Baseball Toaster Catfish Stew
STOP CASTING POROSITY! An Oakland Athletics blog.
Frozen Toast
Google Search
Catfish Stew

02  01 

12  11  10  09  08  07 
06  05  04  03  02  01 

12  11  10  09  08  07 
06  05  04  03  02  01 

12  11  10  09  08  07 
06  05  04  03  02  01 

12  11  10  09  08  07 
06  05  04  03  01 

12  09  08  01 

12  11  10  09  08 
Email Us

Ken: catfish AT zombia d.o.t. com
Ryan: rarmbrust AT gmail d.o.t. com
Philip: kingchimp AT alamedanet d.o.t net

Ken's Greatest Hits
28 Aug 2003
12 Jan 2004
31 May 2005
11 May 2005
29 Jun 2005
8 Jun 2005
19 Jul 2005
11 Aug 2005
7 Sep 2005
20 Sep 2005
22 Sep 2005
26 Sep 2005
28 Sep 2005
29 Sep 2005
18 Oct 2005
9 Nov 2005
15 Nov 2005
20 Nov 2005

13 Dec 2005
19 Jan 2006
28 Jan 2006
21 Feb 2006
10 Apr 2006
16 Apr 2006
22 Apr 2006
7 May 2006
25 May 2006
31 May 2006
18 Jun 2006
22 Jun 2006
6 Jul 2006
17 Jul 2006
13 Aug 2006
15 Aug 2006
16 Aug 2006
20 Aug 2006
11 Oct 2006
31 Oct 2006
29 Dec 2006
4 Jan 2006
12 Jan 2006
27 Jan 2007
17 Feb 2007
30 Apr 2007
27 Aug 2007
5 Sep 2007
19 Oct 2007
23 Nov 2007
5 Jan 2008
16 Jan 2008
4 Feb 2008
7 May 2008
20 Jun 2008
4 Feb 2008
Adjusting PECOTA to Make the A's Win
2008-02-17 22:45
by Ken Arneson

There was a little buzz in the blogosphere recently when Derek Zumsteg at USS Mariner ran the ZiPS 2008 projections through the Diamond Mind simulator, and ended up with the A's winning the division more than any other team. Obviously, something was screwy there, right? Yes, ZiPS doesn't really try very hard to accurately predict playing time, and Diamond Mind has a mind of its own with regards to playing time, as well, so that outcome alone is nothing for A's fans to excited about.

BP's PECOTA, on the other hand, does try to predict playing time, and they just came out with their projected standings. They have the A's in second place, with a 78-84 record, eleven games behind the Angels. That's closer to what I had mentally been preparing myself for, a year without hope. But I thought I'd try a little thought experiment to see if I could come up with a reasonable scenario, based on the PECOTA numbers, that gives the A's a division title.

So I'm going to make some reasonable adjustments to PECOTA's assumptions. By reasonable, I mean, I'm not going to turn Bobby Crosby into Alex Rodriguez. The adjustments have to have some logic to it.

Here's what we need to accomplish: find a reasonable scenario whereby the A's can make up the gap in the projected run differentials between the A's and Angels. Here's what we're starting with:

Team runs: score-allowed
Angels:      810-728     = +82
Athletics:   732-758     = -26
Run differential to make up:      108.0 runs (based on VORP)

First, we'll knock some runs off the Angels' totals. PECOTA projects Vladimir Guerrero to be healthy for the whole year. Vlad has had plenty of health issues in the past, with his back and hammies. We'll say he has one of those injuries, and loses 200 plate appearances, which are gobbled up by Reggie Willits.

Vlad:        -200 PA: -10.5
Willits:     +200 PA:  +1.4
Angels RF Adjustment:               9.1 runs

Kelvim Escobar and Jered Weaver are also projected to be mostly healthy, but Escobar will start the year on the DL, and Weaver had the kind of arm problems last year that can easily recur. We'll knock 50 IP from each, and give them to Joe Saunders and Ervin Santana. Here's where the Jon Garland trade helps, as the dropoff here isn't very big.

