooter sequel to a liked indie. Gunplay, art, music and character variety are praised.

Where it failed

Failed on persistent bugs, one character far stronger than the rest (then nerfed without buffing the others), progression that forces levelling all five characters, solo story missions in a co-op game, few mission types and daily caps.

From the reviews

“balance between the characters is honestly a joke, Raff is hands down the GOAT”

“guns feel very punchy, and the loot system seems like it will get better”

Done right

Let players main one character; every mission playable in co-op; fix blocking bugs before balance passes, and balance by buffing; no daily caps in a game about farming.

Check your own game against the indie action RPGs checklist →

Indie psychological horror games

There are 102 flawed hits on the watchlist for this genre. These ten were picked as clearly psychological horror, indie-priced, and spread across 2013-2026.

Amnesia: A Machine for Pigs

2013
$19.99
69% positive
7,236 reviews

The hook

Bought as the sequel to The Dark Descent. Praised for sound, voice work, the Victorian industrial setting and its themes.

Where it failed

Failed because it removed what the series was loved for (inventory, sanity, scarce light), left enemies rare and harmless, made levels a single corridor, and buried the story in metaphor, at 3-5 hours for $20.

From the reviews

“The lamp also begins to flicker everytime a monster is near”

“Each level is a single, straight, path with no choices on how to proceed”

“The few puzzles that do exist are embarrassingly simple”

Done right

Keep one tension system where the player's state raises the danger; make each of the few encounters a decision (hide, run, spend a resource); tell the main plot plainly and keep the poetry optional.

Fears to Fathom: Scratch Creek

2026
$7.99
75% positive
7,167 reviews

The hook

Bought on the series' name, now with co-op. Atmosphere, voice acting, the texting system and the moments where the two players are separated all land.

Where it failed

Failed on a final chase with invisible walls, darkness, one valid route and no checkpoints, on widespread softlocks, on chores and driving that reviewers read as padding past the refund window, and on a story with no mystery.

From the reviews

“I got stuck there for like 40 min”

“Over half of the game is packing boxes”

“Game pads itself over two hours”

Done right

Chases with several working routes, readable cues and frequent checkpoints; skippable dialogue on retry; both co-op players always have something to do; withhold the antagonist for the first hour.

Observation

2020
$24.99
78% positive
4,590 reviews

The hook

Bought for the premise: you are the AI of a stricken space station. Visuals, sound, acting and the first act are praised.

Where it failed

Failed mostly on controls (heavy drone inertia, slow camera, disorienting zero-g, motion sickness), then vague objectives with pixel hunting, repeated Simon-says minigames, and an ending that explains little.

From the reviews

“the flying controls are absolutely atrocious”

“There is simply no intellectual challenge whatsoever”

“far shallower than it thinks it is”

Done right

Responsive movement and comfort options first; highlight interactables in range; vary minigames instead of repeating them; resolve the central mystery.

Lust for Darkness

2018
$14.99
68% positive
3,980 reviews

The hook

Bought for a rare premise: erotic Lovecraftian horror with Giger/Beksinski visuals. The two worlds look the part and it runs well.

Where it failed

Failed because it is 2-3 hours for $15, not frightening (cheap jumpscares, trivial stealth), a corridor of decorative objects and same-room puzzles, with weak acting and an ending that undercuts the rescue plot.

From the reviews

“Is that it?”

“it's just not a horror at all”

“the game does not have the guts to go full on down the rabbit hole to really disturb the player”

Done right

Commit to the theme as character drama rather than decoration; make the two-world mechanic the core of the puzzles; give stealth real stakes; 6+ hours or a lower price.

Fear Therapy

2021
$4.99
68% positive
3,630 reviews

The hook

Bought as a cheap co-op scare night. Fun with friends for a session, good first twenty minutes.

Where it failed

Failed on jumpscares every minute or two that become noise, a fetch loop across huge dark maps with a forced item order, painfully loud audio, bugs, content that needs exactly four players, and three years in early access.

From the reviews

“I don't think that putting a jump scare moments on every minute increases the fear element, on the contrary, it decreases it”

“searching for stuff in the dark”

“still in early access”

Done right

A small jumpscare budget tied to events; staged objectives with clear markers instead of item hunts; sane audio levels; every mode completable solo or with two.

The Park

2015
$12.99
69% positive
2,534 reviews

The hook

Bought for the setting and the lost-child premise. Atmosphere, art and the lead performance are praised, as is the looping-apartment finale.

Where it failed

Failed because it is 1-2 hours for $13, not scary, fully on rails, with a twist many guessed early, lore that leans on an MMO, and running disabled whenever the character talks.

From the reviews

“The game is so linear it could have been on rails”

“If you haven't figured out the entire plot by the halfway point”

Done right

4-6 hours or a $5-7 price; never lock movement to narration; let the reveal come in layers; make the story stand alone.

Layers of Fear 2

2019
$19.99
71% positive
2,134 reviews

The hook

Bought as more of a loved art-horror game. Looks and sounds superb.

Where it failed

Failed on monster chases that turn exploration into trial-and-error deaths, a story nobody could follow without a video explainer, a black-and-white filter over most of the game that tired eyes and hid the art, and recycled mannequin scares.

From the reviews

“the sheer amount of deaths you go through from cheap chases”

“80% of the time game is black and white old movie which gave me headache”

Done right

No instant-death chases in an exploration game; one visual gimmick per act, not 70% of the runtime; carry the story through the spaces, not only the notes; anchor the protagonist early.

Kholat

2015
$19.99
67% positive
2,036 reviews

The hook

Bought for the Dyatlov Pass mystery, the snowscapes and Sean Bean. Sound and visuals are widely praised.

