Obviously.Any measure of performance must be based upon actual results.Pete's sytem is not meant to be perfect, if it were you would be able now to predict the table in mid May 2010.Yes but this is still entirely reflective of who you have played. Man City haven't had the chance to face a team to get points off while Chelsea have, so it has the exact same flaws, ie it takes ages for a pattern to emerge and the early pattern is entirely based on fixtures.
In other words just use your brain to establish who is on form.
What Pete's system does though,is to weight those early season games according to the relative difficulty of the tie so as to give a better guide just from a few results as to the likely table come the season's end.It's a good guide and easier to understand than Daniel Finkelstein's Fink tank.As the season progresses it should converge with the actual league table.