|
|
|||
|
|
|||
• Wrigley Rooftop Directory
|
Adding Speed to the MixSunday, June 27, 2004I was reading about Rickey Henderson's 130 steal season the other day, and started thinking about the fact that OBP and SLG don't do a good job of explaining the impact of speed on the game. So, I am going to invent three new statistics, OBP-CS, SLG+S, and OPSS (Onbase plus slugging & speed). OBP: On base Percentage = (Hits + BB + HPB)/(AB + BB + HBP + SacFly). I'll modify OBP by subtracting Caught Stealing from the numerator (the top one). If you think about it, if I get a single, and then get caught stealing, its just like getting thrown out by the catcher at first. OBP-CS: On base Percentage adjusted for steals = (Hits + BB + HPB - CS)/(AB + BB + HBP + SacFly). SLG: Slugging % = (Total Bases) / (At bats). I'll adjust SLG by adding stolen bases to total bases, because if a batter gets a single and then steals second, its essentially the same as a double. This should show up in our adjusted slugging. SLG+S: Slugging & Speed % = (TB + SB) / (AB). OPSS: On-base plus Slugging & Speed = (OBP-CS)+(SLG+S) is just like its cousin, Onbase plus slugging, but adds my two new stats. These new statistics should do a better job of factoring in the role of speed in a player's game. Furthermore, if a player is an excellent base stealer, they will be rewarded statistically for not racking up the Caught Stealings. On the flip side, a poor baserunner will cannibalize their OBP-CS each time they are caught stealing. Stats in action: Rickey Henderson Career :
Babe Ruth Career:
Barry Bonds Career through June 26, 2004:
Ryne Sandberg Career:
Vince Coleman Career:
As you can see, factoring speed into the "holy grail statistics" for Sabermeticians (OBP, SLG, OPS) allows speedsters to be evaluated on an even playing field with the big boppers. Guys like Rickey Henderson and Vince Coleman who have stolen significant numbers of bases, with a high success ratio are deeply undervalued by the OPS measuring stick. Henderson and Coleman both have 100+ points added to their OPSS scores if you take into account their speed. This is essentially the difference between Mark McGwire (.982 OPS) and Bernie Williams (.882 OPS) or the difference between Barry Bonds (1.035 OPS) and Albert Belle (.933 OPS). If there is already a similar statistic out there that factors in speed, I'd love to know about it, and I will continue to work on calculating some more OPSS when I have a bit more time. UPDATE: MUCH LATER Posted by Byron at June 27, 2004 10:45 PM
|
|
||||||||||
| Subscribe to The Cubdom - get emails with the latest Cubs info and pictures | |
This post has been tagged: |
|
It seems to me that CS also has to be factored into the SLG side of the formula, since CS outs should logically impact SLG as well as OPS. I would suggest using (TB + SB - CS) / AB for SGL+S. It is true that this formula underestimates the value of runners advanced two or more bases by batters who single (or double) and are subsequently caught stealing. On the other hand, it overestimates the value of stolen bases--which don't advance other runners--by adding them directly to total bases. My suggested formula allows these two factors to cancel each other out, though it still probably somewhat overvalues stolen bases (from a sabermetric perspective).
The benefits of using my formula are easiest to see if we use the case of Babe Ruth. Since his stolen base percentage is so low (51.25%), it makes intuitive sense for his OPSS figure to be lower than his OPS (1.164), not higher as it is with the original formula (which yields an OPSS of 1.167). With the adjustment I suggest, his OPSS will be 1.153, which makes more sense.