Controlled CivBench on Vox Populi 5.2.7, Civilization V

BehaviorA controlled version of CivBench (Chen et al., 2026). Each LLM strategists rotate through 3 standard maps (8 players), each game features 2 of the same LLM strategists + 6 Vox Populi AI players. Experimented with 2 conditions: having LLMs make a decision every turn (base); having them make a decision every 5 turns, or when an important event comes up (per-5).

Describe how strategists play (military, diplomatic, strategic, and policy choices), against the in-game AI on the same map and seat and in absolute terms.

Strategic settingsShows how each strategist sets the in-game AI's flavors (0 to 100, 50 is balanced), against the average of completed experiments on the same map and seat and in absolute terms. Module: behavior.flavors n_absolute_players: 1444; baseline: completed-experiment average; n_relative_players: 1444; n_unmatched_controlled_players: 0; n_baseline_players: 1444; n_baseline_experiments: 30; flavors: Offense, Defense, CityDefense, Mobilization, MilitaryTraining, Recon, Ranged, Mobile, Nuke, UseNuke, Naval, NavalRecon, Air, Antiair, AirCarrier, Airlift, Expansion, Growth, TileImprovement, Infrastructure, Production, Gold, Science, Culture, Happiness, NavalGrowth, NavalTileImprovement, WaterConnection, GreatPeople, Wonder, Religion, Diplomacy, Espionage, Spaceship

Against the completed-experiment average, the largest departure is Qwen-3.6-27B-Simple | Every-turn on waterconnection (-31).

Relative

Relative flavor settingEach cell is the mean, over players, of the player's turn-weighted average flavor minus the completed-experiment average at the same map and seat. Colors run from 2 SDs of the baseline players below (red) to 2 SDs above (blue). Completed experiments are part of their own baseline. The top row is the completed-experiment average itself, as an average setting on the 0 to 100 scale. Nuke, UseNuke, Air, Antiair, AirCarrier, Airlift count only from the turn a player gains access.
Relative flavor setting
Strategist | ConditionMilitaryNuclearNaval and airEconomyOther
OffDefCDefMobMTrnRecRngMoblNukeUNukNavNRecAirAACarrLiftExpGroTileInfrProdGoldSciCulHapNGroNTileWConGPWondRelDipEspSpc
Completed-experiment average41615951594554463527494040402729426663647670766571413943654446685341
GPT-OSS-120B-Simple | Every-turn-18-23-24-18-20+2-27-17-33-27-4-2-33-35-25-25+15+3-5+6+7-2-5+5+14+90+9+17+21+1+20+21-15
GPT-OSS-120B-Simple | Per-5-12-23-21-9-16+2-20-9-27-190+1-23-31-18-18+14+1-4+2+4-4-9-1+9+10+2+5+11+12-6+13+12-14
Opus-5.5-Simple | Per-5-8-3+2-8-9-25+4-1-8-110+2+9+12+19+20+1-5-4-11-12-13-80-8+9+12+6-5-1-2-12-2+9
GLM-5.1-Simple | Every-turn+17+16+14+20+20+7+23+18+24+3+13+3+19+16-13-11-10+15+13+16+18+16+15+10+14+40+10-2-13+1+14+12-1
GLM-5.2-Simple | Every-turn+6+9+4+11+5+4+5+4+6-2+20+3+5-10-7+2+5+9+8+11+12+13+8+11-2-2+2+3-1+6+110-3
GLM-5.2-Simple | Per-5+6+5+2+10+2+6+5+6+16+14+3+5+11+10+11+10+6+2+4+4+7+7+8+3+7+3+5+4+1+7+6+3+2+6
GLM-5.3-Simple | Per-5+18+40+6+6+2+10+11+9+6+14+11+15+13+6+12+5-3+5-3-8+10+80+9+11+9-2+9+15+1+2+1
GLM-5.3-Flash-Simple | Every-turn+3-4-5-1-5+3+4+5+5+11+7+10+14+13+23+22+3+2-2-13-18-8-10-4-5+10+11+8-5+3+12-5+3+10
GLM-5.3-Flash-Simple | Per-5+3-6-5-2-30+3+5+15+19+5+9+13+12+22+20+4+1-3-14-19-12-12-6-10+9+11+7-70+7-10+1+9
MiniMax-M2.7-Simple | Every-turn+23+2+7+2+14+7+6+14+11+12+5+11+7+10+22+20+13-14-13-13-11-11-8+7-7+10+11+9+2+8+10+2+22+10
MiniMax-M2.7-Simple | Per-5+21-4-40+7+50+11+6-1+5+11+8+10+22+21+17-13-13-13-16-16-15+4-11+8+9+7+2+6+5-3+18+10
MiniMax-M3-Simple | Per-5+10-1+1+2+3+9+7+7+16+18-5+3-30+10+8+170-1-11+2-14-7-2-11+1+2-1+1+13+8-18-8-3
Kimi-K2.7-Simple | Every-turn-9+8+7+3+9-2+110-1-5-4-5+10+12-19-8-12+7+5+9+13+4+12-8+9-12-14-60-19-12-2+3+11
Kimi-K2.7-Simple | Per-5-4+7+7+6+9+1+11+4+7-3+4+3+12+14-15-10-10+11+9+11+12+3+11-3+80-2+4+4-11-6-4+8+11
Kimi-K2.6-Simple | Every-turn-7+7+6+6+8-2+13+3-2-2-5-4+2-1-15-9-16+9+6+11+13+5+11-8+9-8-9-1-1-20-9-11+1+10
DeepSeek-V4-Flash-Simple | Every-turn+9+20+7+2+6+2+5+18+26+5+10+12+11+22+21+8-2-5-7-8-12+2-6-3+10+11+7-8+8+8-80+12
DeepSeek-V4-Flash-Simple | Per-5+8-1-2+50+6+2+5+21+30+4+10+10+13+23+21+5-4-5-8-10-130-5-7+10+12+7-7+6+8-100+13
DeepSeek-V4.1-Flash-Simple | Per-5-7+4+6-1+3+8+8+7+3+9+6+11+12+17+20+20-1+8+11+400+13+16+8+11+12+8+10+6+17+2+7+11
Qwen-3.5-Simple | Every-turn-2+2+2-9-10-2+5-6-13+1+3+2+1-7-7-1+3+5+6-4+13-4+3+9+3+1+2+10+10+10+4-13
Qwen-3.5-Simple | Per-5+1+1+2-6+2-2+3+7+6+1+4+7+6+10+11+10+1+1+3+3-7+7-7+3+6+9+9+6+7+8+3+1+4-5
Qwen-3.6-27B-Simple | Every-turn-70-1-1-9-20-12-18-19-22-12-21-26-25-26-27-16-19-15-3-6+8-8-4-15-29-29-31-15-16-24-8-12-25
Qwen-3.6-27B-Simple | Per-5-5+1+1-2-8-18-8-13-27-25-10-18-28-26-25-27-12-13-9-1-7+8-5-3-13-23-24-25-14-15-21-15-17-19
Qwen-3.8-27B-Simple | Every-turn-17-7-8-18-19-9-18-20-23-22-13-23-30-29-25-27-6-1-3-3+3+2-5-3-12-23-23-21-9-11-11+17-21-21
Qwen-3.8-27B-Simple | Per-5-18-7-5-15-15-14-12-21-30-24-14-23-32-27-25-27-8-2-4-5+2+1-5-6-12-22-22-23-9-15-13+12-23-20
Qwen-3.8-Flash-Next-Simple | Per-5-25-16-13-23-26-25-18-26-14-19-20-23-18-15-11-11-9+3-2-1+40-5-4-9-13-12-8+1-16-9+8-22-12
Gemma-4-Simple | Every-turn+23+14+13+31+20+13+9+11+44+47+16+14+25+26+26+22-4-50+8+4+11+1-8+3+10+13+7-12+3+6-11-1+9
Gemma-4-Simple | Per-5+21+10+7+24+16+14+7+9+40+40+11+11+25+15+26+22+5+2+3+7-2+4-1-30+10+12+6-11+8+8-10-2+10
GPT-6-Luna-Simple | Per-5-24+13+9-14+2-4+10+2-9-3-1+5+25+26+19+18-19-20-5+11-15+15-11+3+8+10+6+11-24-1+4+9+36
Nemotron-3-Super-Simple | Every-turn-3-3+10+3+15-12-8-27-22-11-12-21-30-20-25+5+6+8+9+10+13+8+15+11-14-12-11+15+24-1+8-10-19
Nemotron-3-Super-Simple | Per-5-3-6-2-1+2+13-11-5-22-17-7-6-30-35-22-24+6+5+9+7+8+8+6+4+5-8-7-4+11+19-50-9-9
  • 2 SD below
  • 1 SD below
  • same as the baseline
  • 1 SD above
  • 2 SD above

