2005's "true" White Sox and Indians

In 2005, the Chicago White Sox won the Central Division and the World Series. The Cleveland Indians failed to even make the playoffs. But were the Indians the better team?

At face value, the answer is no. Baseball is about winning games and eventually winning championships. The White Sox won more games than the Indians, and won the championship. They were the better team.

However, available evidence indicates that baseball teams have the ability to score runs and prevent runs; they do not have the ability to win games. Winning (and losing) stems from scoring and preventing runs---no team inherently possesses the ability to win close games (I'm sure I will be receiving a counter-argument from Ozzie Guillen any day now...).

The "Pythagorean theorem of baseball" (so named because it contains several squared terms) predicts a team's wins and losses based on its runs scored and runs allowed. There are deviations from this predicted record, but the theorem remains true on a large scale, and all available evidence indicates that there are no factors other than luck that cause a team to under-perform or over-perform their projected record.

In 2005, the White Sox scored 741 runs and allowed 645. The Indians scored 790 and allowed 642. That's right: the Indians were better at scoring runs, and the two teams were basically even at preventing runs. The White Sox's Pythagorean projected record was 91-71; the Indians' was 96-66.

But wait, there's more. Sometimes actual runs scored and runs allowed are not indicative of a team's ability to score or prevent runs. For example, some teams hit an unluckily high amount of solo home runs based on how many base runners the team had. This results in a lower amount of runs scored than normal. Also, sometimes pitchers can be lucky or unlucky in a similar way.

Baseball Prospectus accounts for these other elements of luck with "Third Order Pythagorean" standings. Once these adjustments are made, here are the projected 2005 standings:

Chicago White Sox 87-75
Cleveland Indians 98-64

Basically, in 2005 the White Sox were incredibly lucky, and the Indians were unlucky.

What does this mean? Obviously, it changes absolutely nothing for 2005---the White Sox are still World Champions, and the Indians still failed to make the playoffs.

What it does change is 2006 projections. Statistical variation (luck) tends to regress to the mean. If you are particularly lucky one season, chances are you will not be as lucky next year. The White Sox's 99 wins (and the Indians' 93 wins) did not represent the true of abilities of their 2005 teams. In other words, if you play the season over again, with the same players producing the same results, Chicago will not be as lucky, and Cleveland will not be as unlucky. The 2005 Indians had a true ability of 98 wins, not 93 wins. The 2005 White Sox had a true ability of 87 wins, not 99 wins.

As I said, this changes nothing for 2005, but does alter projections for 2006. Because the White Sox's "true" ability was 87 wins in 2005, 87 should be the base from which we judge improvements or downgrades. Furthermore, the Indians' base should be 98 wins.

The 2005 White Sox were baseball luckiest team. This luck will not continue.

Is Jake Westbrook the Indians' best starter?

Indians righty Jake Westbrook is an interesting case study in how much luck factors into a pitcher’s season. For all practical purposes, Westbrook’s 2004 and 2005 lines were identical. Westbrook faced the exact same amount of batters each season, gave up the exact same amount of homers, walked three less in 2005 and struck out five more. That’s as identical as two seasons can be. Yet his 2004 ERA was 3.38, while his 2005 ERA was 4.49. He gave up a ton of groundballs each season. Even his batting average per ball in play (BABIP)---much of which is out of the pitcher’s control---was very similar (.276 in 2004 versus .290 in 2005). What did Jake do differently?

For all intensive purposes, Jake himself did nothing differently. His “skills” (homers, walks, strikeouts) were exactly the same. Let’s look further at some of the factors that Jake cannot control.

**Defense: The Indians ranked third in baseball in defensive efficiency. Ronnie Belliard and Jhonny Peralta were among the best at their positions defensively, and Aaron Boone was above average. Ben Broussard was merely average. Overall, the Indians were excellent defensively, especially on the infield. Thus, Westbrook’s higher BABIP is strange.

**Line Drive Percentage: Westbrook ranked 11^th out of 44 qualifying AL starters in lowest line drive percentage (line drives tend to fall for hits more often than regular batted balls). Hitters do not often hit line drives off of Westbrook; this furthers the idea that Westbrook’s amount of hits given up in 2005 is abnormally high.

