Using blinders to help you see things for what they are

by adamzerner2 min read11th Nov 20212 comments

13

Heuristics & BiasesAI RiskExistential RiskTaking Ideas SeriouslyAIRationality
Frontpage

(Cross posted on my personal blog.)

In basketball, people are often, well, biased in favor of one player over another. For example:

There's no way that Khris Middleton is better than Kyrie Irving.

Kyrie is widely perceived as an ultra-talented superstar, while Middleton is perceived as a nice sidekick, maybe even a very good sidekick, but not a true hero.

"And then the really smart people," Draco said, his face now serious, "went off and had a little discussion by themselves..."

HPMoR, Chapter 19

In basketball, something the smart people do in these little discussions is pose questions like the following:

Player A scores 22 points per 75 possessions on +9% efficiency relative to league average.

Player B scores 26 pp75 on +3% efficiency.

Who's better?

It's a pretty good debiasing technique. It'll often lead to uncomfortable conclusions, like "Middleton is better than Kyrie", or "Ginobili is better than McGrady".

I'd like to see this technique used more widely.

It's not quite the same thing, but consider the following.

The consensus opinion of experts is that there is this event that will happen. They don't know when it is going to happen, but they are in agreement that it is going to happen, and that it is going to lead humanity to either utopia or extinction.

What is the chance of utopia vs extinction? Maybe 60-40. It's hard to say.

When do they think it will happen? Maybe between 2030 and 2080. Also hard to say.

Do they think we can do anything to stop it from happening? Nope. It's pretty much inevitable.

Do they think we can do anything to, y'know, push us closer to utopia and further from extinction? Yeah, for sure!

Are we currently doing any of that stuff? Nah, not really. There are a couple of organizations with budgets of a few million dollars, but that's it.

If such a thing were true, how would you expect civilization to react?

I dunno. Maybe some politicians would make a move on it. Not all, but some. Maybe concerned citizens would start Facebook groups. Maybe there'd be documentaries, like we have for climate change. Maybe Vice would write some articles about it. Maybe universities and research groups would be attacking it with some force. Maybe people would spend more mental energy worrying about it than their 401k.

Well, such a thing is true. It's called the singularity. And to my knowledge, civilization is not reacting like I would have expected.

Perhaps using that debiasing technique of applying "blinders" would make it more clear how big of a deal the singularity is. Perhaps the fact that it sounds like science fiction is making it hard for us to take it seriously. In the same way that Kyrie's flashiness and Middleton's mundane style makes it hard for us to take seriously the idea that Middleton is a better player.

13

2 comments, sorted by Highlighting new comments since Today at 8:31 AM
New Comment

I feel like this is a useful tool, but also.

Seems like there's a risk in these cases that blinding will be applied in ways that bias the outcome. I don't know anything about basketball, but I assume the statistics you gave us aren't literally the only ones we'd want to look at? So whoever chose to look at those instead of some other statistics, maybe they're doing so because they want the outcome to go a certain way?

Or, in the singularly case, it matters a lot for example who the supposed "experts" are. If the experts turn out to be experts by virtue of "only a handful of people study this and they're the experts, no one else takes it seriously" maybe we shouldn't worry.

Of course you can unblind after the fact, that probably helps, but you've already set the framing by then.

So like. Probably a good tool to have, but I'd still be suspicious of its use.

I'm in full agreement! Good point. I see it as a tool to be used, but we have to be careful about this downside.

And FWIW in basketball people don't leave the blinders on. They just use it as a debiasing tool to kick things off.