The trouble with regret minimization

Once you do accept it, regret doesn't make a lot of sense: you did this instead of that, but so what? There's an infinity of other things you didn't do and never could have done. Instead, focus on the life you have and living it as best as you possibly can. It doesn't last, and the real waste would be to spend your time regretting the past or using every moment to prepare for a future which might not come and won't be what you expected if it does (real life never matches the expectations we create in our fantasy version of the future).

Obviously there are some things that you should never try. Things that you find morally objectionable. Things that have too high a risk of death or permanent disability. Things that will keep you from fulfilling your responsibilities (wife and kids will slow you down). You have to do your own maths on those, but beyond that you should try as many new things as you can manage.

You will fail. You will fail a lot. That's part of the fun.

As a 50 year old man, with four grown kids, in the middle of a divorce after 25 years of marriage, sitting at home with a leg I broke skateboarding last month I can honestly say that I have very few regrets so far. I travelled through a lot of Europe and most of the US before marriage, I've had interesting work, and I've learned how to do a lot of things just by being willing to try.

Remember, failure is always an option and is the expected outcome at least 50% of the time. If you're not failing, then you're probably not trying anything new.

I do like his idea of maximizing gratitude though. What will I be most grateful for on my deathbed? Probably the relationships I had, and a chance to see the world as it is now, in the golden age of our civilization, before it gets wrecked by climate change.

- https://www.theguardian.com/lifeandstyle/2012/feb/01/top-fiv...

The author seems quite focused on his career, but a common regret is to wish "I didn't work so hard"

It needs to be calibrated with the best estimates of one's future regrets. E.g. when I am 40, I will regret not having children. Therefore I should get my life in order during my late twenties and early thirties to avoid this future regret.

First, Bezos' is about a once-in-a-lifetime chance, whereas the mathematical version is more about choices you repeat over and over.

Another difference, relating to the author's point

> Regret is conditioned on what happens

is that, in the mathematical version, while you do update your regret tables on the basis of what eventually happens, you make your decisions now on the basis only of what you know so far. Depending how you look at it, you might be satisfied -- "no regrets" -- so long as you perform this procedure properly, i.e., you can forgive your past self for ignorance. In this way you are attached not to the actual outcome but to your correct choice and execution. This creates a kind of "equanimity". The idea may be very vaguely Jain/Buddhist.

I find that comforting and a great help in the face of option paralysis.

That's not regret minimization, really it's the opposite. Regret minimization does not compare two options, it is a threshold that any option either passes or does not. The whole point is that in general there are many possibile courses of action, and it is often impossible to determine the optimal route, but so long as you pick one of the options you won't regret, you're good, even if it was suboptimal. In Bezos' quote, he's not saying he knew all along that the internet would lead to amazing success and the best of all possible futures, but rather that it was good enough that no matter what happened, he'd be okay with it. Regret minimization is a method to beat analysis paralysis, not produce it.

So in the author's example of deciding whether or not to get a PhD, there isn't a right answer, but someone doing regret minimization might say "I'd be proud of getting my PhD even if it doesn't ultimately put me in a better position than spending that time working" and thus getting the PhD would be the regret-minimizing course of action. Or equally possible, one might say "I'd rather try and fail at doing something in the real world than spend all those extra years in academia" and thus not going for the PhD would be the regret minimizing strategy.

Now it's true that it can be difficult to predict what you'll regret and no one can know all the consequences of their decisions, but that's just life. Regret is not an unhealthy and unnatural state that you should convince yourself to ignore, regret is the means by which we recognize and learn from bad decisions. You will invariably make decisions over the course of your life which will put your future self in undesirable circumstances, but that you may sometimes fail to make the right decision doesn't mean you should never try. To pretend that all choices are equally valid and to tell ourselves that we'd probably regret the alternatives just as much is to deny that we ever made a mistake in the first place, which may feel good in the moment but leads to much worse future pain as no lesson is learned from the experience and the mistake is repeated. Or worse, without proper introspection it is possible to take away the wrong lessons from a mistake - legitimate concerns become irrational fears, caution becomes anxiety, problems are avoided instead of solved, and we endure terrible pain making a life in which we can pretend to be happy.

Regrets are like scars - you shouldn't be trying to get them, but if in your old age you find yourself with a lot, you must have lived a pretty interesting life.