I now introduce you to the 2nd member of the new formed TRDM Sabermetrics team. You thought only the Mets could bring in high quality saber guys?— TRS
Don’t Forget The Leather and the Lumber: Mets Starters’ WAR after Looking at Fielding and Batting
Wins Above Replacement (WAR) is a beautiful and simple statistic for measuring a player’s value to his team. The metric starts with the idea of a replacement player, the kind of guy that every team has in AAA waiting for a call-up in an injury or emergency. Think of Jesus Feliciano, Mike Hessman, or 2009’s Tim Redding. These are often career minor-leaguers who are good enough to hold down the fort in the majors, but who are not talented enough and are abundant enough that pretty much every team can call upon someone of their quality to fill a roster hole. Baseball is all about winning games, and a player’s Wins Above Replacement is simply how many more games a team would win with him on the field as opposed to the generic replacement player. It attempts to take everything he does, and puts it into the one, simple number of wins, the most important number in the game.
The 2010 Mets had four pitchers who were mainstays in their rotation and who threw a significant number of innings a starter: Mike Pelfery (203 IP), Johan Santana (199 IP), Jon Niese (173.2 IP), and The R.A. Dickey (173.1 IP). Pitcher WAR on fangraphs.com uses the Fielding Independent Pitching (FIP) metric combined with innings pitched to come up with the number of wins added. FIP only takes into account plate appearances entirely decided by the pitcher and hitter: strikeouts, walks, homeruns, and hit-by-pitches. This discounts balls put in play, because then luck and defense start taking control. FIP, and by consequence WAR, only tries to measure the things directly under the pitcher’s control. The validity of this approach lends itself to a lively debate, but that’s a-whole-nother article.
Of the Mets starters, Santana led the way with 3.5 WAR, Pelfrey added 2.9, the Dickey also inserted 2.9 WAR into the Mets win total, and Niese contributed 1.9. All of these marks are average or better, since a full season of average starting pitching should lead to about 2 WAR.
While WAR provides good estimates of how valuable these Met pitchers were on the mound, there are two aspects of every pitcher’s game that are not covered in WAR: fielding and batting. Pitchers take the field just like every other position, and they can cost or save a team runs just like any other fielder. Watching Johan Santana pounce like a cat on a slow dribbler, you know that pitchers can help win games with their gloves as well as their arms. And one out of every nine spots in the lineup is reserved for the starting pitcher. They have to bat just like every one else in the starting lineup, and when they ground out weakly with two outs and runners on second and third, those two ducks left on the pond are just as stranded as if a three-hitter left them there. WAR just does not give pitchers enough credit or blame for their bats and their gloves.
So if we want to know how many wins a pitcher truly added to his team, we have to look at their batting and fielding. If we update the WAR statistic to incorporate these things, what happens to these players’ value? Well, the Mets starters as a whole are significantly more valuable once you include all aspects of their game.
To understand the results, a quick note on how WAR works. WAR comes up with its win value by taking the number of Runs Above Replacement (RAR) a player earned and converting that to wins, which makes sense because you win games by scoring runs. So a certain number of runs scored, usually around ten, will equal one win added. I took the pitching RAR, and I added to it the fielding and hitting runs above average. I took that number and then converted it to wins. First, I’ll give you the results; then I’ll give you the methodology.
Here are each player’s adjusted RAR:
| Orig. RAR | Fielding Runs | Batting Runs | New RAR | |
| J. Santana | 33.1 | 2 | 2.0 | 37.1 |
| M. Pelfrey | 27.3 | -2 | -2.5 | 22.8 |
| R. Dickey | 26.9 | 5 | 5.7 | 37.6 |
| J. Niese | 18.5 | 3 | 5.2 | 26.7 |
And here are each player’s adjusted WAR, based on these New RAR:
| Orig. RAR | Orig. WAR | New RAR | New WAR | Change in WAR | |
| J. Santana | 33.1 | 3.5 | 37.1 | 3.9 | 0.4 |
| M. Pelfrey | 27.3 | 2.9 | 22.8 | 2.4 | -0.5 |
| R. Dickey | 26.9 | 2.9 | 37.6 | 4.1 | 1.1 |
| J. Niese | 18.5 | 1.9 | 26.7 | 2.7 | 0.8 |
Look at the R.A. Dickey! He added 1.1 wins to the 2010 Mets just by being a great fielder and hitter. This new WAR total would make him the thirteenth most valuable pitcher in the National League in 2010. He’d catapult above guys like Matt Cain, Cole Hamels and Chris Carpenter. The legend continues to grow.
