Greeting

A Baseball Blog - Scientific and Speculative Thoughts from Third Base

Tuesday, October 22, 2013

Evaluating Postseason Performance, with Carlos Beltran, David Ortiz, and WAR

Many fans outside of New England and the Midwest may be upset with this year’s World Series match-up; no matter which team wins, the Red Sox and Cardinals will have combined for half of the past decade’s titles. One upside of it though is that the world gets a little bit more of two of the game’s greatest postseason hitters in Carlos Beltran and David Ortiz.

Beltran and Ortiz are both special for another reason; a majority of fans still see them as borderline Hall candidates in need of more October glory to stamp their eventual ticket to Cooperstown. Several writers this week have looked at the issue already. I would agree with Dave Cameron’s thinking that Beltran is already a Hall of Famer, but I wanted to look at it in another way.

Wins Above Replacement has taken off in the national consciousness as of late, and for good reason. Few stats can take as all-encompassing a look at on-field results and turn them into something easy to understand and compare. With the spread of the WAR framework, many people have gotten used to the scale it works on as well. For example, by Baseball-Reference’s version of the calculation, players start to get serious Hall consideration around 60 or so Wins and become locks around 70.

Beltran already has a solid case at 67.5, while Ortiz is a little lagging at only 44.2. However, those figures don’t account for their aforementioned post-season prowess. Is there a way we can add that in to the WAR framework?


Well, one important thing to consider about WAR is that it adjusts for context. I’m not sure I have the right combination of time and resources to adjust for individual seasons or parks, but in general, postseason play is a lower run-scoring environment than the regular season. It’s a pretty natural consequence of only taking the top eight-to-ten teams’ pitching staffs, then eliminating fifth starters and extraneous flotsam in the bullpen.

So what does that difference in run scoring environments look like? Well, first I took the average runs per game from 1995 to this season. In case you were wondering, that time saw an average of about 4.7 runs scored per team per game. Then, I entered all the scores from postseason play dating back to the start of the Wild Card era and found average runs per team per game. Since 1995, teams have scored about 4.2 runs per game in October play. That turns out to a ratio of .899, meaning runs are about 9/10 as common in postseason play compared to the regular season. Or, in reverse, multiplying postseason runs scored by 10/9 will get you on a scale similar to regular season offense.

That was how I decided to go about determining postseason WAR. WAR is ultimately only a framework, and I wanted to use it to translate postseason numbers into a regular season context. Rather than trying to set the baseline of all postseason performances to zero (the mythical replacement player), I went with the assumption that postseason rosters would, by design, feature fewer replacement-level players, and that any replacement players playing in the postseason would still be worth 0. In essence, this is just treating the postseason like extra postseason play, just with only the best teams still playing.

So with that reasoning out of the way, I went ahead and looked at Baseball-Reference’s WAR framework. Batting and positional adjustments wouldn’t be too much of an issue. Fielding and base running would be almost impossible to obtain without going through the play-by-play of each and every playoff game. However, I’m only concerned with getting a rough estimate of how the player has done, so I can deal with my thinking in those areas on a case by case basis.

Let’s start with Beltran. Plugging his postseason numbers into the formula given for weighted Runs Above Average, we wind up with 29.21 wRAA across 45 postseason games (to date). To account for the scarcity of runs in the playoffs, I multiplied that by the ~10/9 constant from earlier, resulting in an adjusted wRAA of 39.37. So, just on his batting, Beltran has been worth almost 4 wins in the postseason (remember, 10 runs to a win).

Positional adjustment would be the other easier factor to account for. Baseball-Reference includes the following handy guide for batters:

“ 1. C: +10 runs
2. SS: +7.5 runs
3. 2B: +3 runs
4. CF: +2.5 runs
5. 3b: +2 runs
6. RF: -7.5 runs
7. LF: -7.5 runs
8. 1B: -10 runs
9. DH: -15 runs”

Over a full season (about 150 games), that would be how many runs you would need to add or debit to a player to account for the difficulty of the position. Here, Beltran has spent about half of his games as a center fielder and about half as a right fielder. A full season as a right fielder plus a full season as a center fielder would equate to -5 runs over two seasons, but we’re dealing with about one-sixth that total right now. We’re already dealing with estimates, so we’ll dock Beltran a run.

