the "super sub" or its corollary that "he can't score when he starts" is and always was complete rubbish. The act of scoring a goal is in and of itself such a rare event with so many events leading up to it that an appropriate sample to somehow determine that a player is only good when coming off the bench is probably measured in years, if not decades, and that would only be valid if the player somehow stayed in the same team with the same setup for that entire duration. It boggles the mind that people somehow bought into an idea that a proven International who has scored goals as a starter throughout his career could somehow only be good off the bench for City because of no reason other than small sample size evidence. It was absolutely elementary thinking and it is, to be honest, embarassing.
In order to even try to posit the hypothesis that Dzeko is best off the bench, you'd have to start by reconciling all his work for his national team and at Wolfsburg, then come up with an actual reason as to why his scoring record off the bench is predictive as opposed to simply a fluctuation in statistics. There aren't a lot of variables that change after 60 minutes of a game has been played aside from fatigue and tactic modification, and in pretty much all cases both are probably influential, but neither would completely explain why a player would be dominant off the bench and lousy as a starter.
To put it another way, any player who is capable of scoring a bunch of goals in a run of games is probably capable of doing it as a starter or as a sub. It will, always, manifest itself in different ways that seem to look like a pattern but are entirely a result of random events. It was a really classic case of correlation leading people to think causality was at play.
