Read Understanding Sabermetrics Online

Authors: Gabriel B. Costa,Michael R. Huber,John T. Saccoma

Understanding Sabermetrics (19 page)

2. If you were told that the American League in 1998 had 78,416 at-bats, 21,276 hits, 4248 doubles, 408 triples, 2499 home runs and 7737 bases on balls, how many runs would such a league be expected to score? How does this compare with the league’s actual runs total (see question 1)?
3. Bobby Murcer in 1971 had the following statistics:
 
Calculate his BRC, SBRC, and TECHRC.
4. In the questions following chapter 4, you were asked to compare the contenders for the 1983 AL MVP award, Orioles teammates Cal Ripken Jr. and Eddie Murray, by their Linear Weights scores. Using HDG23, and the correction,
determine who had the better offensive season.
 
5. The Seattle Mariners won an amazing 116 regular season games in 2001. Calculate the expected number of wins based on the Pythagorean Record (with
x
= 1.82), using their RA and RS at the end of each month of the season. The data is listed below:
 
What happens when you change the value of the exponent (
x
) to be exactly 2?
Hard Sliders
 
1. As in chapter 4, the table below lists the statistics for the top five home run hitters of all time, for the season in which each won his first Most Valuable Player Award: Hank Aaron in 1957, Barry Bonds in 1991, Babe Ruth in 1923, Willie Mays in 1954, and Sammy Sosa in 1998.
a. Calculate runs created using the appropriate formula. For Aaron, Bonds and Sosa, use HDG23, for Mays, use HDG22, and for Ruth, use HDG12.
b. Calculate the number of outs for which each player was responsible, and then calculate the number of “games” for each. In the case of Bonds, Sosa, Ruth and Aaron, use a divisor of 27 for games. For Ruth, use 25.5, as all outs are not accounted for in the available statistics.
c. Divide the runs-created total in a. by the games total in b. to obtain RC/27outs figure. How do the players rank?
 
 
2. In question (2) above, you were asked to calculate Bobby Murcer’s runs-created totals. Calculate his RC again, this time using HDG23 and the correction
 
Given that the AL in 1971 averaged 3.87 runs per game, calculate Bobby Murcer’s offensive winning percentage.
 
3. Mark McGwire was voted the American league’s Rookie of the Year in 1987 when he hit a rookie record 49 home runs, drove in 118 runs and slugged .618 Compare Big Mac’s runs created using both the BRC, the RCTECH, and the HDG23.
Inning 7: Win Shares
 
Shortcomings in Runs Created
 
Honest to a fault, Bill James recognized some flaws in his runs-created system and he expounded on them in
The New Bill James Historical Baseball Abstract
. Among the most egregious, in his opinion, was the need for the number to be placed into a team context. Recall that the BRC formula
accurately predicts a team’s runs to within 5 percent, and the adjustments for statistics available in a given season usually moves the estimate to well within 1 percent.
James used the following examples to illustrate what he asserts was one of the shortcomings: Player A and B play on Team 1. Each has 400 AB + BB, 150 TB, and 150 H + BB, for BRC = (150)(150) / 400 = 56.25, or a total for the two of them of 112.25, which matches the sum of their accomplishments 800|300|300 for BRC = (300)(300) / 800 = 112.5.
However, consider Player C and Player D on Team 2. Player C has 400 AB + BB, 100 TB, and 100 H + BB, or BRC = (100)(100) / 400 = 25. Player D has 400 AB + BB, 200 TB, and 200 H + BB, with BRC = (200)(200) / 400 = 100. Their total BRC as a sum of the individuals is 125, but their aggregate totals match the players from Team 1, 800|300|300 and BRC = 112.5. If at all possible, the whole should equal the sum of its parts.
Another observation from James is that, if two players each steal 10 bases, but one is a better hitter than the other, the runs-created formula will compute that the stolen bases created more runs for the better hitter, simply because the multiplier is larger.
To correct these “problems,” the adjustments for the runs-created formula essentially creates a typical team, figures out how many runs the team would create with the player, how many runs the team would create without the player, and subtracts one from the other.
Recall that the correction is:
The numbers for Players A and B remain unchanged by this correction, but the numbers for the Team 2 players are given in Table 9.1.
 
Table 9.1 Runs created with adjustment
 
Thus, the Team 2 players’ individual total adds up roughly to 114, and James contends that the adjustment eliminates the problem of interaction between the run elements.
Other adjustments in the runs-created formula deal with situational statistics. In James’ opinion, the two that have the greatest impact on runs scored are batting average with runners in scoring position and home runs with men on base (as opposed to bases-empty HR). Thus, these are brought into the RC formula.

Other books

The Cloud Maker (2010) by Patrick Woodhead
Beautiful Americans by Lucy Silag
Prisoners in the Palace by Michaela MacColl