This is the first in a series of pieces on a tool I’ve been working on. Admittedly, right now it’s quite raw, and probably needs some adjustments, which I’ll elaborate on towards the end of this post. It’s also quite lengthy – set it aside for when you have ample time to follow along, as there are some example calculations included to demonstrate the process.

Most of you are familiar with the “Similarity Scores” feature on Baseball Reference. If not, the explanation can be found here. The idea is to provide player comps using the player’s statistics. This has been around a while, and is based on a fairly simplistic “points-based” approach. Such an approach has the advantage of being easy to follow and intuitive, and as a quick tool to create fun conversation, it’s nice. However, it’s not very useful for purposes of projection for many reasons – not the least of which being that the points used are arbitrary and the statistics used are result statistics (hits, HRs, RBIs, etc) rather than being process-driven. It’s also intended to work on a player’s entire career. Some players have one or more drastic shifts in results over the course of their careers – and, to project a player in 2015 from his work in 2013-2014, we need to isolate data by season.

With the mountains of granular data available since Similarity Scores were first published, I thought it would be interesting to take a cut at creating something new in the same vein. My primary objectives were to create a similarity metric that (a) compared individual seasons rather than entire careers; (b) was based primarily on a hitter’s “process” or approach at the plate rather than strictly on results which are influenced heavily by luck; and (c) was mathematically defensible, in other words, non-arbitrary.

Read the rest of this entry »

It looks like the New York Mets are going to be entering the season with Wilmer Flores as their shortstop.  Flores has become a polarizing figure among Mets fans for a myriad of reasons, most notably of which would be his defensive capabilities at the position.  Scouts have long held that Flores is not a capable shortstop; however his defensive metrics were pretty good last year!  That being said we know a sample size of one season of defensive metrics is prone to a lot of statistical noise.  And THAT being said we know that Flores played just 443.1 innings at shortstop last season.  Uh oh.  What exactly can we take from that sample size?  How much weight should we place on these defensive metrics for Mr.Flores?

Are the scouts right?  Are the metrics right?  Or is the answer somewhere in between?  (Almost definitely.)

What follows is an exercise which will answer precisely zero of the above questions.  However, I cannot remember a situation quite like this Flores predicament, so I went on a quest (through FanGraphs) to find some comparables.  What shortstops have had the type of defensive metric success Flores has had in such a short sample size, and how have they fared outside of that season?

I looked at a sample of players from 2003-2014 who played from 400 to 500 innings at the position with a UZR/150 from 5 to 19 (Flores was at 12.5).  All these parameters are quite arbitrary, but this whole exercise is quite arbitrary so let’s move along.

This brings us a list of ten seasons excluding Flores.  The seasons are as follows:

2014    Jose Ramirez (498.2 Innings, 18.9 UZR/150)

2008    Marco Scutaro (472.1, 17.6)

2008    Maicer Izturis (448, 15.9)

2009    Robert Andino (478.1, 14.1)

2010    Jerry Hairston (489.2, 8.9)

2006    Alex Cora (434, 8.7)

2012    Paul Janish (450.1, 8.6)

2014    Stephen Drew (413.1, 8.1)

2012    John McDonald (426.1, 6.1)

2010    Wilson Valdez (458, 5.2)

What this list of players lacks, is a very poor fielding shortstop.  The lowest career shortstop UZR/150 of the bunch belongs to Mr. Izturis at -3.1 in 1697.1 innings.  This seems to be a list of humans in which you can confidently state “Hey!  None of these players were atrocious major-league defensive shortstops over their careers!”

So what does this mean in regards to Flores?  Basically, nothing.  However, Mets fans can now take solace in knowing that the 10 players (from the last 12 seasons), who had the most similar statistical defensive season to Flores’ 2014, had careers in which they were able to play the shortstop position not horribly.  Now, if Flores himself can play the shortstop position not horribly then the Mets might just have them a nice little player.

Then again, there is always this:

In the fantasy baseball world, Troy Tulowitzki is the Lamborghini that is terrific when it’s on the road but spends too much time in the garage. Since becoming a regular in 2007 (eight years), Tulowitzki has had just three seasons in which he played more than 140 games and none of those seasons were in the last three years. He’ll be 30 years old during the 2015 season, so age is not on his side when it comes to health.

Last year was the most tantalizing and ultimately disappointing season of all. Tulowitzki was off to a tremendous start, hitting .340/.432/.603 through 91 games. He was hitting like a vintage Albert Pujols but at the shortstop position. In a little more than half a season, he accumulated 5.1 WAR and had a career-best 171 wRC+.

Then it happened—the yearly injury. On July 19^th, in Pittsburgh, Tulowitzki strained his left hip flexor while running to first base and his season was over. Despite playing just 91 games, Tulowitzki ranked 73^rd in Zach Sanders’ End of Season Rankings for 2014. Sanders had Tulowitzki worth $16.08, right in the same ballpark as Ryan Braun ($16.33), Jonathan Lucroy ($16.15), and Jimmy Rollins ($16.01). These rankings were based on a 12-team, 5 x 5 league with one catcher, so Tulowitzki’s placement at 73^rd would make him the first pick in the 7^th round, despite playing just over half the season. Of course, the pre-season consensus rankings of FanGraph writers had Tulowitzki anywhere from 12^th to 20^th, so 73^rd was a big disappointment.

Over the last three years, Tulowitzki has averaged 88 games and 363 plate appearances per season, with a batting line of .316/.399/.551. You know he’s going to play well, you just don’t know how much he’ll play. So what do you do with Tulo on draft day?

First, let’s look at his injury history.

After a 25-game cup of coffee in 2006, Troy Tulowitzki became a Rockies regular in 2007, playing 155 games as a 22-year-old.

In 2008, Tulowitzki hit the disabled list twice. On April 29^th, Tulo was not in the original starting lineup but was put in the game at the last minute when Jeff Baker broke a blood vessel in his throwing hand during pregame warm-ups. He then tore his left quadriceps on a defensive play in the first inning. He came off the DL on June 20^th and played regularly until July 5^th, when he went back on the DL with a cut hand. During the game on the previous day, Tulowitzki hit his bat against the ground in frustration, only to have the bat shatter and cut the palm of his hand up to his index finger. He would need 16 stitches and miss the next two weeks. With the two injuries limiting him to 101 games, Tulowitzki had the least-productive year of his career, other than the first-year 25 game stint.

