Take a look at the ERA leaders at this time (through games of April 19, 2004) last year compared to this year. At the low end we have more pitchers under 1.00 and more pitchers over 9.00. It appears there's a wider spread so far this year. Let's measure it more precisely.

In 2004, the mean ERA is 4.36 and it has a standard deviation of 2.21. In 2005 the mean ERA is 4.32 and the standard deviation is 2.29. So ERAs are lower and the spread is wider. Stephen Gould once theorized that the disappearance of .400 hitters was do to everyone getting better, or to put it differently, the variance between hitters shrinking. As the mean batting average pretty much stayed the same (due to better pitching and defense), no one could be far enough away from the mean to hit .400.

A narrowing of the spread in ERAs would do the same thing, making it more difficult for pitchers to have extremely low ERAs. But if you widen the spread, you might get some really great pitching number from the low ERA end of the tail.

However, a two tail TTest shows a probability of .89 that the two populations are the same. I'll try to revisit this as the season progresses, especially if we continue to have extremely low ERAs among the league leaders.