Persist the latest generated recommendation set for the 20260509 draw so the repository reflects the current final_practice run result.
Co-authored-by: Cursor <cursoragent@cursor.com>
Keep the fixed 11-number set intact while adding a second-stage portfolio selection that caps final recommendations to the 70,000 KRW budget, and update docs/data/scripts to match the current project structure and runtime flow.
Co-authored-by: Cursor <cursoragent@cursor.com>
- Mirrors 3_Practice_22 flow: DataCrawler, optional API crawl, fixed combo,
Monte Carlo filtered samples (default) or exhaustive mode
- Caps total recommendations under 100; saves recommend_ball.final.json
Made-with: Cursor
- final_BallFilter: CSV history loader, TRAIN_ALLOW for 6-sum and week diff,
fix filterOneDigitPattern ball overwrite bug, drop socket call
- final_filter_params: build sum6 and abs_sum_diff from rounds 1-800
- filter_model re-exports BallFilter; train/valid evaluate pass-through counts
- final_filterTest aligned with 1_FilterTest_25 plus optional MC survivors
- README and scripts/run_with_ncue.sh for ncue workflow
Made-with: Cursor
