Unverified50% confidenceFactExact time
Among trajectories that terminated cleanly yet actually failed, field value errors accounted for 77.61%, unexpected additional side effects for 43.30%, and missing required side effects for 25.36% (categories overlap).
1
Sources
50%
Confidence
Long-term
Relevance
10/9/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedScoring with completion-style proxy metrics will classify failures that terminate cleanly but produce incorrect results as successes, which is a flaw in many current agent evaluations.72% similarUnverifiedAs a negative critical moment, failure to redeem entitlements has a negative impact on user retention rates 3-5 times greater than that of ordinary bugs.71% similarVerifiedIn long-chain tasks, if each step has an error rate of approximately 5%, the overall success rate may drop to below 36% after 20 steps.70% similarUnverified21.2% is a relative reduction rather than an absolute one, meaning the new version eliminates approximately one-fifth of the failures relative to the earlier version's failure rate baseline.70% similarUnverifiedTreating every failure as an equally weighted deduction is a serious evaluation pitfall, as an average score of 95% can mask the most costly and irreversible failures within that 5%.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/991244API
curl https://kongchang.com/api/v1/knowledge/claims/991244MCP
get_claim(id=991244)