And Santana and Niese also help out their WAR causes when we consider all facets of their games. And this was a below-average hitting and fielding season for Johan, so in general he’d add even more value with the bat and glove. Big Pelf was the only Met starter to lose ground, but he was still an above-average big-league pitcher. Overall, the Mets starters were worth 1.8 more WAR after taking all their contributions into account.
One thing you might notice is that the conversion factor between RAR and WAR is not exactly the same for each pitcher. That’s because each pitcher pitches in a different run environment, so runs are more valuable for some than for others. The ballparks a pitcher throws in could affect his run environment, as a run in Petco is more valuable than in Citizens Bank. And the pitchers themselves affect their run environment. With an ace pitcher on the mound, runs are harder to come by and at a premium, so each run becomes more important.
You might be wondering how in the hell I came up with these batting and fielding runs above replacement numbers. I’ll tackle fielding first because my technique was a good deal simpler, and then I’ll get into batting.
Since I was using numbers from fangraphs.com for the original WAR totals, it made sense to me to use the site for the additional statistics as well. The only fielding metric that gives a number of runs above or below average for pitchers on fangraphs is Defensive Runs Saved (DRS) from the Plus-Minus system. For more information on how this works, check out this link:
This goes pretty in depth so read at your leisure, but it explains the system much better than I ever could. But what matters for us it the end result. We get a number of runs above or below average for a fielder, and once we know how many runs he was worth in the field, we can add that to our WAR calculation. The number you see as Fielding runs is simply that pitcher’s DRS. It should be noted that defensive metrics are notoriously fallible in small sample sizes. Ideally we would be able to average a few different stats to come up with something more precise, but we had to work with what we have, and it’s certainly better than nothing.
When it came to batting, I based my numbers on the wOBA statistic. wOBA is based on linear weights, which is much simpler than it might sound. Every time you get a hit, it adds to the expected number of runs scored in that inning. If you hit a single you increase the number runs you can expect to score, and if you hit a home run you increase it by a lot more. A linear weight is just the average change in run expectancy for each way to reach base and for stolen bases and caught stealing. To find a batter’s wOBA, you multiply his number of singles by the average change in run expectancy for a single, and then you do that for doubles, triples, etc.. Then you divide that number by the batter’s number of plate appearances. To make wOBA look like something we are more familiar with, it is scaled to OBP, so there is a small adjustment, about a 15% increase, to put it on the same scale as OBP. So around .330 is average, .400 is elite, and .300 approaches (but does not quite reach) Gary Matthews Jr. territory. To find out how a wOBA converts to runs, you subtract the league average wOBA from the player’s wOBA. Then you readjust for the OBP scaling factor. This number will be how many runs above the league average a player provides per plate appearance. If you multiply that number by the batter’s total PAs, you have his batting runs above or below average. It might sound a little bit complicated, but it’s probably the most telling single batting statistic we have right now.
For pitchers though, using traditional batting runs creates a problem. Fangraphs actually does calculate runs above or below average for pitcher batting, but it mistakenly compares them to league average and wisely does not incorporate this number into their WAR. The fact is that batting for position players and pitchers are judged on fundamentally different levels. While pitchers do take up a lineup spot, they take the very specific one reserved for the pitcher, almost always the last in the order. We do not compare Santana’s bat to Jose Reyes’, we compare it to other pitchers’. There is no positional adjustment for pitchers that accounts for the fact that they are such poor hitters, but if we compare a certain pitcher’s wOBA to the average pitcher’s wOBA, then we can really see how many runs above or below average he was with the bat.
