Batter's Box Interactive Magazine Batter's Box Interactive Magazine Batter's Box Interactive Magazine
At about this time last year, I devoted a fair amount of attention to the striking discrepancy between Toronto's won-lost record and what their actual runs scored and allowed should have led us to expect. This year, of course, things returned to normal.

Anyway, here are your Pythagorean Standings. These will differ slightly from what you see on ESPN or other places; being a traditional sort of fellow, I use the traditional formula: runs squared / (runs squared +runs allowed squared). The first winning percentage is the Pythagorean expectation, which we use to produce the expected win-loss records. This will not match the winning percentage of that won-loss record, because we must round off the numbers. I mention this now because I know what happens when we get to the NL Central!

                   Expected                  Actual       Difference
AL EAST W L PCT GB RS RA W L PCT
New York 96 66 .595 -- 930 767 97 65 .599 + 1
Toronto 87 75 .535 9 809 754 87 75 .537 0
Boston 81 81 .497 15 820 825 86 76 .531 + 5
Baltimore 68 94 .422 28 768 899 70 92 .432 + 2
Tampa Bay 64 98 .393 32 689 856 61 101 .377 - 3


AL CENTRAL
Detroit 97 65 .597 -- 822 675 95 67 .586 - 2
Minnesota 94 68 .579 3 801 683 96 66 .593 + 2
Cleveland 90 72 .553 7 870 782 78 84 .481 -12
Chicago 88 74 .544 9 868 794 90 72 .556 + 2
Kansas City 61 101 .378 36 757 971 62 100 .383 + 1

AL WEST
Texas 86 76 .531 -- 835 784 80 82 .494 - 6
Oakland 86 76 .529 -- 771 727 93 69 .574 + 7
Los Angeles 85 77 .523 1 766 732 89 73 .549 + 4
Seattle 77 85 .477 9 756 792 78 84 .481 + 1


NL EAST
New York 92 70 .566 -- 834 731 97 65 .599 + 5
Philadelphia 86 76 .532 6 865 812 85 77 .525 - 1
Atlanta 85 77 .527 7 849 805 79 83 .488 - 6
Florida 80 82 .491 12 758 772 78 84 .488 - 2
Washington 68 94 .423 24 746 872 71 91 .438 + 3

NL CENTRAL
St. Louis 82 79 .512 .5 781 762 83 78 .516 + 1
Houston 83 79 .511 -- 735 719 82 80 .506 - 1
Cincinnati 76 86 .466 7 749 801 80 82 .494 + 4
Milwaukee 70 92 .434 13 730 833 75 87 .463 + 5
Pittsburgh 70 92 .429 13 691 797 67 95 .414 - 3
Chicago 69 93 .424 14 716 834 66 96 .407 - 3
**OK, I know this looks extremely funny. As I warned, it's because a) we're rounding off the integers, and b) the Cardinals played one fewer game. A .512 winning percentage would give St.Louis 82.48 wins in 161 games; a .511 inning percentage would give Houston 82.78 wins in 162 games. Which leads to the bizarre result shown above...

NL WEST    
Los Angeles 88 74 .544 -- 820 751 88 74 .543 0
San Diego 87 75 .537 1 731 679 88 74 .543 + 1
Colorado 81 81 .501 7 813 812 76 86 .469 - 5
Arizona 79 83 .490 9 773 788 76 86 .469 + 3
San Francisco 76 85 .471 12 746 790 76 85 .472 0

You will notice that three teams ended with exactly the same number of wins old Pythagoras predicts: the Giants, the Dodgers, and the Blue Jays. The biggest over-achievers in 2006 were the Oakland Athletics, who went 93-69 when they could reasonably have been expected to go 86-76. Other teams who were five wins better than their Pythagorean expectation were the Red Sox, the Mets and the Brewers. Fortune smiled on these teams, and they would do well to realize it, lest 2007 take them by surprise.

