The gambler’s fallacy and the base rate fallacy should not be believed.
Data correlations and anecdotal evidence are useful for generating hypotheses, but correlation does not imply causation, so solid conclusions still require well-designed experiments.
Look for tried-and-true experimental designs that demonstrate statistical significance, such as randomized controlled experiments or A/B testing.