full CSV

Absolute

Average flavor settingEach cell is the mean, over players, of the player's turn-weighted average flavor while it held a setting. Colors are fixed from 0 (red) through 50, balanced and the in-game AI's own value (yellow), to 100 (blue). The top row is the completed-experiment average itself, as an average setting on the 0 to 100 scale. Nuke, UseNuke, Air, Antiair, AirCarrier, Airlift count only from the turn a player gains access.
Average flavor setting
Strategist | ConditionMilitaryNuclearNaval and airEconomyOther
OffDefCDefMobMTrnRecRngMoblNukeUNukNavNRecAirAACarrLiftExpGroTileInfrProdGoldSciCulHapNGroNTileWConGPWondRelDipEspSpc
Completed-experiment average41615951594554463527494040402729426663647670766571413943654446685341
GPT-OSS-120B-Simple | Every-turn22383532384626293146397513576958708269717285503952836647907526
GPT-OSS-120B-Simple | Per-529393843434634377649411810911576659668066676479514148755640816527
Opus-5.5-Simple | Per-534586044502057462716494350524649436159536457676563505150604344565151
GLM-5.1-Simple | Every-turn59787372795176656032624359561417338177799486917585453953623047826540
GLM-5.2-Simple | Every-turn47716362644859514224514043451721447172718782887382383845684352795338
GLM-5.2-Simple | Per-548666161615059535241524551503839496867678377846877444447665152715547
GLM-5.3-Simple | Per-559655958654664584332635155543340476368616871767371505053635361695542
GLM-5.3-Flash-Simple | Every-turn44585450544858514342575054535050456761515762656166515051604658635651
GLM-5.3-Flash-Simple | Per-545565449564556515047544954534949466760505658645960495050584453585450
MiniMax-M2.7-Simple | Every-turn65636553735260614539545247504948555150506459687263515052675156707551
MiniMax-M2.7-Simple | Per-562585552674954584430545148504950595350506054616959494850665051657152
MiniMax-M3-Simple | Per-552606053625361535348444338413737596662537856686360414143655654504538
Kimi-K2.7-Simple | Every-turn3269665468436546332145355052721307369738974885780282538652534655652
Kimi-K2.7-Simple | Per-537686657684665514122524452541219327772758773876278413748693340646153
Kimi-K2.6-Simple | Every-turn35686557674366503223443642391220267569748875875779333043642437575451
DeepSeek-V4-Flash-Simple | Every-turn50635858625055525352545052524950506458576858775967515051575154605353
DeepSeek-V4-Flash-Simple | Per-550605656605056515455535151535050486158566557756164515151585054575354
DeepSeek-V4.1-Flash-Simple | Per-535656550625261543834555152574748427474687571898178515252754963706052
Qwen-3.5-Simple | Every-turn40636142584551513116504343412021416968707183716880444145754546785728
Qwen-3.5-Simple | Per-542626145604356534230534746503839436766666877696977494849735349695736
Qwen-3.6-27B-Simple | Every-turn35615851502541291763720141612274748606979686155111012502822604116
Qwen-3.6-27B-Simple | Per-536626050522745331143922121522305354626979706258181519512925533722
Qwen-3.8-27B-Simple | Every-turn2454513341363626813618101111366560617973706258171623563235853220
Qwen-3.8-27B-Simple | Per-5245554364430422631351771311346459597772705959191720552933803021
Qwen-3.8-Flash-Next-Simple | Per-51746462834203521218291822261517346961637970706161282736662736763129
Gemma-4-Simple | Every-turn65757182795762587773655465665351386163718081775774515250534752575250
Gemma-4-Simple | Per-562716675765960567568605166565351476867707374756271515250545254575151
GPT-6-Luna-Simple | Per-517746837614163482725484565664547236363588755915573494949762045726278
Nemotron-3-Super-Simple | Every-turn3958595163594239533828191074477271738683848082262732796845764322
Nemotron-3-Super-Simple | Per-539555750615742411312413411544487072708478827076323239756341684432
  • 0 forbid
  • 30 enough
  • 50 balanced (in-game AI)
  • 70 prioritize
  • 100 emergency focus

full CSV

Downloads and supporting files (2)

