id: 389fa673b9d24a978dbc1d11fc07385c
parent_id: 
item_type: 1
item_id: a80f1949d707441683cc651ee38f25e0
item_updated_time: 1782623680228
title_diff: "[{\"diffs\":[[1,\"Gen 4 Deep Sweep Results — 2026-06-27\"]],\"start1\":0,\"start2\":0,\"length1\":0,\"length2\":37}]"
body_diff: "[{\"diffs\":[[1,\"# Gen 4 Deep Sweep Results — 2026-06-27\\\n\\\n17 variants × 5 seeds × 50k hands, 6-max (2 flock + 2 chump + 1 test bot)\\\n\\\n## Results (chips/hand, higher is better)\\\n\\\n### Threshold Scaling Extension (key finding!)\\\n| Variant | Profit/Hand | vs ts07 (145) |\\\n|---------|------------|---------------|\\\n| **ts10_narrow (ts=1.0)** | **153** | **+8** |\\\n| ts09_narrow (ts=0.9) | 152 | +7 |\\\n| ts08_narrow (ts=0.8) | 149 | +4 |\\\n| ts07 (reference from ns) | 145 | — |\\\n\\\n**ts=1.0 is best but plateauing** (152→153 = only +1). Diminishing returns.\\\n\\\n### Narrowing Strength (at ts=0.7)\\\n| Variant | Profit/Hand | Notes |\\\n|---------|------------|-------|\\\n| ns08_ts07 (0.8×) | 145 | |\\\n| ns12_ts07 (1.2×) | 145 | |\\\n| ns15_ts07 (1.5×) | 145 | Doesn't matter! |\\\n\\\n**Narrowing strength has zero effect** — 0.8× = 1.0× = 1.2× = 1.5× all give 145.\\\n\\\n### Center Offset (at ts=0.7)\\\n| Variant | Profit/Hand | Notes |\\\n|---------|------------|-------|\\\n| ncp01_ts07 (+0.1) | 148 | **+3 over neutral** |\\\n| ncm01_ts07 (−0.1) | 140 | **−5, hurts** |\\\n\\\n**+0.1 center offset helps slightly** (shifts ranges wider, less aggressive narrowing).\\\n\\\n### S-Curve Steepness (at ts=0.5)\\\n| Variant | Profit/Hand |\\\n|---------|------------|\\\n| sc03_ts05 (0.3×) | 133 |\\\n| sc07_ts05 (0.7×) | 136 |\\\n| sc10_ts05 (1.0×) | 136 |\\\n| sc15_ts05 (1.5×) | 135 |\\\n\\\nS-curve at ts=0.5 (133-136) consistently worse than linear at ts=0.7+ (145+).\\\n\\\n### S-Curve Floor (at ts=0.5)\\\n| Variant | Profit/Hand |\\\n|---------|------------|\\\n| sf00_ts05 (floor=0.0) | 134 |\\\n| sf05_ts05 (floor=0.05) | 135 |\\\n| sf10_ts05 (floor=0.10) | 136 |\\\n\\\nFloor has minimal effect.\\\n\\\n### Best Combinations (scurve + narrow + high ts)\\\n| Variant | Profit/Hand |\\\n|---------|------------|\\\n| sc05_ns12_ts07 | 144 |\\\n| sc07_ns12_ts07 | 143 |\\\n\\\nStill worse than pure linear (145 at ts=0.7).\\\n\\\n## Summary\\\n\\\n**Best config: ts=1.0, narrowing enabled, linear ranges, center_offset=+0.1**\\\n\\\nProduction recommendation:\\\n```toml\\\nnarrow_enabled = true\\\nnarrow_strength_mult = 1.0  # doesn't matter\\\nnarrow_center_offset = 0.1  # slight improvement\\\nnarrow_floor = 0.02\\\nrange_use_scurve = false    # linear always wins\\\nrange_tightness_scale = 1.0  # maximum, but plateauing\\\n```\\\n\\\n**Key insight:** The threshold scaling parameter is the most impactful Gen 4 parameter. It compensates for the reduced win_prob values that result from narrowed opponent ranges. At ts=1.0, the threshold discount fully offsets the range narrowing effect.\\\n\\\n**Next sweep direction:** ts=1.1 or 1.2 to see if it goes negative (over-compensation), confirming ts=1.0 is optimal.\"]],\"start1\":0,\"start2\":0,\"length1\":0,\"length2\":2517}]"
metadata_diff: {"new":{"id":"a80f1949d707441683cc651ee38f25e0","parent_id":"e13f1845de9b4b6392ad866354fbd562","latitude":"0.00000000","longitude":"0.00000000","altitude":"0.0000","author":"","source_url":"","is_todo":0,"todo_due":0,"todo_completed":0,"source":"joplin-desktop","source_application":"net.cozic.joplin-desktop","application_data":"","order":1782533948774,"markup_language":1,"is_shared":0,"share_id":"","conflict_original_id":"","master_key_id":"","user_data":"","deleted_time":0},"deleted":[]}
encryption_cipher_text: 
encryption_applied: 0
updated_time: 2026-06-28T05:17:38.824Z
created_time: 2026-06-28T05:17:38.824Z
type_: 13