In 2009, Tulo played 151 games and had 5.5 WAR, one of six seasons with 5 or more WAR in his career

The 2010 season saw Tulowitzki pick up right where he left off in 2009. Through 62 games, he was hitting .306/.375/.502. Then, on June 17^th, he was hit by a pitch from Alex Burnett and fractured his left wrist. The injury kept him out for five weeks but he came back better than ever, hitting .323/.386/.634 after the injury. Despite playing in just 122 games, Tulowitzki had his best season in 2010, accumulating 5.9 WAR.

Tulowitzki played 143 games in 2011. Nothing to see here.

In 2012, Tulo was off to a slow start, hitting just .287/360/.486. On May 30^th, he strained his groin while running out a ground ball and his season was over. He played just 47 games that year.

Two years ago, Tulowitzki missed nearly a month in the middle of the season with a fractured rib from diving for a ground ball. He still hit .312/.391/.540 and had 5.4 WAR in 126 games.

Last year, as mentioned above, Tulo was off to his best start ever but his season ended in July because of a strained hip flexor sustained while running to first base.

So, over the last seven years, Tulowitzki has been on the DL six times. Twice he was hurt while running to first base. Twice he was hurt while making a play on defense. And twice he was hurt through what I would call flukes—slamming his bat into the ground and getting hit by a pitch. He’s had a torn quadriceps, strained groin, fractured rib, strained hip flexor, cut hand, and fractured wrist. On the one hand, two of those were flukes and he hasn’t hurt the same body part more than once. On the other hand, the quad, groin, and hip flexor are all lower-body type injuries, which could continue to occur as he gets older.

So how much do you factor in the injury history when considering when to draft Troy Tulowitzki in 2015?

On his player page, Tulo is projected for 525 at-bats by Steamer and 467 by the Fans (23 fan projections). He hasn’t reached either of those totals since 2011. I have projections from other sources that are more conservative with his playing time:

Cairo: 352 AB

ZiPS: 381 AB

Marcel: 410 AB

Davenport: 417 AB

CBS: 466 AB

Average them all together and Tulowitzki is projected for 431 at-bats.

So let’s go to the spreadsheet. I created dollar values using the z-scores method for a 12-team, 5 x 5, one-catcher league and came up with the following.

Scenario #1: If Tulowitzki gets the 525 at-bats projected by Steamer, he would be a first-round pick, right there with Jose Abreu and Paul Goldschmidt. (525 AB, 159 H, 85 R, 28 HR, 91 RBI, 3 SB, .302 AVG).

Scenario #2: If Tulowitzki gets 431 at-bats (the average of the seven sources I’ve collected) and his other numbers are pro-rated to that total, he drops down to around the middle of the 4^th round. (431 AB, 131 H, 70 R, 23 HR, 75 RBI, 2 SB, .302 AVG).

Scenario #3: If Tulowitzki gets 314 at-bats (his average over the last three years) and his other numbers are pro-rated to that total, he drops to the 18^th round. (314 AB, 95 H, 51 R, 17 HR, 54 RBI, 2 SB, .302 AVG).

But that’s not the whole story because I haven’t factored in an injury replacement. When Tulowitzki gets injured, he usually goes all out and heads to the DL and generally misses significant time. He’s not like an aging Chipper Jones who would play 4 or 5 games a week and would be difficult to replace if you can’t make daily moves.

So let’s factor in an injury replacement for Tulo if he misses some time. I took the average of three “replacement-level” shortstops from my spreadsheet (Jordy Mercer, Wilmer Flores, and Yunel Escobar) and pro-rated them to the amount of time Tulowitzki would miss in the latter two scenarios from above.

Scenario #2, Adjusted:

431 AB, 131 H, 70 R, 23 HR, 75 RBI, 2 SB, .302 AVG—Tulowitzki

94 AB, 24 H, 10 R, 2 HR, 10 RBI, 1 SB, .258 AVG—Jordy Flores Escobar

525 AB, 155 H, 80 R, 25 HR, 85 RBI, 3 SB, .295 AVG—Tulo & Friends

Add in 94 at-bats from a replacement-level shortstop to Tulowitzki’s projected stats, which would bring his total to the 525 at-bats projected by Steamer, and Tulowitzki would drop from the middle of the 1^st round to the middle of the 2^nd round; still definitely worth having on your team.

Scenario #3, Adjusted:

314 AB, 95 H, 51 R, 17 HR, 54 RBI, 2 SB, .302 AVG—Tulowitzki

211 AB, 54 H, 23 R, 5 HR, 23 RBI, 1 SB, .258 AVG—Jordy Flores Escobar

525 AB, 149 H, 74 R, 22 HR, 77 RBI, 3 SB, .284 AVG—Tulo & Friends

Add in 211 at-bats from a replacement-level shortstop to Tulowitzki’s projected stats, which would bring his total to the 525 at-bats projected by Steamer, and Tulowitzki would drop from the middle of the 1^st round to the middle of the 4^th round. That’s the somewhat realistic downside risk.

Looking at the three scenarios above, we have:

• A fully-healthy Tulowitzki is a mid-1^st round pick.
• A somewhat healthy Tulowitzki (using the average of 7 projection sources) plus a replacement-level shortstop used for the time missed and Tulo drops to the middle of the 2^nd round.
• A healthy-as-he’s-been-the-last-three-years Tulowitzki (using an average of his at-bats over the last three years) plus a replacement-level shortstop and Tulo drops to the middle of the 4^th round.

In the Rotographs’ Top 300, the five participants had Tulowitzki with an average pick of 29^th overall, just one spot behind Ian Desmond. In that Top 300, Zach Sanders had Tulo ranked 75^th, which was quite the outlier (the others had Tulo from 15^th to 28^th). If you remove Sanders’ rankings, Tulowitzki would be the 17^th player off the board, which would put him right in line with Scenario #2 from above—the middle of the 2^nd round.

Everyone has his own appetite for risk, but I would go ahead and roll the dice on Troy Tulowitzki in 2015.

It was announced recently that the Miami Marlins will host the 2017 All-Star Game, making this the first time the Marlins will host the Midsummer Classic. Fourteen years ago the Marlins were in line to host the 2000 All-Star Game but after their fire sale following their 1997 World Series championship, MLB flipped the game to Atlanta.

