They aren't rubbish. They are based upon what's happened in the past. For example, Andros Townsend is an eye friendly player. Elite dribbler with a high shot volume until you see his shooting locations.
The point is analysing how well Liverpool will do this season. The point is that you can't base a season upon an unsustainable finishing streak. Teams always regress to the mean. The majority of statistical models can predict a league table with a high degree of accuracy. My point is that you can get away with winning a big game if a Jordan Henderson screamer if the overall method is good in the majority of games.
No team in the league barring City at a big stretch have been good enough to show that they will win the league, neither have they played enough games. Liverpool may have beaten Arsenal and Chelsea with a bit of luck but it doesn't matter as we're above you in the table albeit on goal difference.