One of the key questions for the Brewers over the next couple of weeks is: Just how good is Dave Bush?  Or, more specifically: How good can we expect him to be between now and October?

There's no disputing the fact that Bush has pitched poorly thus far.  His ERA is about 7.00, and while his FIP is better, it's only a run less.  If you think his performance with runners on is something that will continue, it's possible that his FIP is overly optimistic.  His xFIP (adjusting for the number of fly balls that leave the park) is better yet, but at 5.29, it's still not something he's about to put on his business card.

All of these numbers, though, refer to 5 starts.  30 innings.  It's certainly troublesome that the only 2008 data we have to go on is so bad, but every league average pitcher is going to have a stretch like this.  Even Ben Sheets, in his tremendous 2004 campaign (season ERA: 2.70), had a stretch of 45+ IP in which his ERA was 5.40.

Do we know that Bush is just going through a rough time, and that he'll turn things around?  No, we don't.  But we are talking about a 28 year old pitcher with over 100 career starts, 60+ of which came on the same team in the same park against the same league.  Very few players fall apart in their late 20s, and a pitcher is much more likely to perform in line with a sample of 400+ IP than with a sample of 30.

ZiPS projected Bush to have an ERA of 4.64 this year.  That's actually on the pessimistic side as far as projections go; my system, MINER, forecast him 4.46, while CHONE, another extremely accurate system, figured him for 4.38.  Those numbers reflect historical data that suggests that pitchers in their late 20s generally keep chugging along at their established rates.  Even in Bush's bad-for-him 2007, he was just about average for a starting pitcher, and that's where the projections put him for 2008.

Of course, a projection isn't a guarantee.  I looked at the set of ZiPS projections from 2007 to see how starters with similar forecasts fared.  Here are 8 guys who were projected to have ERAs between 4.59 and 4.69, and how they did in 2007:

• Jon Garland, 4.23
• Joe Blanton, 3.95
• Paul Maholm, 5.02
• Matt Chico, 4.63
• Esteban Loaiza, 5.79 (in 7 GS)
• Miguel Batista, 4.29
  
• Ricky Nolasco, 5.48 (in 4 GS)
• Orlando Hernandez, 3.72

Of those, closest comp is probably Garland, as he's the only one on the list within a year of Bush's age.  This is all just for illustration -- there's a reason projections are based on huge pools of aggregate data, not simply on a small group of comparable players.  But the point should be clear: pitchers with similar track records to Bush's usually end up somewhere around that level.

Now to switch gears.  It's become fashionable in some stathead circles to say that "scouting knowledge trumps statistical forecasts."  The idea is that if a pitcher changes his arm slot, or loses velocity, or has a nagging injury, then all bets are off.  To some extent that's true, and thanks to the great work of guys like Josh, we can start to analyze some of those things in a more rigorous way than just pointing and frowning.

I'm no professional scout, but I find it hard to identify noticeable differences in Bush's skills this year compared to previous years.  He has never dominated, and even when he's pitching well, he seems to throw a lot of straight fastballs.  To say his 7.00 ERA is due to a bunch of straight fastballs is to ignore where his straight fastballs (plus excellent control) got him in the first place.

And for better or worse, one of the better ways we have of knowing whether something is really wrong is by seeing how a player's team treats him.  Obviously the Brewers weren't thrilled with what they saw in Vargas, or they would've kept him.  We don't know whether Maddux, Yost, and Melvin think that Bush is still as good as he has been for the last four years, but we can be pretty sure that they thought he was better than Vargas, and they now think he's better than DiFelice, Narveson, or McClung.  (That's not an enthusiastic endorsement, I know.)

In May 2006, Bush had one stretch that looks almost identical to his start this year.   Of course, it didn't turn out that he was done--he followed  it with seven quality starts in his next eight outings.  I have no idea whether we'll get anything like that from Bush starting tonight, but the preponderance of evidence we have to go on suggests we'll get something a whole lot more like 4.64 and a whole lot less like 6.98.