FanGraphs Baseball


RSS feed for comments on this post.

  1. Over the last three years, Fangraphs has Nolasco worth basically 3.5 wins per season. B-Pro, however, had him at 2.5 per annum…and B-Ref says a mere 1.5

    Talk to me like a (relatively intelligent) visitor from another planet: Why is B-Pro’s valuation incorrect, and B-Ref’s waaaay off?

    Just a bit of food for thought about Nolasco, and why, perhaps, he *isn’t* better than his ERA suggests. How effective is Nolasco over his career, with the bases empty? An OPS allowed of .718, with a K/BB of 5.34

    With men on base? A .780 OPS. (And just over 3-1 K-BB, adjusting for IBBs.)

    With multiple men on base? Try around .900 OPS permitted.

    I’m honestly not trying to be snide–but maybe, just maybe, for some pitchers, we need to look deeper than FIP, xFIP, or, say, SIERRA, to fairly appraise their performance on the hill.

    Comment by Bob — December 25, 2010 @ 9:54 pm

  2. I agree with you. As soon as I read the analysis, my first thought was “His fWAR is high, I bet his rWAR is low.”

    Nolasco is generally dubbed as “unlucky” because his ERA is always higher (or close to) his FIP.

    However, when you are consistently “unlucky”, I think it renders the term “unlucky” useless. Luck implies something that does not happen year after year. Something that regression to the mean.

    A pitcher cannot be “unlucky” as their mean. Nor do I think you can blame the defense year after year, when the same scenario is not repeated with the rest of the staff. It’s not like the marlins just happen to play bad defense when RN pitches, just as OAK just doesn’t happen to play like 8 glod glovers when Trevor Cahill pitches.

    We have to attribute SOME of the *cough* luck to something the pitcher is doing. We may not know what exactly happens with men on base, or why it happens, but we know that RN consistently underperforms in regards to his FIP. We can’t call it “bad luck” every year. Well, we can but we would not be accurate.

    It could be something mental (crapping his pants, losing focus, distracted by runners), it could be mechanical (such as rushing), etc. What IT is does not matters, that IT exists and materializes in his stats does.

    Again, “luck” is something that regresses to the mean, otherwise it IS the mean. Nolasco’s FIP is misleading. That has to be (or should be)apparent by now.

    Same thing with StL pitchers, when they outperform their FIP 5+ consecutive years, you stop putting so much weight on their FIP, and realize that there is likely more to it than just chuckling and calling it luck.

    My perference and what makes the most sense to me is to average fWAR and rWAR and give the pitcher some credit. Not to name drop, but I read similar commnets from Tnago where he said something like “I’d rather be half right than all wrong” … and this is THE guy that developed FIP. Shouldn’t that carry some weight?

    Comment by CircleChange11 — December 25, 2010 @ 10:03 pm

  3. To answer your question …

    Fangraph’s WAR (fWAR) for pitchers is based on FIP (Fielding Independent Pitching), a formula calculated by multiplying BB, K, and HR by various factors and scaling it to an ERA-like number. FIP and fWAR ignores all batted ball results.

    Baseball reference uses Rally’s WAR (rWAR), which is based on runs allowed and modified by various league/park factors.

    Comment by CircleChange11 — December 25, 2010 @ 10:12 pm

  4. they should do something that combines them. i am of the opinion that a pitcher can control Ks, BBs, and HRs the most, but getting grounders or getting outs in the air is also a pitcher’s skill/style and cannot be ignored. defense plays a part in those outs, so instead go by league average BABIP on the batted ball types. then you can have defense neutral, luck neutral value based on batted ball profile/repertoire and Ks, BBs, HBPs, and HRs. doesn’t tERA do something like this?

    Comment by phoenix — December 25, 2010 @ 10:50 pm

  5. I agree … in fact, when I’m not dead ass tired, Circle, I should send you some results I looked at when comparing pitch values and “luck” metrics. But, in short, the only single “plus” pitches (combos can be more potent) that seem to have some control is fastballs on HR/FB and curves on LOB… and what do you know? Nolasco had a strong curve and LOB in ’08 and a poor of each the last two years (as well as a drop in CB usage), while he has never had a good fastball or a good HR/FB (though his best of each were in ’08).

    I’ve seen you champion your POV here and especially on Tango’s site, and here I think you are spot on. I mean, I was wait and see last year since there wasn’t a clear pattern and ’09 was almost historically “unlucky”, but now you really need to realize that WAR’s choice of FIP because they felt it was important to mirror the contextual-lessness (new word?) of their Batting formula and not because context has been proven never to matter makes me wary when it comes to pitchers.

    While I’ve mainly looked at “elite” pitchers as sources for evidence that control over certain factors is possible, there is no reason to think that certain skills not captureed in K/BB (Nolasco’s really doesn’t change all three years!!!) don’t have an effect on poor performance. Nolasco has thrown 555 IP over the last three years, and is of the ten “least lucky” with HR/FB, BABIP, and LOB. Just taking LOB further, he is sixth worst of 68 qualifying pitchers … and just who are the “lucky” ones?: Johan Santana is first, then Wainwright, Halladay, Cain, Hamels, Lilly, Marcum, Felix, Lincecum, and Lester. LUCK MY ASS.


    Comment by William — December 25, 2010 @ 10:55 pm

  6. Just talking out of interest at this point, and hoping to have something explained …

    [1] In batter WAR, we include BABIP which can include good/bad luck*. We do this because it represents what actually happened on the field, instead of going with FIB (Fielding independent batting) that might include only the TTO’s (BB, K, HR). Batter WAR tells us about the value that the batter had based on the actual on-field results. We can look at BABIP to see whether it is it something that the batter is likely to repeat or not.

    * Luck = random events compared to the mean.

    [2] With fWAR, we ignore BIP data, even though that is actually what occurred on the field.

    [3] FIP can also reward/punish the pitcher based on luck on HRA. For example, this year Lee had a 5-game stretch where he was knocked around the ballpark quite a bit. He zoomed up the fWAR charts during this time. The thinking in some circles was that “he was unlucky on BIP.”. Why wouldn;t the thinking be “He is getting lucky that some of those hits are not leaving the park?”. Same thing with Liriano to a degree.

    [4] In this context, at least rWAR is consistent for both batters and pitchers in that it gives value based on what hapened on the field … rather than what we think “should” have happened based upon our philosophy or preconceived notions.

    So, what I am wondering is why do simultaneously give the batter credit/discredit for good/bad luck on BABIP, but not picthers?

    I understand that FIP has some predictive applications by removing the fielding aspect, but why would we include this into WAR vaues for that specific season. Couldn;t we just look at BABIP and HR rate for the season and decide whether the WAR was repeatable (sustainable) or not, just like we do with batters.

    I ask because it doesn’t make sense to use two different methods for the same value stats for the two groups of hitters. In short, it seems arbitrary to do so. What is the basis for this.

    To not be completely sidetracked …. if we look at Nolascoand average his fWAR and rWAR, we get a 2.5 WAR pitcher …. he’s being paid fairly for the value he produces on the field.

    Comment by CircleChange11 — December 25, 2010 @ 10:57 pm

  7. I am of the opinion that your opponent has a strong influence on your K, BB, and HRA. I feel this way not based on statistical analysis, but because of my experience as a pitcher. When I pitched against top level teams that absolutely refused to swing at junk and sit dead red, my BB rate went up, K rate went down, and HRA rate went up. When I pitched against pretty much everyone else, I could get ahead and get them to chase junk and be successful.

    I will say that when you are an elite pitcher (I was not), your overwhelming stuff allows you to pretty much be more consistent against even the better offenses. I would imagine that elite pitchers are pretty consistent with being in the leaderboards for both FIP and ERA.

    Comment by CircleChange11 — December 25, 2010 @ 11:03 pm

  8. I’m very interested in “luck” data. Thanks for the additional explanations.

    It’s not so much that I’ve championed my POV because I feel I’m a genius … but more like once my brain locks on to something that “doesn’t make sense”, it can’t let it go. It can be annoying, so I try to limit it, but in this case …. it just REALLY does not make sense to me.

    Comment by CircleChange11 — December 25, 2010 @ 11:08 pm

  9. A few thoughts after reading your post, some in agreement some in disagreement.

    LOB% is not supposed to have a mean that all pitchers regress to. A better pitcher should in fact have a higher LOB% than average that he regresses to. If a pitcher is better, than he is more likely to get batters out, whether or not there is a runner on base (assuming the difference between his windup and stretch performance is average). Therefore, with runners on base, a good pitcher will get more batters out than a bad pitcher in a similar situation. The good pitcher will have a higher LOB%.

    Two things are happening when you look at the data for qualified pitchers over the last three years. One, a good pitcher is much more likely to qualify because he has to be effective enough to stick in a rotation for three seasons. So our pool of pitchers is going to be very biased towards pitchers at the higher end of the talent pool. Second, the three-year, qualified pitcher cutoff means that all the pitchers will have a relatively large sample, so their LOB% will be much closer to their true talent than after one or two years of data, for pitchers without many innings. And since a good pitchers true-talent LOB% is going to be better than average (presumably by the same rate at which he is better than average at getting hitters out), and we have a large sample of mostly good pitchers, it should come as no surprise that the top of the leader board is good pitchers.

    Comment by nitro2831 — December 25, 2010 @ 11:17 pm

  10. Pitcher WAR needs to be calculated using a basket of metrics. Say ERA, tRA and xFIP. That would give you a picture that crosses all spectrums of blame for the actual results that occur while a pitcher is on the mound while still giving some credence to the theory that pitchers are not responsible once the ball leaves the bat.

    Comment by mowill — December 26, 2010 @ 12:51 am

  11. You can use a hammer to sink a screw, and have reasonable success doing so … but that’s not using the tool designed for the job.

    From what I understand, WAR is used to evaluate how many wins above replacement a player, pitcher or batter, provided his team.

    By using fWAR, we’re basically figuring a value more aligned with “true talent”. So, we’re assigning wins based on what we think “should have happened” versus what actually happened.

    If we did this with batters, Pujols would win the fWAR MVP every single year based on his low K, high BB, high HR totals. We could rationalize it by explaining that his true talent is above everyone else and he should win the MVP every year.

    It wouldn’t matter so much, but we (Fangraph community) use fWAR to evaluate EVERYTHING …. from writer’s intelligence, to the “real” CYA winner, contract vlue, etc. But fWAR doesn’t measure exactly what happened on the field, but what we figure “should” have happened. That’s the part that I can’t get my mind around.

    If a batter BABIP’d his way to a .412 BA season, we wouldn;t go back and remove the hits that he “lucked” his way into, just as we wouldn’t go back and award Ryan Howard any hits he should have had if the D hadn’t put on the “Williams Shift”.

    It just seems that WAR was intended to calculate the value based on what actually happened, and that’s not what is happening with pitchers and fWAR …. so I don’t see how we have an accurate portrayal of the pitcher’s WAR, but do hav a better idea of “true talent”.

    Does that make sense?

    Comment by CircleChange11 — December 26, 2010 @ 1:02 am

  12. Baseball Reference WAR is based on Run Average. It’s park adjusted, adjusted for quality of opponent batters, adjusted for quality of team defense (based on team total zone). FanGraphs WAR is based on FIP, which is park adjusted. Both are attempts to replicate past value, but they both do it in different ways. BR thinks that pitchers should be given credit for their performance on balls in play (after adjusting for defense) and their timing, while FG, justifiably recognizing that those two things are generally not repeatable skills, thinks that pitchers should not be given credit for those things. I’m not here to tell you which one is right or wrong: that doesn’t matter. What matters is which is more predictive going forward. Based on correlation studies, it’s been shown that strikeouts, walks and home runs – what FIP is made of – are far more predictive than performance on balls in play and timing – the primary additional factors included in RA. Even over 3-4 seasons, you would still use FIP more than RA to predictive future performance

    So in 716.1 innings, Nolasco has been worth 4.4 WAR per BR and 11.8 WAR per FG. That works out to 1.1 WAR and 3.0 WAR per 180 innings pitched by the two metrics respectively. A reasonable assumption is that Nolasco is worse than his FIP, but is closer to his FIP than his RA, therefore the 2.5 WAR projection seems about right.

    Comment by vivaelpujols — December 26, 2010 @ 3:32 am

  13. Didn’t this exact same situation happen with Dave Bush a few years ago where people kept predicting over and over that his ERA would drop, only to watch him post another poor season? I remember reading a column on ESPN (I think it was by Matthew Berry) apologizing for getting the predictions wrong year after year and talking about how he was just terrible pitching from the stretch, so his numbers were inflated whenever runners were on base. I haven’t done the research everyone else here has but I’ve always made a comparison between the two, though Nolasco is clearly more talented. I always saw Dave Bush as the reason to never discount scouting reports fully during analysis, and his story could be an important lesson when dealing with pitchers like Nolasco.

    Comment by EastCoastBall — December 26, 2010 @ 7:23 am

  14. Great conversation. I tend to agree with Circle that if we are we are seeing this error (difference between ERA and FIP) repeatedly for certain players there is something that should be improved to the WAR formula.

    Discussions like this are why I keep coming back to fangraphs.

    As it refers to Nolasco, I would agree that there is a pretty solid chance he earns this contract if he can stay healthy. Health for Nolasco is not something I would bet on though.

    Comment by Socrates — December 26, 2010 @ 10:36 am

  15. Is there any consideration for the axiom “Luck is the residue of design” ?

    Comment by Scout Finch — December 26, 2010 @ 11:48 am

  16. You guys realize Nolasco “consistently” under-performing his FIP is taking place over a total of his last like 330 innings, right? Maybe Nolasco is one of those rare exceptions to the rules, but for now it’s well within the realm of possibility that his ERA is going to fall in line with his FIP.

    Comment by cpebbles — December 26, 2010 @ 12:19 pm

  17. I don’t think the intent is to base it on a pitcher’s true talent level (Else they would have moved to xFIP long ago), just to remove defense and bullpen support.

    Comment by cpebbles — December 26, 2010 @ 12:24 pm

  18. There are a few difference between batters and fielders that help explain the difference.

    Hitters exhibit a much greater control over their BABIP than pitchers do in general. When assigning WAR values, we want to give people credit for things they have done, and its much more likely that a batter’s high BABIP is due to things he actually did on the field to earn a higher BABIP.

    Second, when a batter gets a hit, there is absolutely no way it could have been due to any other offensive performer (barring someone stealing a base and making the second baseman cover the bag, thus opening up a hole on the right side. these situations are rare enough though I will ignore them). A hit occurred, so the offense deserves credit, and there is only one person to give credit to, the person who hit the ball.

    When a team gives up a hit though, it is either because the hitter beat the pitcher or the fielder screwed up. When assigning blame in the WAR calculation, we have to find a way to balance these two possible causes. Since we know that pitchers really do not have much control at all over their BABIP (except in extreme cases), we should assume that their BABIPs would be average. So if they give up more hits than average, its likely due to nothing they did, but rather that the defense behind them screwed up or perhaps some seeing-eye-singles snuck their way into the outfield. Either way, WAR tries to credit people with what they provided their team, and to be honest, a pitcher provides four things in general (K, BB/HBP, HR, BIP). The first three happen and we know the fielders had nothing to do with it. The BIP results are more likely due to fielders than pitchers. That’s why we count fielding in a batter’s WAR. If we also blamed the pitchers for hits on balls in play, we’d be double-counting the blame. We’d be doling out twice as much blame as neccesary.

    You either credit the BIP results to the fielders or pitchers, but not both. Facts and statistics tell us its much more likely not due to anything the pitcher did, so when assigning blame, don’t assign it to him! Give him credit and blame for the things he did, not his teammates. That’s why FIP makes sense. It’s what he did. BIP results are more likely what the fielders did, so we give them credit or blame, not the pitcher.

    Comment by nitro2831 — December 26, 2010 @ 12:32 pm

  19. “LOB% is not supposed to have a mean that all pitchers regress to. A better pitcher should in fact have a higher LOB% than average” (nitro)

    Obviously the two different windups is the wildcard here as you pointed out with the Bush example

    Comment by rick p — December 26, 2010 @ 12:44 pm

  20. “LOB% is not supposed to have a mean that all pitchers regress to. A better pitcher should in fact have a higher LOB% than average that he regresses to. If a pitcher is better, than he is more likely to get batters out, whether or not there is a runner on base (assuming the difference between his windup and stretch performance is average).”

    Just thought I’d post a little more of the quote. I acknowledge that some pitchers might just suck from the stretch, but in general it should not be a huge difference. Not sure if you missed that last part, so I thought I’d post it again. Definitely agree with you though,

    Comment by nitro2831 — December 26, 2010 @ 12:58 pm

  21. “You guys realize Nolasco “consistently” under-performing his FIP is taking place over a total of his last like 330 innings, right? ”

    Thank you!

    Really guys, it’s not as though we’re talking about someone who has made this a career habit. His recent ERA-FIP discrepancy has happened over basically one and a half seasons. That is not nearly enough evidence for all you people claiming Nolasco just isn’t as skilled as his FIP would indicate. In fact, before this stretch, he had never had his ERA and FIP be farther than .25 runs away from each other over a his previous three seasons. That does not seem like someone who has consistently performed at a different level than his FIP would indicate.

    If you are going to claim that someone consistently under performs their FIP, you have to actually, you know, look at their stats. Nolasco has had one and a half years of such results. If you are looking at the numbers, you have to assume he will pitch closer to his FIP than he has the past one and a half seasons worth of pitching. Saying otherwise is just lazy research.

    Comment by nitro2831 — December 26, 2010 @ 1:07 pm

  22. “In fact, before this stretch, he had never had his ERA and FIP be farther than .25 runs away from each other over a his previous three seasons.”

    I’d like to formally apologize for this sentence.

    Comment by nitro2831 — December 26, 2010 @ 1:08 pm

  23. Hitters exhibit a much greater control over their BABIP than pitchers do in general.

    Can someone link me to the research that led to this conslusion? Or an article summarizing the research?

    Pitches in certain zones go for fewer hits than pitches in other zones. Pitches with less movement and/or less velocity are much more likely to be hit hard. Harder hit balls or more likely to go for hits.

    IMO, pitchers influence more than we give them credit for. What might be the problem is even ML pitchers do not hit their spots as often as we think they should and/or even the lower ML hitters are still pretty darn good.

    It’s what he did.

    That’s the part that bothers me about FIP. The batter influences A LOT in the resulting outcome of pitch. If I throw a cockshot, all I did is throw a cockshot. Whether it leaves the park or gets popped up to the catcher is ALL on the batter. If I throw a cutter in on the hands, that’s all I’ve done. Whether the batter grounds or or gets his hands inside the ball and rips it down the line … it’s all him.

    We kep giving pitcher’s credit for things as if they were throwing to a batter painted on to a tarp, where you get “points” if you hit a certain area.

    If a pitcher throws the exact same 10 pitches to Albert Pujols as he does to Skip Schumaker, what is the greatest influence on the outcome? Hint: It’s not the pitcher.

    We reward pitchers for what we think they “control”, but IMO it’s no more arbitrary than awrding them credit for anything else.

    Look at a diagram of a baseball field. Draw a circle around each position player to represent “average range”. There’s a reason why the average BABIP is what is … there just isn;t much green to get a hit in. Everything is favored to the defense. The most consistent way to gt hits (or give up hits) is to hit the ball hard. So, who deserves credit/blame for hitting the ball hard or not? The batter for making great contact? Or the pitcher for throwing balls that can be struck well?

    I like FIP for what it does. But, I disagree with some of the applications.

    Comment by CircleChange11 — December 26, 2010 @ 2:44 pm

  24. What I was getting at is whether a batter BBs and Ks has as much to do with the batter as it does the pitcher (which is why I’ve also requested that opponent strength be a component of pitcher WAR). Lots of things go into whether a ball stays in the park or not, perhaps the least of which is the pitcher. Yet we give him all/full credit for that?

    Comment by CircleChange11 — December 26, 2010 @ 2:46 pm

  25. I was looking at his 2009 and 2010 seasons.

    I give a lot more weight to his last 330 IP, than I do his first 400.

    I would assume that Nolasco would be more probable to pitch like he has the last 2 years rather than all of the sudden in 2011 pitch like he did in 2007 and 2008.

    I don’t know what the term would be, but IMO his mean has moved to be much more like 09 and 10 than 06, 07, 08. I think 09 was pretty drastic. But I don;t see much reason to assume that he won’t have another season where his ERA is significantly higher than his FIP.

    Comment by CircleChange11 — December 26, 2010 @ 2:50 pm

  26. Obviously a strikeout depends on both the batter and the pitcher. All we can do is best assign credit and blame for what happens given the imperfect information we have. We know that pitchers have a certain level of consistency with their K’s BB’s and HRs, but not the case with BIP ( This leads me to believe that pitchers are a good deal responsible for K’s BB’s and HR’s, but not for hits on BIP. We have to pick a certain point to stop giving pitchers credit, and the point at which they exhibit marginal at best control over the outcome seems like the place to stop.

    Comment by nitro2831 — December 26, 2010 @ 2:56 pm

  27. A few nuggets from Nolasco’s career splits:

    On 1-0 counts 1.181 OPS allowed
    On 2-0 counts 1.339
    On 2-1 counts .977
    On 3-1 counts 1.175

    The collective BABIP in those situations is around .330

    And that makes some sense. MLB-wide, every year, BABIP—and more importantly, non-HR extra base hits—is affected by the count.

    For instance, check the BABIP data on 2-0 counts vs. 0-2 counts. Every year, without exception, the BABIP is higher (usually 20-40 points) on 2-0. And every year roughly one in three non-homers are doubles & triples in 2-0 counts. And every year on 0-2 counts, only one in *five* non-homers go for extra bases.

    Why is Nolasco over his career so ineffective with multiple men on base (.900 OPS allowed)? Why is he clobbered soooo badly when he falls behind?

    Luck? Really? I’m guessing he simplies over-relies on get-me-over fastballs in such situations. Just conjecture, of course. But I think it’s inaccurate to attribute Nolasco’s struggles to lucklessness.

    Comment by Bob — December 26, 2010 @ 2:57 pm

  28. Thanks for the link. I have some of Voros’s stuff before. I’ll jump into this after I beat the snot out of my 3 in basketball.

    Again, I am not trying to get rid of FIP-based WAR, but just average it with runs-based WAR.

    IMO, as an amatuer, I don;t think we really know enough about BABIP yet to really declare the “picther’s control” or non-control over it.

    What we also know are BIP data by both [1] count and [2] location.

    So, it would seem by simply throwing to a specific spot (intentionally or unintentionally) in a specific count, the pitcher can influence what the most likely outcome of a situation is.

    The reason to not do this would be if we think that a pitcher hitting his spot is more luck than skill. IMO, it has more to do with lack of skill (command) than it does luck.

    The more I get into this stuff, the more I think “we really don’t know”. It would be nice if were as simple as some of the models postulate, and it sure would make it easier to just have a value number that tells all, but I don;t see it anywhere near that simple … and it’s frustrating. Again, I don;t have the answers.

    Comment by CircleChange11 — December 26, 2010 @ 3:33 pm

  29. Good command of your stuff Bob! That’s a great pitch.

    Those are nice splits to look at. That is a significant notion to ponder: how often do pitchers get behind in the count and how poorly do they perform in those situations?

    Comment by Scout Finch — December 26, 2010 @ 4:22 pm

  30. Sometimes I wonder if we have too much paralysis by analysis going on in Fangraphs.

    As you mentioned, at some point it has to be asked WHY Nolasco’s peripherals ALWAYS seems to be better than his actual performance.

    It’s not luck when it’s a consistent theme.

    He gets hit harder with men on base.


    I think some people here are overreaching with the “luck”.

    It’s like someone wishing you “good luck” on a first date.

    Luck has nothing to do with it…….

    Comment by DIVISION — December 26, 2010 @ 5:52 pm

  31. VEP,

    What matters is which is more predictive going forward. Based on correlation studies, it’s been shown that strikeouts, walks and home runs – what FIP is made of – are far more predictive than performance on balls in play and timing – the primary additional factors included in RA.

    Can you link to those studies?

    I am simply asking because someone conducted and presented research at Tango’s blog that showed that Runs Allowed was a greater predicted of ERA than was FIP from season to season.

    You mention that FIP is predictive? Does that mean FIP is good at predicting FIP the next season (i.e., less variance in FIP than ERA?)

    I’m asking because I want to make sure I am presenting accurate information.

    Comment by CircleChange11 — December 26, 2010 @ 6:10 pm

  32. EastCoastBall – you mean the Dave Bush with the career 4.64 FIP and the correspondingly inflated ERA of 4.66?

    Comment by Toffer Peak — December 26, 2010 @ 7:38 pm

  33. Here’s an excellent example of one such study: .

    Comment by Toffer Peak — December 26, 2010 @ 7:47 pm

  34. the best part about fangraphs is when people legitimately question their methods, and no one ever answers with a competent defense of FIP, fWAR, etc

    Comment by fredsbank — December 26, 2010 @ 8:35 pm

  35. Thanks again nitro. Re-reading information about FIP, and in this case, FIP-based WAR usually does expalin the situation.

    Voros does state that refining BABIP runs into problems with the data available. So, at least for now, FIP, DIPS, etc does a pretty good job of representing what it is intended to do.

    If/When we know more … well, we’ll know more.

    This isn;t life or death (relative) I guess, just something that gets stuck in my head and won;t get out.

    For my own uses, I’ll continue to average fWAR and rWAR, but mostly due to preference and not because the evidence shows that either is a better representation of what the pitcher is actually responsible for.

    I do have a better understanding for the “why” aspect, and I appreciate your guys involvement in the discussion.

    Comment by CircleChange11 — December 26, 2010 @ 8:41 pm

  36. if pitchers have no control over their BIP… then explain why batted ball data exists at all. brandon webb, greg maddux, aaron cook, derek lowe, tim hudson… yeah, these guys probably have no control over their gb rates.

    or even better, fangraphs’ favored son cliff lee:

    2004 33.4
    2005 35.6
    2006 32.7
    2007 35.3
    total IP: ~580

    2008 45.9
    2009 41.3
    2010 41.9
    total IP: ~670

    yeah, he probably had nothing to do with, i’d chalk it all up to luck

    Comment by fredsbank — December 26, 2010 @ 9:00 pm

  37. Groundball and Fly ball rates … yes.

    Hits on those BIP, not so much.

    Voros explains with his Maddux example that in consecutive years he was among the best and then the worst on BABIP.

    Groundballers gain ground in their lower HR rates.

    I was reading earlier about whether GB or FB pitchers give up more extra base hits, and there is not all that much difference. The big difference was that when guys do get airborne against a GB pitcher, there is a greater chance of a HR. But, overall the FB pitcher will give up more HR just do the increased number of BIP that go into the air.

    What will be interesting is when we start having BABIP data for each pitcher’s pitches and/or zones. That may already be available and I just don;t know it.

    This is basically what McCracken is saying …

    There is little if any difference among major-league pitchers in their ability to prevent hits on balls hit in the field of play.

    He goes on to explain that detecting any ability to do so, at this point, gets lost in randomness or noise (lots of factors go into it).

    The 3 things that pitchers (as individuals) are consistent on are K, BB, and HR, which is why they say they have influence on those.

    They also have influence, as you point out, on GB/FB rates. What BABIP data is saying that when those balls are put on the ground, pitchers have not, to a detectable degree, pocess some ability to get batters to hit the ball in the vicinity of their fielders (and not between them).

    I, too, have trouble with the term luck because it can imply that the pitcher is just out doing nothing but closing his eyes, throwing it to the plate, and hoping for the best, when we know they are doing much mor than that.

    Luck refers to a greater/lesser percentage of ground balls going for hits than normally does.

    Everything is luck in some sort. When you throw a ball down the middle,whatever the batter does with it is “luck” in the regard that anything could happen. He could mash it into the upper deck, take it for a strike, foul it off, hit in the ground/air, etc. Whenever it is something that happens above/below the average (mean), it’s called “luck”.

    Comment by CircleChange11 — December 26, 2010 @ 11:02 pm

  38. Nobody says pitchers don’t control the value of their balls in play. Nobody. Has. Ever. Said. That. Ever.

    What people have said is that a pitchers value on balls in play (essentially his BABIP) is highly influenced by luck. This is backed up by common sense: the batter, fielder, and ballpark all influence BABIP to some extent. It is also backed up by year to year correlations, which show that past BABIP needs to be regressed 50% to the mean once you reach 1500 BIP for that pitcher (and 49% for 1450 BIP, 51% for 1550 BIP, etc.).

    (CC I have linked you to this article on numerous occasions and it definitely shows the predictive power of BABIP. This is also a good article:

    So for a guy like Nolasco, who has about 2000 BIP over his career. You would still regress BABIP heavily to the mean.

    Comment by vivaelpujols — December 27, 2010 @ 1:01 am

  39. Thanks.

    What I have been doing is starting with this article ( ), reading all of the information linked at the end of the article.

    So far it’s about 130 pages (C&P’d into Word for saving).

    What I cannot locate is Tom Tippet’s study on …

    Tom Tippett of Diamond Mind Baseball published a lengthy study on pitchers with long careers and found evidence that those pitchers exhibited some ability to affect the rate of hits on balls in play


    Tom Tippet showed that good pitchers allow less hits on balls in play than their teammates

    For fun, I exported the career stats from this site into excel, specifically looking at K/9 and BB/9 and their correlation with BABIP … thinking that either high K guys’ stuff would be so good that they’d be hard to hit and/or guys with outstanding control might be hard to hit.

    The graphs of each relationship is pretty much symmetrical when divided by a line at .290 BABIP. Then while reading I found that this had already been done by guys quite a bit smarter than me.

    Comment by CircleChange11 — December 27, 2010 @ 2:23 am

  40. Found it …

    Comment by CircleChange11 — December 27, 2010 @ 3:05 am

  41. The “luck” argument is bullshit, circle.

    Luck is simply an intangible which can’t be quantified by stats.

    There are too many variables involved to even try.

    Think about it.

    If “luck” played any small part in success, Clay Zavada and his baby mustache would still be pitching in Phoenix.

    It wasn’t a matter of luck.

    His body couldn’t handle the rigors of MLB pitching, thus he got injured.

    Comment by DIVISION — December 27, 2010 @ 1:45 pm

  42. I not disagreeing with that.

    When a pitcher throws a sphere 93mph and the batter strikes it with a cylinder with enough force to send it at 100mph … where it drops or goes is a mere matter of millimeters on the bat.

    My issue with the word “luck” is that it implies it is not the result of skill.

    In the baseball version of luck it is more in line with “random” in the regard that anything in the realm of probability can happen. But, I also point out that everything in life “luck” in that regard.

    It’s all probabilities.


    As for Zavada, he complained his arm hurt in September. He was told it’s all in his head. He kept pitching. He now has a 6-inch scar on his elbow as a result of him imagining his arm hurt. He’ll be back. He’s not the most talented guy on the block, but he can’t be rattled. Baseball can’t throw anything at him that life already hasn’t.

    There aren’t a whole lot of other teams that use their bullpen to the degree Arizona has (nor have traded away so many relievers).

    I’m not the biggest Arizona fan right now.

    Comment by CircleChange11 — December 27, 2010 @ 5:35 pm

  43. Thanks for the link, Circle. Love to see the same study incorporating the (I presume) ability of better pitchers to suppress doubles & triples on balls-in-play—not merely *hits.*

    Sidebar: does FIP or xFIP incorporate a pitcher’s ability to negate the running game (e.g. Carpenter great, Lincecum crappy) or field their position?

    Comment by Bob — December 27, 2010 @ 10:47 pm

  44. fredsbank – What are you talking about? They just did that only a few months ago.

    Comment by Toffer Peak — December 28, 2010 @ 1:15 am

  45. I just read through those posts, and I’ll be danged if I didn;t say the same things in those threads as I did this one. How unlike me? *grin*

    It is weird that “averaging rWAR and fWAR” was never an option.

    DC states that both WAR version have their flaws, but averaging them would seem to “reduce” the magnitude of each’s flaws … rather than choosing just one and hoping that it’s the one that most accurately represents the value of the pitcher.

    I’m repeating myself again. Crap.

    Comment by CircleChange11 — December 28, 2010 @ 3:44 am

Leave a comment

Line and paragraph breaks automatic, e-mail address never displayed, HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Close this window.

0.258 Powered by WordPress