Understanding Sabermetrics (17 page)

Read Understanding Sabermetrics Online

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

Two of the most glaring figures absent from the BRC formula are the stolen base and the caught stealing. There are other elements missing (HBP, for example) that will be considered in other versions of the formula. However, the formula will always take on the basic form of
where A represents the on-base portion, B represents the advancement portion, and C represents opportunities.
Thus, to incorporate stolen bases, we first must consider the impact of a failed stolen base attempt, caught stealing (CS). It seems that CS statistics needs to be subtracted from the on-base portion, but added to the opportunity portion. Thus, A = H + BB - CS, while C = AB + BB + CS. Now, to consider the advancement value of a stolen base, James decided not to count a stolen base as being equal to a base gained via hitting. Instead, B = TB + (0.7 × SB). In his
1982 Baseball Abstract
, James gave two reasons for this adjustment:
1. A stolen base advances only one runner; each base hit advances the batter as a runner and anyone else who happens to be aboard.
2. If you used 1.0 at the value for each SB, the formula would no longer work; it would no longer predict actual run totals.
Let us apply this new version of the runs created formula, which we will call Stolen Base Runs Created (SBRC), where SBRC =
to the National League in 1986. Table 8.2 shows the league numbers with the stolen base considerations.
 
Table 8.2 1986 NL totals, including stolen base considerations
 
 
This adjusted number, is only one quarter of one percent, or 23 runs, higher than the actual total runs scored that year incredibly close, given that there are only 6 elements to the formula.
It is easy to see that with some refinement the formula could be made even more precise; however, not every statistic is available for every season. Nonetheless, if one has access to GIDP, IBB (intentional walks), SH (sacrifice hits), SF (sacrifice flies), and HBP (hit by pitch), then these can be incorporated into
as follows:
• A = H + BB - CS + HBP - GIDP
• B = TB + [0.26 (TBB - IBB + HBP )] + [0.52 (SH + SF + SB)]
• C = AB + BB + HBP + SH + SF
 
James referred to this formula as the technical version of runs created, which we will call TECHRC. The added statistics for the 1986 NL were obtained from the
Stats Inc. Handbook
:
 
Table 8.3 Additional factors for the NL 1986
 
So, for RCTECH,
A = 16643 + 6560 - 858 + 312 - 1360 = 21297
B = 24997 + [0.26(24977 - 803 + 312)] + [0.52(869 + 519 + 1842)] = 33043
C = 65730 + 6560 + 312 + 869 + 519 = 73990
and
which is within four tenths of one percent of the actual run total. It is unusual that the formula that uses less data turns out to be more accurate than the more involved model.
There are now twenty-four versions of the runs-created formula based largely on the statistics available for a particular season. They are all listed in the
Stats Inc. Handbook
.
For the American League’s 1923 season, the H, BB, AB and TB stats are all available:
 
Table 8.4 1923 AL totals
 
so
, within 5 percent of the actual runs total, 5887.
In addition, HBP, K, SB and CS are available, and players were then credited with a sacrifice hit any time a fly out advanced a runner, not just those on third, so that will affect some of the factors. The same data was available for the NL that year, and this formula, HDG12, is used for both leagues in 1915 as well as 1920-1925. HDG12 has the following formula:
A = H + BB + HBP - CS
B = 0.98(TB) + [0.25(BB + HBP)] + [0.46(SB + SH)]
C = AB + BB + HBP + SH
 
The raw data for the AL in 1923 is presented here:
 
Table 8.5 Additional 1923 AL totals
 

Other books

InkStains January by John Urbancik
Ask Me Again Tomorrow by Olympia Dukakis
All Up In My Business by Lutishia Lovely
Domes of Fire by David Eddings
Twelfth Krampus Night by Matt Manochio
A Vampire's Honor by Carla Susan Smith
Swimming with Sharks by Neuhaus, Nele