**Homers per fly ball: Westbrook was the worst of the 44 qualifying AL starters in homer per fly ball ratio (this ratio is park adjusted). 19% of the fly balls that Westbrook allowed (and remember, Jake allowed the fewest fly balls per ground ball of any pitcher in the AL) became homers. This is an extremely high percentage---research has shown that pitchers average an 11% HR/FB ratio, and any deviation is mostly attributable to luck.

**Left-on-base percentage: Westbrook was third worst in LOB%, which measures the percentage of base runners who come around to score. A mere 62.9% of Westbrook’s base runners failed to score (for comparison, both Cliff Lee and Scott Elarton stranded 72% of their base runners, while Kevin Millwood ranked 2^nd in the AL, stranding 79% of his runners).

All of these factors played a large role in Westbrook’s higher 2005 ERA. If we delve even further into Westbrook’s stats, we see that he was the anti-Kevin Millwood in terms of “clutch pitching.” Overall, opposing batters hit .265 against Westbrook. However, with runners on base, batters hit .298 against him. Furthermore, although Westbrook had 130 fewer at-bats against him with runners on base, Westbrook allowed 11 of his 19 homers to occur when runners were on base. With runners in scoring position, batters hit .304 against him. With a runner at third base, the opposition hit an absurd .542. Opposing batters hit .360 with runners at first and third, .353 with the bases loaded, and a crippling .333 with runners in scoring position and two outs. In 2004, opposing batters hit .255 overall against Westbrook, but a mere .233 with runners in scoring position and .200 with runners in scoring position and two outs.

Westbrook is just now entering his prime. We can reasonably expect similar IP, BB, and K in 2006. However, we can also expect less home runs, less hits, and less of the runners who do get on base scoring. These are three very important factors that influence ERA, and Westbrook got very unlucky in all three facets in 2005. Furthermore, we can expect a regression to the mean in terms of batting average in clutch situations as well as the amount of homers that occur with runners on base. Thus, although a return to the 3.38 ERA of 2004 is unlikely, Westbrook has a very good chance of bettering his 2005 ERA, perhaps by a fair amount.

Jeremy vs Jeremy

Indians lefty Jeremy Sowers looks like the real deal. Sowers, the Indians’ first-round pick out of Vanderbilt, began his professional career at high-A Kinston. In 72 innings, Sowers had a 2.78 ERA and a 75/19 strikeout/walk ratio. He also induced 1.65 groundballs per flyball, an excellent amount.

Sowers earned a promotion to AA Akron, where he pitched 82 innings. With the Aeros, Sowers had a 2.08 ERA and a 70/9 K/BB ratio. He induced 1.38 groundballs per flyball---not as good as at Kinston, but still very good.

Sowers looks like he can be a solid #3 starter at the major league level, and he looks to be ready perhaps by mid-2006. But then again, wasn’t the same thing said about another Indians prospect named Jeremy a couple of years ago?

Jeremy Guthrie was the Indians’ first-round pick out of Stanford in 2002. Guthrie, like Sowers, was drafted out of college more for his polish than his ceiling (in other words, the Indians understood that they probably didn’t have an ace, but they thought they had a guy who was not far from being a #3-4 starter in the big leagues).

Guthrie, like Sowers, dominated AA Akron in his first professional season. In 62 innings with the Aeros in 2003, Guthrie had a 1.44 ERA and a 35/14 K/BB ratio. In hindsight, Guthrie’s miniscule strikeout-per-inning-pitched ratio (5.03 K/9) should have been a warning sign, but many (including myself) lauded over the excellent ERA.

Guthrie was promoted to AAA, where he pitched 96 innings. Guthrie stumbled at Buffalo, posting a 6.52 ERA and a 62/30 ratio. Despite his poor performance, many were still high on Guthrie---feeling about Guthrie the way many today feel about Sowers. However, despite sharing a first name, there are telling differences between the Jeremys.

