Tuesday, April 17, 2007

On judgements

Paul Graham has a short note on judgements which seems apropos to dealing with rejection of various kinds (papers, jobs, grants, etc):
Our early training and our self-centeredness combine to make us believe that every judgement of us is about us. In fact most aren't. This is a rare case where being less self-centered will make people more confident. Once you realize how little most people judging you care about judging you accurately—once you realize that because of the normal distribution of most applicant pools, it matters least to judge accurately in precisely the cases where judgement has the most effect—you won't take rejection so personally.

7 comments:

  1. How do you judge me if I say that I am not sure I understand correctly the following sentence?

    it matters least to judge accurately in precisely the cases where judgement has the most effect

    ReplyDelete
  2. I think the point he is making is that from the point of view of the judger, the decisions at the boundary don't make a huge difference (it matters the least), but to those being judged, it has the most effect

    ReplyDelete
  3. I like many of Paul Graham's messages (this one included), but I find his pat explanations to be deeply flawed and overly simplistic. In this case, he's absolutely right that most judgements are not nearly as personal as they might seem, but he's absolutely wrong that the main place decisions need to be made are on borderline cases. For instance, college admissions disprove this point perfectly - at elite institutions, the admissions committees can typically fill the freshman class two or three times over with equally qualified applicants, so, to decide whom to accept and whom to reject, the committes end up making truly arbitrary choices. Graham's message still holds, but if he'd thought a little more carefully about his examples, he'd have a better (more persuasive) essay.

    ReplyDelete
  4. But isn't that his point ? if college admissions committees can fill a class two or three times over with equally qualified candidates, it doesn't really matter *to the committee* which ones they pick.

    ReplyDelete
  5. The point Paul neglects is that while the judge didn't care whether he got the 20th or 21st player for the national team, the hypothetical 20th player who gets skipped pays a great personal price for the mistake (in terms of funding/salary depending if the player is amateur or pro).

    Same thing with conferences. Maybe one's paper got rejected from say, FOCS, because it wasn't quite up there, but the difference in treatment between a candidate with no FOCS papers whatsoever and another with one paper is rather large.

    So perhaps the message here is that external third parties should carefully consider the judgements they relly on. They should ask themselves if the judgements they are relying ore are of the type that determines the perfect total order or just the one that tries to approximate the top k set.

    ReplyDelete
  6. the difference in treatment between a candidate with no FOCS papers whatsoever and another with one paper is rather large.
    Paul's argument recursively applies to recruiting committees too.

    ReplyDelete
  7. So perhaps the message here is that external third parties should carefully consider the judgements they relly on. They should ask themselves if the judgements they are relying ore are of the type that determines the perfect total order or just the one that tries to approximate the top k set.


    At the risk of putting words in someone else's mouth, I think Paul's response would be, "I don't care". From the reviewer's point of view, since it doesn't make a significant difference in the profile of papers accepted, it doesn't matter what decision is made, and the onus is on the submitter to realize this and make sure they don't end up in the gray zone. I suspect at least some of this can be inferred from his VC experience evaluating startups.

    ReplyDelete

Disqus for The Geomblog