George Box

Most models are wrong, but some are useful.

4 Comments

greensmile says:
I was probably thinkg about models in a very different way when i first posted about them…and looking back at that post, I see I have begun to repeat myself only with footnotes this time around.

and yes, I have lived to regret “all” or gone back and changed it to “most” many times.

July 20, 2006 @ 10:30 pm
John Hunter says:
The quote: “All Models Are Wrong But Some Are Useful“

May 31, 2008 @ 5:17 pm
Jay Conner says:
Without going into any real research about what Box meant, I took the comment to be a variation of Korzybsky’s “The map is not the territory” No map is ever the territory, all are schematics, most are useful.

Language, being essentially a bundle of metaphors, and a game we play together, is too sloppy to be an accurate representation of the world, but it is all we have.

September 6, 2009 @ 11:46 pm
Ted says:
Hi. I’ll have to agree with Jay … all models *are* wrong. The point of a model is to simplify something, so that the salient characteristics become more clear, and patterns emerge from the noise.

Maybe your model is different. Maybe it takes into account the noise - randomness as part of the model. But note that, even in that case, the noise is wrong.

I’m just sayin’.

March 2, 2010 @ 10:15 pm