By In Stuff

No Wins, No Errors

My pal Brian Kenny is trying to start a movement to kill the pitcher win, and the general theme is something I heartily endorse. I don’t actually believe n KILLING the win — it has a lot of history — but I do believe that every time someone puts a little bit less stock in the win, an angel gets its wings.

See it isn’t that the win is meaningless. It’s not meaningless. Defective stats are not meaningless. RBIs are not meaningless. Hit by pitches are not meaningless. Checked-swing strikes are not meaningless. Infield singles are not meaningless. But there’s a huge canyon, a large canyon, a big canyon, a (what’s the word I’m looking for) mammoth canyon between “meaningless” and “meaningful.” If someone told me they like Bruce Springsteen, there’s probably a better chance that I would like them than if they told me they despised Bruce Springsteen. But I wouldn’t want to build my friendship base around that.

Yes: In the grand sense (there’s that word!), good pitchers get more wins than bad pitchers. If you are told you have to choose one of two pitchers and the only information you are given is how many wins each had, sure, you would always take the pitcher with the higher win total.

It’s just that if you were given almost any other piece of pitching information — like runs per nine, home runs allowed, strikeout-to-walk, ERA+, some version of FIP, WAR, numerous other cool pitching stats — you would have a better shot at picking the better pitcher.

Give you an example. Take two pitchers — Pitcher A, Pitcher B.

PItcher A wins: 12

Pitcher B wins: 9

OK, obviously you know this is a trick and that Pitcher B is the better pitcher — otherwise why would I bring it up? But I’m not asking you to pick one pitcher or another. I’m trying to make the point that wins (and losses) are pretty much the only statistics out there that do not show CLEARLY that Pitcher B is the better pitcher. Let’s look at those losses:

Pitcher A losses: 10

Pitcher B losses: 12

OK, so Pitcher A is 12-10, pitcher B is 9-12. If that’s the only information you have, you choose Pitcher A.

Then, let’s say, you ask for ERA.

Pitcher A: 4.27

Pitcher B: 3.09

OK, whoa, you chose poorly. It’s obvious now that Pitcher B is actually better. But is he? Maybe Pitcher A is pitching in Colorado and Pitcher B is in San Francisco? How about we look now at strikeouts to walks.

Pitcher A: 90-to-50

Pitcher B: 181-to-37

Yikes. Pitcher B is suddenly looking A LOT better than Pitcher A. How about home runs per nine?

Pitcher A: 1.4

Pitcher B: 0.9

Pitcher A is now looking like the booby prize on Let’s Make A Deal. What about that WAR statistic many people hate so much?

Pitcher A: 0.1

Pitcher B: 5.7

Oh boy. How would you like to get stuck with Pitcher A now? And simple FIP — that is, their ERA as figured only by home runs allowed, strikeouts and walks?

Pitcher A: 4.93

Pitcher B: 3.14

Yeah, pitcher B is WAY, WAY, WAY, WAY better than Pitcher A. But you guessed that already. My point is that the ONLY statistics that don’t show this huge divide clearly is wins and losses. Pitcher A is Kansas City’s Jeremy Guthrie. Player B is Chicago’s Chris Sale. Pitcher A is a 34-year-old veteran pitching worse than league average who has won four games this year when giving up four-plus runs. Pitcher B is a 24-year-old superstar who has lost six games when he gave up two earned runs or fewer.

No, the win is not meaningless. It’s a piece of baseball Americana that has been a part of the game for more than a century. I don’t think the win should go away. I like it the win the same way I like that gas stations still have mini-stores. I NEVER go into one of those stores, unless I need to use the restroom or have to buy some windshield washer fluid or something. I always pay at the pump. But that’s OK, I still like those stores being there. And it’s kind of cool to see Max Scherzer is 19-1.

But as a statistic to assess a player’s value? Uh, no.

Anyway, the pitcher’s win have moved way behind on my “Man do I loathe that statistic” chart … as mentioned, the error is what drives me crazy these days. I have written at length about why I think the error is a moralistic judgmental statistic. Today’s point is different. Today’s point is that the error, as currently constructed, is so imbecilic, it makes no sense even if you LIKE errors

Sunday, the Washington Nationals and Kansas City Royals played, and it was a decent game, and it was tied 4-4 going into the bottom of the eighth inning. The Nationals then played an inning that should be placed in the Museum of Disastrous Defense. My sense is that Washington has played quite a few of those innings this year, which hasn’t helped in their wildly underachieving season.

Kansas City’s Alex Gordon led off the inning by flying out. Then Emilio Bonifacio drag bunted past the pitcher to second baseman Anthony Rendon, who for reasons unknown decided to glove and throw rather than barehand the ball. Replays showed his throw to be late — and it looked late live too — but the umpire called Bonifacio out. So Washington got bailed out on one curious defensive play. Anyway: Two outs.

Eric Hosmer walked on five pitches. That led to Billy Butler who hit a fairly routine ground ball to the right of first baseman Adam LaRoche. There is something you should know about Billy Butler: He is slow. He is not a typically slow Major League Baseball player — he is cable repairman slow. He is couch delivery slow. He is new house construction slow. He is so slow I once invented a concept I called the “Billy Butler Cycle”* which is a single, a double and two home runs … because they guy ain’t gonna hit a triple (my suggestion was he stop at third on the second home run). He’s hit zero triples this year to go along with his zero stolen bases. If he holds on, this will be his third season with 0 triples and 0 stolen bases — only nine players in baseball history have managed to do it three times. Paul Konerko is the all-time leader with five 0/0 seasons.

*In case you are wondering, five players have had four Billy Butler Cycles in their careers — Bob Johnson, Vlad Guerrero, Ken Griffey, Juan Gonzalez and Moises Alou. Frank Robinson, Cal Ripken, Albert Pujols, Joe Morgan, Mickey Mantle, Barry Bonds and Carlos Beltran are among the 14 who have had three in their careers. Other interesting players include:

Babe Ruth: 2

Mike Schmidt: 2

Hank Aaron: 2

Lou Gehrig: 2

Andre Dawson: 2

Kirby Puckett: 1

Two people who never had a Billy Butler Cycle? Willie Mays and, yes, Billy Butler. It’s kind of a quirky thing — Mays hit multiple homers SIXTY THREE times in his career, but only three times did he hit a double in the same game.

Anyway, Billy Butler: Slow. LaRoche had roughly 583 different ways of getting Butler out on the play once he picked up the ball. And best I can tell, there was only ONE possibility to NOT get Butler out. That way required:

  1. Pitcher Craig Stammen not covering first base.

  2. LaRoche being ABSOLUTELY SURE Stammen would cover first base.

  3. LaRoche turning to flip the ball to Stammen, only to see he wasn’t there.

  4. LaRoche being so baffled by Stammen not being there that he found himself too confused to run to first base and step on the bag before Butler arrived.

  5. Stammen finally running to first base to confuse matters even more.

All of these things happened, and Butler was on first base. And do you know what that was ruled? Right. A single. Two major league baseball players (at least — I’m not even sure WHAT second baseman Rendon was doing on the play) played like they were unsure of the general rules of baseball, and it was ruled a single.

Stammen then walked Mike Moustakas on five pitches to load the bases. It wasn’t exactly a banner inning for Stammen.

So bases loaded, and up comes Salvador Perez. He hits a line drive that short-hopped shortstop Ian Desmond, who struggled with it but eventually picked it up. OK, you have probably played Little League Baseball — what is it that every single Little League coach on our little planet yells when the bases are loaded and two outs. Right: “Force at any base!” Or: “Play at any base!” Or: “Out at any base.” The key part is that “at any base.” It’s the one good thing about having bases loaded. You can throw the ball to any base and get an out.

Well, Desmond picked up the ball and, somewhat alertly, prepared to throw the ball to third base. Why? Well, if you’ve been following along, you know: BILLY BUTLER WAS ON SECOND BASE. And we’ve been over this: Billy Butler is very slow. So all Desmond had to do was throw the ball a third baseman covering the bag and the inning would be over.

And it would have worked too except for one thing: Washington’s Ryan Zimmerman forgot to cover third base. The replays are actually somewhat comical — he’s just standing there, apparently mesmerized by watching Ian Desmond try to pick up the baseball. You know, four or five years ago, Ryan Zimmerman was probably the best defensive third baseman in baseball. He was awesome out there, Now? It’s hard to watch. He had shoulder surgery, so his arm looks shot, he seems to have a phobia about throwing across the diamond, it seems to infect just about every part of his defensive game. Hard to watch.

Zimmerman tried to run back to the bag, and Desmond finally threw it to him, but it was too late. Billy Butler, for the second time in an inning, had beaten the defender to the bag. That, I can promise you, is a new Billy Butler record.

And, once more, it was not called an error. A mishandled grounder, a third baseman not covering the bag, and it’s not an error. The Royals scored a run on two of the worst defensive plays you will ever see — it was an earned run with no errors on the books.

I love the line from Coach Ted Lasso (Jason Sudeikis) about Premier League Soccer: “Ties and no playoffs … why even do this?”

If you are not going to give errors when team make embarrassing, gargantuan, almost impossible to believe Little League mistakes … why even have them?

57 Responses to No Wins, No Errors

  1. Richard says:

    I would be perfectly happy to see the statistics of baseball reconstructed to remove all the errors.

  2. Mike says:

    The most amazing thing about Butler’s infield single is that he wasn’t really running until he saw that Laroche and Stammen were discombobulated. He sauntered out of the box, took his time moving down the line then suddenly “hustled” when he realized the play had been completely bungled.

  3. Jerry Esses says:

    To be fair, Billy Butler has 4 triples and 5 stolen bases in 4064 Plate Appearances. Benjie Molina had 6 triples and 3 stolen bases in 5159 PA, including 6 0/0, although he didn’t have enough AB’s to qualify for the batting title. Konerko had 8 triples and 9 stolen bases in 9163 PA’s. Interestingly enough he had 4 triples in 1999. He had 1 in 2000, then went 7 years before his next. So maybe it should be called a “Konerko cycle”

  4. R says:

    When it comes to the topic of Butler cycles, his 4th career triple (in 2012) came in a game where Butler was a single away from a cycle during two consecutive PAs. He struck out both times. A Billy Butler cycle would have broken the universe and probably caused mass chaos.

    It’s kinda criminal that the MLB video highlights didn’t include the Butler triple from that game in Baltimore. Large guys running out triples should be an automatic inclusion for the highlight reel.

  5. Kansas City says:

    This comment has been removed by the author.

  6. Kansas City says:

    I like Butler, but as Mike pointed out, he did not hustle to first base. Then, he also did not hustle to third base on the other play. He was slowing to a jog and looking over his shoulder before he realized the SS was going to make a play on him. He loafed twice on plays that could have cost a team with a seven game losing streak loss number 8. Yost seems to have a good relationship with his players (otherwise being an inept manager), but he tolerates loafing by numerous players on the bases.

    • KHAZAD says:

      This is my pet peeve. You only get 4 or 5 PA’s a game, and some of them are K’s, BB’s, HBP’s and HR’s where you don’t even have to run, so run hard. Sometimes good things happen when you force the infielders to hurry.

      You can find examples of loafing to first base on every team, every day, but I have never seen anyone do it as consistently as Butler. He seriously costs himself at least a dozen hits (or reached on error, depending on how it would be scored) a year out of sheer laziness.

  7. If we’re stuck with the error (how often does a stat ever go away?), but scorers are hesitant to assign blame on a play like LaRouche-Stamens’, then perhaps a new category — team error — could be established.
    Sorta precedent: team rebound in basketball.
    Yeah, I know, hitters won’t like it, and no one really needs, but …

    • David says:

      I can think of one stat that went away: the GWRBI (game-winning RBI) was pretty popular for a while. Now you can’t find it published anywhere, basically.

      As for “team” stats, I was a statistician for a DIII school (football, men’s and women’s basketball, and golf were my sports) when I was in college. When a QB intentionally grounded a pass to stop the clock OR took a knee at the end of a half, we would count it as a “team pass” or a “team rush.” Of course, I was also told not to count the cross-court heaves at the end of the half unless they went in. So these, at least, are counterexamples to Joe’s own theory that only baseball has “judgmental” stats. Other sports do, too – you just might not know about it.

    • Ed says:

      Those aren’t really judgmental though, at least not in the same way the error is judgmental. Those are statistical quirks that only arise in very specific, defined situations. An error can happen on any play at any point in the game, and it’s totally up to the official scorer to decide.

      The only thing I can think of that may be comparable is someone having to decide if a basketball player was trying to change a shot to a pass mid-air and whether to count that as a shot attempt, but 99% of the time it’s clear whether the player was trying to shoot or pass.

    • brhalbleib says:

      Actually the basketball assist is just as judgmental as any baseball statistic. You get an assist if you intended to pass instead of shoot and only if the ball goes to the person you “intended” to pass it to. If it is deflected to someone else who scores off the deflection, no assist. (unlike hockey and soccer, which are very nonjudgmental when it comes to giving out assists)

    • Rob Smith says:

      Yes, and the basketball assist can get padded by a home scorer too. I love Magic Johnson… my favorite player ever… but at home, he got an assist just if he passed the ball and someone scored regardless of how iffy and regardless of the move the player put on the defensive player to score. I doubt that is different anywhere in the NBA. BTW: I’ve kept score in HS basketball. It’s much harder to get an assist in HS vs. the NBA. Why? You make a good pass that should be assist and the kid misses the shot. Happens a lot. Kids can lose as many as 3-5 assists per game because of kids blowing layups and missing wide open looks. In the NBA, if they miss those shots, they aren’t in the NBA long. Make a good pass? It almost always turns into a basket.

    • invitro says:

      FWIW… last season, home teams got an assist on 60.6% of baskets, and away teams got one on 58.6% of baskets. (I hope this calculation is right… I can’t find how to do it on

      The 2008-2009 Denver team had interesting splits: 66.0% home, 52.9% road. That team also got 7.3 blocks/game at home and 4.7 on the road. But 2008-2009 Miami had more assists at home than away.

      Back to baseball… I can’t find out how to get error splits on b-r. But in 2012, home teams allowed 895 reach-on-errors, and away teams allowed 847. Home teams scored 838 unearned runs, and away teams scored… 838 of them.

    • Ian R. says:

      Hmmm. I wonder how much of those splits has to do with the fact that home teams often don’t bat in the bottom of the ninth. Since home teams play slightly more defensive innings than away teams to begin with, it follows that they’d allow slightly more ROE.

    • invitro says:

      Great point. The ROE/IP is .0402 for both home and away for 2012. I am rather shocked at the equalities for this and for unearned runs, and hope I am not doing something wrong :/.

  8. It’s similar to an outfielder losing a fly ball in the sun/lights/stands and having it drop 50 feet in front of him—it’s never called an error. If he attempts to make the play and muffs it, it’s an error, but if he has no clue at all where the ball is he gets off clean.

  9. slaskey says:

    This comment has been removed by the author.

  10. Great summary.
    Reminds me of Bill James’ description about Willie Mays Aikens stretching a triple into a triple.
    Gold. Solid gold. Solid eighty-seven carat gold.

  11. Dana King says:

    Legendary Pirates broadcaster Bob Prince used to advocate for what he called the Team Error. These would be charged when no individual can be singled out, but the play clearly should have been made. Pop-ups that fall between fielders, for example. The LaRoche-Stammen example you cite would surely qualify. The Desmond-Zimmerman mix-up might qualify, as well.

    As for pitcher wins, they don;t mean much within a season, but over a career the factors that makes them less than meaningful in a single season tend to even out. Bad pitchers don’t win 200 games.

    • Ed says:

      That’s true (and a significant reason is that a bad pitcher usually won’t get enough starts to even have a chance to win 200 games), but even in that circumstance other stats would give you a much better indicator of a pitcher’s quality.

    • Rob Smith says:

      Good one. They have the “Team Rebound” in the NBA when somebody comes up with the ball after a missed shot… usually after a bounce or two, and perhaps a tip and/or a carom or two, but nobody really did anything to deserve a rebound. Yeah, it ended up in the hands of the point guard, but all he did was stand in the right spot and the ball was slapped or deflected his way.

    • invitro says:

      “but over a career the factors that makes them less than meaningful in a single season tend to even out.”

      Do they? The main factor that makes pitcher wins not-so-meaningful is run support, which I think is around 40% of what makes up a pitcher win. A pitcher often has much lower than normal run support for an entire career (same with higher). Run support evens out for some pitchers, and does not even out for some pitchers.

    • KHAZAD says:

      Saw an example of the team error Sunday. An infield pop up just behind the pitchers mound that was probably high enough for EVERY infielder to make the play. Pitcher went under it, thought he got called off and moved away, while two infielders (one of whom seemed like he actually did call the pitcher off) peeled away to let the pitcher catch it. Because it was so high it actually allowed for a moment to have all 3 guys look at each other and make belated moves towards the point just behind the mound where the ball fell harmlessly-for a single.

  12. Unknown says:

    Ha! How about naming a guy who, over a stretch of time, hit at least 431 home runs (and possibly many more) without hitting a single triple.

  13. Frank says:

    I’m a Nats fan, and I watched this brutal inning unfold as Joe told it. I advocate for another new stat for these incidents: the BF – brain fart.

    Adam LaRoche would be the all time leader in BF’s. When he was on the Braves years ago, I watched a game (against the Nats) where LaRoche field a little infield chopper about a 1/3 of the way to the mound. He leisurely trotted back to first with his back to the plate. Nick Johnson – almost as slow as Billy Butler – turned on whatever jets he had and beat him to the bag for an “infield hit.”

    After the game, LaRoche blamed it on the fact that he had Attention Deficit Disorder. Classic.

    • Rob Smith says:

      I remember that. LaRoche can hit & has power. He is also typically a good fielder (at least according to traditional stats & from some of the good plays her makes). But the Braves didn’t really want to keep him when his contract expired because of all the mental errors he made. None of that shows up in any stat (that I’m aware of, anyway… not sure if it ends up being rolled up in some Advanced Def. stat, or something). So, from the outside, people probably wondered why the Braves didn’t have much interest in retaining him. Yes, the BF stat would have been helpful in explaining the Braves indifference towards LaRoche.

    • Frank says:

      LaRoche is an intriguing player in a lot of ways. I would say he was the Nats’ MVP last year. Power bat and Gold Glove. But he’s like the little girl with the curl – when he’s good he’s very good, and when he’s bad he’s horrid. His poor play this year, together with Zimmerman’s, is a lot of the reason the Nats are not very competitive this year.

    • FranT says:

      LaRoche blamed his miscue on the same thing Miguel Tejada blamed his failed PED test.

    • invitro says:

      “None of that shows up in any stat (that I’m aware of, anyway… not sure if it ends up being rolled up in some Advanced Def. stat, or something).”

      I would very much hope that would be included in Rfield. Anyway, he was 13 and 11 runs worse than average for fielding in his last two full seasons in ATL (2005 and 2006). His numbers are better than average for more recent seasons (+8 for his Gold Glove season last year).

    • Rob Smith says:

      Makes sense. Bobby Cox wasn’t much for advanced stats, but LaRoche drove him crazy with the mental errors. Whether the numbers showed them, or not, Bobby didn’t like the mental errors at all. I’m sure he kept some unofficial count in his head.

  14. Mark Daniel says:

    Errors as a stat give a sense of fairness, on balance. When a CF drops a routine fly ball, it only seems fair not to blame the pitcher. If some advanced stat exists that builds some type of fairness into the model (such as FIP), then we’re on our way to eliminating the capricious “error”.

    • Frank says:

      I think you have hit upon the issue regarding traditional stats vs. advanced stats when it comes to errors. (1) Trad stats often introduce some area of subjectivity. A scorer has to decide whether “ordinary effort” would have made the play. Fans can argue about H / E, but a supposedly neutral third party makes the final call. (2) Advance stats are not necessarily so subjective and, hopefully, more precise. However, they cannot be calculated or ascertained by the average fan, or, more properly, by ANY fan. Ultimately, information (and lots of it) has to be fed to a computer to generate the stat. I’m not sure I want to live in a world where FIP is all we have.

    • Rob Smith says:

      Well, I think we all realize that Official Scorers are not impartial. There are home town calls all the time. I was at a no-hitter quite a few years ago that Bill Singer pitched (in the 70s). Early in the game, a ball was hit back to Singer, who botched the play, fumbled it and threw off balance to first. Safe. Ruled a Hit by the scorer. Later, when we progressed and it was obvious Singer had a great game going, that call was changed to an Error (this was a few innings later)!! Ironcially Singer butchered a similar play later in the game which we all knew was going to be called, and was called an Error. Today, players will lobby (especially in their home town) for hit calls to avoid being charged with an error, or to be credited with a hit. And sometimes, the scorer changes the call!!! While we want to get away from this subjectivity, I’m just not sure that we can. Overall, I’m not sure that this stuff doesn’t even out over time and really doesn’t matter. Sure, there are some years where things can be pointed out. But in the scheme of things, I think it’s small potatoes.

    • Rob Smith says:

      I should mention that both balls hit to Singer were hit hard. Singer grabbed at both of them, didn’t come up with them cleanly and the batter reached base. Traditionally, scorers don’t expect pitchers to come up with those type of plays(they are about 55, or so, feet away from the batter after delivering the pitch… which is less than the length of little league bases…. & they are ruled hits. Unless, of course, there is context that requires them to call them errors.

  15. Rob Smith says:

    Joe, in thinking about the Win Stat, at first I was unsure. Then I started thinking about wins in the context that they’re used as career numbers for selecting HOFers. Wins are a huge benchmark even today. We wouldn’t even be talking about Jack Morris for the HOF if Wins weren’t weighing heavily on the discussion. I also think Bert Blyleven might have had an easier time if we discounted his Won/Lost record from years with terrible teams.

    But here’s my quiz. Pick the obvious HOFer(s) based on their career win/loss record, yes there could be more than one:

    Player 1 – 213-155
    Player 2 – 254-151
    Player 3 – 205-111
    Player 4 – 219-100
    Player 5 – 212-127
    Player 6 – 254-186
    Player 7 – 266-162

    They all look pretty good.

    Player 1 – John Smoltz (ding, ding, ding)
    Player 2 – Andy Pettite
    Player 3 – Tim Hudson
    Player 4 – Pedro Martinez (ding, ding, ding)
    Player 5 – Chief Bender (voted in by the Veterans Committee.
    Player 6 – Jack Morris
    Player 7 – Bob Feller (ding, ding, ding)

    Not saying that Hudson, Pettite and Morris aren’t good and won’t get some HOF support (Morris obviously has)… but it’s an uphill battle. I don’t think there is a question on Martinez and Smoltz, and obviously Bob Feller is in.

    • Tom G says:

      just spent a few minutes on baseballreferene Elo ratings stopping after I saw 10 pairs of pitchers. Of those 10, the one I picked as the better pitcher had a higher win total seven times. I suspect that could hold up after 1000 pairs. If I’m right, that means if given two random pitchers throughout Major League history, you could pick the better one 70% of the time just by knowing how many wins they have

      I wonder how the success rate for wins (whether 70% or something else) compares to other stats that are often disparaged: batting average, RBI, fielding percentage, et cetera

      Obviously if we used pairs that weren’t random the results could be far different.

    • Rob Smith says:

      70% is not a really good correlation factor. I took a probability/statistics class, that required a minimum of 70% on our assignments… which typically took various factors to try to link to specific results. That’s not to say that coming up with 70% is that easy. I would say try your experiment again using ERA+ or WAR. I’d bet you get up towards 90% correlation (stat to good pitcher… however we want to define that). I think that’s the point. Wins are interesting. They do tell you something. But they can definitely take a pitcher who is either on a bad team, a team that plays poor defense or a team that gives no run support on said pitcher’s rotational day (sometimes that happens on a good team)… and make them look like less. Cliff Lee’s 6-9 record last year in 211 innings with a 3.16 ERA (He’s 10-6 with the same ERA this year) has been noted previously. Poor record, good pitching. On the contrary, the best example of a bad pitcher with a lot of wins (and an inflated reputation) that I can think of is 1970 Jim Merritt on the Reds. He was 20-12 with a 4.08 ERA. His ERA+ was 103. Almost the league average. Yet, he was 4th in the Cy Young voting. Why? The Reds were awesome and he won 20 games. Period. I didn’t realize he was that bad until I saw his Strat card. It was like, wow, this guy was supposed to be good. His card is awful! Hopefully we’re getting more enlightened.

    • Tom G says:

      Definitely agree. Merritt’s ERA was actually below league average, 4.08 to 4.05 — Reds played in hitters park so ERA+ was a bit better. Six times since World War II a pitcher has won 20 and had an ERA+ below 100 (think Lew Burdette is the worst; check out Jack Coombs 1910 and 1911, 31 and 28 wins, ERA+ of 182 then 89, how the hel! does that happen?)

      But you admit the Lee – Merritt comparison wasn’t random. I asked baseballreference for a random page and the first one was a pitcher who had fewer wins than both Lee and Merritt and was also a worse pitcher than both of them (interestingly, while we all agree Lee is the better pitcher, he actually does have more wins than Merritt)

      I think WAR would be about 95%+, but not sure about ERA+ — might be under 90% just because of the times starters match up with relievers. An unadjusted ERA would also probably do better than wins, but of course is biased toward different decades. I strongly believe SO / BB would do better than wins; but I think the better win total would coincide with the better pitcher more often than any of the FIP stats alone (SO, BB, HR, SO/9, BB/9 and HR/9). Yet we all believe those stats are worth something

  16. simon says:

    Until he won three of his last four starts to get to double digits, Sale’s 2013 was looking dangerously like Cliff Lee’s 6-9 2012.

  17. Joe,

    I wholeheartedly agree with everything that has been said by you and Brian Kenny, but neither of you have really touched one of my biggest complaints about W-L records. Try Pitchers C, D, and E on for size:

    Pitcher C: 7-1
    Pitcher D: 6-1
    Pitcher E: 6-1

    They all demolish Chris Sale’s .429 winning percentage. Who are these guys? Brad Ziegler, Brian Duensing, and Justin Wilson. Three middle relievers that have combined to throw three more innings than Sale (175 IP vs 172 IP), but have ‘won’ more than twice as many games.

    W-L for starting pitchers is an extremely flawed statistic, but it’s downright ridiculous for relief pitchers.

    What does his W total tell us about Chris Sale? There were 9 times this year that he was able to hold down the opposition below the lead that his offense was able to build and wouldn’t relinquish.*

    *The fact that it’s that convoluted should be enough to convince everyone that W is a terrible stat.

    What does their W total tell us about Ziegler, Duensing, and Wilson? There were 19 times that they entered the game with their teams losing or tied, but their teams took the lead before they were taken out of the game. It’s as much about when their manager decided to put them in the game as their performance.

    Ok, so fine, we have this stat called W, and we’ve been keeping track of it forever. And even the most ardent supporter of the W knows its limitations. And most everyone knows that W-L record for relief pitchers are always nutty*, but what’s the first thing the commentators tell you when a relief pitcher enters the game? “Brad Ziegler is 7-and-1 on the season, with a 2.26 ERA,” as if that 7-1 means that he’s having a better season than Chris Sale.

    *Mariano Rivera, the undisputed best relief pitcher of all time, has had 8 career seasons with .500 or worse winning percentage.

  18. Phil says:

    This comment has been removed by the author.

  19. Phil says:

    I think the Win is so ingrained that it makes some sense to “fix” win-loss records by using advanced statistics. I simply use annual ERA+ and negative binomial distribution to come up with “pragmatic” career records for pitchers which attempt to neutralize run support. The results square with what we consider “good,” but in win-loss terms that are familiar to everyone. Some cases in point:
    Jack Morris:
    Actual 254-186, Pragmatic 232-208
    Pedro Martinez:
    A 219-100, P 229-90
    Bert Blyleven:
    A 287-250, P 318-219
    Walter Johnson:
    A 417-279, P 476-220
    Tony Mullane:
    A 284-220, P 297-207
    Sandy Koufax:
    A 165-87, P 162-90
    (Chavez Ravine and that two-foot mound were very kind to Mr. K.)

    • Ian R. says:

      Of course, my first reaction to this was “OMG Walter Johnson.”

      I’m confused by your comment on Koufax, though, given that his W-L record barely budged after your adjustment.

    • Phil says:

      Just that Koufax had a better W-L than he “earned,” since the conditions at Dodger Stadium catered to his skills and assigned him more wins than his pitching alone would have garnered. (Drysdale pitched 15 wins “better” than his record showed, so it wasn’t all a run support issue.) I suppose you could take the opposing tack and say that he was actually HURT by the location when assessing his career, as his ERAs were truly not as good as they appeared at the time. Before doing this study, I had assumed that Koufax’s otherworldly ERAs would have implied a better W-L than he actually compiled.

  20. charlied says:

    i believe in the stats of bill james, realist quantifiable truly accurate representations of value since 1985

  21. nscadu 9 says:

    Compare Wins then ERA+, WHIP, WAR, K/BB, and just about any other meaningful stat for Schilling and Morris and then for a lark throw in post season stats and tell me who the better pitcher is. Then go vote on the BR Hall poll.

  22. jim louis says:

    Joe, did you see the KC STAR article about Sunday’s Royals win? The writer seems to totally buy that Butler’s base-running saved the Royals.


    Wow. And Yost’s quote regarding Butler being safe at third is classic Yost:

    “My heart sunk there for a minute as soon as it was hit and I saw him going to third thinking, ‘Maybe I should have pinch run for him there,” Yost said. “In the ninth, it’s a no-brainer, but he ended up making it there and then scored on Lough’s single. Because it worked, the decision was right, but it was a tough decision.”

    • invitro says:

      “Because it worked, the decision was right”

      I wish more people would argue against this incorrect logic. The rightness of a decision depends only on the information available at the time the decision was made. It does not depend on the outcome of the decision. This broken argument happens all the time with trades. If you make a bad decision with a trade, but the player you got does better than expected, you can say you got lucky. You can not say you made a good decision. Logic 101.

    • JRoth says:

      You’re right, of course, but I see people using your point as a defense for having been wrong in the first place:”Player X should have been better than he was, so I was right to want to acquire him. My idea has turned to crap, but secretly I was correct.” This is true sometimes, but sometimes the original premises were bad, and the person should be revisiting his priors, not insisting that the process was right and reality is wrong.

      If Brian Cashman is out there arguing that the results of the 2013 season don’t change the correctness of his decision to let Russell Martin go, then he’s using a logical smokescreen to cover up a poor process.

  23. JRoth says:

    I still haven’t seen you answer the point I raised last time you went off on errors: how many homeruns does NL-leading Pedro Alvarez have? MLB’s official records say 31, but 32 times this year he has hit a ball and crossed the plate during the same play. By your rules, that’s a homerun, because we don’t judge how it was that he scored that run.

    So which is it, Joe? Will you be advocating for Pedro Alvarez’s True Homerun Total?

  24. Richard S. says:

    As long as we determine which teams make the playoffs by their win total and not something like “run differential”, pitcher’s wins will be an important and relevant stat. The One True Stat that tells you precisely how good they are? Of course not, but they are still important nonetheless.

Leave a Reply

Your email address will not be published. Required fields are marked *