That only leaves base running and fielding unaccounted for. For that, I just looked at Beltran’s career rates. His four postseason appearances have come in 2004, 2006, and the past two years. In his prime, Beltran racked up some serious base running value (even in the postseason, he’s 11 for 11 in stolen base attempts overall). He’s always coupled some solid legs with a good sense of when to go for it. In 2004, B-R credits him with 10 runs (or a full win) on the base paths. In 2006, he was still a plus-4 runner. Even as his legs have worn down the past two seasons, he still isn’t a negative runner (-1 last year, +1 this year). We probably can’t credit him with any runs this year, but 1-2 runs from his days with the Astros and Mets probably isn’t out of the question.

Fielding is a little rougher, especially with the variability of year-to-year numbers and disagreement between systems. Just going with the fielding runs component of Baseball-Reference’s WAR, he’s been a mildly below average right fielder since moving there in 2011 (with totals of -3, 5, and -6 runs in that time). Since we’re dealing with about one-eighth the sample size (he has just over 20 postseason games in that time so far), it’s probably fair to call him a net-zero; at worst, he’s maybe cost the Cardinals one run. His heyday with the Royals, Astros, and Mets was much better; from 2003 to 2009, he graded out well every year (with totals of 0, 6, -1, 13, 10, and 17 in that order). Again, over about 22 games, that’s probably about 1 to 2 runs added.

Those are the major components of WAR (any double play runs, which B-R also includes, will likely be insignificant for Beltran, with only 2 such occurrences across almost 200 plate appearances). Altogether, it’s been somewhere in the neighborhood of 40 runs, or 4 full wins. That in and of itself would be impressive for a full season, but when you consider that it’s only over 45 games, it becomes unreal. Over a full season, that comes out to something like 13-14 WAR, something not even peak Barry Bonds managed. If that seems a little extreme, also recall that we’re dealing with a player with a postseason OPS of 1.173.

Also, those 4 wins assume that you weight postseason games the same as regular season games. If you want to assign any extra importance, it only gets better. In any case, it definitely jumps him above the 70-WAR threshold.

What about David Ortiz? The longtime Sox DH has a reputation for his timely postseason hitting. Over 76 games, he’s managed an .899 OPS (coincidentally, Beltran’s OPS in the NLCS this year). Translating that into wRAA gives us 16.64 runs; adjusting for the postseason run scoring gives us 29.56 batting runs.

Unfortunately, the other areas of WAR are not as kind to Big Papi. Looking up to the earlier positional adjustments, we can see that Ortiz would be docked 15 runs for a full season of DH’ing. With 76 games under his belt, we’ll take half of that, leaving us at about 22 runs.

As a designated hitter, Ortiz doesn’t have an opportunity to lose much in the field. He’s only played in NL parks four times (the Sox’s two sweeps in the Series), and I can’t imagine him playing first base in American League parks very much. At absolute worst, he’s cost his team one run; more likely, he’s been worth nothing. Base running will also be a small negative; Ortiz grades out to about 2 runs below average over a full season usually, so half a season’s worth of postseason games will probably cost him only a single run. Altogether, Ortiz comes out to something around 2 wins for his postseason play.

Again, this is more a framework for a process than a hard-and-fast rule. Maybe you think that replacement level should be reset to the baseline of postseason play. Maybe you think value in the postseason should be weighted even more, since it more directly contributes to a championship. There also isn’t a place on Baseball-Reference with career postseason stats just listed, like there is with regular seasons stats or the play index, so I can’t drop a bunch of players in and judge the process’s effects. This isn’t even getting into the mess at comparing pre-Wild Card postseason numbers, which were likely their own separate run scoring environment. However, in the end, I think this is a good starting point for contextualizing and appreciating modern players’ postseason value.

1 comment:

  1. This is great work, Theo. The approaches I've been toying with (that I just published at http://www.hallofstats.com/articles/is-david-ortiz-our-generations-jim-rice

    Rather than re-constructing WAR, I've focused more on the differences in win expectancy and run-expectancy the player was responsible for. So, I've been using WPA and RE24. I haven't settled on an approach yet, though.

    I've thought about re-creating WAR much as you have, but kudos for going ahead and doing it. I wish there were better ways to handle base-running and (especially) defense, though. Of course, WPA and RE24 have these same issues.

    Great stuff, though!

    ReplyDelete