FanPost

Subtraction by Addition

Mark J. Rebilas-USA TODAY Sports

Addition by subtraction is one of those sports phrases that's almost become cliche. When we added Rickie Weeks Scott Weber said:

Weeks isn't going to be a door-buster impact player, but this is addition by subtraction. If he can even be replacement level, he'll improve on what the M's struggled to replace a year before.

At the beginning of 2014, David Schoenfeld of ESPN claimed the Mariners would surprise because dumping Raul Ibanez, Michael Morse and Jason Bay was "addition by subtraction." Later that year I thought it was addition by subtraction when replaced Abraham Almonte (0.2 WAR, 113 plate appearances) with James Jones (-1.0 WAR, 328 PA), and then I thought it was addition by subtraction when replaced Jones with Austin Jackson (-0.3 WAR, 236 PA ). Then I started wondering if the Mariners have unusual issues with negative-WAR players (spoiler: yes) and whether they'll hurt us again this year.

Addition by subtraction should be easy. Zero-WAR replacement level players are easily available by definition. But for addition by subtraction to be real, some player on the team is playing at below replacement level. At some point we did subtraction by addition. How common could that be?

I went to Fangraphs for stats on all players with >100 PAs for a team over each of the past 5 seasons and took a look at their WARs.

WAR Histogram

Well, I don't know about you, but I found that somewhat surprising. Over the last 5 years, a full 24% of position players finished the season with a WAR below zero. An important caveat is that those bad players received fewer plate appearances than their decent peers. But still, 1/6 of the plate appearances went to sub-replacement players. I should also note that most of those at bats went to players with -0.5 to 0.0 WAR, which either makes you feel better about about baseball's ability to identify replacement level, or worse about James Jones.

As to how we end up with this picture, it's clear with a little reflection. Those "replacement level" players are expected to deliver around zero WAR, but sometimes bad things happen to mediocre people. By the time a player has had enough plate appearances to be fairly evaluated, they've also had enough plate appearances to take their team out of the playoffs.

How bout them Mariners?

OK, it sure seems like we've had more than our fair share of these sub-replacement level players, but is that really true? Using the same data, here's some relevant charts.

Negative WAR Seasons Negative WAR Totals

The first chart shows the number of sub-replacement seasons (e.g. Justin Smoak was less than zero for 2010, 2012 and 2014, so he's counted 3 times). The second chart shows the total damage done in those years (-0.9 WAR for Justin). And my intuition was right: the Mariners are second worst on both charts, probably due to bad roster construction compounded by bad luck. Notice that the teams that have been good over the last few years don't get burned by sub-replacement players.

OK sure, that's bad, but our new roster is much better. Jack isn't making the same mistakes anymore, is he? I'm hopeful, but let's take a deeper dive into this cesspool and figure out what kinds of players have delivered negative WAR in the past and whether we might have the same kinds of issues this year. Someday I would like to do a statistical clustering analysis, but for now here are the categories I dreamed up and a few exemplars, mostly from the M's.

Category 1: Faded Stars

Also known as the cure for nostalgia. Griffey cut short his last season, but still managed to put up -1.0 WAR in only 108 plate appearances. Derek Jeter's -0.3 in his farewell tour seems pretty good by comparison.

Season Name Team G PA BABIP AVG OBP SLG wOBA wRC+ WAR
2010 Ken Griffey Jr. Mariners 33 108 0.22 0.184 0.25 0.204 0.215 32 -1.0
2011 Jorge Posada Yankees 115 387 0.262 0.235 0.315 0.398 0.313 92 -0.7
2014 Derek Jeter Yankees 145 634 0.294 0.256 0.304 0.313 0.279 73 -0.3
2014 Carlos Beltran Yankees 109 449 0.252 0.233 0.301 0.402 0.31 95 -0.5

2015 Suspects: None. Ichiro would be the only candidate and though I wish him the best I'll just paraphrase a line from Fiddler on the Roof: "May God bless and keep Ichiro -- far away from us".

Category 2: Faded Mediocrities

These guys were never that good. Then they got worse.

Season Name Team G PA HR AVG OBP SLG wOBA wRC+ WAR
2013 Brendan Ryan Mariners 87 287 3 0.192 0.254 0.265 0.235 45 -0.6
2013 Endy Chavez Mariners 97 279 2 0.267 0.290 0.327 0.272 71 -1.2
2014 Brendan Ryan Yankees 49 124 0 0.167 0.211 0.202 0.190 12 -0.7
2012 Yuniesky Betancourt Royals 57 228 7 0.228 0.256 0.400 0.280 73 -1.0
2013 Yuniesky Betancourt Brewers 137 409 13 0.212 0.240 0.355 0.257 55 -1.9

But hey at least we got rid of Yuni and Ryan before the worst.

