June’s Bonehead Sports Analyst: Tim McCarver, Fox Sports
By Keith Glab
In a nationally televised contest between the Philadelphia Phillies and the Atlanta Braves, Fox Sports’ Tim McCarver decided to engage us in rhetoric regarding how misleading On-Base Percentage is in terms of gauging an effective offense. He cited the fact that Philadelphia led the NL in OBP, but was 11th in Runs Scored at that time. Using this as the sole evidence for his argument, he proceeded to claim that you have to be very careful when looking at a misleading statistic such as OBP, whereas good ol’ Batting Average, Home Runs, and RBI’s don’t require any further research.
What Tim failed to mention was that OBP correlates to Runs Scored much better than Batting Average does. In their outstanding book, Curve Ball, Jim Albert and Jay Bennett test Batting Average, Slugging Percentage, and On-Base Percentage for the one that best correlates to team run production for every season between 1876 and 2000. Batting Average only had the lowest Root Mean Squared Error (or best correlation to a team’s Runs Scored) in 16 of the 124 tested seasons (13%), while On Base Percentage was highest in 58 of them (47%). "The probability of this happening if AVG, SLG, or OBP were equally effective measures is less than 1 in 5 million," writes the duo. But McCarver does not merely state that they are equal predictors of run production; he has the audacity to claim that Batting Average, which hasn’t had the lowest RMSE since 1955, was better than OBP!
Tim also forgot that while team RBI’s by definition correlate well with team Runs Scored, that an individual’s RBI totals are more indicative of where a player hits in the lineup, and the OBP’s of the players hitting in front of him than any measure of skill for that player. Hack Wilson’s 191 RBI’s in 1930 was largely a product of his hitting cleanup behind Woody English (.430 OBP), Kiki Cuyler (.428 OBP), and either Riggs Stephenson (.421 OBP) or Gabby Hartnett (.404 OBP). Bill James writes in his revised Historical Baseball Abstract that "Given Wilson’s hitting stats and the hitting stats of the other players on the 1930 team, we would expect him to drive in 190-195 runs, even assuming that he has no tendency to hit in the clutch." I wish that James had provided his methodology, but inductively, this statement makes a lot of sense. Assuming that clutch hitting does exist, an RBI total might not reflect it. The eighth place hitter on the 2004 Diamondbacks might not amass 100 RBI’s even if he hit over .500 with runners in scoring position because he doesn’t appear in that situation very often. Therefore, Runs Batted In is a statistic that requires much more context than OBP, contrary to McCarver’s claims.
As for Home Runs, well, they do correlate reasonably well with Runs Scored. However, since 2000, the team with the league’s best OBP also led the league in Runs Scored seven times, while the team with the most homers led in scoring just once over that span (The 2003 Braves, who were second in team OBP (.349) to St. Louis (.350)). It seems to me that if you’re interested in scoring runs, getting on base is the best place to begin.
That thought still hasn’t occurred to our pioneer Bonehead in the Broadcast Booth.
Related Content: Take a Pitch; A Walk Can Win a Game