So I calculated the average pitcher’s wOBA from the major league’s last season, and it came out to .160. This is close to a whopping .170 points less than the average hitter (if the average wOBA includes pitchers, this difference would likely be even bigger if we just compared it to average position-player wOBA). I then did the same process as usual for converting wOBA to runs above average. Only this time, I compared it to the average pitcher’s wOBA as opposed to league average wOBA. The Dickey led the way with a .268 wOBA in 61 PA, Niese had a .250 wOBA in 66 PA, Santana had a .195 wOBA in 67 PA, and Pelfrey had a .119 wOBA in 70 PA. The difference between these numbers and .160, then readjusted from the OBP scale, multiplied by the pitchers’ PA, gave the batting runs above average total in the charts above.
So there you have it. We now have a new, more encapsulating version of WAR that shows our Mets starters are better than the traditional WAR stat would have us believe. And I’m convinced that anytime we improve upon our statistics, it will only further reveal the greatness that is The R.A. Dickey.
42 comments
kingman 26
11/9/2010-8:58am at 8:58 am (UTC -4)
Really nice piece, and it is so refreshing to see a stat which actually does measure what we clearly see. Dickey is an excellent hustling fielder, and on more than one occasion was the best hitter in the lineup when he pitched.
Prismo
11/9/2010-9:09am at 9:09 am (UTC -4)
WAR, huh – yeah, what is it good for?
Oh, that.
Ceetar
11/9/2010-9:14am at 9:14 am (UTC -4)
Does wOBA take into account runners on? It must right?
One thing that always gets me here is the Runs->Wins conversion. There’s an assumption of average luck there that doesn’t usually apply, especially for the smaller sample sizes that pitchers have. So when you apply it into WAR, it mistakenly implies that it’s talking about the team’s over all win record, when really it’s just a number scale evaluating pitchers.
“How much do they contribute to the success of a winning ballclub”" versus “how many wins did they help the team achieve?”
A pitcher with no run support could get let up 1 run a game for 30 starts, lose every game 2-1, and still have an amazing WAR.
All runs are not equal.
reyordonezrevival
11/9/2010-9:15pm at 9:15 pm (UTC -4)
I think that WAR takes every event out of context is actually one of its great strengths. It only rewards you for what you do, not what your teammates do. If you’re batting third and the top of the lineup is always getting on base in front of you, you will naturally have a lot of RBIs. If you are no more or less skilled than the number eight hitter, you will have more RBIs, but that doesn’t reflect any skill on your part. You were just in the right place at the right time. WAR tries to account for this fact and only reward you for things you have done, regardless of your situation.
“A pitcher with no run support could get let up 1 run a game for 30 starts, lose every game 2-1, and still have an amazing WAR.” And should we not give this pitcher credit? He pitched amazingly, so he should get credit. The batters who didn’t hit and who are responsible for the losses will have their WAR hurt, so it will even out. Only give credit or blame to who is really deserving of it, or at least that’s what WAR does.
Obviously there’s a good counterargument to this which you brought up, but I thought I’d present the rationale. There is a stat called Win Probability Added (WPA) that takes context into account. Basically, it looks at what your team’s odds were of winning before your at bat versus after your at bat. If you improve your team’s chances, you get credited by exactly that amount, and vice versa. There’s definitely a lot of room for debate on which is a better approach though, I just happen to prefer the context-neutral approach because you cannot control your context, only what you do.
oleosmirf
11/9/2010-9:16am at 9:16 am (UTC -4)
the question is can RA Dickey be even somewhat near as good as last season????
Ceetar
11/9/2010-9:16am at 9:16 am (UTC -4)
Sure, why can’t he?
oleosmirf
11/9/2010-9:23am at 9:23 am (UTC -4)
can you tell me the last time a mid 30s career minor leaguer with not even an average season under his belt comes out of nowhere to have a Cy Young caliber season and then goes out the next year and repeats that performance???
Ceetar
11/9/2010-9:29am at 9:29 am (UTC -4)
Knucklerballers? I imagine it’s happened.
