Breaking 100: An Historical Review

By Dan Fox

Breaking 100…hmmm….as I thought about this blog and its quest to see the Royals wind up with only 99 losses, what immediately came to mind was the historical question as to how frequently does a team that loses 100 games in one season loose fewer than 100 the next? And what might that mean for the Royals?

Some quick queries using the excellent Lahman database yielded the following answers:

Since 1901 there have been 129 teams that have lost 100 games. The average winning percentage of those teams was .334 equivalent to a 54-108 record in a 162 game season The most recent 100-loss seasons (since 2000) include:

Year	Team	Lg	W	L	Pct
2001	TBA	AL	62	100	0.383
2001	PIT	NL	62	100	0.383
2002	TBA	AL	55	106	0.342
2002	DET	AL	55	106	0.342
2002	KCA	AL	62	100	0.383
2002	MIL	NL	56	106	0.346
2003	DET	AL	43	119	0.265
2004	ARI	NL	51	111	0.315
2004	KCA	AL	58	104	0.358
2005	KCA	AL	56	106	0.346

Of the 128 teams 35 of them or 27% lost 100 games the following year. That seems like pretty good news for Breaking 100 since historically it would appear that losing 100 games one season does not mean that you’ll lose 100 the next. Excellent news for Breaking 100.

Hold on though…

In the following season for those 128 teams (since we won’t include the 2005 Royals) the teams averaged a winning percentage of .405 or a 66-96 record over 162 games. So teams that lost 100 didn’t exactly become world beaters. Probably the largest single explanation for the 12 game improvement is the “Plexiglas principle”.

This principle as defined by Bill James in the 1982 Baseball Abstract says that

“all things in baseball have a powerful tendency to return to the form which they previously held. If a player's batting average jumps in one year, it will usually decline in the next. If his HR total drops sharply in one season, bet on him to improve it the next."

Essentially this means that players and teams tend to regress towards the mean because the element of chance tends to even out. Teams that lose 100 games do so not only because they’re bad, but also because they’re unlucky. In 2005 the Royals scored 21 fewer runs than predicted by the sum of their offensive elements, gave up 19 more runs than would have been expected given their defensive statistics, and won 4 fewer games than would have been expected given their ratio of runs score to runs allowed. In total, when evening out these factors (which can loosely be correlated with luck) the Royals “should” have won 64 games (http://www.hardballtimes.com/main/article/second-look-at-luck/).

The other reason that teams who lose 100 games one season usually don’t do so the next is simply that losing 100 games comes close to the wall of how many games a team can reasonably lose and still be in the major leagues. Most analysts agree that even if you had a team populated with replacement level players (players you can sign for the league minimum and that are therefore “freely available” talent) you would still win somewhere between 40 and 50 games. Simply having a handful of major leaguers above replacement level gets a team to 55 or more wins.

That said, the Royals of 2005 were among the 27% as they went from 104 losses in 2004 to 106 in 2005.

But we’ll end with a happier note and list those teams that performed the best after their 100 loss season.

Year	Team	Lg	W	L	Pct
1905	PHI	NL	83	69	0.546
1967	CHN	NL	87	74	0.540
1989	BAL	AL	87	75	0.537
1974	TEX	AL	84	76	0.525
1918	PIT	NL	65	60	0.520
1986	CLE	AL	84	78	0.519
1980	OAK	AL	83	79	0.512
1986	SFN	NL	83	79	0.512
2003	KCA	AL	83	79	0.512
1947	PHA	AL	78	76	0.506
1963	CHN	NL	82	80	0.506
1962	PHI	NL	81	80	0.503
1909	NYA	AL	74	77	0.490
1971	CHA	AL	79	83	0.488
1990	DET	AL	79	83	0.488
1997	DET	AL	79	83	0.488
1994	NYN	NL	55	58	0.487
1988	CLE	AL	78	84	0.481
2005	ARI	NL	77	85	0.475
1992	CLE	AL	76	86	0.469

What you see here is that a best case scenario for any team would be around 85 wins but the 2003 Royals (“We Believe!”) came in 9th.

Let’s go get ‘em!