Talk:Bayesian model comparison

about "true models"
Maybe we should avoid using the phase the "true model"? Because all models are, by definition, abstractions and thus a "true model" is an oxymoron. Josuechan 06:21, 26 June 2007 (UTC)