I was skeptical through about June, but when he continued to dazzle and make adjustments, when he beat the same teams over again, I believed.
R.A. Dickey from 2010 is not the same pitcher he was in 2005, so it’s nto even fair to compare. He owns different pitches now, he pitches differently. If you equate his use of the knuckleball to coming up as a rookie, you’d say 2010 was him coming into his own.
oleosmirf
11/9/2010-9:43am at 9:43 am (UTC -4)
I am still skeptical, if he still pitching this well or close to it at June 15th, i will no longer consider him a fluke…
TRS86
11/9/2010-9:42am at 9:42 am (UTC -4)
You tell me another mid 30′s career minor leaguer with not even an average season out of his belt came out of nowhere to have a Cy Young caliber season and perhaps I can answer that question.
oleosmirf
11/9/2010-9:48am at 9:48 am (UTC -4)
which is basically my point. i would absolutely love Dickey to be a perennial all-star as he might be the most likeable Met since John Franco but i cant help but think 2010 was just a fluke.
TRS86
11/9/2010-10:06am at 10:06 am (UTC -4)
You are welcome to that thought BUT if it happen last year and not previously that does not write the book for the future. It does as you say make that return less likely than someone who had done it for 5 years.
TRS86
11/9/2010-10:09am at 10:09 am (UTC -4)
It could be said, while certainly successful, that Wakefield did not reach his stride until his early 30′s either.
Mr North Jersey
11/9/2010-9:30am at 9:30 am (UTC -4)
My comment yesterday started a huge debate so I will attempt to say this without causing too much fan fare.
To me personally speaking only for myself what prevents me from accepting this War statistic is the following.
“Wins Above Replacement is simply how many more games a team would win with him on the field as opposed to the generic replacement player.”
This is an estimate of the maker of this stat of how many more wins a team would win as opposed to a generic replacement (in other words fictional). Which is something I always have had trouble with because it is not a hard numbered statistic.
What I mean is I know that slg% is based on a hitters total bases divide by at bats for example or that OBP is without reciting the formula is for the most part a result of a players combined Hits and Walks.
WAR is something that can be used to reference a player’s value as suggested but personally for me I just find it difficult to accept.
Ceetar
11/9/2010-9:39am at 9:39 am (UTC -4)
You have to look at it as a scale, and try not to overapply it to anything. Since it’s calculated the same way, the a higher number for Dickey means he’s theoretically ‘better’.
It’s harder to actually apply it to a team value. Also, that replacement value is an average, and hardly an easy to grab piece. Mike Hessman, Jesus Feliciano..sure..but it’s not a given. You could grab the minor league replacement guy, and he could suck worse. So Carlos Beltran could be “worth” more to the Mets because the difference in value between him and the _Mets_ replacement guy is greater than the average replacement value calculation, but that doesn’t factor into Beltran’s WAR.
But please keep up the debate. I’m hoping to write a couple of skeptical sabermetric articles this offseason, and the debate and questions help.
oleosmirf
11/9/2010-9:41am at 9:41 am (UTC -4)
not to mention how i find it hard to believe David Wright would only add 4 wins to the Mets roster over Fernando Tatis
TRS86
11/9/2010-9:47am at 9:47 am (UTC -4)
Again I think if you get too caught up in any stat it becomes faulty. As Cheetar mentioned it’s really most useful when comparing players of similar positions to each other. It’s never going to actually account for runs scored or wins over a fictional player as NJ suggested nor is it going to factor in other variables beyond it’s control. However, it is just another in a long line of stats that tries to compare a players overall impact.
oleosmirf
11/9/2010-9:50am at 9:50 am (UTC -4)
exactly. I think its a good measure when comparing 2 players but i wouldnt value it more than OPS
TRS86
11/9/2010-10:04am at 10:04 am (UTC -4)
Hmmm, once you factor in fielding it may be more valuable. Depending on how reliable those fielding stats are. For example is Angel Pagan’s ultimate worth completely measured by OPS or do you also have to factor in defense and speed?
stickguy
11/9/2010-10:13am at 10:13 am (UTC -4)
you have to take all factors of the game into account. that is why an aggregate measure like WAR is handy (if you trust it).
