michael_nielsen

Interesting in part because there's many almost-Nobelists - people who reasonably could have been given a Prize, but didn't get it because of politics, bad luck, the rule that only 3 people can be given the Prize, etc.

Evidence that citation counts are a better measure? (That's a very unexpected assertion from my point of view.)

If you look at the list of bad Nobels, think about the censoring from deaths, the many-nominated-but-not-winners etc, a large fractions of Nobels should've been awarded to other people. Plus the sampling error is *massive* due small n: Most 'cells' are just empty. Terrible.

Strange, from my point of view. I looked in depth at the lists recently, and was surprised by how good the discovery / Laureate selection was, especially the recent prizes in Physics (where I can best judge). There are errors, but I thought it was surprisingly good.

You might reconsider if you were reminded of everyone who died and could see the nomination lists. Anyway, any measurement which returns False for 999,999 out of 1 million scientists and a dubious True for the last datapoint, is not a good measurement compared to continuous ones

Well, maybe. But the social context of physics is that people think a _lot_ about this in advance. So, yes, I'm aware of a lot of the people who missed out because they died, and while nominations aren't exactly public, there's a lot of gossip. The committee does a good job.

Certainly, citations simply _aren't_ a good measure. It's pretty standard on hiring committees to have discussions of highly-cited duds, and low-cited gems. Citations are certainly _interesting_ to consider, but high (or low) citation in itself means nothing.