Sometimes it's better to know you don't know than to think you know. If it's midnight and you're standing at the edge of a cliff you're better off thinking that you have no idea what's ahead of you than if you listen to a guy with "better than average night vision" telling you that there's a sidewalk ahead.

If you are, as you say, "in the market for a guess" that would imply that you are actively looking at different guesses and comparing the credentials of the guessers. Then you would _probably_ be right to choose eg. the guess from the virologist over the proctologist or the mechanic when you need a guess on how the Covid situation will unfold. But the issue might be that an "educated guess" in itself could be worse because you wouldn't base any major decisions on a mechanic that present an "uneducated guess" that "the virus will die down in a month" but if you have a crackpot chiropractor presenting themselves as a "certified doctor with an expertise in everything concerning the body" that says the same thing, you might be worse off in saying that this "educated guess" is worth basing your travel plans on.

The logic parent is mentioning only works for linear, smooth functions. Any non-linearity and discrete behaviour and "slightly better" is just noise.

Being less wrong is probably as good as it ever gets.