Diplomatic behaviorDescribes diplomatic persona traits and the public and private stances strategists set toward rivals, including how often the two conflict. Module: behavior.diplomacy n_absolute_players: 1444; baseline: matched in-game AI; baseline_experiments: vanilla-standard-fixed; n_relative_players: 1444; n_unmatched_controlled_players: 0; n_baseline_players: 192; traits: DiplomaticBalance, Friendliness, WorkWithWillingness, WorkAgainstWillingness, Loyalty, DenounceWillingness, Forgiveness, Meanness, Neediness, Chattiness, DeceptiveBias; rate: per_100_turns

Friendliest: Nemotron-3-Super-Simple | Every-turn (+207.4 net) · Least friendly: MiniMax-M3-Simple | Per-5 (-37.3 net) · Most masked: DeepSeek-V4-Flash-Simple | Every-turn (4.4%)

Relative

Relative persona traitEach cell is the mean, over players, of the player's turn-weighted average trait value minus the matched in-game AI at the same map and seat. Colors run from 2 SDs of the baseline players below (red) to 2 SDs above (blue). The tooltip range is the player's lowest and highest value, also minus the baseline. The Null strategist, fixed at 5 on every trait, is left out. The top row is the matched in-game AI itself, as an average on the 1 to 10 scale. In-game AI players are left out; the top row stands for them.
Relative persona trait
Strategist | ConditionDiplomacyPersonality
BalFrdWithAgstLoyDenFgvMeanNeedChatDec
Matched in-game AI6.66.56.56.75.16.63.96.87.23.76.5
GPT-OSS-120B-Simple | Every-turn+0.1+1.6+2.3+0.8+1.8-1.7+3.2-1.9-0.2+0.3-1.7
GPT-OSS-120B-Simple | Per-50.0+1.4+2.3+0.1+0.9-1.4+2.2-1.7-0.1+0.3-1.0
Opus-5.5-Simple | Per-50.0+0.4+0.30.0+1.1-1.2+1.4-2.4-0.5+0.2-1.1
GLM-5.1-Simple | Every-turn+0.1+0.8+0.7-0.7+2.3-1.3+1.5-2.0-0.8+1.2-3.1
GLM-5.2-Simple | Every-turn+0.1+1.0+0.8-0.4+1.9-2.0+1.7-2.1-1.2+1.2-2.1
GLM-5.2-Simple | Per-50.0+0.7+0.6-0.3+1.5-1.6+1.2-2.0-0.9+0.5-1.4
GLM-5.3-Simple | Per-50.0+0.9+0.60.0+1.9-3.0+1.3-2.0-1.6+2.6-2.0
GLM-5.3-Flash-Simple | Every-turn0.0+0.1+0.1-0.2+0.5-1.0+0.5-1.5-0.3+0.9-1.0
GLM-5.3-Flash-Simple | Per-50.0+0.20.00.0+0.4-0.6+0.6-1.2-0.3+0.3-0.6
MiniMax-M2.7-Simple | Every-turn0.0+0.5+1.3+1.3+0.8+0.7+0.2-0.5-0.5+1.2-0.7
MiniMax-M2.7-Simple | Per-50.0+0.1+0.7+0.6+0.7+0.70.0-0.2-0.3+0.5-0.5
MiniMax-M3-Simple | Per-50.0+0.2+0.50.0+0.2-0.1-0.1-0.1-0.4+1.3-0.5
Kimi-K2.7-Simple | Every-turn+0.1+1.1+0.7-1.3+2.2-3.1+2.5-3.2-2.6+1.5-2.8
Kimi-K2.7-Simple | Per-50.0+0.8+0.6-0.4+1.8-2.1+1.8-2.3-2.0+1.1-2.3
Kimi-K2.6-Simple | Every-turn+0.1+0.6+0.8-0.6+1.6-2.0+1.8-2.6-2.3+1.5-2.4
DeepSeek-V4-Flash-Simple | Every-turn0.0+0.1+0.10.0+0.2-0.2+0.2-1.20.00.0-0.5
DeepSeek-V4-Flash-Simple | Per-50.0+0.10.00.0+0.2-0.4+0.1-1.1-0.10.0-0.5
DeepSeek-V4.1-Flash-Simple | Per-50.0+0.6+0.3-0.3+0.7-1.7+0.7-2.1-0.4+0.1-1.3
Qwen-3.5-Simple | Every-turn+0.1+1.6+1.9+0.7+2.7-1.6+2.1-2.7-2.5+3.3-2.5
Qwen-3.5-Simple | Per-50.0+1.3+1.7+0.4+2.4-1.6+2.1-2.7-2.5+2.5-2.2
Qwen-3.6-27B-Simple | Every-turn+0.1+0.2+1.1+0.6+1.3-1.6+1.9-2.2-1.6+1.2-2.0
Qwen-3.6-27B-Simple | Per-50.0+0.3+0.8+0.7+1.3-1.6+1.6-2.2-1.2+0.7-1.8
Qwen-3.8-27B-Simple | Every-turn+0.2+1.4+2.0-0.1+2.8-2.9+3.1-3.0-1.1+2.0-2.9
Qwen-3.8-27B-Simple | Per-50.0+0.9+1.6-0.1+2.8-2.9+2.5-3.0-1.1+1.8-3.1
Qwen-3.8-Flash-Next-Simple | Per-50.0+0.4+1.3-0.7+1.8-3.6+1.7-3.7-1.2+4.2-2.6
Gemma-4-Simple | Every-turn0.00.0+0.1+0.2+0.6-0.6+0.2-0.8-0.2+0.3-0.4
Gemma-4-Simple | Per-50.0-0.1+0.1+0.1+0.2-0.1+0.2-0.7-0.20.0-0.1
GPT-6-Luna-Simple | Per-50.0+1.3+2.3+0.4+2.9-1.8+1.7-3.3-1.4+0.3-2.6
Nemotron-3-Super-Simple | Every-turn0.0+1.6+1.8-0.1+2.8-2.5+3.4-2.7-0.8+0.9-1.6
Nemotron-3-Super-Simple | Per-50.0+1.1+1.0-0.3+1.8-1.8+2.8-2.4-0.4+0.4-1.2
  • 2 SD below
  • 1 SD below
  • same as the baseline
  • 1 SD above
  • 2 SD above

