Yesterday watching the baseball game, I was trying to predict what each batter would do. My son was as well, but his predictions were far more accurate.
I would say, "Okay, this guy is going to tie the game in one swing." My son would counter with, "No. He's going to fly out."
Needless to say, my son's predictions were better. But they were also less interesting. Even the great ballplayers get out 70% of the time.
In the social world, if you want accuracy, you should predict no change, a continuation of the status quo. Any given day, the odds are overwhelmingly in favor of tomorrow being like today. You'll wake to the same weather, the same aches and pains, the same habits, the same news. But of course if that were always true, life would fade into all static, no stations, a buzz of monotony. What defines life - what is really interesting - is what is novel and, thus, improbable.
For this reason alone, it seems to me, predictions that force you to consider the world differently are worth more than predictions that prove accurate. The idea should be to provoke a new possibility, not confirm an old fact. The best predictions are the ones that change the odds rather than just report them.