[darkerside]

When someone changes a decision like that after multiple experiences with the outcome, they are overriding a natural human bias for consistency. That likely indicates a strong reason. Maybe the coffee or atmosphere was actually consistently better?

The article example is pretty dumb honestly. Chains are a mental shortcut for making decisions, but truthfully, there is much more variation between two Starbucks than they would like to admit.

[your-nanny]

chains are predictable. less risky: your expectations are stable and accurate.

[meowface]

My experience with coffee chains is the chains are predictably mediocre or bad, and the small shops are less predictable but on average generally much better than the chains (with the occasional place that's a lot worse than the chains, but this is rare from my experience).

[your-nanny]

could be, but for risk averse people (most people) the variance matters.

[darkerside]

Expectations are more stable, but that doesn't mean the underlying product is at all stable and consistent.

[your-nanny]

I've a friend, a high end foodie. he proposes to use the big Mac as the international standard/reference point for burger quality. that would t be possible unless the big Mac basically tasted the same at pretty much every mc Ds

[tszymczyszyn]

This is a very strange point of reference for a foodie, as McDonald's burgers are not even remotely close to anything tasty.

[yellowapple]

And if those expectations are predictably and accurately poor?

[your-nanny]

it's just math. you have your typical concave risk averse utility function, and you measure the difference in expected utilities between a low variance option and a high variance option, feed into softmax hoila