The Curse of... oh, let's say, Clay Bellinger:



RSS 2.0 Atom

*ADVERTISEMENT*

*ADVERTISEMENT*

RedSox (50-37) @ Yankees (45-40), 7:05pm **Game Chatter**
(136 Comments - 7/4/2008 1:43:30 am)

NY POST: FINALLY BATS MORE LIKE IT!
(97 Comments - 7/3/2008 6:43:21 pm)

Rangers (44-41) @ Yankees (44-40), 7:05pm **Game Chatter**
(132 Comments - 7/3/2008 10:03:45 am)

NY Daily News: Joba Chamberlain, Yankee offense come up short in loss to Rangers
(93 Comments - 7/2/2008 5:49:15 pm)

Rangers (43-41) @ WOE (44-39), 7:05pm **Game Chatter**
(445 Comments - 7/2/2008 10:25:02 am)

Yankees.com: Yanks edged by Rangers in opener
(121 Comments - 7/1/2008 6:32:08 pm)

Rangers (42-41) @ Yankees (44-38), 7:05pm **Game Chatter**
(362 Comments - 6/30/2008 11:31:39 pm)

The Times and Democrat: Holly Hill’s Brett Gardner called up to New York Yankees
(102 Comments - 6/30/2008 6:18:41 pm)

Easy Schedule Checkpoint - June 29
(16 Comments - 6/30/2008 4:54:52 pm)

Yankees (44-37) @ Mets (39-41), 1:10pm **Game Chatter**
(194 Comments - 6/30/2008 1:32:45 am)

Player

 Current Projected
Jonathan Albaladejo
7 G
21 G
13.7 IP
42 IP
Chris Britton
4 G
4 G
7 IP
7 IP
Kyle Farnsworth
27 G
69 G
28.3 IP
73 IP
Dan Giese
2 G
5 G
6.3 IP
16 IP
LaTroy Hawkins
22 G
57 G
26.7 IP
69 IP
Ross Ohlendorf
19 G
49 G
33.3 IP
86 IP
Edwar Ramirez
16 G
41 G
17 IP
44 IP
Mariano Rivera
30 G
68 G
32 IP
73 IP
Jose Veras
15 G
39 G
16.3 IP
42 IP
Total
171 G
440 G
204.3 IP
527 IP
Look what people have to say about Larry Mahnken's commentary!

"Larry, can you be any more of a Yankee apologist?.... Just look past your Yankee myopia and try some objectivity."
- Bernal Diaz

"Mr. Mahnken is enlightened."
- cordially, as always,
rm

"Wow, Larry. You've produced 25% of the comments on this thread and said nothing meaningful. That's impressive, even for you."
- Anonymous

"After reading all your postings and daily weblog...I believe you have truly become the Phil Pepe of this generation. Now this is not necessarily a good thing."
- Repoz

"you blog sucks, it reeds as it was written by the queer son of mike lupica and roids clemens. i could write a better column by letting a monkey fuk a typewriter. i dont need no 181 million dollar team to write a blog fukkk the spankeees"
- yan

"i think his followers have a different sexual preference than most men"
- bob

"Boring and predictable."
- No Guru No Method

"Are you the biggest idiot ever?"
- Randal

"I'm not qualified to write for online media, let alone mainstream media."
- Larry Mahnken

This site is best viewed with a monitor.

Disclaimer: If you think this is the official website of the New York Yankees, you're an idiot. Go away.

Wednesday, October 17, 2007

CAIRO Projections v0.1

One of the areas of baseball analysis that I've taken an interest in are projections. Some of the people I consider the best analysts out there have spent a lot of time and effort devising various projection systems. With Dan Szymborski's ZiPS, Sean Smith's CHONE, Tango Tiger's Marcels, Nate Silver's PECOTA, and some others I'm surely forgetting there's no shortage of systems out there.
I have a stubborn side to me though. I have this burning need to understand how all these numbers work instead of just taking them as presented. So I've taken to trying to calculate a lot of stuff myself. That's why I decided to come up with my own projection system, which I've code-named CAIRO after my favorite bad baseball player. It will eventually stand for something, but I'm not quite sure what yet.

I don't expect CAIRO to be any better than any of the systems mentioned above, but at least now I don't have to wait for the others to come out. The other thing I want is to make this a reasonably open-source system. The heart of CAIRO is Tango Tiger's Marcel system, but with a few of my own tweaks. This is the methodology I am currently using.

