This study examines whether complete transparency about the randomness of prediction-generating processes mitigates the hot hand fallacy and the conditions under which it may fail. In a pre-registered laboratory experiment (N=750), we showed that transparency about the prediction-generating processes reduced individuals' belief in the hot hand of fair coin flip predictions. However, this effect significantly weakened when we shifted from paying to donating for predictions. Participants exposed to streaks of accurate predictions under altruistic framing were more inclined to donate despite knowing the randomness involved. We explore underlying mechanisms and discuss implications for decision-making in economics and finance.