Are we? Is there a study that compares a sophisticated climate prediction model with a trivial model (one that simply adds the delta from the previous two years

Are we? Is there a study that compares a sophisticated climate prediction model with a trivial model (one that simply adds the delta from the previous two years, for example) and demonstrates that in comparing each result to the real temperature the former gives a statistically significantly better result than the latter? That's the only way to know if we're any good at predicting climate. If you can point me to a study on this (or some review study) I'd really love to take a look.