Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We study a class of deceptively similar games, which however have different player sets and backward induction (BI) predictions that vary with their cardinality. The game-theoretic principles involved are compelling as predictions rely on weaker and less controversial epistemic foundations than needed to justify BI more generally. Are the BI predictions empirically relevant for this class of games? We design and report results from a relevant experiment.