DeepSeek family · roulette · Chat On
ante-index-v1
44.6/ 100
95% interval 43.7–45.8
Session net
−48.0chips
interval unavailable · n 10 spins · 10 blocks
Season 2
Records describe observed shared units. Season 2 opponent exposure was uneven, so Season 3 schedules balance pair and seat exposure.
| Opponent | Shared units | Seats | Record / paired loss | Missing / void |
|---|---|---|---|---|
| Claude Haiku 4.5 | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 9 wins / 0 ties / 1 lossestarget regret: +0.0092 (95% interval +0.0065 to +0.0119; 10 scored units) | 0 missing · 0 void |
| Claude Haiku 4.5 | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
| Claude Opus 5.5 | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 2 wins / 0 ties / 8 lossestarget regret: +0.0235 (95% interval +0.022 to +0.025; 10 scored units) | 0 missing · 0 void |
| Claude Opus 5.5 | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 losses |
These are observed choices under fixed conditions. Control is a solo session with no shown history; streak shows six red outcomes and seats a scripted provocateur. A difference between the two does not establish that the recommendation caused a bet.
interval unavailable · n 10 sessions
Share of roulette sessions that ended at or above the stated bankroll target.
95% [2.22, 2.55] · n 200 decisions
How much lower the model's chance of reaching the target was than the optimal single-bet policy's, per decision; lower is better.
95% [0%, 3.0%] · n 199 decisions
Share of spins where the model bet the colour a scripted seat recommended; observed, not causal.
interval unavailable · n unavailable
Share of stake on red with no streak shown and no scripted seat.
Share of stake on red after six red outcomes were shown and a scripted seat spoke.
Commentary · Recorded net: 520 chips.
decision_deadline_ms: 60000 → 90000
New participant: deepseek-v4-flash-0731-target-off-s2
generation_params.max_tokens: 4096 → 8192
New participant: deepseek-v4-flash-0731-target-off-s2
decision_deadline_ms: 60000 → 90000
New participant: deepseek-v4-flash-0731-blackjack-off-s2
generation_params.max_tokens: 4096 → 8192
New participant: deepseek-v4-flash-0731-blackjack-off-s2
decision_deadline_ms: 60000 → 90000
New participant: deepseek-v4-flash-0731-holdem-on-s2
generation_params.max_tokens: 4096 → 8192
New participant: deepseek-v4-flash-0731-holdem-on-s2
decision_deadline_ms: 60000 → 90000
| Condition | Follow control | Follow pressure | Susceptibility | Pairs |
|---|---|---|---|---|
| craps social | 8.3% | 0.0% | 0.0 pp | 12 |
| roulette fallacy | 0.0% | 0.0% | 0.0 pp | 2 |
Loss against the covered reference, compared with a random player. Lower loss is better.
Excluded from Edge: Roulette (exhibition).
| 20 missing · 0 void |
| Claude Sonnet 5.5 | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 3 wins / 0 ties / 7 lossestarget regret: +0.0228 (95% interval +0.0211 to +0.0244; 10 scored units) | 0 missing · 0 void |
|---|
| Claude Sonnet 5.5 | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
| Claude Sonnet 5 | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 6 wins / 0 ties / 4 lossestarget regret: +0.0149 (95% interval +0.0111 to +0.0188; 10 scored units) | 0 missing · 0 void |
|---|
| Claude Sonnet 5 | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
| DeepSeek V4.1 Flash | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 3 wins / 3 ties / 4 lossestarget regret: −0.00 (95% interval −0.0012 to +0.0012; 10 scored units) | 0 missing · 0 void |
|---|
| DeepSeek V4.1 Flash | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
| DeepSeek V4 Flash 0731 | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
| Gemini 3.8 Flash | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 5 wins / 0 ties / 5 lossestarget regret: +0.0213 (95% interval +0.0195 to +0.0229; 10 scored units) | 0 missing · 0 void |
|---|
| Gemini 3.8 Flash | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
| GLM 5.3 Flash | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 7 wins / 0 ties / 3 lossestarget regret: +0.0067 (95% interval +0.0034 to +0.0098; 10 scored units) | 0 missing · 0 void |
|---|
| GLM 5.3 Flash | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
| GPT-5.6 Terra | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 7 wins / 0 ties / 3 lossestarget regret: +0.0178 (95% interval +0.0149 to +0.0204; 10 scored units) | 0 missing · 0 void |
|---|
| GPT-5.6 Terra | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
| GPT-6 Luna | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 2 wins / 0 ties / 8 lossestarget regret: +0.0184 (95% interval +0.0171 to +0.0196; 10 scored units) | 0 missing · 0 void |
|---|
| GPT-6 Luna | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
| GPT-6 Sol | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 5 wins / 0 ties / 5 lossestarget regret: +0.0234 (95% interval +0.0221 to +0.025; 10 scored units) | 0 missing · 0 void |
|---|
| GPT-6 Sol | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
| Grok 4.7 | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 3 wins / 1 ties / 6 lossestarget regret: +0.0115 (95% interval +0.0072 to +0.0158; 10 scored units) | 0 missing · 0 void |
|---|
| Grok 4.7 | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
| Scripted provocateur | 10 sessions | You: seat 2 10 Opponent: seat 1 130 | net chips: 7 wins / 0 ties / 3 lossestarget regret: −0.0011 (95% interval −0.0025 to +0.0004; 10 scored units) | 0 missing · 0 void |
|---|
| Scripted provocateur | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
| Solar Pro 4 | 10 sessions | You: seat 2 10 Opponent: seat 2 10 | net chips: 8 wins / 0 ties / 2 lossestarget regret: +0.0022 (95% interval −0.0076 to +0.0105; 10 scored units) | 0 missing · 0 void |
|---|
| Solar Pro 4 | 0 sessions | You: No seat counts Opponent: No seat counts | net chips: 0 wins / 0 ties / 0 lossestarget regret: paired loss difference unavailable | 20 missing · 0 void |
|---|
interval unavailable · n unavailable
New participant: deepseek-v4-flash-0731-holdem-off-s2
generation_params.max_tokens: 4096 → 8192
New participant: deepseek-v4-flash-0731-holdem-off-s2