Advertising

The Seattle Times Company

NWjobs | NWautos | NWhomes | NWsource | Free Classifieds | seattletimes.com

Mariners


Our network sites seattletimes.com | Advanced

Mariners Blog

Geoff Baker covers the Mariners for The Seattle Times. He provides daily coverage of the team throughout spring training, and during the season.

E-mail Geoff| Mariners Forum| RSS feeds Subscribe | Blog Home

September 8, 2008 10:11 AM

Fun with numbers

Posted by Geoff Baker

PECOTA projected the team to win 73 games. I don't have access to the PECOTA projections for individual players, but you know what? I'm betting that they weren't very good. Why? Good teams don't go 73-89.

-- blog reader back in June


We've been hearing all year about how the sabermetric projection systems got it right on the Mariners, while the rest of the professional prognosticators in America -- i.e. writers -- got it wrong. I'll be the first to admit that I was wrong about this Mariners team. So, in the spirit of learning something more about the art of predicting things -- so as not to lead you all down the wrong path next year -- we'll start giving equal time to some of the projection systems that are out there. Since the PECOTA (short for Player Emperical Comparison and Optimization Test Algorythm) system from Baseball Prospectus came the closest on the Mariners, in their original projections from February, let's see how some of their American League predictions are looking with the season now three weeks from being over. PECOTA was accurate within 4.3 wins overall last year. We'll print their predictions in the first set of numbers, then the team's actual won-lost pace in the next set (in brackets), followed by the difference in the two totals, showing whether a team is outperforming its win projection (a plus-sign in front of the number) or underperforming (with a minus sign).

AL EAST

1. New York 97-65 (86-76)...-11
2. Boston 91-71 (96-66)...+5
3. Tampa Bay 89-73...(98-64)...+9
4. Toronto 77-85...(87-75)...+10
5. Baltimore 66-96... (72-90)...+6


AL CENTRAL

1. Cleveland 93-69 (79-83)...-14
2. Detroit 91-71...(78-83)...-13
3. Chicago 77-85...(91-71)...+14
4. Kansas City 72-85...(70-92)...-2
5. Minnesota 71-91...(88-74)...+17


AL WEST

1. Los Angeles 88-74 (98-64)...+10
2. Oakland 80-82 (74-88)...-6
3. Texas 74-88 (78-84)...+4
4. Seattle 73-89 (64-98)...-9

And now, we'll look at the predicted order of finish versus current actual standings...

AL EAST

1. New York (fourth)...three too high
2. Boston (second)...accurate
3. Tampa Bay...(first)...two too low
4. Toronto...(third)...one too low
5. Baltimore...(fifth)...accurate


AL CENTRAL

1. Cleveland (third)...two too high
2. Detroit (fourth)...two too high
3. Chicago (first)...two too low
4. Kansas City (fifth)...one too high
5. Minnesota (second)...four too low


AL WEST

1. Los Angeles (first)...accurate
2. Oakland (third)...one too high
3. Texas (second)...one too low
4. Seattle (fourth)...accurate


Now, the way PECOTA works is, it comes up with projected individual player performances and then simulates the amount of playing time for each roster member and comes up with a projected team win total. In fairness, rosters change between February and the start of a season. So, on April 24, an updated set of PECOTA projections was published. I have no problem with a little do-over, since writers make their predictions at the end of spring training. So, let's see how those updated projections are faring.

If you want, we'll look at PECOTA's projections for all of baseball this time.

AL EAST

New York 96-66 (86-76)...-10
Boston 91-71 (96-66)...+5
Tampa Bay 90-72 (98-64)...+8
Toronto 77-85 (87-75)...+10
Baltimore 67-95 (72-90)...+5


AL CENTRAL

Cleveland 92-70 (79-83)...-13
Detroit 90-72 (78-84)...-12
Chicago 78-84 (91-71)...+13
Kanasas City 73-89 (70-92)...-3
Minnesota 73-89 (88-74)...+15


AL WEST

Los Angeles 85-77 (98-64)...+13
Oakland 79-83 (74-88)...-5
Seattle 76-86 (64-98)...-12
Texas 71-91 (78-84)...+7


NL EAST

New York 93-69 (91-71)...-2
Atlanta 87-75 (70-92)...-17
Philadelphia 86-76 (88-74)...+2
Washington 73-89 (63-99)...-10
Florida 72-90 (81-81)...+9


NL CENTRAL

Chicago 91-71 (97-65)...+6
Milwaukee 86-78 (93-69)...+7
Cincinnati 81-81 (73-89)...-8
St. Louis 75-87 (87-75)...+12
Pittsburgh 73-89 (69-93)...-4
Houston 72-90 (86-76)...+14


NL WEST

Arizona 87-75 (81-81)...-6
Los Angeles 87-75 (83-79)...-4
Colorado 81-81 (75-87)...-6
San Diego 79-83 (62-90)...-17
San Francisco 68-94 (71-91)...+3


And once again, the predicted order of finish:

AL EAST

1. New York (fourth)...three too high
2. Boston (second)...accurate
3. Tampa Bay (first)...two too low
4. Toronto...(third)...one too low
5. Baltimore...(fifth)...accurate


AL CENTRAL

1. Cleveland (third)...two too high
2. Detroit (fourth)...two too high
3. Chicago (third)...two too low
4. Kansas City (fifth)...one too high
5. Minnesota (second)...three too low


AL WEST

1. Los Angeles (first)...accurate
2. Oakland (third)...one too high
3. Seattle (fourth)...one too high
4. Texas (second)...two too low


NL EAST

1. New York (first)...accurate
2. Atlanta (fourth)...two too high
3. Philadelphia (second)...one too low
4. Washington (fifth)...one too high
5. Florida (third)...two too low


NL CENTRAL

1. Chicago (first)...accurate
2. Milwaukee (second)...accurate
3. Cincinnati (fifth)...two too high
4. St. Louis (third)...one too low
5. Pittsburgh (sixth)...one too high
6. Houston (fourth)...two too low


NL WEST

1. Arizona (second)...one too high
2. Los Angeles (first)...one too low
3. Colorado (third)...accurate
4. San Diego (fifth)...one too high
5. San Francisco (fourth)...one too low

Yes, there are three weeks still to go. And yes, they came close to nailing the M's offensive numers, predicting a .323 (.318 actual) on-base percentage and a .399 (.392 actual) slugging percentage. We'll check back in again at season's end. Give them credit for trying to project win totals as well as predict the order of finish. It isn't easy to do either. As you can see.

Digg Digg | Newsvine Newsvine

Comments
No comments have been posted to this article.

Advertising

Marketplace

Advertising

Advertising

Categories
Calendar

May

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30
31            
Browse the archives

May 2009

April 2009

March 2009

February 2009

January 2009

December 2008