ANd as noted, it is supposed to be a relative scale, not an absolute “real” #.
njstuckintx
11/9/2010-10:28am at 10:28 am (UTC -4)
Cheetar?
Ceetar
11/9/2010-9:48am at 9:48 am (UTC -4)
Another quirk of fielding for pitchers is that they stand in front of guys. A pitcher will get rewarded for making a stop and throwing to first when not making the stop could mean the ball gets to the SS and they turn a double play. or get the lead runner instead of the guy at first.
stickguy
11/9/2010-9:58am at 9:58 am (UTC -4)
Can Murphy Pitch?
Ceetar
11/9/2010-10:03am at 10:03 am (UTC -4)
I think everyone should learn to pitch. Can you reall ybe worse than the average crappy relievers? and all pitchers should learn to play the outfield. This creates awesome versatility.
stickguy
11/9/2010-10:12am at 10:12 am (UTC -4)
Well, from my experience with youth BB, all the pitchers do play other positions. Often SS, but also the OF.
so really, there is no reason that a pitcher shouldn’t be able to play in the field too when needed. And I bet a ton of the position guys have plenty of pitching experience.
Ike was a stud pitcher when he wasn’t playing 1B in school (college too I believe). And I saw someplace that Duda could hit 90 when pitching in HS.
Ceetar
11/9/2010-10:13am at 10:13 am (UTC -4)
These are probably innovations that even LaRussa wouldn’t attempt though.
njstuckintx
11/9/2010-10:29am at 10:29 am (UTC -4)
LaRussa had Canseco pitch once, didn’t he?
metsfan4decades
11/9/2010-10:31am at 10:31 am (UTC -4)
I theory maybe but can you see any team taking it’s multi million dollar ace and sticking him in a fielding position? Everyone would be holding their breath when that said pitcher tried to turn a double play and someone came in spikes flying, or attempted a diving catch in the OF.
Yikes….
Ceetar
11/9/2010-10:32am at 10:32 am (UTC -4)
These guys are supposed to be athletes, not coddled throwers!
Although even Babe Ruth said it was too hard to pitch one day and play OF the rest. In fact, he may be one of the only guys to really do it, or the last anyway.
GravediggerHebner
11/9/2010-11:14am at 11:14 am (UTC -4)
With the caveat that I don’t fully understand all of this after reading it one of my 1st thoughts was that if he has any trade value at all the Mets should consider trading Pelfrey.
He’s a pitcher who due to an absence of total strikeouts relies more on his team to be successful (or not) and then his hitting and fielding detract further.
A useful pitcher when he’s hitting his spots and getting good defense, yes, but seems at least on the surface to require a lot more help from 8 other guys to achieve positive results than Santana, Niese or Dickey and it’s easier to sub out 1 pitcher than it is to ensure the entire team is good enough to hold that pitcher up.
Could be completely misunderstanding and/or misapplying but that’s what struck me after reading this.
Ceetar
11/9/2010-11:24am at 11:24 am (UTC -4)
Pelfrey’s groundball rate is high though, which tends to be a positive ‘controlled’ thing. If he hadn’t had that horrific mid-season disaster, even if it was just poor, he’d havea ton of value. But right now I dont think you’d get ar eturn, especially given that you’d want at least a pitcher of his caliber back, because that’s what we need.
GravediggerHebner
11/9/2010-11:43am at 11:43 am (UTC -4)
That makes sense and given the hole it would leave in the SP staff I agree with you that trading him at this juncture doesn’t seem wise.
But I read something when Alderson was being wooed about how he values offense over defense because offense is more clearly assigned to one person (the batter) whereas defense was the combined efforts of the pitcher and the 8 defenders so if this is true about Alderson I could see his team moving toward defensively challenged quality hitters and K pitchers.