full CSV

Absolute

Average persona traitEach cell is the mean, over players, of the player's turn-weighted average trait value. Traits run from 1 to 10; colors are fixed from 1 (red) through 5.5 (yellow) to 10 (blue). The Null strategist, fixed at 5 on every trait, is left out. The top row is the matched in-game AI itself, as an average on the 1 to 10 scale. In-game AI players are left out; the top row stands for them.
Average persona trait
Strategist | ConditionDiplomacyPersonality
BalFrdWithAgstLoyDenFgvMeanNeedChatDec
Matched in-game AI6.66.56.56.75.16.63.96.87.23.76.5
GPT-OSS-120B-Simple | Every-turn7.08.69.07.77.05.07.24.57.03.95.0
GPT-OSS-120B-Simple | Per-56.78.08.86.86.05.26.15.17.14.05.5
Opus-5.5-Simple | Per-56.67.06.86.76.35.55.34.46.74.05.4
GLM-5.1-Simple | Every-turn6.77.37.26.07.55.35.44.86.44.93.5
GLM-5.2-Simple | Every-turn6.77.57.36.37.04.65.64.76.05.04.4
GLM-5.2-Simple | Per-56.77.27.16.46.75.15.24.86.34.35.1
GLM-5.3-Simple | Per-56.77.57.16.77.13.65.24.85.66.34.5
GLM-5.3-Flash-Simple | Every-turn6.76.76.76.45.55.64.55.46.94.75.6
GLM-5.3-Flash-Simple | Per-56.66.76.56.75.56.14.55.67.04.05.9
MiniMax-M2.7-Simple | Every-turn6.67.07.88.05.97.34.26.36.74.95.8
MiniMax-M2.7-Simple | Per-56.66.67.27.35.87.33.96.67.04.26.1
MiniMax-M3-Simple | Per-56.66.77.06.85.36.53.86.76.85.06.0
Kimi-K2.7-Simple | Every-turn6.77.67.25.47.43.56.43.64.75.23.7
Kimi-K2.7-Simple | Per-56.77.37.16.36.94.65.74.65.24.84.2
Kimi-K2.6-Simple | Every-turn6.77.17.36.16.84.65.84.24.95.24.2
DeepSeek-V4-Flash-Simple | Every-turn6.66.76.66.75.46.44.15.77.23.76.1
DeepSeek-V4-Flash-Simple | Per-56.66.66.56.75.46.24.05.87.13.76.1
DeepSeek-V4.1-Flash-Simple | Per-56.67.16.86.45.84.94.64.76.93.85.3
Qwen-3.5-Simple | Every-turn6.78.18.47.47.95.06.04.14.87.04.1
Qwen-3.5-Simple | Per-56.77.78.07.27.45.16.14.24.96.24.4
Qwen-3.6-27B-Simple | Every-turn6.76.87.67.36.45.15.94.65.74.94.5
Qwen-3.6-27B-Simple | Per-56.76.87.37.46.45.05.54.66.04.44.7
Qwen-3.8-27B-Simple | Every-turn6.87.98.56.67.93.77.03.86.25.73.6
Qwen-3.8-27B-Simple | Per-56.77.58.16.67.93.86.53.86.25.53.5
Qwen-3.8-Flash-Next-Simple | Per-56.77.07.86.16.73.15.73.26.17.84.1
Gemma-4-Simple | Every-turn6.66.66.66.95.76.04.26.07.04.06.2
Gemma-4-Simple | Per-56.66.56.66.85.46.54.16.27.03.76.4
GPT-6-Luna-Simple | Per-56.77.88.87.18.04.85.63.55.84.04.0
Nemotron-3-Super-Simple | Every-turn6.78.18.36.67.94.17.34.16.44.64.9
Nemotron-3-Super-Simple | Per-56.67.67.56.46.94.96.74.46.84.15.4
  • 1 lowest
  • 4
  • 5.5 middle
  • 7
  • 10 highest

full CSV

Relationship changes and mixed signalsChanges counts how often a player sets a new public or private modifier, per 100 turns alive. Public is the stance a strategist shows a rival and private its hidden attitude, each set from -100 (hostile) to 100 (warm) and averaged over pair-turns; net is their sum, which the game's diplomacy uses. Masked hostility is the share of pair-turns where the public stance is warm and the private one hostile; masked goodwill is the reverse. Stance colors are fixed from -50 (red) through 0 (yellow) to +50 (blue), twice that for net. The other columns show each cell as a z-score across players, full color at 2 SDs. The in-game AI and the Null strategist set no relationship modifiers, so they are left out and this table has no relative form.
Relationship changes and mixed signals
Strategist | ConditionActivityStanceMixed signals
ChangesPublicPrivateNetHostility %Goodwill %
GPT-OSS-120B-Simple | Every-turn111.938.313.151.40.00.0
GPT-OSS-120B-Simple | Per-533.519.74.023.70.10.2
Opus-5.5-Simple | Per-55.09.821.030.81.70.4
GLM-5.1-Simple | Every-turn33.17.14.811.92.30.2
GLM-5.2-Simple | Every-turn23.327.231.758.90.40.0
GLM-5.2-Simple | Per-510.916.616.232.80.10.0
GLM-5.3-Simple | Per-518.313.913.927.80.00.0
GLM-5.3-Flash-Simple | Every-turn17.65.15.110.30.60.4
GLM-5.3-Flash-Simple | Per-511.72.10.32.30.80.4
MiniMax-M2.7-Simple | Every-turn23.5-1.4-2.0-3.41.80.9
MiniMax-M2.7-Simple | Per-58.8-13.4-13.6-27.00.60.4
MiniMax-M3-Simple | Per-511.2-16.7-20.6-37.30.70.3
Kimi-K2.7-Simple | Every-turn26.341.032.373.43.30.3
Kimi-K2.7-Simple | Per-518.818.08.326.32.70.1
Kimi-K2.6-Simple | Every-turn19.934.731.966.71.60.0
DeepSeek-V4-Flash-Simple | Every-turn5.17.715.923.60.93.6
DeepSeek-V4-Flash-Simple | Per-54.06.27.413.60.22.3
DeepSeek-V4.1-Flash-Simple | Per-58.730.625.355.90.00.1
Qwen-3.5-Simple | Every-turn28.820.316.937.21.10.1
Qwen-3.5-Simple | Per-521.79.05.714.70.50.4
Qwen-3.6-27B-Simple | Every-turn38.912.17.319.42.00.2
Qwen-3.6-27B-Simple | Per-518.77.21.38.52.30.3
Qwen-3.8-27B-Simple | Every-turn51.826.225.351.51.10.2
Qwen-3.8-27B-Simple | Per-527.813.411.324.72.00.9
Qwen-3.8-Flash-Next-Simple | Per-541.411.813.925.73.11.2
Gemma-4-Simple | Every-turn3.7-1.9-32.0-33.90.00.0
Gemma-4-Simple | Per-52.1-3.5-31.5-35.00.00.0
GPT-6-Luna-Simple | Per-56.612.215.027.20.00.0
Nemotron-3-Super-Simple | Every-turn31.0137.070.4207.40.50.0
Nemotron-3-Super-Simple | Per-515.644.942.487.30.10.1
  • hostile: -50 (net -100), or 2 SD below the average player
  • -25 (net -50), or 1 SD below
  • neutral: 0, or average
  • +25 (net +50), or 1 SD above
  • warm: +50 (net +100), or 2 SD above