2015 Suspects: Looks pretty good. Maybe Rickie Weeks, but I'm not too worried.

Category 3: Broken Down Sluggers

Seems like this has been Jack Z's crack habit, but the Jack Custs of the world didn't play enough to do much damage. The players below did damage, though notably a number of them had at least one decent year after crashing.

Season Name Team G PA HR AVG OBP SLG wOBA wRC+ WAR
2013 Michael Morse Mariners 76 307 13 0.226 0.283 0.41 0.303 92 -1.1
2014 Kendrys Morales Mariners 59 239 7 0.207 0.285 0.347 0.282 81 -0.9
2014 Corey Hart Mariners 68 255 6 0.203 0.271 0.319 0.266 70 -1.2
2014 Kendrys Morales Twins 39 162 1 0.234 0.259 0.325 0.255 59 -0.8
2011 Raul Ibanez Phillies 144 575 20 0.245 0.289 0.419 0.306 90 -1.7
2011 Adam Dunn White Sox 122 496 11 0.159 0.292 0.277 0.268 60 -3.0

2015 Suspects: Please stay healthy Mr. Cruz.

Category 4: Failed Replacements

These guys were never expected to be much more than zero-WAR players, but they turned out to be much less.

Season Name Team G PA AVG OBP SLG wOBA wRC+ WAR
2010 Matt Tuiasosopo Mariners 50 138 0.173 0.234 0.307 0.243 51 -0.8
2010 Adam Moore Mariners 60 218 0.195 0.23 0.283 0.227 40 -1.1
2011 Trayvon Robinson Mariners 44 155 0.21 0.25 0.336 0.258 63 -1
2014 James Jones Mariners 108 328 0.25 0.278 0.311 0.262 68 -1
2014 Stefen Romero Mariners 72 190 0.192 0.234 0.299 0.239 51 -1.2
2012 Eric Thames Blue Jays 46 160 0.243 0.288 0.365 0.285 74 -0.6

Oh, Hello Eric Thames.

2015 Suspects: I don't know if you can see these guys coming in advance. Could be James Jones again?

Category 5: Struggling Prospects

Smart people thought they would be good. Some of them are.

Season Name Team G PA BABIP AVG OBP SLG wOBA wRC+ WAR
2011 Michael Saunders Mariners 58 179 0.212 0.149 0.207 0.217 0.192 19 -0.5
2012 Justin Smoak Mariners 132 535 0.242 0.217 0.29 0.364 0.288 85 -0.5
2010 Alex Gordon Royals 74 281 0.254 0.215 0.315 0.355 0.303 85 -0.5
2010 Michael Brantley Indians 72 325 0.271 0.246 0.296 0.327 0.28 73 -0.6

2015 Suspects: I suppose Miller could fall here, but seems more likely to break out. Either way we have good insurance in Taylor so he's unlikely to be allowed to subtract much.

Category 6: Inconsistent Fringey Guys

Season Name Team G PA BABIP AVG OBP SLG wOBA wRC+ WAR
2013 Logan Morrison Marlins 85 333 0.281 0.242 0.333 0.375 0.312 96 -0.6
2010 Casey Kotchman Mariners 125 457 0.229 0.217 0.28 0.336 0.27 69 -1.5
2012 Casey Kotchman Indians 142 500 0.233 0.229 0.28 0.333 0.271 70 -1.6

2015 Suspects: See above - I'm a bit worried about LoMo. Kotchman had a couple years of 2+ WAR. Last year's 1.0 is the best Logan has been able to muster. But the new Montero will step in if needed, right?

Category 7: Figgins

The Figgins thing is a mystery to me. In 2009 he put up over 6 WAR for the Angels. 2010 was worse than expected, but in 2011 at age 33 he was way worse than replacement.

Season Name Team G PA BABIP AVG OBP SLG wOBA wRC+ WAR
2011 Chone Figgins Mariners 81 313 0.215 0.188 0.241 0.243 0.220 37 -1.2
2012 Chone Figgins Mariners 66 194 0.237 0.181 0.262 0.271 0.242 54 -1.1

2015 Suspects: Austin Jackson is only 28 but Jake Mailhot outlined his struggles last year. Last year was a fluke for him right? I'm not worried. Actually, I am worried.

Summary

This year's roster seems both better in the middle and less fragile at the bottom than years past. There's no way sub-replacement players are going to put up -5.6 WAR like they did last year. On the other hand, almost every team ends up with someone underperforming replacement level even if Steamer thinks everyone is above replacement . I'm happy where we are, but after doing this analysis I'm a slightly more nervous. For your viewing pleasure, here's a link to underperforming Mariners for the last 5 years.