My postseason awards

I guess we all have our own approach to the MVP debates, but I like how Joe Posnanski recently grouped things into two approaches: the “narrative” approach and the “best” approach. The official MLB guidelines don’t really get into this type of nuance, but sportswriters are paid to find stories on (and off) the field, so my guess is that many of them are partial to the “narrative” approach.

Have you ever noticed that the MVP voting tends to differ from the Hall of Fame voting? MVP awards are more likely to go to slugging first basemen but, when it comes to inclusion into the Hall, voters are more likely to consider overall skills and position. It’s interesting, isn’t it? The “Fame” votes tend to follow the “best” approach while the “Valuable” votes are more often influenced by the “narrative” approach. Talk about confusing.

Unlike most sabermetric types of baseball fans, I am a fan of the narrative approach. Carl Yastrzemski was terrific in 1967, but it was his role in that incredible pennant race that really cemented his MVP and made his year memorable for himself, his team and his fans. That was a year in which both approaches—the narrative and the best—really came together. Unfortunately, things don’t always work that way.

Take Don Baylor in 1979. Baylor had 139 Runs Batted In (his second-best season total was 99) and the Angels finished first, so Baylor won the MVP. But you can make a very good case that Baylor wasn’t even the best player, from a narrative or value perspective, on that team. The problem isn’t the “narrative” approach, it’s the stats we use to capture the narrative.

Enter Win Probability Added (WPA). You may know that WPA takes every event on the field and attributes a positive or negative “win impact” to that event, based on relatively simple win expectancy tables. That’s the short version (here’s the long one), but I ask you, why don’t all those narrative lovers use WPA? It is, after all, the “story stat.”

Don Baylor didn’t lead the Angels in WPA, not even close. Why? Because he batted .256 in “high leverage” situations and .328 in “low leverage” situations. Many of his RBI’s came in non-crucial situations.

Compare that to, say, Bobby Grich (full disclosure: Grich is one of my all-time favorite players). Grich batted .327 in “high leverage” situations and .287 in “low leverage” situations. As a result, his WPA was 2.9, two wins (each “win” is half a WPA) better than Baylor. Shoot, Dan Ford had a 3.9 WPA, two more wins better than Grich. Grich was 8th in MVP voting. Ford was 26th. These votes didn’t really reflect the “narrative” of the season because sportswriters paid attention to the wrong stat.

Which leads me to my selections for this year’s National League MVP and Cy Young (not that I get a vote. I’m just sayin’). According to Baseball Reference, Prince Fielder led the league in batting WPA at 7.7. He batted .305 /.411/.648 in high-leverage situations. Matt Kemp, who appears to be the frontrunner for the MVP, had a WPA of 6.2. Pretty good, but Fielder’s was better and his team was involved in a pennant race which, in my book, is an important part of the narrative.

Now, you may say that WAR doesn’t agree, and you’re right. But WAR is a “best” stat, not a narrative stat. Simply put, who cares about replacement level when you’re trying to capture a narrative? Personally, I don’t. A batter is a batter, regardless of what position he fields. If a batter delivers in high-leverage situations and during a pennant race, that’s what gets my attention.

You may also say that WPA doesn’t capture fielding, and you’re right again. Fielder is, at best, an average fielder at a “low impact” fielding position. If you want to say that Kemp closes the gap of 1.5 WPA through his fielding, I won’t disagree. You may be right. But I’ll still say that Fielder was in a pennant race and Kemp wasn’t. He still gets my vote.

As for the Cy Young, guess who led the National League in pitching WPA? Not Roy Halladay, not Clayton Kershaw. Arizona’s Ian Kennedy took that particular title (actually, he tied with Halladay), and that’s why I’d give him my Cy Young vote.

Kennedy had a great season, going 21-4 with a 2.88 ERA. Yes, other pitchers had better ERA’s and more impressive secondary stats. And we’ve all learned to ignore won/loss records. But sometimes–not always, but sometimes–those won/loss records are telling. WPA is the confirmation.

Kennedy’s WPA was 4.3, tied with Halladay and trailing only relievers Tyler Clippard and Jonny Venters (WPA particularly favors relievers who pitch in high leverage situations, but that’s not really appropriate for Cy Young consideration). WPA gives a tie to Halladay and Kennedy, and I’ll pick Kennedy just to be contrary.

In high leverage situations, batters batted just .171 against Kennedy. To put another spin on it, batters batted .127 against Kennedy in “late and close” situations. Kennedy was a mainstay on the surprising Diamondbacks, and he “pitched to win.” He had a season that many sportswriters should love. I have no problem “granting” him the Cy Young award.

Thanks to WAR (which really is a fine stat) and the easy access given by Baseball Reference, Baseball Prospectus and Fangraphs, there have been lots of people promoting its use for MVP voting, sometimes ad infinitum. But look, MVP is still in the eye of the beholder. And if everyone were to march in lockstep to one particular stat (and WAR, regardless of who calculates it, still has its flaws), life would be pretty boring. Keep an open mind, folks. Don’t give up on the narrative.