Where it failed

Failed on navigation as punishment (a map with no player position, identical terrain), long empty hikes, tiny stamina and no jumping, one-hit deaths that undo 15 minutes of walking, and a story that abandons the real mystery for a tangle nobody could follow.

From the reviews

“doesn't say where the hell you are”

“leaves the map almost useless”

Done right

Getting lost must cost minutes, not the session; interiors and safe rooms between outdoor stretches; traversal as a tool, not a handicap; one mystery per act.

Remothered: Tormented Fathers

2018
$19.99
75% positive
1,774 reviews

The hook

Bought as a Clock Tower successor with a giallo flavour. The mansion, lighting, music and premise are praised.

Where it failed

Failed because the stealth cannot be learned: spatial audio that does not tell you where the stalker is, inconsistent detection, plus basement-to-attic fetch loops, unreliable QTEs with distant checkpoints, and unrebindable, fussy controls.

From the reviews

“whether he's right next to you or upstairs”

“go to second floor to use X”

“Stalker mechanics are rather clunky”

Done right

In stalker horror, audio positioning and fair, learnable detection are the game; unlock shortcuts as the house opens up; frequent checkpoints; no reflex QTEs.

Autopsy Simulator

2024
$24.99
49% positive
1,337 reviews

The hook

Bought, after years on wishlists, as a forensic simulator with horror. The anatomy and procedures impress even pathologists.

Where it failed

Failed because it is not a simulator: a scripted story where every click is highlighted and nothing can be got wrong, a protagonist who sighs and whines constantly, horror with no stakes, a rushed ending, softlocks, 4-6 hours for $25.

From the reviews

“contains neither simulation and barely any autopsies”

“You click when and where you are told to click”

“sighing constantly”

Done right

Let the player deduce and be wrong, with consequences; cut the narration; ship the free-play mode on day one; name the game for what it is.

Check your own game against the indie psychological horror games checklist →

How far to trust this

For each game, the 142 most helpful Steam reviews were read in full: about 90 not recommended, 40 recommended with criticism and 12 plainly recommended. An AI model did the reading (one Claude Haiku subagent per game) and wrote a post-mortem of that game. The genre tables were then put together by hand from the ten post-mortems.

The n-of-10 figures are judgments, not counts. Each was tallied by hand from the ten post-mortems of a genre, and a post-mortem records an impression of what reviewers said, not a number of reviews.

Each genre report carries its own note on how far to trust it, including how its quotes were checked against the review text. The three notes are reproduced whole here. Where one says "below", it means the report's own quotes, which are the ones this page draws from.

Souls-likes report · September 21, 2026

Indie action RPGs report · September 20, 2026

Themes are reliable where several of the ten independent analyses converge, and most of them also show up in the structured extraction of recent games. Frequencies are the subagents' impressions, not counts; the "n of 10" below were counted by hand from the ten reports. Quotes were a weak point this time: hitfinder verify-quotes found only 31 of the 114 quotes the subagents cited (27%) verbatim in the cited review, because several agents paraphrased. Every quote below is verbatim: 21 come from the subagents and passed an exact-match check, and 30 were found by phrase search in the same exported reviews and then verified (30 of 30). Two reports repeated reviewers' accusations about named developers' personal conduct; those are not reproduced here, only the moderation and ban problems players describe.

Indie psychological horror games report · September 20, 2026

Themes are reliable: the ten independent analyses converge with each other and with the structured extraction of the last-6-months games. Frequencies are the subagents' impressions ("most", "about half"), not counts. Of 63 quotes checked against the database, 48 (76%) were verbatim in the cited review; only verified quotes appear below.

Every game here found an audience

Every game on this page found an audience. All 30 passed 500 reviews, so nothing here says what gets a game noticed in the first place, and avoiding every failure on this page does not mean a game will sell. What it does show is why games that already had buyers lost their recommendation.

Two quotes in the source reports are not in English and are left off this page, because most readers could not check the wording. Game names are given as the reports give them, and no reviewer is named anywhere.

The benchmarks elsewhere on this site are built differently, from store-page facts and not from reviews. The methodology page covers those.

Questions about this study

What is a flawed hit?

A Steam game with 500+ reviews and a 40-79% positive score. The review count shows the hook sold; the score shows the execution let players down. Every one of the 30 games on this page fits that definition.

Which failure was most common in each genre?

Souls-likes: "Combat that does not feel fair or reliable", in 8 of 10 games. Indie action RPGs: "Content runs out or repeats", in 9 of 10 games. Indie psychological horror games: "Story does not land", in 9 of 10 games. These are judgments, not counts.

Are the n-of-10 figures statistics?

No. The n-of-10 figures are judgments, not counts. Each was tallied by hand from the ten post-mortems of a genre, and a post-mortem records an impression of what reviewers said, not a number of reviews. Read them as a ranking of what came up most across ten games, and do not quote them as percentages of players or of reviews.

Are the review quotes real?

Yes. Every quote appears word for word in a Steam review of that game. The source reports checked each quote against the exported review text. A quote that did not match exactly was dropped, or replaced with wording found in the same reviews and verified. Quotes are anonymous, and no reviewer is named or linked.

If my game avoids all of these failures, will it sell?

This page cannot tell you that. All 30 games here passed 500 reviews, so every one of them had already solved the problem of being noticed. The post-mortems show why a game with buyers lost its recommendation. They say nothing about getting buyers in the first place.

Check your own game against these patterns

The Genre Readiness Check turns each genre's post-mortems, together with a larger set of recent reviews, into questions about what you have seen in your own playtests. It is free, and it flags any recurring combination from this page that your answers resemble.

The full genre reports (PDF, $49 each) are sold from each genre's page. See all genres.