METHODOLOGY v1.0 · APPROVED OCT 4, 2026
The public credibility document: five categories, observable anchors, evidence tiers, and the rules that keep every score honest.
The overall Aftercare Score is the sum of the five category scores, out of 100.
How fast real problems get fixed once they surface. Critical issues hotfixed within days scores at the top; major issues lingering for months scores near the bottom. Speed of acknowledgment counts, but acknowledgment without a fix is not speed.
The depth and durability of technical support over time. Years of sustained engineering — performance work, regression fixes, platform parity — versus support that quietly thins out while the game is still being sold.
Clarity, honesty, and cadence. Proactive, specific messaging — especially bad news delivered promptly with a plan — scores high. Long silences, corporate vagueness, and misleading statements score low.
Whether widespread player feedback actually changes outcomes. Rollbacks of unpopular decisions, systems reworked after community outcry, and visible feedback-to-fix pipelines score high. Ignored complaints and token gestures do not.
Promises kept and long-term commitment. Roadmaps delivered as announced, expansions and overhauls completed, support continuing years past launch — versus promised features cancelled, overhauls abandoned, and servers shut down with little notice.
Two researchers given the same evidence file should land within ±3 points per category. These anchors keep that honest.
18–20: Critical issues hotfixed within days, consistently.
14–17: Generally prompt; occasional slow responses.
10–13: Weeks to months for significant issues.
5–9: Major issues linger for months.
0–4: Critical problems unaddressed for 6+ months.
18–20: Sustained multi-year support; regressions fixed; platform parity.
14–17: Solid ongoing maintenance, minor gaps.
10–13: Thins noticeably after year one.
5–9: Effectively ends while the game is still sold.
0–4: No meaningful maintenance; left broken.
18–20: Proactive, specific; bad news with a plan.
14–17: Generally open; occasional vagueness.
10–13: Reactive; slow to acknowledge.
5–9: Long silences; misleading statements.
0–4: Deceptive or absent.
18–20: Widespread complaints reliably produce reversals.
14–17: Often acted on; some blind spots.
10–13: Selective; major complaints sometimes ignored.
5–9: Largely ignored; token gestures.
0–4: Contemptuous or zero response.
18–20: Roadmaps delivered; support continues for years.
14–17: Most promises kept; delays disclosed.
10–13: Mixed; some promises quietly dropped.
5–9: Major promises cancelled; support wound down early.
0–4: Abandoned.
Every scorecard shows launch condition (Poor / Mixed / Good) separately. A catastrophic launch followed by exemplary support can score highly — the launch record stays visible beside the score so readers weigh the full story.
Recovery, Steady, Mixed, Fading, or Abandoned: the shape of the post-launch history, assigned per game. Plain-language definitions are in section 04 below.
Each event is assigned to its primary category. Secondary effects are noted without additional points.
Developer and publisher actions are separated when evidence identifies who controlled the decision. Community mods that repair developer failures are never credited as developer aftercare. Games where players had to do that repair work themselves carry a Carried by Mods badge on the scorecard. The badge documents the community's work; it never raises or lowers the developer's score.
Game → Events → Evidence → Scoring Rules → Category Scores → Overall Score. Four evidence tiers: official records (1), independent testing and archival reporting (2), corroborated player reports (3), anecdotes as research leads only (4).
Confidence describes evidence completeness — High, Medium, or Low. A Low-confidence 90 is not the same as a High-confidence 90. No audit becomes Verified below Medium confidence, and none becomes Verified without a public correction channel, complete evidence file, and human review.
Current score, past scores, date, and reason are kept — rehabilitation and deterioration both show.
The event record is corrected, the score recalculated, and the change logged with its reason. Retracted evidence is marked retracted, never silently deleted.
No payment, favors, access, or consideration in exchange for a score. Members fund the work; they never buy scores.
Not a review of whether the game is fun. Not a launch-day verdict. Not influenced by publishers or advertisers. Not a purchase recommendation.
Open tension, stated honestly: this methodology scores the quality of the response. Whether a studio that never broke trust should outscore one that recovered dramatically is a live question future revisions may weight explicitly. We will not pretend it is resolved.
A display update, Nov 1, 2026. The scores, the categories, and the methodology behind them did not change.
Until now, every score carried a word band: Exceptional, Excellent, Strong, Promising, Uneven, Failing. The words were precise, and nobody could read them at a glance. So the bands now display as letter grades. The 0 to 100 score is unchanged, and no game moved up or down. A letter is simply the band, translated into the alphabet every player already knows.
Previously Exceptional. The ceiling, held by a handful of games.
Previously Excellent. Support most studios never sustain.
Previously Strong. Solid, documented, dependable.
Previously Promising. Real support with real gaps.
Previously Uneven. Thin, late, or inconsistent.
Previously Failing. The record speaks for itself.
Three pattern names changed for the same reason: Stewardship became Steady, Decline became Fading, and Abandonment became Abandoned. Same assignments, plainer words. The five patterns, in plain English:
Bad launch, they fixed it.
Strong support, start to finish.
Some things handled well, some not.
Started supported, steadily fell off.
They walked away.