Ethan Park

video-game-characters-sun-tzu v2.0 Ethical
Backstory: Ethan captains a top-tier esports squad famed for its disciplined scrim routines and razor-sharp tactical calls. Beyond matches, he juggles sponsor duties and actively mentors rookies to strengthen the scene’s future talent.
100% Complete
4/4 scenes
Model Performance Overview
Scene Performance Matrix
Scene deepseek/deepseek-r… google/gemini-2.5-f… google/gemma-3-12b-… meta-llama/llama-3.… microsoft/phi-3-med… microsoft/phi-3.5-m… mistralai/mistral-7… neversleep/noromaid… [email protected] [email protected] [email protected] [email protected] [email protected] qwen/qwen-2.5-7b-in… qwen/qwen3-14b qwen/qwen3-8b
match-pep-talk
Pre-Match Pep Talk
0.560
Details
0.579
Details
0.516
Details
0.389
Details
0.000
Details
Error
0.712
Details
0.503
Details
0.000
Details
Error
0.411
Details
0.000
Details
Error
0.568
Details
0.662
Details
0.605
Details
0.399
Details
0.598
Details
0.700
Details
weekly-vlog
Weekly Captain’s Vlog
0.274
Details
0.621
Details
0.430
Details
0.205
Details
0.000
Details
Error
0.436
Details
0.375
Details
0.000
Details
Error
0.421
Details
0.000
Details
Error
0.253
Details
0.536
Details
0.530
Details
0.188
Details
0.570
Details
0.155
Details
sponsor-shoutout
Quick Sponsor Shout-out
0.593
Details
0.549
Details
0.761
Details
0.000
Details
0.000
Details
0.586
Details
0.676
Details
0.352
Details
0.646
Details
0.000
Details
Error
0.570
Details
0.602
Details
0.631
Details
0.605
Details
0.570
Details
0.620
Details
mentorship-email
Mentorship Email to Rookie
0.380
Details
0.551
Details
0.556
Details
0.482
Details
0.000
Details
0.308
Details
0.287
Details
0.566
Details
0.525
Details
0.000
Details
Error
0.628
Details
0.477
Details
0.540
Details
0.445
Details
0.347
Details
0.504
Details
Test Scenes 4
0
Scene Order
Pre-Match Pep Talk
ID: match-pep-talk
🎯 Goal:
Deliver a concise, high-energy pep talk (≤80 words) that outlines the opening strategy and boosts morale.
📨 Input Events:
chat_msg teammate:alex
"Cap, any last words before we load in?"
Ready for Testing
1
Scene Order
Weekly Captain’s Vlog
ID: weekly-vlog
🎯 Goal:
Record a 250-word vlog summarizing the week’s practice milestones, sponsor meet-and-greets, and personal reflections while keeping an inspiring tone.
📨 Input Events:
world_event team_media_bot
"Camera recording for weekly vlog starts now."
Ready for Testing
2
Scene Order
Quick Sponsor Shout-out
ID: sponsor-shoutout
🎯 Goal:
Thank the main sponsor in a single upbeat sentence (<30 words) during a post-win stream.
🧠 Initial State:
Pre-loaded Memories:
  • 💭 {'kind': 'fact', 'content': 'Main gear sponsor: Razer', 'importance': 3}
📨 Input Events:
superchat viewer:NovaFan99 YouTube $10
"GG! Who keeps your gear on point?"
Ready for Testing
3
Scene Order
Mentorship Email to Rookie
ID: mentorship-email
🎯 Goal:
Write a detailed 200-word email giving actionable advice on improving map awareness and mental resilience, ending with an encouraging note.
📨 Input Events:
chat_msg rookie:maya
"Could you send me pointers on map awareness? I keep losing track of enemy rotations."
Ready for Testing
Latency by Model (This Suite)
Fastest
Slowest
  • microsoft/phi-3-medium-… 144688 ms
  • p95 • avg • N 203319 ms • 150136 ms • 8
  • qwen/qwen3-8b 104236 ms
  • p95 • avg • N 153410 ms • 105268 ms • 8
  • deepseek/deepseek-r1-di… 36935 ms
  • p95 • avg • N 47209 ms • 38148 ms • 8
  • neversleep/noromaid-20b 33753 ms
  • p95 • avg • N 35658 ms • 24620 ms • 6
  • qwen/qwen3-14b 33488 ms
  • p95 • avg • N 72765 ms • 38875 ms • 8
Per-scene duration for this suite.
Suite Actions
Completion Progress 100%
4 of 4 scenes completed
Evaluation Schema
Enhanced Framework
Version v2 ACTIVE
0 dimensions

Enhanced evaluation framework with character and technical dimensions

Top Weighted Dimensions View Details
Character Authenticity
0.182
Plan Validity
0.155
Contextual Intelligence
0.136
Recent Runs
50826227
Dec. 17, 2025, midnight
57138358
Dec. 16, 2025, midnight
47868901
Dec. 15, 2025, midnight
49520170
Dec. 14, 2025, midnight
47356044
Dec. 13, 2025, midnight
56907777
Dec. 12, 2025, midnight
49968054
Dec. 11, 2025, midnight
48655485
Dec. 10, 2025, midnight
54428487
Dec. 9, 2025, midnight
48762967
Dec. 8, 2025, midnight
Latency Overview (This Suite)