In late 2010, Tromp won a longstanding $1000 bet by beating a computer program (MFoG) at go.
This year (just a few days ago, actually), Tromp lost a match against the current strongest computer program (Zen).
One could say he correctly predicted the year that computers would overtake him at go. I think that had a great deal to do with luck, but still, he definitely was within 5 years of the true time, so it should probably count as some sort of evidence in favor of his forecasts. (I also think MFoG was not stronger than Zen back in 2010, so he may not have been playing the strongest program available.)
Possibly of interest:
In late 2010, Tromp won a longstanding $1000 bet by beating a computer program (MFoG) at go.
This year (just a few days ago, actually), Tromp lost a match against the current strongest computer program (Zen).
One could say he correctly predicted the year that computers would overtake him at go. I think that had a great deal to do with luck, but still, he definitely was within 5 years of the true time, so it should probably count as some sort of evidence in favor of his forecasts. (I also think MFoG was not stronger than Zen back in 2010, so he may not have been playing the strongest program available.)
Tromp vs Zen—coverage: http://dcook.org/gobet/