Comments on: Overconfidence is Stylish

By: Alan Gunn

Alan Gunn — Fri, 19 Sep 2008 18:19:53 +0000

This passage shows about as well as anything why Strunk & White is a bad book. It mixes meaningless advice about style with statements about usage that are flat wrong (like telling people never to use “which” to introduce a restrictive clause). It is popular, though, so White (the book’s real author) may have had a point about how to score points with the public.

By: John Vaughan

John Vaughan — Fri, 19 Sep 2008 17:17:24 +0000

Strunk was talking about writing. I could be wrong, but I don’t think he would like your interpretation. It may have been that his philosophy concerning “resolution” extended to things other than writing, but the way I read it, he was saying that it is important to clearly state your ideas. As a writer this is a very difficult thing to do. It is about learning what is really in your mind and then being able to put that on the page. I think that is the confidence he was looking for. Once one’s ideas are on the table we can discover whether there is anything to be confident about. To be irresolute means you have not even come that far.

By: Lara Foster

Lara Foster — Fri, 19 Sep 2008 16:54:32 +0000

There would seem to be a perponderance of evidence that most people trust the overconfident more than those who use qualification. This may have nothing to do with the current efficacy of actions based on overconfident claims or predictions, though it does indicate that somewhere along human evolutionary history, the tendency to favor and reward the overconfident provided some fitness advantage. This could be causative, i.e. overconfident actions tended to have better outcomes than underconfident ones, or correlative, i.e. extra testosterone lead both to overconfidence and stronger, more fit babies.

From our modern perspective, the relevant questions are:
1) How can we personally avoid biasing towards the arguments of the overconfident, and if we should bias slightly towards overconfidence in general, by how much.
2) How can we prevent others from such bias on a large scale (very difficult I assume).
3) How should we present our own arguments to various audiences as to achieve a desired outcome. ie, With whom should we display overconfidence and with whom due humility.

I really think specific research in this field is needed to actually answer these questions.

By: Jay Levitt

Jay Levitt — Fri, 19 Sep 2008 16:38:08 +0000

I understand the allusion to the financial crisis, but there’s a larger point here:

It’s impossible to be both precise and concise.

Or, more correctly, I should say that I personally find it difficult on many occasions (though certainly not all) to make my overall point – or points, if I have more than one – in a way that acknowledges all the inherent qualifications, shadings and reservations I might place on that concept in my mind, while still maintaining the attention of my listener (or reader) to the extent necessary to fix said point in their memory, let alone to induce agreement.

See what I mean?

If your goal in making a statement is to create for posterity a pure model of reality, then yeah, be irresolute. If your goal is to persuade – or even to inform – you’ll have to make some trade-offs.

By: Peter de Blanc

Peter de Blanc — Fri, 19 Sep 2008 15:52:32 +0000

Thom, William, frelkins:

You are not your best point estimate. Falsifiability is the virtue of scientific theories, but the penalty for being falsified is death. Concentrate all of your probability mass in one theory and you will die with it, for you will have become that theory.

If you possess many theories, then you may prosper while your theories die in droves.

By: Ken Sharpe

Ken Sharpe — Fri, 19 Sep 2008 15:47:23 +0000

Wow, I forgot it’s talk like a pirate day. I keep my head low this day because I really do, in real life, wear an iconic piece of pirate paraphernalia. The present period is predominantly populated by pirates so people presume I’m posing… pesky.

Anyway, it’s very true that people prefer (stop that) confidence in the written word because it’s easy for most people to visualize a clear strong argument — even if it’s actually nebulous and weak — when it’s strongly and decisively stated.

No one wants to read or hear equivocation because it introduces mental overhead by demanding that we integrate the caveats and the new information more tightly into what we already know, instead of accepting the complete, glossy information packet passively.

Most people don’t have those well-developed critical facilities, hence the preference to avoid using them.

See — that whole statement above is at best a hypothesis that needs to be defined clearly to be possibly tested… but it’s coherent and it’s strongly stated, so most people would simply accept it.

There I go again… this must be one of Eli’s stranger meta loops.

By: frelkins

frelkins — Fri, 19 Sep 2008 15:22:52 +0000

Avast, ye mateys, clear thought results in yer clear proposition which be givin’ ya the clear simple statement, says Orwell. Ya ken smell the privateers, scallywags, scurvy dogs & bilge rats by such refusin’ to give a clear, simple, positive statement, yar. Make a definite simple assertion, me hearties, an’ yer signallin’ trustworthiness.

By: William Newman

William Newman — Fri, 19 Sep 2008 15:21:50 +0000

The “falsifiability” thing mentioned by Thom Blake is roughly makes me disagree with you when you say “you couldn’t ask for a clearer demonstration that we prefer overconfident people.”

At least from the quotes you gave, it doesn’t seem clear how much Strunk’s attitude might be like Pauli’s famous “it isn’t even wrong.” To be opposed to statements that aren’t even wrong isn’t necessarily to favor overconfidence. And in a writing manual, it would be pretty reasonable to take an attitude like “I can’t give you universal rules for being correct, you’re on your own there, but I *can* give you universal rules for expressing yourself clearly.” Believing that correctness is out of scope in a writing manual isn’t the same thing as believing that it’s unimportant.

I fully agree that people too often react favorably to overconfidence. But I’m not sure this example is terribly strong.

By: Thom Blake

Thom Blake — Fri, 19 Sep 2008 14:49:01 +0000

Aye, but if ye be able to confidently assert yer proposition, then landlubbers from Popper’s school be givin’ ya higher marks for falisifiability. Yar.

By: Aron

Aron — Fri, 19 Sep 2008 14:27:14 +0000

Confidence does seem stylish as does risk-taking. The banker who says ‘I’m not really sure whether I can truly quantify the risk on these CDO’s, but let’s do it anyway’ would seem a rare combination.