What Went Wrong with the Tribe? (Part 5)by J.P. McIntyre
November 10, 2006
Serious students of baseball often treat baseball statistics as a puzzle in which everything eventually fits nicely. When used properly, baseball stats can do just that—or at least give us an almost complete picture with a few pieces missing. The Technological Age has spawned a great deal of nearly completed puzzles that offer baseball fans many answers, often obtainable through a few mouse clicks. As a result, baseball fans, both serious and casual, have grown accustomed to receiving answers in compact packages.
The 2006 Cleveland Indians did not offer answers for their disappointing season in neat packages. There were a great deal of things inherently wrong with the team, and there is no simple answer that will offer a magic pill to make things immediately better or make what transpired easily dismissed. That is not to say the team is a long way from being good; many important pieces are in place for the Indians to be successful in the future. However, unless the pieces are successfully complemented, the results could be similar, which is something Tribe fans do not want to hear.
A "Where It Went Wrong" article employs hindsight, but I want to make it clear that I am not using hindsight to point blame (although, in some cases, blame should be assessed), but to evaluate what the shortcomings of the team were. To dismiss the season as a fluke is to bury one's head in the sand and run the risk of the same results.
This year, the Indians finished eleven games under what could be expected from their Pythagorean Record, which is derived from Run Differential (runs scored minus runs allowed). Before we get into the reasons why, allow me to iterate some important facts because the Pythag might be the most misused tool in baseball analysis. First off all, for a full season of games, the standard deviation in the team's W-L record can be estimated as half of the square root of 162. The square root of 162 is about 12.7, so the standard deviation for a full season is about 6 wins. In other words, any difference of six wins or less from the Pythag and actual record could very well just be random statistical "noise." Also, the Pythag can under-predict home wins.
Secondly, it is a truism that substantial standard deviation from the Pythag is correlated to unexpectedly high or low success rates in close games. Often, success rates in close games are associated with:
1. Bullpen effectiveness or ineffectiveness
2. Tendency (plus or minus) towards "clutch" performance, especially scoring late in close games
3. Managerial in-game acumen
Points one and two can be measured with statistical data, but are by no means absolutes. Yes, “clutch” is an elusive concept to measure, but we can certainly note when a team does better than expected in “clutch” situations over a particular season. Point three may or may not be true; no one has invented a good way to evaluate managers' performances. It certainly makes sense though. A good or bad bullpen doesn't necessarily equate to large differentiation of the Pythag; the run support the bullpen gets plays a large factor. A bad pen that is getting strong run support often is bailed out. Let's take a look at Run Support and Runs Allowed in relief and "Close and Late" (C&L) situations:
RS C&L RS REL* RA/9 IP* RA/C&L RS/Starters MIN 6.39 4.47 3.18 2.39 5.24 TEX 5.36 4.01 4.18 3.91 5.99 NYY 5.29 4.20 4.57 3.87 6.67 BOS 5.26 4.38 4.92 4.16 5.52 KC 5.27 4.83 5.74 4.65 4.74 CHI 5.17 4.18 4.78 3.53 5.86 BAL 5.25 4.04 5.60 3.65 5.34 TOR 5.19 4.36 4.37 3.20 5.53 TBR 5.01 4.61 5.51 4.41 4.23 SEA 4.39 4.88 4.24 3.29 4.61 LAA 4.38 3.59 4.23 3.10 5.29 CLE 4.31 4.30 5.04 4.25 6.01 DET 4.17 4.14 3.95 3.46 5.58 OAK 3.60 3.75 3.87 3.01 5.26*all relief situations
The Yankees, Red Sox, and Royals had lackluster bullpens overall, but those pens received very healthy Run Support and did not have negative Pythags. The Indians displayed a disturbing pattern of scoring a great deal early (their RS for starting pitchers was second in the AL), then not scoring late as the bullpen gave up a great deal of runs. In other words, the Indians often scored early, only to watch their lead disappear while the offense was often incapable of answering the comeback. While the Indians' bullpen was extremely problematic, the A's, on the other hand, had terrible Run Support late in the games, but their pen was excellent, and their Pythag was +8. The Indians were 18-26 in one run games; the A's 32-22. Expand that to close games, and the Indians were 29-40 while the A's were 47-37.
So what happened to the Indians' offense late in the game? How did the second highest scoring team in the AL overall drop to the thirteenth in Close and Late Situations? One reason was the Indians' strikeout rate increased dramatically in these situations, jumping from 18% to 24% (League Average in C&L was 19%). Still, that increase doesn't account for all of the drop in scoring.
The Indians weren't very good at manufacturing runs, something that is very important in low-run scoring environments. In the 2007 Bill James Handbook, a system for counting Manufactured Runs is introduced. Essentially, James' methodology revolves around the general definition that a manufactured run is a run "that is at least one half of the offense doing something other than playing station-to-station baseball." The methodology is very detailed, so you should click on the link and purchase the book if you're interested in the details. The Indians were tied for 11th in the AL with the Blue Jays with only 139 Manufactured Runs (MR). In comparison, the Twins led the majors with 224 MR, and the AL Average was 163.
