I think this goes down to a very similar thing as the hot hand theory. The problem with that, though, is that it fails to take in to consideration the particular dynamics of a game or situation. Studying the hot hand theory would just aggregate all data to see if there are patterns, but you're not just simply wanting to know if making one shot increases the likelihood of making another, but you want to isolate actual instances where the setting of the game would dictate a particular flow. For instance, there's often a palpable feeling during a game of what may be pivotal events, but watching that devoid of any understanding of the psychology of the given game it may not appear significant at all.
What do you do? Particularly with therapy, it's hard because there's this huge dogmatic push for being "evidenced-based," and we tend to view evidence as some type of approval that bestows some divine power into something. For instance, FDA approval for medications is treated as if something becomes effective when the FDA has an indication for it, and if the FDA doesn't have an indication for it, then it doesn't help (if the tree falls in the forest and nobody heard it, then it never happened). In therapy, I find it funny because a lot of people (particularly younger generations) are on a high horse about being "evidence-based" (as opposed to those "morons" who don't understand evidence-based practice), and thus malign something like psychodynamic psychotherapy because it appears it "doesn't work" because we can't demonstrate good data for it. In reality, it's an issue of measurement. We're trying to fit problems into particular boxes by giving it a label like "depression" and then using a measure of depression to see if there's evidence, whereas many issues in psychodynamic psychotherapy that you're working on may cause symptoms of distress and aren't simply "depression," but are nevertheless important to the individual, so because the "evidence base" doesn't show psychodynamic psychotherapy showing being effective for "treating" what we would call "depression" [or at least not treating the number on a depression rating scale], we then just conclude "it doesn't work!"
It's like trying to measure Rudy's defensive impact with a rule or a wrist-watch.
Bottom line: evidence isn't proof and vice versa.