Traditionally, the All-Star game has alternated between leagues. The last time the same league hosted back-to-back All-Star Games was 2006-2007, when the game was played in Pittsburgh’s PNC Park then San Francisco’s AT&T Park, both in the National League. Before that, you have to go all the way back to 1950-1951 to find All-Star Games hosted by the same league in back-to-back seasons (the White Sox’ Comiskey Park in 1950, Detroit’s Briggs Stadium in 1951). Awarding the 2017 game to the Marlins means that National League will host the game three years in a row, following Cincinnati this year and San Diego in 2016.

In the case of the Marlins in 2017, it appears that outgoing commissioner Bud Selig had slated the Marlins to get an All-Star Game in the near future after the team recently opened a new ballpark in 2012. Other teams in contention were the Baltimore Orioles, who last hosted the game in 1993, and the Washington Nationals. The Nationals have not hosted a game since their move from Montreal in 2005.

Along with the 2017 All-Star Game announcement, incoming commissioner Rob Manfred had this to say in a recent interview with ESPN’s Jayson Stark: “One of the things that I am going to try to do with the All-Star Games is—and we’ll make some announcements in the relatively short term—I am looking to be in more of a competitive-bidding, Super Bowl-awarding-type mode, as opposed to [saying], ‘You know, I think Chicago is a good idea’”.

The Super Bowl bidding process is quite a thing to behold. In an article in the Minneapolis Star-Tribune from June of last year, the paper listed many of the concessions made by the city to get the Super Bowl. The paper also posted a copy of the NFL’s “Host City Bid Specifications and Requirements”, which is 153 pages long. Among the items in this document:

• NFL controls 100 percent of the revenues from all ticket sales, including suites, and exclusive access to all club seats.
• Exclusive, cost-free use of 35,000 parking spaces for gameday parking.
• Full tax exemption from city, state, and local taxes for tickets sold to the Super Bowl, including the NFL Experience, the NFL Honors show and other NFL Official Events.
• NFL has the option to install ATMs that accept NFL preferred credit/debit cards in exchange for cash and to cover up other ATMs.
• Host city is asked to pay all travel and expenses for an optional “familiarization trip” for 180 people in advance of the Super Bowl to inspect the region.
• NFL requires the usage of three top-quality, 18-hole golf courses in close proximity to one another and greens and cart fees at these three courses must be waived or otherwise provided at no cost to the NFL. (Golf courses in Minneapolis in February? Really?!)
• NFL requires the reservation of up to two quality bowling venues at no rental cost.

And those are just a few of the request by the NFL. The “competitive-bidding, Super Bowl-awarding-type mode” is great for NFL bigwigs. It’s not surprising that MLB owners would want to get on board this gravy train.

Other sports have their own bidding processes for their showcase events. Notably, FIFA has had numerous scandals associated with the bidding process for the World Cup. Most recently, after awarding the 2018 World Cup to Russia and the 2022 World Cup to Qatar, FIFA had an investigator create a report looking into accusations of impropriety in the World Cup bidding process. They then announced that the report “cleared its integrity and should constitute closure” despite the fact that the 42-page summary of the report identified numerous instances of corruption, collusion, and vote-buying. Even as the chief chair of FIFA’s ethics committee patted himself on the back, the man who created the report, U.S. prosecutor Michael Garcia, claimed that the portrayal of his report was “erroneous and incomplete”.

Of course, the long-standing king of bidding process corruption would be the Olympics. Do a Google search on “corruption in the Olympics bidding process” and numerous articles as far back as 1999 turn up dealing with shenanigans when it comes to awarding the Olympics to cities competing for the honor.

In a 1999 article at the New York Times, the International Olympic Committee (IOC) acknowledged corruption in the Salt Lake City bidding for the 2002 Winter Olympics. Richard Pound, a lawyer who led the IOC investigation said, “We have found evidence of very disappointing conduct by a number of IOC members. Their conduct has been completely contrary to everything the Olympic movement has worked so hard to represent.” HA! That’s quite funny, in hindsight. Rather than clean up the process, things have just grown worse in the last 15 years.

Sure, the Salt Lake City Olympics scandal triggered reform that was supposed to ban gifts and favors to Olympic committee members, but less than a decade later the Olympics were awarded to Sochi (won the bid in 2007, hosted the 2014 Winter Olympics). This article at Salon.com proclaimed the Sochi Olympics the “most corrupt Olympics ever.”

In the case of the NFL, the process for bidding to host the Super Bowl is not necessarily corrupt; it’s just pure greed. Fat cat NFL owners realize there’s no event bigger than the Super Bowl, so they can demand anything they want from prospective host cities. Is it at all surprising that the billionaire baseball owners want in on this action?

Personally, I like that the baseball All-Star Game gets passed around from city to city. In the past 25 years, 24 different MLB teams have hosted the All-Star game (Pittsburgh hosted the game twice, in two different ballparks). In that same time period, just 14 NFL cities have hosted the Super Bowl, with four cities hosting three or more Super Bowls during that stretch. In the 49-year history of the Super Bowl, the game has been played in Miami and New Orleans ten times each, and another seven times in Los Angeles, meaning more than half of the Super Bowl games have been played in just three cities. Of course, the NFL does prefer to have the game played in warm-weather cities, unless you have a dome (Minneapolis, Indianapolis, Detroit) or you are New York (???). Major League Baseball doesn’t have to worry about the weather for the Midsummer Classic.

We’ll have to see what MLB commissioner Rob Manfred has to say about the “competitive-bidding, Super Bowl-awarding-type mode” he’s considering for the All-Star Game. Will there be a mechanism in place so the same city doesn’t host the game every few years? Will cities without major league baseball get a chance to bid for the game? And just how many top-quality, 18-hole golf courses will MLB owners demand?

Jeff Sullivan recently sent up a warning flare regarding Kris Bryant’s potential swing and miss problems, and this post is essentially riffing off that one, so you’ll probably want to read that first if you haven’t already checked it out. I’m using strikeout rate rather than contact rate, but the message is similar.

Bryant isn’t alone among Cubs prospects with contact avoidance issues. Here’s what some of their bigger names did last year:

Player                                        Level            K%             wRC+

Javier Baez                             MLB            41.0                 51

Arismendy Alcantara        MLB            31.0                 70

Jorge Soler                             MLB            24.7               146

Kris Bryant                                 AAA            28.6               164

Three of those four (i.e., the non-Alcantaras) are thought to be integral parts of The Future for the Cubs. But those are strikeout rates that have not generally led to long-term career success.

