I think that feeling good is mostly about not feeling bad.

I think that, in general, there are more ways to feel bad than to feel good, and so switching is often best done by identifying a way to feel good and moving towards it. There are certainly deviations from feeling good which demand special attention, like psychological illnesses with clearly physiological origins, where this plan won't apply.

The cognitive mechanisms that healthy, satisfied people use are not magic, are often transferable, and are continually useful. What I mean by that is it looks like many successful treatments for, say, depression are more like brushing your teeth than drilling cavities- even people with healthy teeth brush their teeth. Similarly, even secure people deal with insecurities and failures; what makes them "secure people" is that they cope effectively with those challenges, not that they don't have them.

I think a large part of figuring the expected value is quick comparisons of the current situation to past situations, especially to see whether it resembles any past situations that resulted in painful outcome. Implicit emotional memory isn't very analytical.

Agreed. The helpfulness of this model is that it suggests that you 1) get more past situations that are similar to the current situation and positive and 2) focus on the expectations explicitly. (One of the treatments for anxiety is basically contingency planning, which helps because it both actually reduces the chance of negative outcomes and can sate the fear of negative outcomes by taking them seriously and working against them.)