There are “reliabilist” accounts of what makes a credence justified. There are different accounts, but they say (very roughly) that a credence is justified if it is produced by a process that is close to the truth on average. See (this paper)[https://philpapers.org/rec/PETWIJ-2].
Frequentist statistics can be seen as a version of reliabilism. Criteria like the Brier score for evaluating forecasters can also be understood in a reliabilist framework.
There are “reliabilist” accounts of what makes a credence justified. There are different accounts, but they say (very roughly) that a credence is justified if it is produced by a process that is close to the truth on average. See (this paper)[https://philpapers.org/rec/PETWIJ-2].
Frequentist statistics can be seen as a version of reliabilism. Criteria like the Brier score for evaluating forecasters can also be understood in a reliabilist framework.