Among the more notable under-achievers were the Atlanta Braves. The run of division championships finally ended, as the Braves stumbled all the way to a 79-83 mark. But Atlanta scored 849 runs - second best total in the NL, while allowing 805. They finished 18 games behind the Mets, but they're probably a lot closer than that. The Texas Rangers were also very likely a better team than this year's record suggests. Having since fired Buck Showalter, they are now the obvious choice to win the 2007 World Series because that's what happens to teams after they fire Buck Showalter. And then there are the Indians...

Ah, yes. The Cleveland Indians.

In the course of last year's investigation, I discovered, to my shock, astonishment, and delight, that just ten teams in history had outscored the opposition by more than 50 runs while posting a losing record. One of them was the 2005 Toronto Blue Jays, one of just three teams in history to outscore the opposition by 70 runs while losing more often than they had won.

Well, move over. Now there are four.

The 2006 Cleveland Indians went 78-84 even though they scored 870 runs and allowed just 782. The Indians scored more runs and gave up fewer than the defending champion White Sox, who went 90-72. The Indians were 88 runs to the good, and no team with a losing record in the history of the game has outscored the opposition by so many runs. Ever. Here's how that list looks now:

Year	Team	Lg	G	W	L	PCT	RS	RA	Diff

2006 CLE A 162 78 84 .481 870 782 88
1955 CIN N 154 75 79 .487 761 684 77
1958 CIN N 154 76 78 .494 695 621 74
2005 TOR A 162 80 82 .494 775 705 70
1995 BAL A 144 71 73 .493 704 640 64
1984 HOU N 162 80 82 .494 693 630 63
1967 BAL A 161 76 85 .472 654 592 62
1951 BOS N 154 76 78 .494 723 662 61
1964 MIN A 162 79 83 .488 737 678 59
1908 BOS A 154 75 79 .487 564 513 51
1968 PIT N 162 80 82 .494 583 532 51

Last year, I noted that the ten teams on my list had all played very well in blowouts and not particularly well in close games. (This of course is pretty well a mandatory condition for outscoring the opposition by a significant amount while still having a losing record.) This year's Indians went an unimpressive 18-26 in one-run games - not particularly good, although certainly better than the 2005 Jays (who went 16-31 in one-run games.)

In Cleveland's other 118 games, they went 60-58 despite outscoring the opposition by 80 runs. So what happened in their one-sided games?

The Indians went 16-9 in games decided by seven runs or more. That's a .640 winning percentage, which certainly qualifies as playing well. But in those 25 games, they scored 231 runs and allowed 132. Which suggests that even here they should have won more games than they actually did. The Indians had a real knack for piling on additional runs when no more were needed: they won games by margins of 18 runs, 14 runs, 13 runs (twice), 12 runs (three times), and 11 runs (three times.)

The good news for Cleveland? Of the other 10 teams on this list, most bounced back the following year. These 10 teams collectively played .489 ball (768-801); the next year, they played .553 ball (876-708). That's a pretty impressive mark for ten teams who all had losing records the previous year. In fact, only one of the ten would have a losing record the following season (the 1959 Cincinnati Reds). Best of all were the 1965 Twins, who won 102 games, and went all the way to the seventh game of the World Series. Two other teams won more than 90 games; four more won 87-88 games.

So if Eric Wedge can survive the next few weeks... better days are almost surely coming.

Checking in With Pythagoras | 17 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.
Wildrose - Thursday, October 19 2006 @ 09:27 AM EDT (#157033) #
I didn't realize the Tribe's under-achievement, reached such  a historical precedent . It prompted me to read this series at THT,  the author blames defence and the bull-pen( thank God we have B.J. Ryan for one).

Also I note Boston in these type of measurements really over-achieved. It'll be an interesting winter in Beantown. In many respects they have more questions and work (and probably more resources) to do than the Jays.

Magpie - Thursday, October 19 2006 @ 06:41 PM EDT (#157047) #
Does anyone know the predictive validity of Pythagorean records?