Manufactured Runs don't necessarily correlate to team success; the Tigers were last in the AL in MR. However, the Tigers, A's, and the Indians, the bottom three of RS in C&L, were all poor run-manufacturing teams. In the Indians' case, it appears they didn't manufacture a great deal of runs because they did not play a great deal of "small ball." The Indians were near the bottom of the AL in Sacrifice Bunt Attempts, Stolen Base Attempts, and putting runners in motion (another cool thing that is tracked in this year's Handbook), plus they were near the bottom in Sacrifice Flies and were not the type of team that strung a bunch of singles together to score.
This is not necessarily a bad thing; after all, the Indians were second in runs scored. However, in a low scoring environment late in close games, this inability to manufacture runs bit them about eight inches below their numbers. Plus, the Indians appeared to shy further away from small ball in C&L situations. They only attempted eight stolen bases in C&L, at an abysmal 50% success rate. Only thirteen of their 30 successful sac bunts came in C&L situations.
The lack of the Indians' small ball, while a weakness in low-scoring environments, is not a criticism of strategy. The team that was assembled was not made to play small ball; it was rather slow and not filled with a great deal of high contact hitters. Attempting to play small ball with a team that lacks those skills is a recipe for disaster. Plus, when the Indians did play small ball, they did it with a modicum of success in certain areas. They were above average in stealing and successful sac bunt attempts, so manager Eric Wedge either chose when to employ small ball tactics wisely or was fortunate.
What the Indians were not was a good base running team. Bill James improved upon his base running metric this year in the Handbook. While this metric is by no means perfect, we can use it to examine another weakness of the 2006 Tribe. James’ metric measures a few things for guys who reached based fifty or more times: runners going from first to third on a single; scoring from second on a single; scoring from first on a double, bases taken (advancing on defensive miscues); base running outs, and runs scored as a percentage of times on base. The last category is extremely context driven; teammates and order of the line up greatly influence this. Here’s how the Indians fared (M stands for "Moved" and C stands for "Chances" to move):
1st to 3rd 2nd to Home 1st to Home Player OB Scored M C M C M C Sizemore 271 39% 12 37 25 29 1 3 Michaels 188 36% 8 17 12 17 7 11 Belliard 193 25% 11 23 8 19 2 3 Hllndswrth 65 30% 1 6 5 5 0 1 Boone 127 33% 4 14 7 9 1 8 Choo 72 27% 2 4 5 8 0 0 Broussard 151 26% 7 21 11 15 1 2 Peralta 209 33% 2 20 10 25 3 11 Blake 166 26% 4 24 14 17 4 5 Perez 66 19% 1 9 5 8 0 0 Garko 74 28% 0 10 2 8 0 2 Hafner 212 27% 3 7 4 6 1 1 Martinez 261 25% 3 34 9 20 1 9Here's the rest of the info. O.A. stands for the number of times a baserunner was out attempting to take an extra base on a hit. Doubled off is when the runner is caught off base on a ball hit in the air, and "BR" is the number of time a runner was out attempting to advance on a wild pitch, passed ball or sacrifice fly. The final Rank column takes all of these elements into account.
Player Bases Taken O.A. Doubled Off BR Rank Sizemore 29 0 6 6 19 Michaels 14 0 1 1 15 Belliard 18 1 1 2 4 Hllndswrth 4 0 0 0 3 Boone 9 1 1 2 -1 Choo 4 0 1 1 -1 Broussard 14 1 3 4 -2 Peralta 15 1 0 1 -3 Blake 7 1 1 2 -4 Perez 2 0 0 0 -4 Garko 3 0 0 0 -3 Hafner 8 2 3 5 -20 Martinez 15 1 2 3 -20 Total 142 8 19 27 -17Grady Sizemore and Jason Michaels did very well under this metric. Of course, guys at the top of the lineup fare better under this metric, but Sizemore ranked twelfth overall in the majors, despite being doubled up six times. Sizemore ranked first in the majors with guys over twenty chances in scoring from second on a single. However, most of the Indians did not fare well on this metric. The Indians were doubled up nineteen times, a figure that seems very high. I would venture most of those happened when Wedge put runners in motion, indicating bad luck or foolish strategy. Also, we don’t know how many times putting the runners in motion benefited, so we can’t tell if the Indians experienced enough success outweighed the negative. Travis Hafner and Victor Martinez were detriments on the base paths, which is not that surprising considering their speed.
Interestingly, Bill James notes that David Ortiz, another slow footed slugger, fared decently with this metric, a zero ranking, much better than Hafner. James notes that Ortiz is “slow, but he is alert; he reads the ball well off the bat; he hustles, and he knows what he can do.”
The Indians lost a great deal of close games because of a combination of their bullpen’s inability to prevent runs and the Indians’ offense inability to score runs late. Greatly attributing to the Indians’ late game offensive woes was their inability to manufacture runs and run the bases in an advantageous manner. This was more of a result of the construction of the team rather than strategy. The Indians could use some more speed and smarts on the basepaths, something that will hopefully be addressed this offseason.
References and Resources
References: 2007 Bill James Handbook; and my good friend, Ken Rhodes, who explained standard deviations of the Pythag in a language that is accessible.
J.P. McIntyre is a transplanted Ohioan residing in California and authors the blog Many Go; Few Understand and co-authors Reds and (Blues).com. He welcomes comments via email.