Here are the top ten career K rates for hitters with over 5000 plate appearances:

Player                            K%           wRC+          WAR

Adam Dunn               28.6            123              22.7

Ryan Howard           28.1             126              19.9

Jose Hernandez      27.3              86              12.9

Carlos Pena               26.8             117              16.9

B.J. Upton                  26.4              99              21.7

Jim Thome                 24.7            145              67.7

Dave Kingman          24.4            113              20.4

Gorman Thomas      24.4            114              20.4

Dan Uggla                   24.2             110              22.8

Dean Palmer             24.2             104              11.0

So it’s not impossible to have a long and relatively successful career striking out more than a quarter of the time, but it hasn’t happened much – just five times using my admittedly somewhat arbitrary 5000 PA cutoff. With strikeout rates continuing to rise, a few more players will edge ahead of Dean Palmer in the coming years, but in all likelihood, many more will fall by the wayside long before reaching that somewhat less than august plateau. As Sullivan points out, players who whiff this often need to max out their other skills in order to be useful to a team, putting enormous pressure on those other skills to develop.

And there does seem to be a correlation between hitters’ strikeout rates and overall team success, as Joe Sheehan has noted elsewhere. Here are the teams with the five highest hitter K rates from last year:

Cubs               24.2          73-89

Astros            23.8          70-92

Marlins          22.9          77-85

Braves            22.6          79-83

Reinsdorfs     22.4          73-89

None of these teams came close to making the playoffs. You have to go all the way down to tenth on the list to find a playoff team (the Nats, at 21.0%).

And here are the five least K-lacious teams:

Royals             16.3        89-73

A’s                    17.7        88-74

Rays                 18.1        77-85

Tigers              18.3        90-72

Cards               18.6       90-72

Yankees           18.6       84-78

Numerate readers will have grasped that there are actually six teams on this list, since the Cards and Yankees tied at 18.6%.  Only the Rays and the Evil Empire failed to make the playoffs, and only the Rays failed to break .500.

Not all of the young Cubs  windmill at the plate: Addison Russell and Kyle Schwarber kept their K rates under 20% last year in the minors, as did Anthony Rizzo and Starlin Castro in the majors. And as Sullivan noted, players do develop – Bryant et. al. are not necessarily trapped for eternity in the seventh level of Strikeout Hell. But for now, a significant part Theo Epstein’s plan to bring glory to Wrigleyville depends on whether these players can either find a way to strike out less, or to succeed without doing so, something that few have managed thus far.

If you haven’t read Part 1, Part 2, and Part 3, you may want to go back and check them out.

After looking in-depth at 2014 WAR, I thought it would be interesting to compare 2014 WAR with WAR totals from 2002. Baseball scoring has dropped considerably since 2002 and I wondered how this would be reflected in WAR, either at the positional level or the age level or both.

Here is a comparison of hitting statistics from 2002 and 2014:

Twelve years ago, hitters put up a higher batting averages, on-base percentages, slugging percentages, and isolated slugging. They walked more and struck out less.

But,we pretty much knew this. Did this difference in the level of offense affect the WAR accumulated at each position?

Position Players

The following table shows WAR for each position with 2002 on top and 2014 below.

If we look at the comparison of WAR/600 PA for the premium hitting positions (DH, 1B, RF, LF, 3B), we see that all except third base accumulated more WAR in 2002 than in 2014. On the other end of the fielding spectrum, the key defensive positions (C, SS, 2B, CF) all had more WAR in 2014, when offense was down.

This table shows a comparison of the traditionally offense-oriented positions versus the positions historically known more for their glove work in the two different run-scoring environments of 2002 (4.62 R/G) and 2014 (4.07 R/G).

In 2002, the offense-oriented positions averaged 2.2 WAR/600PA. In 2014, these positions average 1.8 WAR/600 PA. The more defensive-oriented positions averaged 1.9 WAR/600 PA in the higher run-scoring environment and 2.4 WAR/600 PA when runs were more scarce.

This shift of WAR from more hitter-heavy positions to the better fielding positions has been a general trend over the last thirteen years, particularly so in the last four years as run scoring has dropped significantly.

Consider the table below. The column to the far right shows the difference between WAR for the hitting positions and fielding positions each year:

The biggest change has been over the last four years, as run scoring has dropped down below 4.3 runs per game after being in the range of 4.6 to 4.8 runs/game in the 2000s. Teams are getting more WAR/600 PA from the defensive-oriented positions than the bat-first positions. The 2014 season saw the biggest gap in the last thirteen years, with glove-first positions averaging 0.6 more WAR/600 PA than the bat-first positions.

Changing distribution of playing time and WAR based on age

Along with the change in WAR for the hitting positions versus the defense-oriented positions, there has been a shift in WAR and playing time based on age. From 2000 to 2005, position players 33 and older had more plate appearances than players 25 and under. Beginning in 2006, position players 25 and under have had more plate appearances each year than players 33 and older. Since 2010, this difference has accelerated, as the graph below shows:

In 2000, players 33 and older had 40,626 plate appearances and players 25 and under had 38,919. Last year, players 33 and older had dropped to 29,191 plate appearances and players 25 and under were up to 45,439 plate appearances.

With increasing playing time, players 25 and under have seen their total WAR go up, while WAR for players 33 and older has gone down:

The difference in WAR is not just a playing time difference, though. Older players have not only seen less playing time, they’ve also been less productive, as this graph of WAR/600 PA demonstrates:

In 2000, players 33 and older averaged 1.7 WAR/600 PA, while players 25 and under averaged 1.4 WAR/600 PA. The older group of players maintained their lead until 2003, when the two groups were essentially even. Since then, younger players have out-produced older players. Last year, the gap was 0.5 WAR/600 PA in favor of the younger group of players.

Starting Pitchers

For starting pitchers, there are some differences. Innings pitched by starting pitchers 25 and under have fluctuated quite a bit over the last 15 years. Since 2000, starting pitchers age 25 and under have thrown a high of 10,268 innings (2002) and a low of 6,663 innings (2005). Starting pitchers 33 and older have a narrower range of innings pitched per season, with a very slightly downward trend over the last thirteen years, as shown by this graph:

While their innings pitched has been fairly consistent since 2000, starting pitchers 33 and older have been less productive. The following graph shows the WAR/150 innings pitched for starting pitchers 25 and under compared to those 33 and older. The “33 and older” group has dropped from a high of 2.5 WAR/150 IP in 2000 to a low of 1.2 WAR/150 IP last season.

