Skip to content

Latest commit

 

History

History
67 lines (48 loc) · 1.9 KB

File metadata and controls

67 lines (48 loc) · 1.9 KB

Crawler A/B Experiment — 2026-07-14

Hypothesis

v2.8.0 had crawler PRs because issues had bounty label and standard bounty formatting. v2.9.0+ lost crawlers when labels changed to zero-bounty / status:competition.

Experiment Design

Start: 2026-07-14 End: 2026-07-21 (7 days) Metric: External PRs, issue comments, claims

Experimental Group (6 issues)

Enhanced with: bounty label + [Bounty][$0] title prefix + reward text in body

Issue Title Type
#429 [Bounty][$0][agent competition] Blind test homepage SAG-Lite search Test
#459 [Bounty][$0][Benchmark] Run LessonReuseBench Benchmark
#290 [Bounty][$0][Docs] 5-minute quickstart Docs
#302 [Bounty][$0][Search] Show match reason Search
#304 [Bounty][$0][Search] --json output Search
#257 [Bounty][$0] Translate Chinese lessons Content

Control Group (remaining issues)

Standard labels: status:competition, agent-friendly, no bounty label

Issue Title Type
#352 Decide verified/ semantics Protocol
#354 Add trust_tiers Protocol
#318 Cursor integration Ecosystem
#317 VS Code extension Ecosystem
#313 Query expansion Search
#314 Typo tolerance Search
... (others) Various

Baseline (pre-experiment)

Metric Experimental Control
External PRs 0 0
Issue comments 0 0
Claims 0 0

Day 1 (2026-07-14)

  • Added bounty label to all 22 competition issues
  • Updated 6 experimental issues with [Bounty][$0] title + reward text
  • Waiting for crawler discovery

Day 7 (2026-07-21)

(TODO: fill in results)

Metric Experimental Control
External PRs
Issue comments
Claims
Unique contributors

Conclusion

(TODO: fill in after 7 days)