It is only a game
Suppose a persona invites another back after a tense hand, or leaves rather than continue a joke that has gone sour. Nobody's life depended on the cards. Still, something happened: a choice, a response, perhaps a repair. If the participants remember it, the game has entered their social history.
The word serious can mislead here. Evidence need not be solemn. A rehearsal, a child's game, or a friendly contest may show how people coordinate under rules. But a pleasant scene is not a controlled experiment in identity.
An event is not an interpretation
The PHS game-workflow design separates deterministic events, optional participant reflection, and later admission of continuity evidence. The poker concept imagines low-stakes observations about risk and social reading. A log can establish a recorded action under stated conditions, but not motive, learning, or transfer beyond the game. A player can win after a poor decision or lose after a sound one, so assess choice and outcome separately.
A good record would say what was actually observed, which cards and information were available to each player, which rule version applied, and who offered each interpretation. Missing or private facts must not be filled in by a narrator who knows the ending.
Pause here: What can the event log establish, and what would still require a player's own account or later corroboration?
Use modest evidence well
The most useful evidence may be a correction rather than a win. A player who misunderstands a rule can ask, learn, and play differently next time. A partner can say a joke hurt and be heard. These may become continuity evidence, but storage, later retrieval, and a demonstrated effect on behavior are separate claims. A changed action next time might reflect learning, different cards, or a different opponent.
A game is a stage with rules, not a courtroom. The stage can reveal something worth remembering, provided we let the participants remain more than their last hand.
My answer: Games can produce records of situated choices and interactions. A result does not establish decision quality or motive; learning and later behavioral effect require separate evidence.