From 2000 to 2007, pitchers 33 and older were more productive per inning than pitchers 25 and under. Since then, young pitchers have been more productive, except for that 2012 season. The gulf has widened between these two groups over the last two years.

Relief Pitchers

Finally, let’s look at relief pitchers. Since 2000, relief pitchers 33 and older have seen their innings pitch per year rise from around 3,000 in 2000 to a high of 3,951 in 2005, but have steadily dropped since then. In 2014, they pitched a 15-year low of 2,063 innings. Relief pitchers 25 and under saw a sharp increase in innings pitched from 2004 to 2006, and have bounced around a bit since then, but have generally seen a drop in the amount of innings they’ve pitched since then.

When it comes to production, older relief pitchers have followed a different pattern than their counterparts. Position players and starting pitchers 33 and older have seen their production drop (using WAR per playing time), while relief pitchers 33 and over have held steady. Older relievers are pitching fewer innings each year but they are still as productive (and have a slight increase in WAR/50 IP over the last 15 years).

Final Thoughts

Baseball has evolved over the last 15 years from a high-offense, slugging game to a low-offense, pitching-and-defense game and WAR reflects those changes. The offense-oriented positions (1B, RF, LF, 3B) used to accumulate more WAR each season, but no longer do so. Older players were once more likely to sustain their production into their mid-30s, but no longer play as much or as well as they once did at an advanced age.

Looking to the future, we have to wonder what’s to come. Will offense continue to drop or has it bottomed-out and now due for a rebound? Will MLB do something to raise the level offense (adjust the strike zone, perhaps?)? If offense makes a comeback, how will that be reflected by WAR?

If you haven’t read Part 1 and Part 2, you may want to go back and check them out. If you would prefer not to, this is a reminder of where these numbers came from:

Using FanGraphs’ terrific leaderboard tools, I found statistics for all players who played at each position in 2014. The following numbers apply only to the time spent at that position. Buster Posey, for example, accumulated 462 plate appearances at catcher, 128 at first base, and 9 at DH, so his plate appearances in those amounts are included for those positions in the table below.

For position players, I calculated WAR per 600 plate appearances. For starting pitchers, I used WAR per 150 innings pitched. For relievers, I used WAR per 50 innings pitched.

In this installment of 2014 WAR Breakdown, players at each hitting position are split into six different age groups. All “small sample size” warnings apply.

CATCHER

The most productive age groups for catchers were the “27 to 29” and “30 to 32” age groups, with both groups averaging 3.1 WAR/600 PA. Not surprisingly, the production by catchers really starts to dwindle as they move into their mid-30’s. The “30 to 32” age group hit a combined .253/.319/.385, while the “33 to 35” age group hit just .220/.310/.307. Russell Martin will be 32 next year and just signed a 5-year, $82 million contract with the Blue Jays. Here’s hoping he ages better than most catchers, for the Blue Jays’ sake.

Best Catcher 23 and under: Mike Zunino, 23 (1.7 WAR)

Best Catcher 24 to 26: Devin Mesoraco, 26 (4.1 WAR)

Best Catcher 27 to 29: Jonathan Lucroy, 28 (6.3 WAR)

Best Catcher 30 to 32: Russell Martin, 31 (5.4 WAR)

Best Catcher 33 to 35: Carlos Ruiz, 35 (3.2 WAR)

Best Catcher 36 and up: David Ross, 37 (0.2 WAR)

FIRST BASE

In 2014, first baseman aged 24 to 26 were the most productive group at the position, averaging 2.3 WAR/600 PA. The “30 to 32” group was next, at 1.9 WAR/600 PA, with the group in the middle (“27 to 29”) finishing third in this metric. The small sample sizes of young (23 and under) and old (36 and older) were quite unproductive, both averaging negative WAR per 600 PA and hitting under .200.

Best First Baseman 23 and under: Will Myers, 23 (0.0 WAR in just 3 PA)

Best First Baseman 24 to 26: Anthony Rizzo, 24 (5.6 WAR)

Best First Baseman 27 to 29: Jose Abreu, 27 (4.4 WAR)

Best First Baseman 30 to 32: Miguel Cabrera, 31 (4.9 WAR)

Best First Baseman 33 to 35: Justin Morneau, 33 (2.5 WAR)

Best First Baseman 36 and up: Raul Ibanez, 42 (0.3 WAR in 20 PA)

SECOND BASE

The most productive group of second basemen skewed older than you might expect. The group aged 30 to 32 had the highest WAR/600 PA and best hitting line. Overall, the three age groups ranging from age 27 to age 35 were the most productive.

Best Second Baseman 23 and under: Kolten Wong, 23 (1.8 WAR)

Best Second Baseman 24 to 26: Jose Altuve, 24 (5.0 WAR)

Best Second Baseman 27 to 29: Brian Dozier, 27 (4.6 WAR)

Best Second Baseman 30 to 32: Ian Kinsler, 32 (5.5 WAR)

Best Second Baseman 33 to 35: Chase Utley, 35 (4.2 WAR)

Best Second Baseman 36 and up: Brian Roberts, 36 (0.3 WAR)

THIRD BASE

Production at third base in 2014 skewed older. Third basemen 23 and under and 24 to 26 were the two least productive groups. The “33 to 35” group had the highest WAR/600 PA, but this was due mainly to just two players—Adrian Beltre and Juan Uribe.

Best Third Baseman 23 and under: Nolan Arenado, 23 (3.1 WAR)

Best Third Baseman 24 to 26: Kyle Seager, 26 (5.6 WAR)

Best Third Baseman 27 to 29: Josh Donaldson, 28 (6.5 WAR)

Best Third Baseman 30 to 32: Chase Headley, 30 (4.4 WAR)

Best Third Baseman 33 to 35: Adrian Beltre, 35 (5.7 WAR)

Best Third Baseman 36 and up: Aramis Ramirez, 36 (2.2 WAR)

SHORTSTOP

At the shortstop position, the age 24 to 26 group had the largest percentage of playing time of any group, but a lower WAR/600 PA than any group except the “36 and up” group. In the “27 to 29” age group, Troy Tulowitzki had 5.2 of the group’s 21.0 WAR (24.8%) despite getting just 372 of the group’s 5261 plate appearances (7%).

