5 ms·
A decision is either easy, or it doesn't matter
by niccl 6y ago
A decision is either easy, or it doesn't matter
- falcor84 6y agoI'm surprised by this one. What about things like marriage or a career change?
- lethologica 6y agoI kind of take the quote to mean “if it isn’t a hell yes! Then it’s a no” sort of a thing. If I don’t say “hell yes!” To the prospect of marrying my girlfriend, then perhaps proposing to her now is the right move.
- rocqua 6y agoIf you are doubting whether to get married, then apparently, both options are roughly equal. Hence the difference between either decision is small, not in outcome but in expected 'desirability'. Hence the decision 'doesn't matter much' insofar as there are no real wrong choices, just a slightly more right one.
- falcor84 6y ago> ... apparently, both options are roughly equal. But that's the thing, they aren't really roughly equal in any meaningful way. I think Anscombe's quartet [0] is a good illustration of the issue, only in this case there are very many dimensions involved, with huge uncertainty in each, leading to vastly different risk profiles for the different alternatives. > ... no real wrong choices ... This is a matter of philosophy of course, but more objectively, there are many many people actively regretting their choices. [0] https://en.wikipedia.org/wiki/Anscombe%27s_quartet https://en.wikipedia.org/wiki/Anscombe%27s_quartet
- rocqua 6y ago> This is a matter of philosophy of course, but more objectively, there are many many people actively regretting their choices. I think this is a part of the power of the quote. It essentially says, don't judge your decisions with hindsight. And remember that having made the other choice could equally well have turned out bad as far as you knew back then.
- dragonwriter 6y agoThat's a saying that I find incredibly wrong viewed in crisp binary logic but better through the lens of fuzzy logic (that is, on its face it's a giant case of the fallacy of the excluded middle, but if you phrase it as “a decision’s fuzzy membership in the set of easy decisions is the fuzzy negation of it's fuzzy membership in the set of consequential decisions” it's not as flagrantly wrong (it still ignores curable informatiom problems, though.)