Unverified50% confidenceFactExact time
A 2019 survey of NeurIPS papers found that fewer than 25% of papers provided code and configurations sufficient to fully reproduce their results.
1
Sources
50%
Confidence
Long-term
Relevance
9/4/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedA systematic survey by Pineau et al. 2021 found that only about 20–30% of ML papers provided sufficiently complete experimental details to be reproducible.67% similarUnverifiedResearch shows that approximately 15–25% of papers experience score changes after rebuttal, with positive changes slightly outnumbering negative ones66% similarUnverifiedApproximately 23% of papers could not be fully reproduced as claimed, and at least 496 papers had at least one claim that was falsified or disputed.66% similarUnverifiedSome research estimates that fewer than 20% of strategies selected through backtesting within quantitative funds remain consistently effective in live trading.65% similarUnverifiedThe Open Science Collaboration's replication study of 100 psychology experiments found that only about 36% of results were successfully replicated.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/854232API
curl https://kongchang.com/api/v1/knowledge/claims/854232MCP
get_claim(id=854232)