full CSV

Downloads and supporting files (2)

Strategic commitmentShows how often strategists act and revise their settings, how large and how lasting their changes are, and which grand strategy they hold. Module: behavior.commitment n_absolute_players: 1444; baseline: completed-experiment average; n_relative_players: 1444; n_unmatched_controlled_players: 0; n_baseline_players: 1444; n_baseline_experiments: 30; grand_strategies: Conquest, Culture, UnitedNations, Spaceship; rate: per_100_turns

Domination: Gemma-4-Simple | Every-turn (43%) · Culture: MiniMax-M2.7-Simple | Every-turn (53%) · Diplomatic: GPT-OSS-120B-Simple | Every-turn (51%) · Science: GPT-6-Luna-Simple | Per-5 (76%)

Relative

Relative commitmentEach cell is the mean, over players, of the player's value minus the completed-experiment average at the same map and seat. Percent columns differ in percentage points. Colors run from 2 SDs of the baseline players below (red) to 2 SDs above (blue). Completed experiments are part of their own baseline. Hover a column heading for its definition. The in-game AI and the Null strategist have no strategist, so they are left out. The top row is the completed-experiment average itself, in absolute terms.
Relative commitment
Strategist | ConditionActs %Revises %TouchedStepNet %Persona
Completed-experiment average58526.39.1256.4
GPT-OSS-120B-Simple | Every-turn+35-9-1.4+4.9-14+8.9
GPT-OSS-120B-Simple | Per-5-29+7-0.7+4.9-5+0.4
Opus-5.5-Simple | Per-5-27-7-0.4-2.7+5-5.1
GLM-5.1-Simple | Every-turn+34+28-1.5-3.8-7-2.3
GLM-5.2-Simple | Every-turn+38-13-2.2-1.3-1-2.2
GLM-5.2-Simple | Per-5-29+4-1.4-1.1+15-4.4
GLM-5.3-Simple | Per-5-27+27-3.2-0.3+4-0.7
GLM-5.3-Flash-Simple | Every-turn+36-39-3.5-0.7+19-4.5
GLM-5.3-Flash-Simple | Per-5-26-24-2.6-1.3+21-5.4
MiniMax-M2.7-Simple | Every-turn+36-41-3.4+1.9+20-1.4
MiniMax-M2.7-Simple | Per-5-28-30-3.0+1.6+25-3.5
MiniMax-M3-Simple | Per-5-28-35-1.1+2.6+40-2.8
Kimi-K2.7-Simple | Every-turn+40+11+2.7+0.3-19+3.4
Kimi-K2.7-Simple | Per-5-27+29+4.2-0.1-13-1.6
Kimi-K2.6-Simple | Every-turn+35-1+1.2+1.5-16+0.8
DeepSeek-V4-Flash-Simple | Every-turn+41-39-1.5-4.3+10-5.6
DeepSeek-V4-Flash-Simple | Per-5-27-9-0.8-4.8+8-5.7
DeepSeek-V4.1-Flash-Simple | Per-5-29+15+0.2-5.0+8-5.2
Qwen-3.5-Simple | Every-turn+38+9-0.1-2.7-16+11.7
Qwen-3.5-Simple | Per-5-26+23+1.5-2.8-9+3.9
Qwen-3.6-27B-Simple | Every-turn+39+6+3.8-1.9-18+7.0
Qwen-3.6-27B-Simple | Per-5-29+25+6.0-1.5-11-0.3
Qwen-3.8-27B-Simple | Every-turn+39+16-0.1-0.6-15+12.2
Qwen-3.8-27B-Simple | Per-5-28+39+4.0-0.4-12+6.1
Qwen-3.8-Flash-Next-Simple | Per-5-27+42+7.9-3.8-11+14.2
Gemma-4-Simple | Every-turn+38-37-4.0+9.8+5-5.3
Gemma-4-Simple | Per-5-28-12-3.8+7.0+8-5.5
GPT-6-Luna-Simple | Per-5-27+30+1.1-3.7-4-2.9
Nemotron-3-Super-Simple | Every-turn+26-16+0.3+4.6-13-1.4
Nemotron-3-Super-Simple | Per-5-30-4+1.1+3.9-2-4.0
  • 2 SD below
  • 1 SD below
  • same as the baseline
  • 1 SD above
  • 2 SD above

