[dekhn]

I think the right phrase here is not 'not even wrong' but 'wrong but for the right reasons' or 'all models are wrong, but some models are useful (possibly by being less wrong'