I think some RAWKite said that Alternate EPL table has a better prediction rate (based on the last few seasons) than the real table. I haven't verified that claim but I think it is very likely the case that if you look at both tables after half of the matches are played, then Alternate table gives better predictions than the real table.
Of course, if we have a few teams doing what Leicester and Chelsea did last season, then that doesn't happen, but more often than not, I would guess that Alternate table beats the real table after 10-20 matches when it comes to predicting the final table.