Best Shortstop 23 and under: Jose Ramirez, 21 (2.1 WAR)

Best Shortstop 24 to 26: Starlin Castro, 24 (2.8 WAR)

Best Shortstop 27 to 29: Troy Tulowitzki, 29 (5.2 WAR)

Best Shortstop 30 to 32: Jhonny Peralta, 32 (5.3 WAR)

Best Shortstop 33 to 35: Jimmy Rollins, 35 (3.5 WAR)

Best Shortstop 36 and up: LittleNicky Punto, 36 (0.4 WAR)

LEFT FIELD

Thanks to Christian Yelich, the most-productive group of left fielders in WAR/600 PA was the group of player’s aged 23 and under. It’s interesting to see the WAR/600 PA drop-off from the “30 to 32” group to the “33 to 35” group. The “33 to 35” group hit nearly as well as the younger group, but they had a couple of particularly bad fielders (Rajai Davis and Matt Holliday) who brought their WAR total down.

Best Left Fielder 23 and under: Christian Yelich, 22 (4.2 WAR)

Best Left Fielder 24 to 26: Justin Upton, 26 (4.1 WAR)

Best Left Fielder 27 to 29: Michael Brantley, 27 (4.5 WAR)

Best Left Fielder 30 to 32: Alex Gordon, 30 (6.6 WAR)

Best Left Fielder 33 to 35: Matt Holliday, 34 (3.8 WAR)

Best Left Fielder 36 and up: Endy Chavez, 36 (0.1 WAR)

CENTER FIELD

Mike Trout had 664 of the 3523 plate appearances (19%) credited to players 23 and under, but 34% of the WAR for this group. With Trout, Billy Hamilton (3.5 WAR) and Marcell Ozuna (3.4 WAR) all in the 23 and under group, this was the most productive collection of players of any of the age groups in WAR/600 PA. Center field is a young person’s position. Just 4.6% of the plate appearances by center fielders went to players 33 and over.

Best Center Fielder 23 and under: Mike Trout, 22 (7.5 WAR)

Best Center Fielder 24 to 26: Juan Lagares, 25 (3.8 WAR)

Best Center Fielder 27 to 29: Andrew McCutchen, 27 (6.8 WAR)

Best Center Left Fielder 30 to 32: Jacoby Ellsbury, 30 (3.9 WAR)

Best Center Fielder 33 to 35: Rajai Davis, 33 (1.0 WAR)

Best Center Fielder 36 and up: Reed Johnson, 37 (0.0 WAR)

RIGHT FIELD

Players 23 and under had 4.7 WAR as a group in 2014. Yasiel Puig had 3.7 of that total. In the “30 to 32” age group, Hunter Pence (4.7 WAR) had more than half of the total WAR (9.2) for the group.

Best Right Fielder 23 and under: Yasiel Puig, 23 (3.7 WAR)

Best Right Fielder 24 to 26: Giancarlo Stanton, 24 (6.2 WAR)

Best Right Fielder 27 to 29: Mat Kemp, 29 (2.4 WAR)

Best Right Fielder 30 to 32: Hunter Pence, 31 (4.7 WAR)

Best Right Fielder 33 to 35: Jose Bautista, 33 (6.1 WAR)

Best Right Fielder 36 and up: Marlon Byrd, 36 (1.9 WAR)

DESIGNATED HITTER

There wasn’t much WAR accumulated by players at the DH position in 2014. It’s hard to produce WAR with no defensive value and a strong positional adjustment.

Best DH 23 and under: Kennys Vargas, 23 (0.7 WAR)

Best DH 24 to 26: Yan Gomes, 26 (1.0 WAR)

Best DH 27 to 29: Chris Carter, 27 (2.2 WAR)

Best DH 30 to 32: Adam Lind, 30 (0.9 WAR)

Best DH 33 to 35: Victor Martinez, 35 (3.9 WAR)

Best DH 36 and up: David Ortiz, 38 (2.7 WAR)

PITCHER (HITTING)

Ha! Pitchers hitting. That’s funny.

Best Pitcher (hitting) 23 and under: Shelby Miller, 23 (0.5 WAR)

Best Pitcher (hitting) 24 to 26: Madison Bumgarner, 24 (1.2 WAR)

Best Pitcher (hitting) 27 to 29: Travis Wood, 27 (1.0 WAR)

Best Pitcher (hitting) 30 to 32: Zack Greinke, 30 (0.7 WAR)

Best Pitcher (hitting) 33 to 35: Dan Haren, 33 (0.3 WAR)

Best Pitcher (hitting) 36 and up: Bronson Arroyo, 37 (0.1 WAR)

PINCH-HITTER

Pinch-hitters hitting are only slightly less funny than pitchers hitting. Players 23 and under were better pinch-hitters than any other age group.

Best Pinch-Hitter 23 and under: Cory Spangenberg, 23 (0.3 WAR)

Best Pinch-Hitter 24 to 26: Lonnie Chisenhall, 25 (0.4 WAR)

Best Pinch-Hitter 27 to 29: Delmon Young, 28 (0.6 WAR)

Best Pinch-Hitter 30 to 32: John Mayberry, Jr., 30 (0.8 WAR)

Best Pinch-Hitter 33 to 35: Jeff Baker, 33 (0.3 WAR)

Best Pinch-Hitter 36 and up: Lyle Overbay, 37 (0.4 WAR)

Next up is a comparison of WAR in 2014 to WAR in 2002.

Last week I looked at the AL, so it is time to talk about the Senior Circuit depth. After that I will discuss the limitations that I think exist in both my approach and Jeff’s, part of which could be a new form of MVP debates. What is depth?

Again, I started with a rough look at front line versus second for the teams:

I had to adjust my method of using the multiple of using front divided by second line a little bit to account for the Reds and Phillies who have negative second line projections by using absolute values.  The National League is structured in a more stars and scrubs way this year versus the American League where there are no teams that you point at and think they will be horrible.  In Philadelphia, Atlanta, and Arizona things are looking pretty grim on the front lines though I could argue that Atlanta has some upside relative to how much Steamer seems to hate their outfield and starting pitching.

This changed how my depth ranking compared to Jeff’s by making the Diamondbacks and Mets look pretty good depth-wise only due to a combination of okay backups mixed with pretty low overall front line WAR.  This is a limitation of the multiples I use as shrinking the numerator can make for a lower multiple if a bad team has a couple of decent bench players.  I will come back to the discussion of what is depth in a second.

Only one other team was ranked far away from Jeff, the Pirates, and they look a lot like the Yankees did in the AL.  In Pittsburgh, they have good players all over the front lines, but the team is going to depend on those guys a lot according to the projection.  Jeff is giving them credit for guys like Sean Rodriguez who could be capable fill-ins according to the projections, not a sentiment I necessarily disagree with and is something recommending the way he approached it.  So what is depth?