1) Park adjust and league adjust the component stats for each season from 2003-2007. I am including 2006 and 2007 MLEs (major league equivalencies) but for now I've only projected people who appeared in the majors in 2007 except for a couple of Yankee farmhands.
2) Weigh each season using a 5/4/3/2/1 weight (most recent season weighed most heavily) for batters. For pitchers I use a 7/5/3/2/1 weight as I believe pitchers' most recent performance should be weighed a little more heavily since pitchers are more likely to change their true talent level both positively and negatively.
3) Add in a percentage of plate appearances or innings based on league average performance (regression towards the mean)
4) Adjust the appropriate components for the player's age (hits, xbh, HR, BB, K, SB)
5) Park-adjust the final stat line for the player's expected park/league. For pitchers I include an adjustment for the projected defense behind them.
6) This is an almost entirely objective system, but I have made/will make playing time adjustments for some players
6) Last, wait for everyone to tell me how wrong the projections are.

I have included defensive projections using Zone Rating as well. I used a weighted average from 2002-2007 with regression and aging factored in. I included a lot of players who don't have significant playing time at specific positions so small sample size caveats apply, even with the regression that I added in.

This is all very much a work in progress and I expect to continue tweaking it during the offseason, so if anyone sees any numbers that look wrong or has any questions about the methodology feel free to let me know.

Since this is supposed to be a Yankee site, here are the Yankees' projections.

NAME AGE TEAM POS G PA AB H 2B 3B HR RBI BB SO SB CS DP AVG OBP SLG
Alberto Gonzalez 25 NYA ss 75 259 237 60 11 1 4 28 18 34 0 0 0 .255 .311 .363
Alex Rodriguez 32 NYA 3b 157 686 580 174 27 1 42 124 89 127 19 4 17 .300 .402 .570
Andy Phillips 31 NYA 1b 63 198 180 48 9 1 4 24 13 34 1 1 4 .266 .310 .403
Angel Chavez 26 NYA ss 75 280 264 68 14 1 7 34 14 54 0 0 0 .259 .297 .395
Bobby Abreu 34 NYA rf 155 676 565 161 37 2 19 89 101 120 23 6 10 .284 .392 .462
Brett Gardner 24 NYA cf 83 331 297 78 11 3 2 31 31 64 0 0 0 .262 .332 .341
Bronson Sardinha 25 NYA rf 85 322 294 69 12 2 9 39 26 62 0 0 0 .234 .296 .378
Chris Basak 29 NYA 3b 69 239 215 56 11 1 6 30 21 42 0 0 0 .263 .329 .411
Derek Jeter 34 NYA ss 151 688 608 193 35 3 16 90 62 102 18 5 17 .318 .387 .463
Doug Mientkiewicz 34 NYA 1b 79 291 254 66 15 1 7 35 29 39 1 1 7 .262 .338 .409
Eric Duncan 23 NYA 1b 79 303 274 63 14 1 8 36 26 56 0 0 0 .231 .299 .379
Hideki Matsui 34 NYA lf 137 579 506 147 30 2 22 86 65 73 2 1 12 .291 .370 .490
Jason Giambi 37 NYA dh 107 422 339 87 15 0 22 66 70 82 1 0 6 .256 .398 .498
Johnny Damon 34 NYA cf 137 607 539 154 27 4 16 78 60 75 21 5 7 .286 .357 .438
Jorge Posada 36 NYA c 141 553 473 139 31 1 21 82 71 93 2 1 14 .294 .391 .499
Jose Molina 33 NYA c 60 195 179 45 10 0 4 22 10 39 1 0 4 .248 .281 .369
Kevin Reese 30 NYA lf 61 234 212 52 7 1 5 26 18 40 0 0 0 .245 .309 .361
Melky Cabrera 23 NYA cf 139 564 502 144 24 5 9 64 48 64 6 3 7 .287 .345 .407
Robinson Cano 25 NYA 2b 152 624 581 181 41 4 18 90 31 76 3 3 17 .312 .347 .490
Shelley Duncan 28 NYA dh 83 311 281 72 13 1 16 50 26 64 0 0 1 .256 .317 .479
Wil Nieves 30 NYA c 42 147 136 32 6 0 2 15 8 18 0 0 1 .233 .274 .340
Wilson Betemit 26 NYA 1b 109 308 272 71 14 1 12 44 30 70 1 1 4 .263 .332 .451


