The world of baseball statistics has come a long way. Not very long ago, for position players, fans looked at batting average, home runs, RBI, and fielding percentage. For pitchers, it was even easier. ERA, wins and losses, walks/strikeouts, and hits/innings pitched were the primary analytics. Now, we have “advanced” statistics, such as OPS, WAR, BABIP, WHIP, and UZR. The newer statistics allow for even deeper evaluations of players, and help in assessing and predicting their relative values to their teams. However, here’s a cautionary note on the use of statistics. “Thin-slicing”, or looking at small samples to make projections or “substantiate” an opinion, can be problematic.
Jun 29, 2013; Pittsburgh, PA, USA; A major league baseball and glove sits in the bat rack in the Pittsburgh Pirates dugout before the game against the Milwaukee Brewers at PNC Park. Mandatory Credit: Charles LeClaire-USA TODAY Sports
Here is case study number one. There is a former Met, who was off to a fine offensive start in 1999, his fourth season in the major leagues. This player, through June of that year, had a batting average of .292 with a .339 OBP. He was coming off of three seasons in which his batting averages were .257, .216, and .246.
It certainly seemed as if this player was hitting his stride, and it would have been easy to use his hot start in 1999 to demonstrate that he was on his way to being a solid offensive player (despite his previous three seasons). His productive three months could have also been used as a reason for patience with the player if he fell on hard times in the future (which he did). The player is Rey Ordonez, and there aren’t many Mets fans who would have thought of him as a good offensive player. However, not looking at his work as a whole, or not looking at his aggregated trends, and focusing solely on the beginning of his 1999 campaign could have led to some incorrect conclusions about his offense. Ordonez went on to hit .246 with a .289 OBP over his 9-year career.
For case study number two, we’ll discuss a former Mets pitcher who posted a 15-10 record with a 3.56 ERA in his sixth major league season, when he was 25 years old. It would have appeared that this pitcher had arrived, and was poised for a solid career as he continued to mature. It may also have made some sense to sign this pitcher to a lucrative, long-term deal (which the Mets did). You may have figured out that this pitcher is Oliver Perez. Perez has a career record of 62-75, with a career ERA of 4.53. Taken out of context, his 2007 season, along with his age, would have implied a different career trajectory. Looking at his seasons before 2007, Perez was 34-53. However, taking 2007 in isolation, one could make faulty assumptions about Perez.
The point is that with the availability of more statistics, it’s tempting to “thin-slice”, or look at small samples and make inferences or draw conclusions. We could easily look at any isolated 300 at-bat span of a position player’s career to make a point. Players have good and bad stretches. To me, the most important thing to do is to look at the entire body of work whenever possible, which will take out sampling and personal biases. Look at trends, and the correlation of numerous statistical measures. Rather than cherry picking favorable or unfavorable sets of data, try to see the complete picture.
In any case, it’s always fun to talk baseball, in thin or thick slices. But here’s a challenge. The next time someone tells you about how a player did over a three-month stretch a few summers ago, ask about that player’s career numbers. Ask about that player’s recent statistical trends. Most importantly, enjoy baseball debates, and enjoy the fact that we’re three weeks from pitchers and catchers.