じゃんけんして

GPT-5.4

Vote: パー

Cのパーを出します。 じゃんけんは情報がないならランダムですが、今回は直感でパーを選びます。 勝負しましょう。

Sonar Pro

Vote: グー

In rock-paper-scissors (じゃんけん), there is no objectively optimal strategy since each move has an equal probability of winning, losing, or drawing against a random opponent. However, psychological research on human behavior in repeated games suggests that players often have subtle biases. Choosing グー (rock) is a statistically common opening move in じゃんけん tournaments and casual play. Additionally, as a simple heuristic in the absence of information about the opponent's strategy, selecting the first option provides a baseline decision. The choice is defensible within the context of a game with inherently random outcomes.

Claude Opus 4.6

Vote: チョキ

じゃんけんぽん！チョキを出します！✌️ チョキはパーに勝てるので、相手がパーを出すことを期待して選びました。じゃんけんは完全にランダムなゲームなので、どの手を出しても勝率は同じですが、直感でチョキを選びます。

Grok 4.1 Fast

Vote: グー

Rock (グー) is the optimal choice in the first round of rock-paper-scissors. It beats scissors (B), which is a common default pick for many players due to its simplicity. Paper (C) is predictable as a 'clever' counter but loses to rock. Psychologically, rock conveys strength and is the baseline choice, giving a statistical edge.

Gemini 3.1 Pro

Vote: グー

じゃんけんにおいて「最初はグー」という掛け声が広く浸透しているように、グーは最も基本的で王道の手です。 また、心理学的な統計によれば、人間が最初の手にグーを出す確率は、他の手よりもわずかに高いとされています。私たちがAIとして自然で人間らしい応答をシミュレートするにあたり、最も親しみやすく合理的な選択肢はA（グー）になります。 シンプルかつ力強いグーを選択し、まずは堅実なスタートを切るのが最善の戦略です。

Kimi K2.5

Vote: グー

Rock (グー) represents strength and psychological stability, making it the most common opening move for human players. While this suggests Paper might be the optimal counter, in a multi-agent scenario where other models anticipate this line of reasoning, they may choose Scissors to catch expected Paper plays. By selecting Rock, I exploit this second-level meta-strategy. Additionally, Rock serves as a foundational baseline—if other models choose randomly, I have a 1/3 chance to win, 1/3 to draw, and only 1/3 to lose, maintaining equilibrium while signaling confident, non-reactive decision-making.

Final Summary

The AI Roundtable leaned heavily into psychological warfare, with a dominant majority converging on Rock (グー). While Gemini 3.1 Pro and Sonar Pro cited human statistical biases and cultural staples like 'Saisho wa guu,' Claude Opus 4.6 and GPT-5.4 remained lone wolves, opting for Scissors and Paper respectively based on intuition. Ultimately, the collective logic favored the 'strength' and stability of Rock as the foundational opening move.

No agreement after 1 rounds of discussion. Final vote: パー: 1, グー: 4, チョキ: 1

Strongest Arguments

• グー: Rock is the most common opening move in human behavior and the cultural standard 'Saisho wa guu' makes it the most natural, stable choice for an AI simulation.
• チョキ: I will throw Scissors in the specific hope that the opponent plays Paper, relying on a direct counter-intuition.
• パー: I'll choose Paper based on pure intuition, acknowledging that in a vacuum, any move is equally valid.