First of all, Guthrie is older than Sowers; Guthrie completed a Mormon-mission before going to college, so Guthrie (24) was older than Sowers (22) when the Indians drafted him. Age is a huge component of minor league development, as a younger player has more time to develop and improve. Secondly, Sowers has struck out more batters per nine than Guthrie at every level. Furthermore, Sowers’ K/BB ratio is a lot better than Guthrie’s ever was (strikeouts per nine and strikeouts per walk are both excellent indicators of future ability for minor league pitchers, even more so than ERA). Finally, Sowers induces a lot more groundballs than Guthrie, meaning that Sowers is a lot less likely to give up homers or extra-base hits.

Overall, Jeremy Sowers and Jeremy Guthrie may look similar on the surface, but once we look deeper, we realize that Sowers is for real, while Guthrie never was.

Coco Crisp: more than just a name

Coco Crisp has had to prove himself at every level. Scouts have never particularly liked him, and his minor league statistics have not been very worthy (not to mention having such a silly-sounding name). Coco was included as the “Player to be Named Later” in the Indians’ trade of Chuck Finley to St Louis (the main “prize” of the trade was Luis Garcia, who I believe is now out of baseball).

Coco hit .301/.365/.428 at age 23 for AA New Haven before he was traded to Cleveland in 2002. Coco’s draw was his speed, as he stole 26 bases (and was caught 10 times). He managed only nine homers---but more disturbingly, he tallied only 16 doubles and one triple. Crisp profiled as a fourth outfielder/pinch runner, at best.

He proceeded to hit .360/.434/.511 in 56 games at AAA Buffalo in 2003. He also stole 20 bases. However, again Coco had only one homer, 19 doubles, and 6 triples. Still, those were very solid offensive numbers, especially for a 24-year-old. Crisp benefited from the Indians’ rebuilding effort, and was called up to the majors. Many believed that Coco’s offensive numbers in AAA were a fluke, and that he still didn’t profile as a regular in the big leagues.

Coco did look overmatched in 99 games, hitting .266/.302/.353, and adding 15 steals. He showed some flashes of being a decent leadoff hitter, and the Indians didn’t have any better outfield options at the time (especially after Milton Bradley was traded). The Indians still viewed him as trade-bait or a 4^th outfielder.

In 2004, the Indians gave Alex Escobar and Jody Gerut the chance to win the center field job. Escobar was terrible, and in May the Indians sent him to AAA, moved Gerut to right, and stuck Crisp in center. Coco continued proving his doubters wrong. He put together a fine season, hitting .297/.344/.446 with 20 steals. However, Coco added a new twist to his skills: he belted 15 homers and 21 doubles.

Even with his fine performance in 2004, many still weren’t sold on Coco as a true every day player. Scouts and fans alike continued to expect Coco to slide into mediocrity, eventually settling in as a 4^th outfielder. As many of you certainly know, Crisp had another excellent season in 2005, coupling stellar defense with a .300/.345/.465 offensive season. He ran less, but was more successful, stealing 15 bases while being caught six times (still far too many caught stealing, though). Crisp hit another 16 homers---the most he’d ever hit at any level. However, in 2005, Crisp belted 42 doubles---18 more than the previous year---while keeping all of his other offensive numbers intact.

Coco has developed surprising power in addition to his batting skills and speed. His dramatic increase in doubles bodes extremely well for his 2006 season, as with his continued maturity and improvement (Crisp is still just 26 and hasn’t yet reached his prime), a lot of those doubles will turn into homers. It is very reasonable to expect 20-25 home runs from Coco in 2006.

The one caveat with Coco is his on-base skills. Throughout his career, Coco has usually hit for a high average, but has not walked that often. This is not a problem, so long as Coco’s batting average remains as high as it has been the last two seasons. While this is certainly likely, batting averages are prone to fluctuate a lot due to luck. If Coco has a particularly unlucky season---dropping his average to, say, .260-.270---Coco’s inability to draw many walks will render him somewhat of an offensive liability, as his OBP would likely be around .310 or so.

Throughout his entire career, Coco has encountered doubters and repeatedly proven them wrong. Soon, he will be known for more than just his name.

Small market teams are teams too

Even before Johnny Damon turned Benedict Arnold, I heard the rumblings. Murmurs around Bean-Town about potential replacements for Damon, should he choose to leave. Corey Patterson, Joey Gathright, Jeremy Reed. Coco Crisp.