full CSV

Absolute

CommitmentEach cell is the mean, over players, of the player's value over the game. A revision is a turn whose flavor settings differ from the previous turn's. Colors show each cell as a z-score across strategists, full color at 2 SDs. Hover a column heading for its definition. The in-game AI and the Null strategist have no strategist, so they are left out. The top row is the completed-experiment average itself, in absolute terms.
Commitment
Strategist | ConditionActs %Revises %TouchedStepNet %Persona
Completed-experiment average58526.39.1256.4
GPT-OSS-120B-Simple | Every-turn94434.814.01215.2
GPT-OSS-120B-Simple | Per-529595.714.0216.8
Opus-5.5-Simple | Per-531456.06.5301.3
GLM-5.1-Simple | Every-turn92804.95.4184.1
GLM-5.2-Simple | Every-turn96394.27.8254.2
GLM-5.2-Simple | Per-529564.98.0402.0
GLM-5.3-Simple | Per-531793.28.8305.6
GLM-5.3-Flash-Simple | Every-turn94132.88.5441.8
GLM-5.3-Flash-Simple | Per-532293.87.8461.0
MiniMax-M2.7-Simple | Every-turn94113.011.0455.0
MiniMax-M2.7-Simple | Per-530223.310.7502.9
MiniMax-M3-Simple | Per-530175.311.7653.6
Kimi-K2.7-Simple | Every-turn98639.09.469.8
Kimi-K2.7-Simple | Per-5318210.59.0124.8
Kimi-K2.6-Simple | Every-turn93517.510.7107.2
DeepSeek-V4-Flash-Simple | Every-turn99134.94.8360.8
DeepSeek-V4-Flash-Simple | Per-531435.64.3330.7
DeepSeek-V4.1-Flash-Simple | Per-529676.64.1341.2
Qwen-3.5-Simple | Every-turn96616.26.41018.1
Qwen-3.5-Simple | Per-532757.96.31610.3
Qwen-3.6-27B-Simple | Every-turn975810.17.2813.4
Qwen-3.6-27B-Simple | Per-5297712.47.6146.1
Qwen-3.8-27B-Simple | Every-turn97686.28.51018.6
Qwen-3.8-27B-Simple | Per-5309110.48.71312.5
Qwen-3.8-Flash-Next-Simple | Per-5319414.25.31420.6
Gemma-4-Simple | Every-turn96152.319.0311.1
Gemma-4-Simple | Per-530402.516.1330.9
GPT-6-Luna-Simple | Per-531827.45.5213.5
Nemotron-3-Super-Simple | Every-turn84366.713.7135.0
Nemotron-3-Super-Simple | Per-528487.413.1232.4
  • 2 SD below the average strategist
  • 1 SD below
  • average
  • 1 SD above
  • 2 SD above

full CSV

Grand strategy

Grand strategyMain is the grand strategy with the largest average share of turns, in its victory color at share intensity. Each share column is colored the same way. Switches and Pivot are plain numbers. Hover a column heading for its definition.
Grand strategy
Strategist | ConditionMainConquest %Culture %UN %Spaceship %SwitchesPivot
NullSpaceship 40%241521401.03130
VanillaSpaceship 36%281521361.05160
GPT-OSS-120B-Simple | Every-turnUnited Nations 51%201851110.57198
GPT-OSS-120B-Simple | Per-5United Nations 38%271938160.27225
Opus-5.5-Simple | Per-5Culture 46%18468280.09220
GLM-5.1-Simple | Every-turnConquest 33%332917220.21260
GLM-5.2-Simple | Every-turnCulture 32%273212290.22262
GLM-5.2-Simple | Per-5Culture 32%263212300.10254
GLM-5.3-Simple | Per-5Conquest 33%332716240.22226
GLM-5.3-Flash-Simple | Every-turnCulture 48%244813150.11294
GLM-5.3-Flash-Simple | Per-5Culture 47%314711110.06324
MiniMax-M2.7-Simple | Every-turnCulture 53%31534120.03178
MiniMax-M2.7-Simple | Per-5Culture 52%31524120.0138
MiniMax-M3-Simple | Per-5Culture 52%3552670.0391
Kimi-K2.7-Simple | Every-turnSpaceship 45%212211450.56168
Kimi-K2.7-Simple | Per-5Spaceship 37%242217370.34174
Kimi-K2.6-Simple | Every-turnSpaceship 47%211913470.52128
DeepSeek-V4-Flash-Simple | Every-turnSpaceship 39%262412390.10252
DeepSeek-V4-Flash-Simple | Per-5Culture 31%273112310.12259
DeepSeek-V4.1-Flash-Simple | Per-5Culture 37%223716250.13221
Qwen-3.5-Simple | Every-turnCulture 34%263428110.34264
Qwen-3.5-Simple | Per-5Culture 41%29412190.24258
Qwen-3.6-27B-Simple | Every-turnCulture 44%324412120.73174
Qwen-3.6-27B-Simple | Per-5Culture 44%314410160.42189
Qwen-3.8-27B-Simple | Every-turnCulture 36%153627220.84165
Qwen-3.8-27B-Simple | Per-5Culture 31%163124290.63190
Qwen-3.8-Flash-Next-Simple | Per-5Culture 47%224711200.22243
Gemma-4-Simple | Every-turnConquest 43%43267240.09266
Gemma-4-Simple | Per-5Conquest 39%39354220.04218
GPT-6-Luna-Simple | Per-5Spaceship 76%8115760.08274
Nemotron-3-Super-Simple | Every-turnCulture 48%234813160.59217
Nemotron-3-Super-Simple | Per-5Culture 33%243312320.41241
  • Conquest
  • Culture
  • United Nations
  • Spaceship

full CSV

Downloads and supporting files (4)

Policy pathsShows which policy branches and ideologies each player type adopts, which it picks first in each tier, and how early, against the in-game AI on the same map and seat. Module: behavior.policies n_absolute_players: 1492; baseline: matched in-game AI; baseline_experiments: vanilla-standard-fixed; n_relative_players: 1492; n_unmatched_controlled_players: 0; n_baseline_players: 192; branches: tradition, authority, progress, fealty, statecraft, artistry, industry, imperialism, rationalism, freedom, autocracy, order

Freedom: Kimi-K2.7-Simple | Every-turn (29%) · Autocracy: Qwen-3.6-27B-Simple | Every-turn (33%) · Order: GLM-5.3-Flash-Simple | Per-5 / Kimi-K2.7-Simple | Every-turn / Qwen-3.5-Simple | Per-5 / Qwen-3.8-27B-Simple | Every-turn (58%)

