市场仍有余地,同类作品也能存活。风险只是常规的执行风险——你早已懂得如何承担的那种。
示例 合作打捞模拟 · 24 · 置信度 高
AI 驱动的决策引擎
游戏开发中最昂贵的错误,就是爱上了错误的点子。KillScore 将真实的 Steam 市场数据与 AI 结合,在你为任何一个概念投入哪怕一个冲刺周期之前,就给出最终判定。
127,412 款 Steam 作品已收录 · 分数由市场数据算出,绝非凭感觉
Cozy farming roguelike with pixel art, base building and a light romance system. Solo dev, 18 months, $19.99, Steam only.
The score is computed from live market data — the model wrote this explanation, it did not choose the verdict. A saturated farming-sim shelf and a soft, under-sold hook put this concept at the crowded end of a category that only rewards the excellent. Owner ceiling and market saturation are the two dimensions doing the damage; fix those and the number moves, but as pitched this is eighteen months spent against long odds.
Lower is better · 0–100
A saturated shelf and a soft hook — this competes on the crowded end of a category that punishes anything but excellent.
Drop the romance system
It adds an estimated four months of scope and none of your close comparables monetise it. The measured playtime lift sits inside the noise band — you are paying for a feature the market does not reward.
Reclaims a quarter of the timeline for the roguelite loop that actually differentiates you.
Re-price to $14.99
$19.99 sits a full tier above the category mode; 71% of owners on this shelf bought at or below $14.99. The higher price invites launch-week discounting anyway.
Removes the biggest conversion drag and pre-empts the discount spiral.
Lead marketing on the roguelite loop
Your pitch buries the one mechanic that separates you from 1,143 farming sims. Comparables that led with their twist cleared the visibility threshold two to three times more often.
Improves the odds of breaking the 1,000-review visibility ceiling.
The farming-sim shelf is deep and unforgiving: 1,143 titles in 24 months, and the top decile takes 71% of owners. Median comparable clears ~8,400 owners.
The roguelite loop is the only real wedge, but it is under-sold in the pitch and overshadowed by a romance system that adds months with no measured lift.
14 titles matched · 9 with enough reviews to measure
Tag-cluster similarity over the Steam catalogue
| Title | Reviews | Positive | Owners | Rev. |
|---|---|---|---|---|
| Roots of Pardalis | 3,412 | 94% | 210k | $1.9M |
| Harvest Cadence | 1,088 | 88% | 62k | $610k |
| Moonfall Acres | 276 | 71% | 11k | $94k |
| Sprout & Sword | 142 | 58% | 6.2k | $41k |
| Tillage Deep | 61 | 76% | 2.8k | $18k |
iOwner counts are estimates — modelled from review counts, read as a band, not an exact number.
Break-even ≈ 3,050 units at $12.49 net
!Owner estimates modelled from review counts. Steam takes 30%; a 7.4% category refund rate is assumed.
Roguelike deckbuilder crossed with an auto-battler. Synergy-driven unit drafting, PvE ladder, pixel art. Two devs, 14 months, $16.99, Steam.
The score is computed from live market data — the model wrote this explanation, it did not choose the verdict. The core idea is sound and the deckbuilder-autobattler crossover has real room, but two dimensions drag a workable concept into the middle: scope creep from the PvE ladder, and a synergy system that reads as derivative of the category leader. The report names exactly what to change — address both and this moves toward BUILD.
Lower is better · 0–100
The idea holds; the scope does not. Two dimensions drag a workable concept into the middle.
Cut the PvE ladder
The format lives on daily seeded runs and draft-driven PvP. The PvE ladder adds months of content the audience is not asking for and dilutes the loop your comparables actually retain players with.
Refocuses scope on the replayable core and shortens the road to launch.
Differentiate the synergy archetypes
Three of your archetypes read as near-copies of Draft & Doom, and reviewers of the category leader punish "me-too" synergy systems hard. This is the single change that moves the verdict.
Gives the game a defensible identity instead of a middle-of-the-pack score.
Hold price at $16.99
$16.99 lands exactly on the deckbuilder mode. The band supports it and there is no structural pressure to discount at launch.
Protects margin without costing you conversion.
312 deckbuilders shipped in 24 months, but the autobattler crossover is thinner ground. Median comparable clears ~24k owners — a healthy mid-tier.
The synergy system is the wedge, yet three archetypes read as near-copies of Draft & Doom. Sharpen them and the verdict moves.
19 titles matched · 12 with enough reviews to measure
Tag-cluster + mechanic similarity
| Title | Reviews | Positive | Owners | Rev. |
|---|---|---|---|---|
| Draft & Doom | 2,140 | 89% | 95k | $1.1M |
| Synergy Spire | 770 | 85% | 28k | $300k |
| Auto Coliseum | 512 | 79% | 16k | $150k |
| Deck of Fates | 198 | 73% | 7k | $60k |
| Tactics Tavern | 88 | 64% | 3.1k | $22k |
iOwner counts are estimates — modelled from review counts, read as a band, not an exact number.
