That begs the question, why is it hard for you to comprehend that perhaps the correct explanation here is that Marquis' stats are skewed by his July numbers (6.29 ERA in 6 starts), and the guy we saw from April through June (3.46 in 16 starts) is the "real" Marquis? If you want to throw out a stretch of starts as being anomalous, and look at the rest as being more representative of what everyone should expect going forward, then why is it that you choose to throw out three months and 16 starts' worth of results, and keep one month and 6 starts' worth? And while I'm at it, why should we expect Marquis' 2006 season (6.02 ERA) to be the proper benchmark for his future production, and not his 2005 (4.13) or 2004 (3.71) seasons? Because there are tons of secondary numbers that indicate that his 2004 and 2005 seasons were flukes. while he certainly isn't as bad as he was in 2006, he's not a very good pitcher.