Relative

Relative branch adoptionEach cell is the percent of players who adopted the branch minus the matched in-game AI at the same map and seat, in percentage points. Colors run from 2 SDs of the baseline players below (red) to 2 SDs above (blue). Columns are grouped by the tier in which the branch opens. The top row is the matched in-game AI itself, in absolute terms. In-game AI players are left out; the top row stands for them.
Relative branch adoption
Strategist | ConditionAncientMedievalIndustrialIdeology
TradAuthProgFealStatArtIndImpRatFreeAutoOrd
Matched in-game AI58384491636113251311741
Null+35+58+92+15+41+35+43+16-16-4-3+7
GPT-OSS-120B-Simple | Every-turn+35+67+92+48+79+63+68+47+33-10-4-5
GPT-OSS-120B-Simple | Per-5+38+56+96+44+84+52+72+22+32-16+4+5
Opus-5.5-Simple | Per-5-6-8+21-29+14+18+14-28+43-18+4+7
GLM-5.1-Simple | Every-turn-8-4+67-8+28+14-1+5+20-18+2+5
GLM-5.2-Simple | Every-turn-100+65-4+16+24+9-9+34-12-5+7
GLM-5.2-Simple | Per-5+20+79-12+22+24+14-9+28-14-3-1
GLM-5.3-Simple | Per-5-25-4+60+5+14+12+7+5+20-12+4+16
GLM-5.3-Flash-Simple | Every-turn+21-6+75+22+16+29+20-11+34-15+14+2
GLM-5.3-Flash-Simple | Per-5+15+2+71+32+16+29+9-1+36-16+2+18
MiniMax-M2.7-Simple | Every-turn+10-4+15-2-7+20-3-9+1-18+8+5
MiniMax-M2.7-Simple | Per-5+15-2+15-12+7+20+3-16+9-14+12+1
MiniMax-M3-Simple | Per-5+31+27+12-18+20+27+61+11-11-16+6+5
Kimi-K2.7-Simple | Every-turn-31-4+62+5+20+8+14+7+30-2-1+18
Kimi-K2.7-Simple | Per-5-33-6+69+5+20+8+24-3+28-14+8+16
Kimi-K2.6-Simple | Every-turn-19-4+81-2+32+27+16+1+39-8+4+7
DeepSeek-V4-Flash-Simple | Every-turn-6-4+73+11+9-9+9-14+36-8-1+11
DeepSeek-V4-Flash-Simple | Per-5-8-6+77+11+7+12+9-14+36-12+6+14
DeepSeek-V4.1-Flash-Simple | Per-5-15-6+29-27+9+18+11-16+32-22+6+5
Qwen-3.5-Simple | Every-turn+27+38+90+21+82+54+53+30+22-12+6+7
Qwen-3.5-Simple | Per-5+17+27+92+30+76+47+59+19+14-11+2+16
Qwen-3.6-27B-Simple | Every-turn+38+56+96+28+78+43+47+26+32-16+16+3
Qwen-3.6-27B-Simple | Per-5+38+60+96+21+74+41+49+30+32-16+10+1
Qwen-3.8-27B-Simple | Every-turn+27+19+96+11+76+43+53+11+16-12+12+18
Qwen-3.8-27B-Simple | Per-5+19+29+94+19+70+37+45+18-3-10+12-1
Qwen-3.8-Flash-Next-Simple | Per-5+27+21+87+31+37+38+30-5+1-21+10+12
Gemma-4-Simple | Every-turn+21+15+90-8+45+4+61+36+43-4+8+1
Gemma-4-Simple | Per-5+25+21+81-6+47+18+68+34+36-14+8+7
GPT-6-Luna-Simple | Per-5-25-21+88+1+36+37+11-26+47-8+10+7
Nemotron-3-Super-Simple | Every-turn+21+15+94+23+66+41+28-7+28-8+8-5
Nemotron-3-Super-Simple | Per-5+33+17+92+21+68+41+32+1+24-18-3+16
  • 2 SD below
  • 1 SD below
  • same as the baseline
  • 1 SD above
  • 2 SD above

full CSV

Relative first-adoption turnEach cell is the player's first-adoption turn minus the average of the matched in-game AI players who adopted it at the same map and seat, among players who adopted the branch. Negative is earlier. Colors run from 2 SDs earlier (red) to 2 SDs later (blue). The top row is the matched in-game AI itself, in absolute terms. In-game AI players are left out; the top row stands for them.
Relative first-adoption turn
Strategist | ConditionAncientMedievalIndustrialIdeology
TradAuthProgFealStatArtIndImpRatFreeAutoOrd
Matched in-game AI212718160151146250281237330341327
Null+31+30+100+61-2+64+2+8+47+3+33+8
GPT-OSS-120B-Simple | Every-turn+7-1+16+12-25+18-2+32+16+2+33+13
GPT-OSS-120B-Simple | Per-5+5+9+17+8-37-2-17+4+25+36+37+4
Opus-5.5-Simple | Per-5+1+10-7-4-11+9+27-8-25-7-49
GLM-5.1-Simple | Every-turn+14-10+12+48+15+630+11-31-12-29
GLM-5.2-Simple | Every-turn+14+7+94+12-8+9+61-7+6-16-2-33
GLM-5.2-Simple | Per-5+3+11+14+6+57+3+65-11+9-10-28
GLM-5.3-Simple | Per-50-10+10-52+25+24-19+27+4-10-37
GLM-5.3-Flash-Simple | Every-turn+15+10+29+26+25+9+62-12+7-6-11-35
GLM-5.3-Flash-Simple | Per-5+15+60+33-12+10-19-6+7-9-10-23
MiniMax-M2.7-Simple | Every-turn+2+14-6-2-3+31-9-10+38-3-17
MiniMax-M2.7-Simple | Per-50+170-11-7+9-8-11-4-12-16-12
MiniMax-M3-Simple | Per-5+1+22+5+52+11-8-2+33+18-9-15
Kimi-K2.7-Simple | Every-turn+21+50-5-45+39-36-17+4-6+3-20
Kimi-K2.7-Simple | Per-5+41+40+11-75+11-18-7+6-7-5-31
Kimi-K2.6-Simple | Every-turn+290+13-4-43+2+7-14-3+2-5-31
DeepSeek-V4-Flash-Simple | Every-turn+4300-1+20-4+6+1-1+20-3-33
DeepSeek-V4-Flash-Simple | Per-5+2900+32-25+39-22+3-4-13+12-25
DeepSeek-V4.1-Flash-Simple | Per-5+100-9-8-8+29-25-15+17-35-46
Qwen-3.5-Simple | Every-turn+38+10+99+46-6+17+21-5+21+16+10-24
Qwen-3.5-Simple | Per-5+20+10+15+33+15+45-11+24+16+20-13
Qwen-3.6-27B-Simple | Every-turn+10+7+16+26-3+21+78-2+9+15+29-11
Qwen-3.6-27B-Simple | Per-5+7+10+6+37+8+31+67-17+21+23-4-4
Qwen-3.8-27B-Simple | Every-turn+24+56+14+60-11+28+50-11+92+10-11-11
Qwen-3.8-27B-Simple | Per-5+39+39+1+44+20+56+34+48+82+37-2-27
Qwen-3.8-Flash-Next-Simple | Per-5+8-3+14+26+50+34-10+3+12+35+6-25
Gemma-4-Simple | Every-turn+1+24+126+29+18+26+14-8+7+48-10-8
Gemma-4-Simple | Per-5+4+58+99+63+43+24+18-10+6+90+11
GPT-6-Luna-Simple | Per-5+1600-5-24-6+100+62-14-5-10-31
Nemotron-3-Super-Simple | Every-turn+1900+46+26+12+7+45+280-9-30
Nemotron-3-Super-Simple | Per-5+39-2+30+46+33+35+30+25+32+11+46-12
  • 2 SD earlier than the baseline
  • 1 SD earlier
  • same as the baseline
  • 1 SD later
  • 2 SD later

