Recently, a member of Brewerfan.net projected 16 Brewers players. My idea was to see how well fans could predict future performance compared to widely used systems lke PECOTA and Zips. This can be done at the end of the season with regression analysis, but that's a long way off. In the mean time, our Brewers have a division to battle out with the Chicago Cubs.
The main reason for projections is to provide a look at how well you can expect players and teams to perform. So I thought it'd be interesting to see what our predicted numbers think. But to get a grasp of where we stand in the division I asked Cubs fans over at nsbb.com to project 16 players south of us.
A baseline projection for both teams was needed. For this I combined two different systems - Chone and Zips, for an average between the two. Zips came out pessimistic and Chone optimistic, so using both should provide a more accurate view. More importantly, though, it allows me to call them "Chips."
The first order of business was playing time. I originally used Chone's PA and IP numbers, but couldn't believe how inaccurate they were. So I adjusted them for both teams to a more realistic situation. Secondly, I had to adjust so that both teams had the exact same number of PA and IP. By the time I was done adjusting the playing time was barely recognizable form the originals and they were too subjective to my opinion.
So I fixed that by applying a set number of PA to starters and bench, and a set number of IP for starters and bullpen. This strategy has its flaws, but since it's a comparison between the two teams, it won't matter. The numbers were 675 PA for the 8 position players, 300 PA to all bench, 200 IP to starters and 65 to bullpen.
The projections then needed to be converted into runs. I used Tangotiger's formula of Runs Above Average = 1.7*OBP+SLE-1*.25*PA for offense. Once I got a RAA for the entire team, I added that to the average number of runs scored per team in the NL last year, 763.
Pitching was easier, as I just divided IP by 9 and multiplied by ERA, and then added all those runs up. But unearned runs aren't included in ERA. Last year, the average NL team gave up 45 unearned runs. The Cubs gave up 40, and the Brewers 68. I added the same # of runs for the Cubs, but regressed the Brewers number by 15 runs due to defensive changes.
For both teams I used the most likely 25 man roster, using 12 pitcher and 13 positions. It should be noted that the total IP and PA totals are exactly at the league averages from last year, 1453 IP and 6167 PA. Below are the projections for all the players.
Ok, done with the boring process. Here's the results:
|Cubs||824 RS||827 RS||702 RA||710 RA|
|Brewers||852 RS||815 RS||710 RA||743 RA|
Wow. Cubs projections are right in line with the fans. Brewers... not so much. Chips projects us at 88 wins, while we have them at 95 wins. Cubs are at 93 wins from both the fans and Chips.
Why the discrepancy? I think it's pretty obvious. The fans are high on the Brewers pitching because our defense is drastically improved from last season, perhaps by 40+ runs. Do the projection systems know that? No. That's why I'd expect our RA to be closer to 710 than 743.
Secondly, all of our star offensive performers were regressed considerably from their 2007 performance, while the Cubs impact players had a relatively down year and are projected for improvement.
Overall, projections are a fun tool to play around with in the off season. They're enjoyable to look at, think about, and they provide good fodder for baseball discussion in the long months of winter. And although most systems see the division going to the Cubs, the Brew is almost a unanimous decision to make the playoffs through the wildcard.
But here's the best news: baseball is right around the corner. And it there's one thing we can be sure of, it's this: We're in for an exciting year of baseball in Milwaukee.