NAME AGE TEAM ERA G W L IP H ER HR BB SO
Andy Pettitte 36 NYA 4.21 36 13 9 201 199 94 19 57 149
Brian Bruney 26 NYA 5.20 32 2 2 33 31 19 4 22 30
Carl Pavano 32 NYA 4.49 10 3 3 50 53 25 6 11 31
Chase Wright 25 NYA 4.76 17 3 3 51 57 27 6 23 30
Chien-Ming Wang 28 NYA 3.93 35 13 9 202 210 88 12 56 95
Chris Britton 25 NYA 3.79 46 4 2 57 57 24 5 19 39
Colter Bean 31 NYA 4.44 13 2 1 24 23 12 2 16 22
Darrell Rasner 27 NYA 4.77 13 3 2 45 51 24 6 13 26
Edwar Ramirez 27 NYA 3.73 39 4 2 55 45 23 6 25 70
Ian Kennedy 23 NYA 4.19 38 11 9 181 179 84 21 72 143
Jeff Karstens 25 NYA 5.82 13 2 4 51 61 33 9 18 27
Jim Brower 35 NYA 5.88 22 1 2 26 28 17 4 11 19
Joba Chamberlain 22 NYA 3.68 55 10 7 149 136 61 16 48 150
Jose Veras 27 NYA 4.29 27 2 2 38 38 18 4 15 29
Kei Igawa 28 NYA 5.72 25 6 9 138 158 88 29 46 102
Kyle Farnsworth 32 NYA 4.23 58 4 2 57 53 27 7 24 58
Luis Vizcaino 33 NYA 4.46 71 4 4 73 69 36 8 33 62
Mariano Rivera 38 NYA 2.74 66 6 2 75 67 23 4 15 69
Matt DeSalvo 27 NYA 6.82 15 2 4 53 63 40 8 37 26
Mike Mussina 39 NYA 4.26 32 11 8 173 177 82 19 40 129
Philip Hughes 22 NYA 3.73 35 11 6 157 152 65 14 50 124
Roger Clemens 45 NYA 3.69 22 7 4 103 96 42 9 30 85
Ron Villone 38 NYA 4.60 45 3 4 59 56 30 7 29 47
Ross Ohlendorf 25 NYA 4.77 16 4 3 64 73 34 9 15 39
Sean Henn 27 NYA 5.90 19 2 2 32 36 21 5 19 21
Tyler Clippard 23 NYA 5.81 19 4 5 81 92 52 16 35 58


Player Age Team LG Pos G Innings PO A E DP PM CH ZR PM +/- RS RS/162
Doug Mientkiewicz 34 NYY AL 1B 86 669 626 53 4 63 119 137 .870 4 3 7
Jason Giambi 37 NYY AL 1B 60 459 402 32 5 38 72 90 .796 -4 -3 -9
Andy Phillips 31 NYY AL 1B 66 428 342 45 4 41 80 94 .846 1 1 2
Wilson Betemit 27 NYY AL 1B 36 255 88 63 3 20 69 84 .816 -1 -1 -5
Shelley Duncan 29 NYY AL 1B 31 218 51 62 3 18 61 73 .834 1 1 6
Robinson Cano 26 NYY AL 2B 140 1213 320 397 13 100 367 441 .833 4 3 4
Wilson Betemit 27 NYY AL 2B 11 84 53 13 0 7 18 22 .813 -1 -1 -9
Alex Rodriguez 33 NYY AL 3B 156 1337 115 273 16 31 299 394 .758 -3 -3 -3
Wilson Betemit 27 NYY AL 3B 41 288 23 63 4 9 64 81 .786 0 0 -2
Johnny Damon 35 NYY AL CF 117 980 284 16 3 3 280 318 .881 0 0 0
Melky Cabrera 24 NYY AL CF 62 505 153 23 3 4 153 173 .880 2 1 4
Hideki Matsui 34 NYY AL LF 105 892 204 6 4 1 199 238 .833 -7 -6 -10
Melky Cabrera 24 NYY AL LF 72 599 139 7 1 1 134 159 .840 -3 -2 -6
Johnny Damon 35 NYY AL LF 52 423 116 3 2 0 112 129 .869 0 0 0
Bobby Abreu 34 NYY AL RF 128 1089 240 7 4 1 237 275 .864 -2 -2 -2
Derek Jeter 34 NYY AL SS 152 1300 225 373 15 85 369 457 .806 -11 -8 -9
Alberto Gonzalez 25 NYY AL SS 35 225 49 18 2 3 56 66 .838 -1 0 0
Wilson Betemit 27 NYY AL SS 18 114 22 21 2 5 29 36 .793 -2 -1 -19


The full spreadsheet is available here.

Update: Version 1.3 is now available. I added more minor league data and changed some of my pitching algorithms. Link is here.

It's important to remember that any projection system is inherently limited. We're dealing with athletes playing games, and their true talent can change in ways that can be forecasted. In addition fluke seasons happen, both good and bad. I think that on a team level projections are a useful tool for understanding probabilities, but at the end of the day that's all they are. Probabilities, not predictions.
--Posted at 6:47 am by SG / 37 Comments | - (3157)


Page 1 of 1 pages: