The Most Consistent Pitcher of All Time
Lest this thread turns into another “Dave hates the Mariners!” whinefest, let me get this out of the way – I’m glad we won, I’m glad we’re over .500, and I’m looking forward to today’s game. Go M’s.
Now, Jarrod Washburn’s three true outcome stats for each of the last three years:
2005: 2.6 BB/G, 4.9 K/G, 0.98 HR/G
2006: 2.6 BB/G, 4.9 K/G, 1.19 HR/G
2007: 2.6 BB/G, 4.9 K/G, 1.15 HR/G
His walkrate and strikeout rate this season are exactly the same as they were each of the last two years. That’s consistency of a ridiculous nature. Almost every pitcher has some year to year variation in their peripherals, but not Washburn – he’s cemented himself as the 2.6 BB, 4.9 strikeout, 1.1 HR guy. No regression analysis needed here. Of course, his ERA’s have bounced all over the place – 3.20, 4.67, and now 2.88. Why is that? Let me throw some more numbers at you.
2005: 81.8% LOB%, .289 BABIP
2006: 69.6% LOB%, .290 BABIP
2007: 78.9% LOB%, .208 BABIP
As we’ve discussed ad nasuem, Washburn threw up an unsustainable rate of stranding runners in 2005, leading to a superficially low ERA. He wasn’t getting hitters out any more than usual – he was just getting them out with guys on base. Some people called it clutch pitching – we called it an outlier, based on all kinds of historical evidence that there’s not some kind of clutch pitching gene that Washburn has, allowing him to pitch out of jams like he’s Johan Santana. Not surprisingly, his percentage of runners left on base tumbled last year, despite being the exact same guy he was in 2005, and his ERA went up. It turns out that we didn’t irrationally hate Jarrod Washburn – we just recognized that his success was built on a house of cards, and it was going to tumble.
So, now, I guess we get to have this same conversation all over again. Jarrod Washburn is still the exact same guy he was the last three years – seriously, look at those rate stats – but so far, in 2007, hitters are hitting the ball at his defenders and making outs. I shouldn’t have to tell you guys that a .208 batting average on balls in play is unsustainable. When he starts facing hitters who aren’t getting themselves out at every opportunity, his ERA will rise fairly significantly.
I wish Jarrod Washburn was having some kind of breakthrough season where he established himself as a quality #2 starter. He’s just not – he’s establishing himself as the most consistent pitcher of all time, and a walking example of why using ERA to evaluate pitchers is a bad idea.
So, when you read stuff like this quote:
“The big thing now is that he’s got command of all his pitches and he’s hitting his spots,” Hargrove said. “He’s been doing that in all his starts. He’s 2-2 now, and he could very well be 4-0.”
Remember that Mike Hargrove doesn’t know what he’s talking about. Jarrod Washburn is Jarrod Washburn – the outside influences that have a substantial effect on a pitcher’s ERA don’t change that.
Comments
135 Responses to “The Most Consistent Pitcher of All Time”
Leave a Reply
You must be logged in to post a comment.
Ugh, how did a well formed discussion of Washburn’s periferal stats and how sustainable his success can be against good compitition turn into a debate about how ‘clutch’ Robert Horry is?!?!?
The M’s owe Weaver 99.9% of his $8 million no matter what. If another team signs him after he’s DFA’d, the M’s would be off the hook for the pro-rated league minimum for the duration of his stay for another team. Or, probably somewhere in the neighborhood of $200K.
96 Thanks for wasting 1 minute of my life…I really appreciate looking at stats from 1 year 2005 that did not include the playoffs…Wonderfull link there that proved nothing. Having a hard time finding much detailed playoff stat break down but over the course of his career 15 seasons:
Horry playoff 3 pt% is 38.2%
Reg season is 34.9%
Havent found 4th quarter stats but I am confident they would be higher in the fourth. (I took out the stats from the 2002 season both for playoffs and regular season because he had a hand injury that year and shot very poorly all year long)
Karen,
If someone else picks him up, they pay him the league minimum and the mariners make up the difference.
Dave,
How much of Weaver’s .537 BABIP do you think is luck and how much do you think is due to the fact that Weaver has lost the ability to be a major league pitcher?
So Wash is somewhere betwee his 2.88 ERA “good pitcher” and the terrible one with the 8-14 record last year that got no run support at all. Regardless, you have to acknowledge that he pitches good sometimes(regardless of luck), and if the Bats can score on those games maybe he goes 12-12. At this point, with my low standards, Ill take that from wash-b.
How much of Weaver’s .537 BABIP do you think is luck and how much do you think is due to the fact that Weaver has lost the ability to be a major league pitcher?
I’d probably say his true BABIP, while throwing that horrible two seam fastball over the plate all the time, is probably .350 or .360. It’s bad enough for him to not be an effective major league starter, since he doesn’t miss enough bats to survive that kind of BABIP. But even if the M’s gave him 30 more starts, that .537 mark would regress very heavily.
Dave:
One thing that hasn’t been addressed so far in the talk about Washburn’s year-to-year consistency is what he should be expected to concede to age. Several of his PECOTA comps saw their careers drive off a cliff in their early thirties; is there any real expectation that Washburn can defy that fate and just keep posting the same Three True Outcome rates through the life of his contract?
Also, what interest me about Washburn is that, in his interviews, he seems to have a murky understanding of his strengths and weaknesses. What do you think, overall, about ballplayers and their own awareness of what makes them good or not so good? Are they mere mindless conduits for their tools, for the most part, or are there some in your view who can analyze themselves as well as we do?
I think it’s worth pointing out Jarrod Washburn has had a BABIP consistently lower than average. While his .208 BABIP this year (along with the rest of his peripherals from this year) is unsustainable, but his career BABIP is .270. Part of that is defense, part of that may be park (although his career totals are .288 at home and .255 on the road).
One thing that hasn’t been addressed so far in the talk about Washburn’s year-to-year consistency is what he should be expected to concede to age. Several of his PECOTA comps saw their careers drive off a cliff in their early thirties; is there any real expectation that Washburn can defy that fate and just keep posting the same Three True Outcome rates through the life of his contract?
I think Safeco’s going to hide a lot his decline. It’s really the perfect park for him to pitch in. Even if he loses a bit of his command and starts sitting at 85, righthanders can only do so much in Safeco, and most managers aren’t smart enough to put lefties in the line-up against him when they play in Seattle.
Also, what interest me about Washburn is that, in his interviews, he seems to have a murky understanding of his strengths and weaknesses. What do you think, overall, about ballplayers and their own awareness of what makes them good or not so good? Are they mere mindless conduits for their tools, for the most part, or are there some in your view who can analyze themselves as well as we do?
I think most pitchers inherently understand the DIPS theory. You hear comments all the time about how balls just fell in here or there and how the results don’t necessarily match up with how they felt they pitched.
Of course, then there’s Joe Morgan, who seemingly has no understanding of why he was a hall of fame player. So I’d say some get it, some don’t. I think Washburn does.
97: If some players are better in the clutch, it should be measurable vs their ordinary performance. If statistics can’t measure it, why not?
party4marty:
What does the Mariners’ offense have to do with how good/bad/average Washburn is at pitching?
chrisisasavage,
Of course a great number of washburn’s road starts are in Safeco, the Oakland Coliseum, and Dodger Stadium when he was in Anaheim. And now that he’s a Mariner he gets Oakland, Petco, and Angel Stadium. His road parks are pretty friendly as well.
david h,
They only affect his winning %.
I think it’s worth pointing out Jarrod Washburn has had a BABIP consistently lower than average. While his .208 BABIP this year (along with the rest of his peripherals from this year) is unsustainable, but his career BABIP is .270. Part of that is defense, part of that may be park (although his career totals are .288 at home and .255 on the road).
I wouldn’t say he’s had a consistently lower than average BABIP. He’s had a couple of seasons where he was the best in the league at preventing balls in play, and they bring his career average way down. He had a .203 BABIP in half a season in 2000, and then a .271 BABIP in 2002, and a .268 in 2003.
However, those Angels teams were some of the best defensive teams in recent memory. MGL had the ’02 Angels team that won the World Series as being over 100 runs above average with the glove that year.
Washburn’s run BABIPs of .269, .290, .203, .290, .271, .268, .293, .289, .290, and .208 in his career. He’s never been below average, but that’s not consistently above average either. That’s a lot of average and a few unbeliveable years.
HI
does anyone know if there is anyone doing in-depth statistical analyis of soccer? As a baseball fan who lives in the part of south america where almost no baseball games are played, let alone televised, i am being forced to concentrate on a sport i know almost nothing about. help?
Dave —
OK, one last clarifying question. If I’m following your reasoning, with all the other peripherals being remarkably consistent to what he’s done in his career, the low BABIP is a signal that what’s going on with Washburn’s other stats — runs allowed, LOB, etc. — is much, much more a function of what bad hitters are doing to get themselves out (by not converting more of their balls in play into hits) than what Washburn is doing with his pitches. Right?
#115 thanks Dave, I was assuming defense player a major role in it.
Looking at his numbers, I see 5 (including this year) better than average, and 5 average years. That includes 2 completely blatant outliers If you pull 2000 and 2007 (blatant outliers), he averages more like 0.283, which is like 0.008 lower than average. His flyball tendancies, and the fact that he has played in friendlier parks has led to a HR/F of 8.8% (unadjusted, since 02 at least, per fangraphs), leading to a few more flyballs that can be caught. Non HR flyballs have had a .167 (and probably lower w/ good OF defense) babip, also per fangraphs. Off the cuff calculations say that would shave (in conjuction w/ his 43.7% FB rate) ~.0015 off his babip during that time. If you apply a .0015 adjustment the 0.283 (.2825) BABIP, it’s more like .285, which is only ~.006 lower than average, which is well within the realm of good defense, and park factors.
OK, one last clarifying question. If I’m following your reasoning, with all the other peripherals being remarkably consistent to what he’s done in his career, the low BABIP is a signal that what’s going on with Washburn’s other stats — runs allowed, LOB, etc. — is much, much more a function of what bad hitters are doing to get themselves out (by not converting more of their balls in play into hits) than what Washburn is doing with his pitches. Right?
Bingo. He’s faced some horrible offenses, and those horrible offenses have been predictably horrible at hitting him.
Thanks, Dave and arbeck for clarifying that (Weaver’s $$$ if he’s DFA’d).
I like my misconception about who pays better…. 🙂
David H.
Washburn has pitched good(lucky) enough to win 4 of 5 this year. But similar to last year, the offense does not like to score when he pitches. IF he allows 2 runs, and the M’s score 1- they lose. He doesnt deserve to lose those games. But the Ms love to not score when he pitches well(or when he pitches bad I guess)
Dave – Wouldn’t we expect Washburn to have a BABIP that is above average given the number of games played against horrible offenses? (not just now, but for the rest of the year too) If I’m understanding the theory correctly, pitchers don’t impact BABIP and we would, thus, expect two different pitchers to have about the same BABIP. However, this conclusion relies, it seems, on those two different pitchers facing about the same level of competition as hitters do impact BABIP. Given that many of Washburn’s starts will be against poor offensive teams, what would you expect Washburn’s BABIP to finish at? If he were in the AL East instead?
116 – Great question. I’m not aware of anything in soccer outside of fairly rudimentary pythagorean stuff using goals for and against.
It’s a MUCH harder sport to break into binary outcomes, and the depth/breadth of baseball stats are due largely to the high number of data points (ABs/Games/innings/ks whatever) and the ability to make binary outcomes out of them. You just can’t do that in a sport that’s so free-flowing. I keep thinking that you could do something like what BIS does in baseball and have a huge database of ‘plays’ in each game, from one-on-one challenges to saves (this would look a lot like baseball’s defensive metrics) to crosses. But even so, you’re isolating little bits of information that are nowhere near as important to the game as the pitcher/hitter match-up in baseball. It’d be tough, but you’d have to set your sights somewhat lower.
122- I think we could probably expect Washburn to have a BABIP that’s a bit above average because 1) he’s more of a flyball pitcher and 2) he’s a flyball pitcher in Safeco. I think that may matter more than the quality of opposition, but I’d be interested to see the numbers on that after a few more months.
Take a look at: http://www.fangraphs.com/blogs/index.php/batted-ball-splits.
Here’s the BABIP part:
Type BABIP
FB .167
GB .236
IFFB .003
LD .716
I don’t think that flyball pitchers have signifigantly better babip than high gb pitchers, but probably have a somewhat lower babip. One thing is high flyball pitchers do have more LD than high GB pitchers. It’s not a linear relationship though. Higher GB% means less balls in the air, but from my experience, a higher percentage are LD, and a lower percentage are IF than a low with GB% pitcher. The really high GB guys still have slightly lower LD #’s, than the really low GB, just not proportional to the number of balls in the air.
One way to look at it (or at least a way I’ve looked at it) is too take the THT data, and use those BABIP #’s and add everything together. 11.47% of all FB are HR, so remove that percentage of OF , and .97% of LD are HR, so remove those too. Add the remainder, multiply by the corresponding number from the fangraphs post, add it all together to get an expected BABIP. The correlation to GB% is .35 for >=162 IP (qualified starters) which is signifigant, but not strong. The correlation to actual BABIP is .06, which by itself doesn’t tell you anything except the correlation is so weak, it could be noise. I find that (not adjusting LD or IF #’s) is low gb guys have a lower expected BABIP (but more damage on batted balls because of extra XBH, especially HR). I’ve used GB% as a predictor for LD and IF, numbers, which may not be correct, but when I do, and do the same experiment, I get a range of .285-.296 BABIP (based on 1 SD = .005) for almost all pitchers, which is probably fairly close to most pitchers true talent levels. If I dont adjust LD and IF% to GB%, the range is quite a bit higher.
Correction to #124 that was .286-.296 expected range, and 70% of the popultation would be .289-294, with .294 being a 52% GB guy (HoRam), and .289 being a 38% GB guy (Johan Santana). That’s not the actual true talent, but I think the model should come up close, since it corrects for most luck factors.
I realized Safeco will help Washburn look good though. Possibly a lot.
Surely Washburn’s LOB% is an expression of the success he’s having in terms of his general GB%. The three statistics you use when determining what you would expect someone’s LOB% to be are walk and strikeout rates and the amount of flyballs. If you just look at his BB/G and K/G then it looks like he’s exactly the same pitcher, but if you look at his GB% then the numbers are –
2005 – 39.3%
2006 – 39.9%
2007 – 49.9%
LOB% has always puzzled me as a statistic because it’s treated as an indicator of luck, yet the expected LOB% for a given pitcher is determined by his skill set. This is fine when you’re analysing a performance after it’s already happened, but you can’t predict what a pitcher’s BB/G, K/G and GB% will be before the season’s actually taken place and consequently, it’s impossible to predict what the expected LOB% will be beforehand. It’s entirely possible that Washburn is performing better (his GB% certainly indicates this) and this has raised what we would expect his LOB% to be – which in turn makes his performance less of an outlier.
LOB% is partially skill, based on K, BB, BABIP, flyball tendancies, etc, but like everything else, outliers, that are playing above or below what their statistics would indicate are going to regress, like any other statistic. I don’t know any formula perse to predict LOB%, but I’m sure it wouldn’t be hard to figure out.
Check this out:
http://www.hardballtimes.com/main/article/ten-things-i-didnt-know-a-couple-of-weeks-ago/
Read: Better pitchers leave more runners on base.
Talks about using xFIP to predict LOB%.
I’ve heard that, in addition to hating the Mariners, Dave also hates freedom.
True story.
Ok if you use Washburn’s xFIP to predict his LOB% it would be a lot higher this year than last year. The lower a pitcher’s xFIP is, the higher we would predict his LOB% to be. Washburn’s xFIP for the last three years is –
2005 – 5.01
2006 – 5.35
2007 – 4.70
Consequently we would expect (under that reasoning) that Washburn’s LOB% will be closer to the 2005 level than 2006.
aaaaaaand, leading off the Gaspump tonight — how much better Washburn is pitching this year.
I dont think it was meant to imply that xFIP should be used as an accurate predictor of LOB%, but that it they have a somewhat signifigant correlation. That’s intuitive, since xFIP is composed of some of the same peripherals as LOB%. Both LOB% and K% use K and BB, and OF% (HR adjustment in xFIP) correlates strongly w/ hit and XBH allowance which an accurance LOB% model would use.
Dave,
Great write-up as usual. While I won’t dare argue with your reasoning on the three outcomes, and other factors, I have to admit that Jarrod Washburn has been lucky enough to not be the firebreathing horror show I thought he would be when we signed him two years ago to his instant-albatross contract.
May he be so horrible in the future? That is certainly possible. And if he is, then you’ve made your position well known and you can always point to this thread (and others) and say “told you so.”
And I agree that just because you don’t think a Mariners player is the bee’s knees doesn’t make you anti-Mariner. I guess that makes 99% of Phillies and Yankees fans anti-fans.
RE: LOB%
Per THT’s website, the formula is (H+BB+HBP-R)/(H+BB+HBP-(1.4*HR)). It’s the percentage of runners that do not score. Replace R w/ xR, H w/ xH, HR with xHR, and use properly regressed BB+HBP #’s. xR, xH, and xHR can come from whatever formula you want. I’d probably create an “expected line” (H, R, HR) from their batted ball data and run the numbers into a BaseRuns formula, then use the above formula. I’m not sure the value of that though, since the expected BaseRuns data would be more useful in determining a pitchers value. LOB% and BABIP correlate strongly with pitcher performance, but each one measures one aspect of a pitchers performance, there are much better metrics for evaluating pitcher talent.
Agreed however I think the premise of the article here could be called into question a little if you consider that xFIP/FIP is a better predictor of a pitcher’s performance than simply his BB/G + K/G. Surely the fact that Washburn had a lower xFIP/FIP in 2005 and 2007 than he did in 2006 goes some way to explaining his success as opposed to just the luck theory derived from LOB%. I don’t think it’s altogether fair to say that he was exactly the same pitcher in 2005 and 2006 when his FIP/xFIP numbers were lower. Surely at least some of his success was down to improved performance (even if luck does factor into the equation).