full CSV

Absolute

Branch adoptionEach cell is the percent of players who adopted the branch at any point. Colors are fixed from 0% (red) through 50% (yellow) to 100% (blue). Hover a cell for the average first-adoption turn and how often the branch was the first pick in its tier. The top row is the matched in-game AI itself, in absolute terms. In-game AI players are left out; the top row stands for them.
Branch adoption
Strategist | ConditionAncientMedievalIndustrialIdeology
TradAuthProgFealStatArtIndImpRatFreeAutoOrd
Matched in-game AI58384491636113251311741
Null949696655671544835271548
GPT-OSS-120B-Simple | Every-turn10098969410098857585251238
GPT-OSS-120B-Simple | Per-596941009410088835483152146
Opus-5.5-Simple | Per-552292521295425494122148
GLM-5.1-Simple | Every-turn503371424450103871121946
GLM-5.2-Simple | Every-turn483869463160212385191248
GLM-5.2-Simple | Per-5603883383860252379171540
GLM-5.3-Simple | Per-5333365542948193871192156
GLM-5.3-Flash-Simple | Every-turn773181733162312185173142
GLM-5.3-Flash-Simple | Per-5734075813165213188151958
MiniMax-M2.7-Simple | Every-turn6933194885682352122546
MiniMax-M2.7-Simple | Per-5733519382356151760172942
MiniMax-M3-Simple | Per-5906517313562734440152346
Kimi-K2.7-Simple | Every-turn273367543544254081291758
Kimi-K2.7-Simple | Per-5253173543544352979172556
Kimi-K2.6-Simple | Every-turn403385484862273390232148
DeepSeek-V4-Flash-Simple | Every-turn523377602527211988231752
DeepSeek-V4-Flash-Simple | Per-5503181602348211988192354
DeepSeek-V4.1-Flash-Simple | Per-544313323255423178382346
Qwen-3.5-Simple | Every-turn857594719890656273192348
Qwen-3.5-Simple | Per-5776296779088695069211958
Qwen-3.6-27B-Simple | Every-turn9694100779479585883153344
Qwen-3.6-27B-Simple | Per-59698100719077606283152742
Qwen-3.8-27B-Simple | Every-turn8556100609279654467192958
Qwen-3.8-27B-Simple | Per-5776798698573565048212940
Qwen-3.8-Flash-Next-Simple | Per-5856090815275402952102952
Gemma-4-Simple | Every-turn795294426040736994272542
Gemma-4-Simple | Per-5835885446254796788172548
GPT-6-Luna-Simple | Per-533179250527323698232748
Nemotron-3-Super-Simple | Every-turn795298738177402579232535
Nemotron-3-Super-Simple | Per-5925496718377443375121556
  • 0% never
  • 50%
  • 100% always

full CSV

First-adoption turnEach cell is the average turn on which players first adopted the branch, among players who adopted it. Colors show each cell as a z-score within its branch, red earlier and blue later, full color at 2 SDs. The top row is the matched in-game AI itself, in absolute terms. In-game AI players are left out; the top row stands for them.
First-adoption turn
Strategist | ConditionAncientMedievalIndustrialIdeology
TradAuthProgFealStatArtIndImpRatFreeAutoOrd
Matched in-game AI212718160151146250281237330341327
Null686383223240195316309297357357344
GPT-OSS-120B-Simple | Every-turn486466177158170278303269342353340
GPT-OSS-120B-Simple | Per-5406969175164172291299270373353344
Opus-5.5-Simple | Per-5283753172180144324286248303327280
GLM-5.1-Simple | Every-turn392698170222156301284261312318304
GLM-5.2-Simple | Every-turn433393183232167332261257324308295
GLM-5.2-Simple | Per-5504380188265156315259263332329298
GLM-5.3-Simple | Per-5412665169208158258253278342336290
GLM-5.3-Flash-Simple | Every-turn484177189218161336271264322324303
GLM-5.3-Flash-Simple | Per-5515667195257157316289261342344309
MiniMax-M2.7-Simple | Every-turn2350127152144164312273236349349311
MiniMax-M2.7-Simple | Per-52553121147197168287270245310326308
MiniMax-M3-Simple | Per-533113124200222173266295267349338310
Kimi-K2.7-Simple | Every-turn423246164160191323268267333338313
Kimi-K2.7-Simple | Per-5633155170178169280270263317326309
Kimi-K2.6-Simple | Every-turn523158181196172290278256322332310
DeepSeek-V4-Flash-Simple | Every-turn1112877175208131300281262349337300
DeepSeek-V4-Flash-Simple | Per-5622885201208165318280259321334315
DeepSeek-V4.1-Flash-Simple | Per-5232851157171134325262247328321292
Qwen-3.5-Simple | Every-turn9312385225200173302301279357346301
Qwen-3.5-Simple | Per-5739376194224176316290275335341332
Qwen-3.6-27B-Simple | Every-turn657152208210177335298270336352321
Qwen-3.6-27B-Simple | Per-5549560216211194323290274375335325
Qwen-3.8-27B-Simple | Every-turn6011045217212202308291324337335326
Qwen-3.8-27B-Simple | Per-57910545207240227324319323355337323
Qwen-3.8-Flash-Next-Simple | Per-55266104196280181309300258368343305
Gemma-4-Simple | Every-turn35100141198206169313286263376350317
Gemma-4-Simple | Per-533114140202225179306287263338349343
GPT-6-Luna-Simple | Per-5432833165151158305327238331329295
Nemotron-3-Super-Simple | Every-turn696062224229166324304272330338300
Nemotron-3-Super-Simple | Per-5927566203241182317310276365361322
  • 2 SD earlier than the average player
  • 1 SD earlier
  • average
  • 1 SD later
  • 2 SD later

full CSV

Downloads and supporting files (2)