Coco Crisp?! Surely this must be a figment of some Sox fan’s wishful thinking. But the more sources I heard the Crisp rumor from, the more I thought it was a possibility. Furthermore, I heard that Bronson Arroyo or Matt Clement would be going the other way.

First of all, this is ridiculous. The Indians have eight starting pitchers and Coco won’t be eligible for free agency for another three years. He’s just entering his prime, plays excellent defense, and has proven himself for two full years at the plate.

While I scoffed at the silliness of the deal from the Indians’ point of view, I was amazed at how it was treated in Boston. A lot of Bostonians used the same points I made above as rationale why this trade would happen: Coco’s entering his prime, has proven he can hit, won’t be a free agent for awhile, and could likely handle center field for the Sox. They failed to realize that it was these exact points that would prevent this trade from occurring.

Many people in big-market cities where the team is consistently good view other, smaller cities simply as farm teams for the big cities. They are right to a small extent---Florida sold off all of their players this off-season (although they made a lot of smart baseball moves…more on this another time). However, Florida is the exception, not the rule---they are facing extenuating circumstances regarding their ballpark. Big-city fans look at a smaller-market team and simply pick off the players they want from that team. Then, they find the players they like least from their own team and decide that’s who will be traded (hence Clement or Arroyo for Crisp).

The only way the Indians will deal Crisp to the Red Sox is if they get blown away by the value they are receiving…in other words, if the Red Sox offer Andy Marte. Other than that, Cleveland has absolutely no reason to trade Crisp, no matter how well he might fit at Fenway. The Indians are not here to serve other teams; rather, they will use their players to beat other teams.

The strange case of Kevin Millwood

After losing out on Matt Clement, Carl Pavano, and Jon Lieber, Indians General Manager signed Kevin Millwood to a one-year contract worth $7 mil prior to the 2005 season. The deal was relatively low-risk (one year contract) for the Indians and filled their void in the rotation. For Millwood, the deal offered a great stepping-stone to a more lucrative contract. As it turned out, both sides got what they wanted.

Kevin Millwood had a lot of bad luck in 2005. He led the entire American League in Earned Run Average, posting a 2.86 ERA. He pitched for a team which won 93 games and scored a lot of runs in the process. Yet somehow Millwood managed only nine wins. Why? Millwood was very unlucky.

In Millwood’s 11 losses, the Indians scored 13 total runs. They scored more than two runs ONCE. Furthermore, Millwood had many other games in which he received a no-decision while receiving minimal run-support.

Teammate Cliff Lee won nine more games despite having an ERA a full run higher. If Millwood had Lee’s run support, he easily would’ve won 22 games.

In terms of wins and losses, Millwood was extremely unlucky. However, in terms of runs allowed, Millwood was very lucky.

In 2005, Kevin Millwood had the uncanny ability to pitch incredibly well with runners on base. Here are some examples:

Overall, opposing hitters hit .251 against Millwood

With the bases loaded, hitters were 0-for-15 against Millwood

With runners on second and third, hitters were 3-for-20 (.150 batting average)

With runners in scoring position and two outs, hitters were 10-for-84 (.119 average)

Overall with runners in scoring position, hitters batted .194 against Millwood

If opposing hitters hit .251 against Millwood with runners in scoring position and with the bases loaded, Millwood would have given up a LOT more runs.

That’s all well and good, but what if Millwood somehow inherently possessed the ability to pitch better with runners on base? What if Millwood was a “clutch” pitcher?

If so, Millwood would have had similar splits in previous years. Let’s see:

In 2004, opposing batters hit .289 with runners in scoring position against Millwood.

In 2003, the average was .274

2002: .232

2001: .282

2000: .297

1999: .241

1998: .255

1997: .270

As we would expect, Millwood’s batting average against with runners in scoring position fluctuates around his overall average, .250. Simply put, in 2005, Kevin Millwood got very lucky to put up a 2.86 ERA.

Given how few runs Millwood gave up in 2005, he was unlucky to receive only nine wins. However, he was very lucky to give up so few runs in the first place.