If you're talking about climate models, then they fit the observed historical data pretty well. lzw is wrong on that point.

Nope.

I'm saying that the models should fit the past data. I think I misunderstood lzw. :)