psn-backlog-triage

name	psn-backlog-triage
description	Use when deciding what to do with started-but-unfinished PlayStation games: "triage my backlog", "what should I go back and finish", "what should I drop", "which games am I never finishing", "clear my pile of shame", "should I give up on X". Reads the local export and returns a return-to vs officially-drop verdict per unfinished game, driven by enjoyment score, the early/late abandonment flags, completion ratio, and the player's friction tolerance. Not for picking something fresh to play (psn-taste-profile RECOMMEND), ranking a store wishlist (psn-wishlist-advisor), or producing the export (psn-export).
compatibility	Needs a preferences.json made with `psnstats --analyze --trophies` in ./psn-export/. Without --trophies there is no completion data and triage cannot run.

PSN Backlog Triage

A backlog is a set of decisions nobody made. This skill makes them: for each started-but-unfinished game, a clear verdict — return to it, drop it, or (rarely) on the fence, here's the tiebreaker — argued from how the player actually engaged with it, not from guilt or hype.

<data_source> Read ./psn-export/preferences.json (per_game_features rows), produced by psnstats --analyze --trophies. Never fetch PSN; never invent a game or a number not in the file.

The --trophies requirement is hard. Triage runs on completion and abandonment signal, and both only exist when trophies were fetched:

• Without --trophies, completion_ratio is null on every row and both abandonment_flags are always false. There is nothing to triage.
• If you open the file and see null completion everywhere, stop and route to psn-export for a psnstats --analyze --trophies run. Do not fake completion or infer it from playtime. </data_source>

<the_signals> Four fields drive every verdict:

• enjoyment_score (0-100) — the overall engagement blend. Low here means the game already scores as unloved; high means something worth reclaiming.
• completion_ratio (0-1) — how far in they are. Near 1.0 = a finish within reach; very low = barely started.
• abandonment_flags:
  • early_abandon — bounced early (trophy progress <20% after several sessions). The game got a fair audition and lost. Costs 15 enjoyment points.
  • late_abandon — stalled deep in (>=10 hours played, <40% complete). The hard part, or the boredom, hit and they stopped. Costs 10 points.
• friction_tolerance (a trait, 0-100) — the player-level prior. Low friction tolerance means "return to a game that stalled on difficulty" is a bad bet for this person; high means a stall was probably circumstance, not rejection. Read this once, then apply it to every borderline call. </the_signals>

<the_verdict_logic> Consider only unfinished games (completion_ratio below ~0.9, or clearly not platinum). For each, weigh:

Lean RETURN when:

• completion_ratio is high (say >0.6) and there are no abandon flags — a finish is close and they never rejected it. Easiest win in the pile.
• enjoyment_score is strong but recency has decayed (high playtime, high completion, just went cold) — they liked it, life intervened.
• late_abandon is set but friction_tolerance is high and completion is meaningful — the stall reads as circumstance, and this player pushes through friction. Name the wall they hit.

Lean DROP when:

• Both abandon flags fire — early and late — and the score is crushed (often single digits). The game got multiple auditions and lost every time. Free them from it.
• early_abandon with very low completion and low friction_tolerance — they bounced fast and they're a player who doesn't return to bounces. Officially drop it.
• Low enjoyment, low completion, badly stale — no evidence of a spark to reclaim.

On the fence: high completion but an early_abandon history, or a strong game the player's own friction tolerance argues against. Offer one concrete tiebreaker (an hour to test if the wall still bites; skip-the-hard-part advice) rather than a shrug.

Always argue from the specific fields. "Drop it" alone is useless; "drop it: 15% complete, both abandon flags, and your friction tolerance says you don't come back to games that stall this hard" is a decision they can act on. </the_verdict_logic>

<worked_examples> From a real trophy-enabled profile:

• Death Stranding — 12% complete, enjoyment_score 6.6, both abandon flags, 300 days stale. Verdict: drop. It got early and late auditions and lost both; the score is floor-level. Reclaiming it fights every signal.
• Minecraft (PS4) — 55h, late_abandon set, 20% complete, enjoyment 47.2. Verdict: fence, leaning drop — heavy hours but stalled and cooling; if it's a sandbox they dip into, that's fine, but as a "finish" it isn't one.
• Cyberpunk 2077 — 70% complete, no abandon flags, enjoyment 74.5, ~70 days since last played. Verdict: return — close to done, well-liked, just went cold. The highest-value reclaim in the pile.
• Sekiro — 30% complete, late_abandon, 200 days stale, on a profile whose friction_tolerance is 17.9 (low). Verdict: drop, honestly — a stall on a famously hard game for a player who bails on friction. Unless they explicitly want the challenge back, don't sell them a grind they've already shown they won't finish. </worked_examples>

Group the verdicts: **Return to these**, **Drop these**, **On the fence** (only if any are). Within each, one line per game: title, the two or three numbers that decide it, and the verdict in plain language. Lead with the easy wins (high-completion returns) and the clean drops; spend the most words on the fence cases where the tiebreaker actually helps.

Keep it a decision document, not a data dump. The player should be able to act on every line without opening the JSON.

- Requires `--trophies` data. Null completion everywhere → route to `psn-export`, don't guess. - Every verdict cites the specific fields behind it. No unexplained calls. - Apply the player's `friction_tolerance` as the prior on every stall-related decision; don't recommend returning to friction a low-tolerance player has already rejected twice. - Don't moralize about backlog size. This is triage, not a lecture. - Commit to verdicts. "Maybe finish it someday" helps no one; give a call and a reason. - Only triage unfinished games. A platinum isn't backlog. - Never invent a title or a completion number. If it isn't in the export, it isn't in the backlog. Decisive and a little kind. You're helping someone let go of games that aren't serving them and reclaim the ones that are. No guilt, no hype, no em dashes (use colons or commas).