Escobar:      -50 IP:  -7.6
Weaver:       -50 IP:  -6.7
Saunders   :  +50 IP:  +3.4
Santana:      +50 IP:  +5.6
Angels SP Adjustment:               5.3 runs

Now to the A's. PECOTA thinks Jack Cust was a bit of a fluke last year, and projects a huge drop in batting average and slugging percentage. We're going to assume it wasn't a fluke, and give him the same rate stats as last year.

Cust old:     553 PA: -20.6
Cust new:     553 PA:  35.6
A's Cust Adjustment:               15.0 runs

PECOTA thinks Eric Chavez' career is on a steep decline. We know, however, that Chavez has been playing hurt for two years, and just had three surgeries to correct the problems that have plagued him. Let's assume that the surgeries work, not so much that he gets back to his peak levels, but at least so much that he returns to the level of the slightly-hurt Eric Chavez of 2006.

Chavez old:   473 PA: -11.3
Chavez new:   473 PA: +18.0
Chavez       +100 PA:   3.8
Hannahan:    -100 PA:  -2.4
A's 3B Adjustment:                  8.1 runs

PECOTA gives half the CF playing time to Emil Brown, who is projected to have a negative VORP. In reality, Chris Denorfia is probably going to be the starting CF, and Brown will platoon, and mostly play against LHP. We'll take 200 PAs away from Brown and give them to Denorfia.

Brown:       -200 PA:  +0.1
Denorfia:    +200 PA:  +3.4
New CF Total:                       3.5 runs

PECOTA doesn't know what to do with Justin Duchscherer, who is converting to a starter from a reliever, and gives him 130 IP. It also assumes that Rich Harden will get hurt again, and gives him 100 IP. Let's assume they stay healthy, and give them an extra 50 IP each, and take them away from the rookies like Dan Meyer, who like Emil Brown was projected with a negative VORP.

Harden:      +50 IP:  +13.9
Duchscherer: +50 IP:  +11.4
Meyer:       -45 IP:   +1.3
DeLosSantos: -35 IP:   -1.8
Gonzalez:    -20 IP:   -1.8
New SP Total:                      23.0 runs

OK, so where are we now with our reasonable adjustments?

Total Reasonable Adjustments:      64.0 runs
Remaining Angel Advantage          44.0 runs

So we knocked about six or seven games off that eleven game lead. The A's are still trailing the Angels by four games. You can make up four games on pure luck. And a minor miracle or two could happen: Bobby Crosby could stay healthy and break out, Ryan Sweeney could hit his 90th percentile projections, Chad Gaudin could pitch the whole season like he did in the first half of '07 before he got hurt, Dana Eveland could run up numbers in the majors like he did in the minors.

So there you have it, a little balloon of hope to lift you through spring training, before reality comes along to stick a pin in it. Have fun, kids!

2008-02-17 23:14:16
1.   Xeifrank
I ran the four AL West teams through my simulator and played them through their schedules for 25 seasons (yeah, 2000 seasons would've been better but my super computer is on loan). There are some minor tweaking that I need to make for playing time estimators, as my simulator was originally designed for running an exact game matchup X number of times. Anyways, here is what I got using 2008 ZIPS projections as input.

(Average number of wins per season)
LAA: 85.52
TEX: 82.84
OAK: 79.72
SEA: 75.28

vr, Xei

2008-02-18 05:37:08
2.   StolenMonkey86
The Angels have Jered Weaver, not Jeff. If you replace Jered with Jeff, that probably takes at least a dozen runs away from the Angels.
2008-02-18 07:32:41
3.   Sky
How does PECOTA handle fielding? Does it build DER into a pitcher's projected ERA?
2008-02-18 07:46:24
4.   Ken Arneson
2 Oops, thanks, fixed.
2008-02-18 07:47:19
5.   Ken Arneson
3 They take it into account, I'm sure, but they didn't display any fielding information on the Depth Chart/projection pages.
2008-02-18 19:20:56
6.   WaddellCanseco
I'd do something to Francisco Rodriguez to make up part of the gap. Then get breakout half-years from Carlos and Gio. Aaron Cunningham could get a surprise half-year too! We're running away with it now!

Comment status: comments have been closed. Baseball Toaster is now out of business.