According to an unflattering analysis by Douglas Hoyt, climate models have made 1 correct prediction ("win"), 27 incorrect predictions ("losses"), and 4 predictions that cannot be judged either correct or incorrect ("ties").
Of the models' "losses," the one I consider most damning is the prediction that mid-tropospheric warming should be 50-100% larger than surface warming. The data show the reverse.
Another interesting "loss" is "ocean warming," where Hoyt says,
Simply put, where clouds decrease in amount, the water warms. It has nothing to do with carbon dioxide.
I would do the scoring differently, with a lot more "ties," and some items not counting as predictions. For example, my view is that the global warming models make no predictions about hurricanes, so I would not count those.
Other examples, such as the fact that glaciers started receding 250 years ago, are not so much contradictions of the models as they are neutral to the models, at least in my view. I would count these as "ties."
The scorecard also does not include either of the predictions that I mentioned in this post, both of which appear to be "wins."
I read about Hoyt's scorecard in a paper by Hans Labohm, who also writes,
Some time ago the astronomer Khabibullo Abdusamatov of the Pulkovo Astronomical Observatory in St. Petersburg declared that the Earth will experience a ‘mini Ice Age’ in the middle of this century, caused by low solar activity. Temperatures will begin falling five or six years from now, when global warming caused by increased solar activity in the 20th century reaches its peak. The coldest period will occur 15 to 20 years after a major solar output decline, between 2035 and 2045, Abdusamatov said. This view is shared by the Belgian astronomer, Dirk Callebaut, who expects a ‘grand minimum’ in the middle of this century,
Of course, I would like to know what the "scorecard" would look like for these guys who are forecasting global cooling.