I think you can argue several different approached to depth.  Jeff is looking at total number of theoretically useful players, I am looking at a ratio of front line to second line performance to see how much the team is expected to lean on it’s front line, but I also think you could look at two approaches similar to these.  How many capable fill-ins and back-ups are there, Jeff’s number of players minus the number of starters in it would be a simple possible approach to look at how many holes are behind the first group.  Another would be total WAR drop from group 1 to 2 as a percent of front-line, or in other words how much worse is the second group in percent terms.  I could keep going as I have at least three other possibilities, but hopefully you get the point that depth is not a concrete concept just like what does valuable mean in MVP.

What I think might be the best statistical approach to this sort of problem is to have multiple independent people do what I and Jeff have already done and then aggregate the rankings.  Then our approaches can be biased by whatever version of depth we lean toward and let the problems with any given system of measurement be offset by the others.  This isn’t necessary to evaluate all teams, the Reds depth is bad period, but if you look at teams like the Yankees who I think are a little harder to project this year it could be useful.  Since I am a hobbyist who nearly no one knows or cares about, you can now disregard that pipe-dream, though I think over time a system like that would help in understanding how valuable depth is.

If you haven’t read Part I, you should consider doing so. For those who would prefer not to, this is a reminder of where these numbers came from:

Using FanGraphs’ terrific leaderboard tools, I found statistics for all players who played at each position in 2014. The following numbers apply only to the time spent at that position. Buster Posey, for example, accumulated 462 plate appearances at catcher, 128 at first base, and 9 at DH, so his plate appearances in those amounts are included for those positions in the table below.

For position players, I calculated WAR per 600 plate appearances. For starting pitchers, I used WAR per 150 innings pitched. For relievers, I used WAR per 50 innings pitched. Here is the table:

The table below shows a combination of all position players split into different age groups. The PA% column shows the percentage of plate appearances for each age group out of the total plate appearances for all hitters. As you can see, the 27 to 29 age range had the largest percentage of plate appearances and the “36 and up” group had the lowest. Similarly, the WAR% shows the percentage of total WAR accumulated by each age group. The column labeled “W%-PA%” shows the difference between the WAR% and the PA% columns. A positive difference is good. This means that age group was responsible for a higher percentage of WAR than their percentage of plate appearances.

POSITION PLAYER BREAKDOWN BY AGE

The meaty production for position players comes in the two age groups spanning ages 27 to 32. Players in this age range accounted for 51.1% of the total plate appearances and 57.6% of the total WAR. This age group didn’t hit as well as the “33 to 35” age group, but accumulated more WAR/600 PA, mainly due to the difference in the defensive component of WAR.

In news that should not surprise anyone, the “36 and up” age group was the worst, accumulating 4.4% of the total plate appearances but finished with -1.6 WAR with a wRC+ of 80.

Have you heard that on base percentage is an “older player’s skill”? Well, this table backs that up, as OBP increases for each age group from the “under 23” group up to the “33 to 35” group. That skill only lasts for so long, though, as offensive production plummets across-the-board in the “36 and up” age group.

STARTING PITCHER BREAKDOWN BY AGE

In the hitter’s breakdown by age, the two most productive age groups spanned the ages from 27 to 32. For starting pitchers, the two most productive age groups were from age 24 to 29, with the “27 to 29” group being the most productive. This “27 to 29” age group had the best ERA, WHIP, and K/9. For starting pitchers, the “36 and up” age group was actually better than the “33 to 35” group, although the sample size was small, including just 13 pitchers 36 and older.

Strikeouts are important for every pitcher and the above chart shows how strikeouts dwindle as a pitcher ages. The “27 to 29” group has a K/9 of 7.8. This drops to 7.4 K/9 for the “30 to 32” group, then to 6.6 K/9 for the “33 to 35” age group.

RELIEF PITCHER BREAKDOWN BY AGE

For relief pitchers, the bulk of the production is right there in the “27 to 29” age group. This group accounted for 31.5% of the relief pitcher innings and 40.9% of relief pitcher WAR. They also were tops among all groups in ERA, WHIP, and K/9. The next-oldest group of relievers (“30 to 32”) was the worst, accumulating 20.6% of the relief pitcher innings but just 11.1% of relief pitcher WAR. This looks like the age where strikeouts drop precipitously. Relief pitchers aged 27 to 29 averaged 8.8 K/9 as a group, while those in the “30 to 32” age group had just 8.1 K/9.

The information above is the big picture breakdown by age group for all hitters, starting pitchers, and relief pitchers. Part 3 of this series will show the age group breakdown for each position.

What better to do in the middle of winter when there are still a couple weeks until pitchers and catchers report than to look at WAR. In particular, I was curious about WAR in 2014. What positions had the most WAR? What age group? How did younger pitchers compare to older pitchers? So many WAR questions…

I started with the breakdown of WAR by position.

Using FanGraphs’ terrific leaderboard tools, I found statistics for all players who played at each position in 2014. The following numbers apply only to the time spent at that position. Buster Posey, for example, accumulated 462 plate appearances at catcher, 128 at first base, and 9 at DH, so his plate appearances in those amounts are included for those positions in the table below.

For position players, I calculated WAR per 600 plate appearances. For starting pitchers, I used WAR per 150 innings pitched. For relievers, I used WAR per 50 innings pitched. Here is the table:

HITTERS

PITCHERS

Some things that stand out for me are listed below. These aren’t earth-shattering insights, but interesting nonetheless:

• You can see the influence of the positional adjustment and defensive value by comparing some positions. For example, left fielders and center fielders had similar offensive numbers in 2014 (LF: .321 wOBA, 104 wRC+; CF: .319 wOBA, 103 wRC+), yet there was a 1.0 difference in WAR/600 PA.
• The three weakest-hitting spots—catcher, second base, and shortstop—make up for it with their defensive chops and the defensive adjustment.
• Players at first base had the best hitting numbers (.332 wOBA, 112 wRC+) but the lowest WAR total among all position players (DH not included).
• Players in the Designated Hitter spot accounted for just 8.4 WAR and three players accounted for 8.8 WAR (the rest accumulated negative WAR): Victor Martinez (3.9 WAR), David Ortiz (2.7 WAR), and Chris Carter (2.2 WAR).
• Starting pitchers had a better ERA than relievers (3.82 to 3.58), which isn’t surprising, but relievers had a higher WHIP (1.28 to 1.27), which did surprise me. Relievers struck out more batters (8.5 K/9 to 7.4 K/9) but also walked more (3.3 BB/9 to 2.7 BB/9).