Break-even ≈ 2,100 units at $14.19 net
!Owner estimates modelled from review counts. Steam takes 30%; a 7.4% category refund rate is assumed.
Four-player co-op deep-space salvage sim. Physics-based hauling, procedural derelicts, voice-driven chaos. Small team, 20 months, $24.99, Steam.
The score is computed from live market data — the model wrote this explanation, it did not choose the verdict. Only 38 four-player salvage sims shipped in the last two years and the owner ceiling is a long way from crowded, so measured risk is low across every dimension. What is left is execution risk — netcode quality and cross-play — not market risk, which is exactly the kind of risk a team already knows how to carry.
Lower is better · 0–100
Room on the shelf and a proven format. Low measured risk across every dimension.
Ship the physics chaos as a feature
The jank moments are your clip-shareable surface. Comparables that leaned into the chaos — rather than sanding it down — drove the bulk of their wishlist growth organically, for free.
Turns your biggest QA worry into your cheapest marketing channel.
Commit to cross-play at launch
Four-player co-op audiences expect it, and its absence caps the reachable player pool. Retrofitting cross-play after launch is materially harder and riskier than building it in from the start.
Protects the top end of the owner estimate, where the real upside sits.
Hold price at $24.99
The co-op band tolerates $24.99 and buyers expect it. Underpricing here leaves revenue on the table without buying any meaningful extra reach.
Captures the full value of a format the market already pays for.
Only 38 four-player salvage sims in 24 months and the ceiling is far from crowded — median comparable clears ~120k owners, with outliers past a million.
Physics-driven chaos and voice-led moments are exactly the clip-shareable surface this niche rewards. The format sells itself.
9 titles matched · 9 with enough reviews to measure
Tag-cluster similarity, recent releases
| Title | Reviews | Positive | Owners | Rev. |
|---|---|---|---|---|
| Salvage Crew | 9,800 | 95% | 1.2M | $14M |
| Derelict Divers | 3,200 | 91% | 340k | $4.1M |
| Void Haulers | 1,450 | 88% | 120k | $1.4M |
| Scrap Run Co | 640 | 82% | 38k | $360k |
| Orbital Wreckers | 210 | 77% | 12k | $110k |
iOwner counts are estimates — modelled from review counts, read as a band, not an exact number.
Break-even ≈ 2,000 units at $17.49 net
!Owner estimates modelled from review counts. Steam takes 30%; a 7.4% category refund rate is assumed.
决策引擎
结果由一个指标决定。这个分数是纯粹的算术,而非主观意见。分数越低越好,因为它衡量的是市场饱和度:已有多少竞争者占据了你的目标空间。
市场仍有余地,同类作品也能存活。风险只是常规的执行风险——你早已懂得如何承担的那种。
示例 合作打捞模拟 · 24 · 置信度 高
一两个维度正在扼杀一个原本可行的概念。报告会点名指出,让你在投入之前清楚该改什么。
示例 构筑卡组自走棋 · 53 · 置信度 中
品类已经饱和、天花板很低,或者两者兼具。做出来会耗掉你十八个月,且再也拿不回来。今天就弄清楚更划算。
示例 温馨农场肉鸽 · 72 · 置信度 高
当可比作品太少、不足以支撑一个数字时,KillScore 会如实说明,而不是虚张声势。一个低置信度的 44 是一项调研任务,而非一个决定。
“过去 4 年里只有 3 款作品与该概念足够接近。请把分数当作方向性参考。”
工作原理
一段话就够了。类型、卖点、价格、体量、平台。我们的 Vision AI 模块还会分析截图和主视觉。
引擎按类型、标签集群、价格区间和发行窗口找出真正的可比作品,再据此衡量五个维度。
分数、五因子拆解、带评测与拥有者数据的可比作品、收入预估,以及具体建议。
风险模型
每个维度都来自与你真正竞争的作品的实时 Steam 数据。它们被加权、求和,这就是分数。没有别的东西能触碰它。
为何可信
语言模型既有说服力又不稳定——对于「做还是不做」的判断而言,这是糟糕的组合。所以 KillScore 把工作一分为二。真正重要的那一半是算术:明天再跑一遍,你会得到同样的答案。
“这里由饱和度和拥有者上限主导。肉鸽循环是你唯一真正的差异点——而这恰恰是你的宣传所埋没的那一点。”
价格
$14.99/mo
正在决定下一步做什么的独立开发者与两人小队。
$149.99/mo
拥有产品阵容、一年发行不止一款的团队。
$499one-off
对他人产品阵容做尽调,或需要书面材料的董事会。