I don't know, but I'd guess that they'd be... slightly better than the actual won-loss record? Which no one would use as a basis for making predictions anyway.

What I do think can safely be predicted is that large discrepancies between the Pythagorean projection and the actual won-loss record will not repeat themselves; they tend to be weird random events, that settle upon teams during the course of a season and fly off somewhere else the next year.
daryn - Friday, October 20 2006 @ 08:07 AM EDT (#157054) #

It seems to me that if your win ratio is 48%, and you win 48% of the close games, AND 48% of the blow-outs, that your pythagoean and actual results would be alignment.

So, ... Can one simplify the meaning of the difference results down to some performance metric that represents the teams weakness?

There was a line went, you are going to win 60 and lose 60 and its what you do in the other 40 that determines a winner...   ok, what if we modify that to, "you are going to win some games by 5, and lose some games by 5, but its what you do in the close games that will determine if you out play your pythagorian prediction" 

is that true?? Does the Cleveland result imply that they split the blow-outs but lost more then their share of the close games?

Or is it the reverse, ie  more a matter of "You are going to win about half the close games, but its how many of the blow-outs you are on the top side of that shows up in the the pythagorean",

then the meaning would be "The games they won they dominated and the games they lost were close"  because those two analysises speak of very different teams...

What info would support this analysis?  Is "average difference for win, and average difference for loss" a relevant stat?  or is "total wins 5 run lead or more, total losses 5 run deficit or more" a better metric?

If its a matter of losing too many of the close ones and having a normal percentage of blowout wins, that sounds like a core problem, everyone is just "almost good  enough" on an average day with some great days mixed in.

If its a matter of doing about average on the close ones, and winning more than your share of the blow-outs, that sounds like a team ready to dominate if they just added some LOOGys and ROOGys to hold those close games....

If I've screwed this all up, then write it off to the early hour and lack of coffee... but it seems to me that a difference like you've found, 'means something'

 

daryn - Friday, October 20 2006 @ 08:15 AM EDT (#157055) #

Magpie

I'm sorry, after I posted I went back and re-read your post and thought, "Hey dummy, he DID tie the results into real world performance already",  I got myself muddled up in deep thought, and one disadvantage of this interface is you can't SEE the original post while you are composing the reply.

but I think my commentary stands on it own too, maybe the point is, when a team underachieves its Pythagorean, is there a TYPICAL reason...

Magpie - Friday, October 20 2006 @ 09:10 AM EDT (#157057) #
If its a matter of losing too many of the close ones and having a normal percentage of blowout wins,

Bear in mind that the only teams I've looked at closely is a pretty narrow group - but I would think that's more or less a defining characteristic of the 11 teams (now!) that have outscored the opposition by 50 runs over a season but still had a losing record. (There are numerous other teams who significantly underperformed their Pythagorean expectation, but I was just interested in this select bunch). Anyway. there's probably no other way to achieve that type of result.

I did find, when I first looked at all of this a year ago (there's a link to last year's story up there somewhere), that while all of these teams played well in the blowouts and struggled in the close ones, with individual teams one of these two casues frequently was much more significant than the other. In the case of the 2005 Jays and the 1958 Reds, it was their wretched record in close games. In the case of the 1984 Astros and the 1995 Orioles, outstanding performance in blowouts skewed their runs scored and allowed.

The 2006 Indians share both traits - they played .409 ball in the close games and .640 ball in the blowouts. While in both sets of games, they scored and allowed runs at a rate that would suggest a better win-loss record (something like 21-23 in the close games and 19-6 in the blowouts), the large number of blowouts is probably why their W-L is so far out of line with expectations.

The good news for Cleveland, as it was for all of these teams, is that the ability to beat the other fellow senseless is generally the mark of a quality team. Whereas a poor record in close games is normally the mark of an unlucky team, and luck almost always changes.
Checking in With Pythagoras | 17 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.