Oh so if you told someone 1 year ago that the xbox would beat they ps3 after the ps3 won for something like 6 months in a row they would believe you based on your obviously vast knowledge ;)

does it really matter?  You could argue that since the ps3 and xbox 360 came out they have gone back and forth every year, neither one really gaining or loosing ground.  The point is, a prediction is a guess about something that could happen in the future, it doesn't have to be accurate.  Predictions can be wrong.  

It's like one of those games played at parties where you try to guess the number of m&m's in a jar.  You don't have any way of knowing the right number of m&m's without counting them, but you are suppose to guess the total beforehand to see who is closest.  

Using your same example, 1 year ago when the ps3 had been winning for a while, we knew about Kinect (then natal).  Somebody could then make a prediction that Kinect would drive 360 sales past the ps3.  That same thing happened many times, with people predicting Kinect would have a huge impact or it would be a complete flop.  Some were right, some were wrong.  There are a lot of reasons somebody could think that xbox 360 might end up outselling the ps3 at that point in time.  You, apparently only want to look at current sales trends, without looking at the whole picture.  We already know tons of information about the console sales this generation.  We have 5 years worth of data in which to extrapolate information to help us make informed predictions.  We have games and initiatives that the console manufacturers have announced as coming in the coming year or two.  

Not that we really need that in order to make predictions.  As I showed with the definitions, a prediction is merely a statement about some future event.  As you show with your vague interpretation of the most vague definition, it can be grounded in whatever kind of reasoning the person making the prediction wants.  Simple as that.