Alright, this particular project is going to take a few posts, but it's something I've been meaning to do all offseason, anyway. The reason why is because a certain discrepancy I've been noticing with pre-season projections for the Rockies and my view of reality. Of course I'm a homer, and I expect the Rockies to do well, even when everybody else and their mother know that they won't, so it might be easy for you to just look the other way and say there goes silly Rox Girl blabbering away again, but hear me out for a second:
Check out this extremely early forecast using Sean Smith's CHONE projections again and notice the 2007 Rockies results:
66.5-95.5 740 RS, 900 RA
Alright, if this was a singularity, I wouldn't think anything of it. Several projection systems thought the Rockies would have a similar record last year, and of course they underestimated us by several wins. Only the 2006 ZiPS projection seemed to have us figured out, getting within four runs of our actual run scored total, and 33 of our runs allowed number. The latter discrepancy could almost entirely be explained away by ZiPS' missing Jason Jennings' breakout season, projecting him to have a 5.39 ERA last season.
So what about ZiPS this year? Well, Dodger fan Xeifrank has been taking each NL West team's 2007 ZiPS projections and playing them against each other. His results really piqued my interest, and were probably the primary impetus of this. This is why: using these results, let's say we could clone our best starter, Aaron Cook. We'll make four more of him, and our rotation will consist of five Aarons. What's more, for some reason we're in a new league where we only play the Dodgers for 162 games. In this scenario, according to the sim, our rotation of five Aaron Cooks and our regular lineup would still only manage a 64-98 record against a regular Dodger rotation (Schmidt, Penny, Lowe, Wolf and Billingsley) and lineup.
Our record in the sim doesn't get any better if you substitute another one of our starters (and his clones) for Cook. Before we go further, let me just say that I can accept that the Dodgers have a better team than the Rockies this season, but are they really that much better that they can beat our best pitcher 60% of the time? Cook, remember, has been a top 30 starter in baseball over the last season and a half. He's actually been remarkably consistent since returning from going down with blood clots in his lungs in 2004 with an RA+ in 115 range. This shouldn't be a 60% loser, even to a rotation as stacked as the Dodgers is, particularly when you consider the strength of our lineup.
So what's the deal then? Let's go back to the CHONE projection. Why are the Rockies projected to score 75 runs less in 2007 than they did in 2006 despite upgrading at two or three positions and remaining pretty much intact elsewhere? Are Randy Johnson and Barry Zito going to make that much difference on our output? Why are the pitchers going to give up 90 more runs? Were Jason Jennings and Josh Fogg that much better than Rodrigo Lopez and Jason Hirsh? Ray King and Jose Mesa that much better than Taylor Buchholz and LaTroy Hawkins? Okay, so we take these projections with a grain of salt anyway, but when they start coming out with numbers as far off base from what we'd expect as they seem to be this offseason, it might be wise to take a closer look in case we as fans have missed something about how terrible our team is supposed to be this year.
Okay. Closer look done. I still don't get it. As far as I can tell, ZiPS projects us to score about 845 runs, and to give up a few more, depending on how quickly we can extricate ourselves from the clutches of Josh Fogg. However, this seems to be counting on one particularly untrustworthy monster playing a little more tame than it has in the past, but not quite as kitty like as it did last season. And this is when it occurred to me that before we can make any sort of accurate projection for the Rockies in 2007, we have to project what will happen in the Stadium they play in.
Pointless? Yeah, maybe. Futile? Absolutely, but it would still be fun to try.
First of all, I hope it's clear that Coors Field in 2006 played differently than it has since its inauguration in 1995. To demonstrate, here are the historical park factors for Coors in all 12 of its seasons (represented as a percentage to a neutral league scoring environment):
1995: 128
1996: 129
1997: 123
1998: 119
1999: 129
2000: 131
2001: 122
2002: 121
2003: 112
2004: 120
2005: 113
2006: 107
Now before you get all huffy about how the Rockies cheated to get to that number last season, let's isolate the years since the humidor has been in use:
2002: 121
2003: 112
2004: 120
2005: 113
2006: 107
The average for those five seasons is 114.6, meaning that the 7% spike in 2002 is actually almost as much of an outlier than last season's drop. Oooh, take that you doubters! Burn on you Jeff Cirillo! Zing!
No, actually, when you look a little closer at the numbers from last season, there are a couple of fishy things that still pop out, namely the park factor for doubles which was only at 97, and also the strange journey to get to that final acceptable threshold, which required some eye popping scores the last two months of the season, but more on that in the next part of this series. This first part is just an introduction to try and scare up some techie type statisticians to look at this rather than us fluffy writer types.
Will Coors play like it did last year? Neutral in the beginning, ferocious at the end? Is another 2004 on the horizon, where scoring is still up 20% despite the humidor? Or will the field settle back into it's tolerable 14% mode? I think the last one is the safe bet, but what are some other opinions on this? How will this affect the Rockies and the teams that come here this year? Anyway, it's time to start the community projection for the stadium that defies projection...