The discussion revolves around the concept of randomness in the placement of points in a plane, specifically addressing the criteria that define a set of points as "random." Participants explore the implications of independence and uniform distribution of random variables, as well as the perception of randomness in various contexts, including sports statistics.
Participants express differing views on the definition of randomness and the implications of clustering in random events. There is no consensus on whether the left or right set of points is more accurately described as random, and the discussion remains unresolved regarding the interpretation of randomness in various contexts.
Participants acknowledge the complexity of defining randomness, particularly in relation to independence and the effects of prior events on subsequent outcomes. The discussion highlights the potential for misinterpretation in statistical definitions and the influence of clustering on perceived randomness.
Dale said:
I disagree. The video is using the term "random" to mean "random and independent of each other". As the video says, those points start out random, but then criteria are applied based on the prior dots to determine if each new dot will be included or not. So the dots are not independent of each other.Spinnor said:
It certainly is not. There is more clustering in random independent events than people expect in general. When they see a random cluster, they very often think that something non-random is going on. For instance, a baseball batter on a hitting streak is often thought to be better in general. But it might just be the clustering of random hits.
I think that it illustrates a good point.
FactChecker said:
Sure. That is why I said that the hot streak might just be a random cluster of hits. But, IMHO, people tend to underestimate the amount of clustering in independent random events. Therefore, they tend to attribute a hot streak to physical and mental ability more often than they should.Spinnor said:
There was an article I read years ago, in the Royal Statistical Society magazine I think, talking about accident blackspots. They were defined as places on the roads where more than five accidents had happened in the preceding year, and are marked as dangerous by road signage. Fine. But they've kept the definition, so any time a place has more than five accidents in a year it gets classified as an accident black spot. The article was pointing out that everywhere will eventually have five accidents in a year just by bad luck. Apparently this kind of clustering-by-chance escaped whoever framed the definition.FactChecker said: