Inside the SABR Analytics Conference — Part 4

SABR metrics vs. traditional stats — which can we trust more?

When it comes to baseball statistics, which can we trust more: traditional stats or SABR metrics? This question has been debated ever since Bill James coined “SABR metrics.” WAR (wins above replacement) or homers? BABIP (batting average on balls in play) or RBIs? LIPS (late-inning pressure situations) or saves?

Traditional stats can have gaps in them. Take Yankees pitcher Phil Hughes’ 2010 season, for example. Despite having a rather high ERA of 4.19, Hughes “earned” 18 wins. This contradiction probably happened because his offense gave him and the rest of the starting rotation tons of offensive support that year, which led the Yankees to an ALCS appearance against the Texas Rangers.

But those stats can work together with sabermetrics information to present a more complete picture of a player. "I've always believed the best way to assess talent is through sabermetrics, traditional stats and scouting,” Vince Gennaro, the president of SABR, told me. “They all contribute to giving us a complete picture of the true talent of a player. Sabermetrics tend to dig a bit deeper by uncovering what's behind a player's performance. Traditional stats tend to tell you 'what' a player did, while sabermetrics tend to tell you 'why'."

Players can benefit from the saber stats by helping their teams answer the “why” question. For example, the Pirates’ Andrew McCutchen has helped Pittsburgh stay near the top of the NL Central this season. But his BABIP is well ahead of the average and shows he’s no everyday centerfielder. It also helps explain why he’s so valuable to the Pirates.

But sabermetrics can also work against players – badly. Take the example of Derek Lowe. Lowe is one of only six players in MLB history to participate in a no-hitter and record at least 40 saves in a season. He is also one of four players out of the previous six to pitch the entire no-no instead of coming in as a reliever. Lowe’s WAR was very low this season, so Texas GM Jon Daniels must have thought that stats showed more than career accomplishments – Lowe was released by the Rangers in May.

Working off of traditional stats, McCutchen might be considered average and Lowe might still be pitching for the Rangers. But when sabermetrics are factored in, their careers can look the opposite. Sabermetrics and traditional stats might work best together, but taken by themselves the pictures they paint of players can be very, very different.

Do you have an opinion about this topic? Do you think one is more helpful than the other? Leave your thoughts in the comments! 

Max Mannis is an 11-year-old special correspondent for sikids.com and a member of SABR. Catch his posts on advanced baseball statistics. To learn more about SABR and to join, visit www.sabr.org.