Of course trading Pelfrey goes directly against the notion expressed directly by Alderson that having home grown players is important and desired so I’m certainly not trying to predict that Pelfrey is going to be traded, just thinking he’s far from ideal and nothing he actually provides on the field would make him untouchable IMO.
stickguy
11/9/2010-12:33pm at 12:33 pm (UTC -4)
So, does that mean there is hope for my man Duda to win the RF job next year?
He might just be the best pure young hitter they have on the roster (not counting the established ML guys) when you factor in plate discipline, power, etc.
well that, and how keith raved about his swing.
TRS86
11/9/2010-12:49pm at 12:49 pm (UTC -4)
Stick, how can we know that about Duda based on one real year of production at AAA? I am all for Duda but it’s not like he has established those numbers as a career work in the minors. He had one very good year with a SLG over 100 points higher than any other year in any location.
stickguy
11/9/2010-1:29pm at 1:29 pm (UTC -4)
the only thing that changed last year was he added to the slugging. Basically, he started hitting more HRs, while keeping a high OBP and excellent K and BB rates.
he is obviously extremely strong (and big), with a nice qick bat.
not the first guy where the power developed last, and in his case, it seems to be a bit of an approach and learning to trust his wrist.
one of the best things about him is plate coverage, and using the whole field (having been an opposite field hitter for a while). And those are 2 things you often don’t get with “pure” HR hitters (aka all or nothing hackers).
He could easily be a guy like Werth, where an injury set him back a number of years, so he ended up being a bit of a late bloomer
of course, I can’t guatantee anything about him. But, I was impressed by what I saw last year, and intrigued by his supporting numbers.
stickguy
11/9/2010-12:38pm at 12:38 pm (UTC -4)
trading Pelf is an interesting idea. Seems counter intuitive though, given that they are already short of SP, and the general dearth of pitching in MLB.
But, that same lack is why you might trade him, if the return is worth it. And the big finger licker should bring a ton of talent back.
He is young, durable, and cheap and under control for a number of years to come. All prime attributes. he is also at the age and ML service time that any breakout (of career years) would likely be coming up.
Of course, these are all good reasons to keep him too! And frankly, he is quite comparable to Garza, who most posters seem to think will return a huge haul of talent. And Pelf is actually (I beleive) cheaper right now.
anyway, I think Neise is easily the best young pitcher on the Mets, and if you need to keep 1 of the 2 (or 3, counting Mejia), you keep Neise.
TRS86
11/9/2010-12:44pm at 12:44 pm (UTC -4)
Actually perhaps I underrate our own players but I do not think the return for Pelfrey would be anywhere near the same as Garza.
wannybackstra
11/9/2010-9:51pm at 9:51 pm (UTC -4)
whip – garza = 1.315, pelfrey = 1.455
h/9 8.7 v. 9/8
hr/9 1.1 v. .6
k/9 7.1 v. 5.1
so/bb 2.23 v/ 1.56
bb/9 3.3 v. 3.2
era 3.97 (3.86 with tb in the AL East) v. 4.31
The only thing pelf does better is give up fewer HRs. Garza is better in a landslide really.
stickguy
11/9/2010-10:27pm at 10:27 pm (UTC -4)
pelf #s are really skewed by 2009. if you just look at 2008 and 2010 the discrepancy is much smaller.
and yes, I know that they all count, but keep in mind when you are evaluatng players, what they did in the past doesn’t count. Only what you expect them to do in the future. So from that standpoint, trends are important.
wannybackstra
11/9/2010-10:50pm at 10:50 pm (UTC -4)
since we dont have a crystal ball i;d say the best predictor we have is past performance.
stickguy
11/9/2010-10:52pm at 10:52 pm (UTC -4)
point is, which past? 1 year? 2 Years? 3 years?
wannybackstra
11/9/2010-9:44pm at 9:44 pm (UTC -4)
I understand and appreciate FIP to a degree. And I certainly understand and appreciate the concepts it values.
But while ignoring batted balls that rely on fielders it ignores the fact that a great many batted balls could not be fielded by anyone. In other words, a guy can throw meatballs but as long as they don’t leave the park they are no harm no foul.