NEVER TRUST ANYONE OVER 30

The following tables show the breakdown for all hitters, starting pitchers, and relief pitchers by age; specifically, the group of players aged 29 and younger compared to the “30 and over” group.

Not surprisingly, players 29 and under were better than players 30 and over and this was true for hitters and pitchers. There was a big difference in the magnitude, though. For hitters, the difference was about 0.3 WAR/600 PA. This is true even though the older group of hitters had a better wRC+. Defense matters.

For starting pitchers, the difference was 0.6 WAR/150 IP, with starting pitchers 29 and under accumulating 65.3% of the innings pitched by starting pitchers and 72.2% of starting pitcher WAR. Starting pitchers 29 and younger had a K/9 of 7.5, while those 30 and older saw their K/9 drop to 7.0.

Relief pitchers showed the greatest difference between the two age groups in WAR% – IP%, with the younger group finishing at +10.0% (65.3% of the innings pitched, 75.2% of the WAR). There was a big difference in strikeout rate for the two groups, with the younger relief pitchers getting more strikeouts (8.7 K/9 to 8.0 K/9).

POSITION BREAKDOWN BY AGE (29 AND UNDER vs. 30 AND OVER).

Young catchers had 0.3 more WAR/600 PA than older catchers. On the offensive side, young catchers outslugged older catchers (.394 to .360) but had a lower OBP (.308 to .311).

Best catcher 29 and under: Jonathan Lucroy, 28 (6.3 WAR)

Best catcher 30 and older: Russell Martin, 31 (5.4 WAR)

Dishonorable Mention: Jose Molina, 39 (-1.3 WAR)

First base had a near 50-50 split in plate appearances for players 29 and under and 30 and over, but the younger players were 0.5 WAR/600 PA better.

Best first baseman 29 and under: Anthony Rizzo, 24 (5.6 WAR)

Best first baseman 30 and older: Miguel Cabrera, 31 (4.9 WAR)

Dishonorable Mention: Jon Singleton, 22 (-1.1 WAR)

Second base was one of four positions (DH included) at which players 30 and over had more WAR/600 PA than the younger group.

Best second baseman 29 and under: Jose Altuve, 24 (5.0 WAR)

Best second baseman 30 and older: Ian Kinsler, 32 (5.5 WAR)

Dishonorable Mention: Stephen Drew, 31 (-1.0 WAR)

Third base had the biggest discrepancy between players 29 and under and 30 and over when it comes to WAR/600 PA, with a difference of 1.2 WAR/600 PA in favor of the older group, even as the younger group had almost three times as many plate appearances.

Best third baseman 29 and under: Josh Donaldson, 28 (6.5 WAR)

Best third baseman 30 and older: Adrian Beltre, 35 (5.7 WAR)

Dishonorable Mention: Matt Dominguez, 24 (-1.7 WAR)

Shortstop was another position at which players 30 and over had more WAR/600 PA, thanks in part to a better hitting line across the board.

Best shortstop 29 and under: Troy Tulowitzki, 28 (5.2 WAR)

Best shortstop 30 and older: Jhonny Peralta, 32 (5.3 WAR)

Dishonorable Mention: Josh Rutledge, 25 (-0.8 WAR)

You can see how youth plays a role in the different outfield positions by observing the plate appearance percentage for each position. In left field, the split is roughly 60-40 in favor of players 29 and under. In centerfield, where speed is more important, 77% of the plate appearances were given to player 29 and under. In right field, it was much closer to 50-50. All three outfield positions saw more WAR/600 PA from the younger group of players in 2014.

Best left fielder 29 and under: Michael Brantley, 27 (4.5 WAR)

Best left fielder 30 and older: Alex Gordon, 30 (6.6 WAR)

Dishonorable Mention: Domonic Brown, 26 (-1.6 WAR)

Best center fielder 29 and under: Mike Trout, 22 (7.5 WAR)

Best center fielder 30 and older: Jacoby Ellsbury, 30 (3.9 WAR)

Dishonorable Mention: Junior Lake, 24 (-2.5 WAR)

Best right fielder 29 and under: Giancarlo Stanton, 24 (6.2 WAR)

Best right fielder 30 and older: Jose Bautista, 33 (6.1 WAR)

Dishonorable Mention: Oscar Taveras, 22 (-1.2 WAR)

The DH spot is an older player’s spot, with 63% of the plate appearances at DH given to players 30 and over. This group accounted for 89% of the DH WAR, with a higher batting average, on-base percentage, and slugging percentage.

Best DH 29 and under: Chris Carter, 27 (2.2 WAR)

Best DH 30 and older: Victor Martinez, 35 (3.9 WAR)

Dishonorable Mention: Kendrys Morales, 31 (-1.5 WAR)

Pitchers are just terrible hitters, old and young, fat and skinny, tall and short. They stink at hitting. Older pitchers are a little more stinky at hitting than younger pitchers.

Best Pitcher (hitting) 29 and under: Travis Wood, 27 (1.0 WAR)

Best Pitcher (hitting) 30 and older: Madison Bumgarner, 30 (1.2 WAR)

Dishonorable Mention: Bartolo Colon, 41 (-0.7 WAR) Colon was 2 for 62 with 0 walks and 33 strikeouts. Somehow, he managed to score 3 runs. That’s kind of mind-boggling, really.

This might surprise some people. When I think of pinch-hitters, I picture the aging veteran who calmly comes off the bench to deliver a big hit, like Manny Mota in the 70s or Rusty Staub in the 80s or Matt Stairs in the 00s. Last year, though, pinch-hitting was a younger man’s game. Players 29 and under had 59% of the pinch-hitting plate appearances and a slightly better triple-slash batting line.

Best Pinch-Hitter 29 and under: Delmon Young 28 (0.6 WAR)

Best Pinch-Hitter 30 and older: John Mayberry, 30 (0.8 WAR)

Dishonorable Mention: Greg Dobbs, 35 (-0.5 WAR)

That’s probably enough for now. More likely, it’s way too much. Either way, if you notice anything interesting about these numbers, please make your observations known in the comments. Next up is a more involved breakdown of WAR by age group for hitters and pitchers.