Were the park adjustments done on all games played in each park, then there would be one park-adjusted league average ERA for a given league and year. That is not the case, however, as Ervin Santana (4.56) and Tim Wakefield (4.74) and Boof Bonser (4.33) each have different league ERAs for 2007 despite the fact that each of them pitched in the AL last year. The league ERA for the AL in 2007 by this link says it was 4.50. I understand that the 4.50 number is not park-adjusted, but there still should be one park-adjusted figure for a league ERA in any given year.

I didn’t say that inducing groundballs was a bad strategy, but a less than ideal one. Strikeouts are ideal.

The top dozen in K/9IP for 2008 are TIM FREAKING LINCECUM, Edinson Volquez, Chad Billingsley, Jonathan Sanchez (0.93 G/F), C.C. Sabathia, Randy Johnson, Javier Vazquez (0.87 G/F), Josh Beckett (0.94 G/F), Jake Peavy, A.J. Burnett, Matt Cain (0.70 G/F), Ted Lilly (0.74 G/F). Matt and Ted have the “worst” GB/FB ratios among MLB pitchers with 80+ IP this year.

The top dozen in K/9IP for 2007 were Erik Bedard, Scott Kazmir, Jake Peavy, Johan Santana (0.92 G/F), A. J. Burnett, TIM FREAKING LINCECUM, Javier Vazquez, Oliver Perez (0.70 G/F), Daisuke Matsuzaka (0.92 G/F), Josh Beckett, Cole Hamels, and Chris Young (0.57 G/F).

There are no Dan Cabreras or Paul Maholms on these K/9IP lists. What all of the outliers on our lists have in common is that good results (ERA) can be achieved in more than one way, but the results won’t be above average if one’s WHIP is 1.40+.

There has been an underlying assumption that groundballs are good and flyballs are bad, but that’s not necessarily true either. When has a groundball been called a “can of corn” or a “lazy groundball”? More groundballs become hits, and every one of them is not necessarily the fault of the defense.

it needs to be used in tandem with other statistics or you’re comparing apples and oranges. There are some crazy different pitchers on this list, groundballers or no. You must recognize that saying "Roy Halladay’s rookie year was one of his three or four best" and "Pedro Astacio’s rookie year was one of his three of four best" aren’t like terms. Halladay, with the exception of one injury prone year, was head and shoulders above Astacio for their entire careers.

All of their career stats are available via linkage to compare whatever data anyone else cares to compare. Things like Roy Halladay’s or Pedro Astacio’s rookie years were among the best seasons in their careers are meant to suggest that it is not at all unfair to expect a 24-year old first year starter or rookie to succeed. Halladay was indeed a better pitcher than Astacio. I wasn’t ranking them. It would be really wrong to suggest that Astacio was as good as Halladay, and I never said that, but both of them essentially were who they were out of the box, as more than 75% of them were. ERA+ does tell us that much.

Besides, NYRoyal had already compared Hoch to 8-9 other (VERY cherrypicked) current-day sinkerballers. Hoch wasn’t very similar to them in terms of quality results (ERA or ERA+) or consistency (Quality Starts, WHIP), and that is apparent in the links here also. Hochevar had the “all-important” g/f ratio, statistically, but in the real world Luke has been more of a hit parade of groundballs rather than a producer of ground outs.

Regarding Webb, this isn’t much of a conversation. You’re having one with yourself about whether Webb was statistically more valuable in any given year, but I don’t really care about that. ERA+ is simply reflecting the degree to which Webb outperformed the average pitcher in terms of fewest runs allowed while getting batters out. When a team has two aces and a third emerges from a rookie in the back of the rotation and moves to the front of the rotation after his success has made trading one or both of the older aces for future considerations, that is a gold mine for a GM. When that rookie is anchoring the staff later, he is still extremely valuable, but not in the same way. I’ll bet his performance relative to his salary will always be highest in his rookie year. So yes, this is a very relative deal. To the GM, Webb’s rookie year may have been the most valuable to him.

At worst, a few of the rookies and first-year starters wouldn’t have had the very best seasons of their careers and moved to the had one of their 4 best seasons right away bin. That would still prove the point that very, very few struggle so badly in year one and make a quantum leap to become significantly above average afterward.

Bonderman isn’t a comp so much as an example of one young pitcher’s career path. You reiterated many of the reasons why he belongs in this conversation tangentially. Both were rushed to the majors and forced to learn on the job. Organizational expectations were high. Pitchers do get hurt at terrible rates, as I’d mentioned before. If Hochevar takes that long to improve, the odds of him being hurt along the way before ever becoming good are high. Several of the 40 above had injury issues at some point in their careers also. Maybe that shouldn’t be discounted altogether, but I wouldn’t feel comfortable trying to quantify that risk.

A) only only statistic, which is always pretty hard to prove anything with B) the age that these pitchers had their "first full year" as a starter varies quite a lot so as to limit the significance (in RE to Hochevar) of this exercise further and C) the high variance of the career paths of these pitchers isn’t really taken into account in your argument…I don’t believe you’ve necessarily proven that Hochevar needs to put up or shut up this year to be a good mid-rotation pitcher.

A) I think we’ve already covered the ERA+...the question was whether the sinkerballers were effective relative to the league average from the get-go and what % of the first-year below average ones improved to #3 or better starters. You’ve criticized and complimented this twice without suggesting what else should have been used, so I don’t see how this is productive.

B) I agree that the variance in age means that some of them are more meaningful than others. I personally think that the younger pitchers were better and more talented prospects than Hoch in the first place, and that some of the older rookies who were late-round picks or free agent signees were probably less hyped, but still outperformed Hochevar as a group. Seeing so many of them begin their careers in the bullpen suggests that the Royals may have made the wrong decision in that regard. However many years they spent in the pen could translate to the number of years that Hochevar disappoints in the rotation. It doesn’t mean that all examples not age 24 were invalid, though. NYRoyal’s exercise didn’t have many 24-year-olds at all, and Zambrano is not the same sort of groundball pitcher as Wang or Cook.

C) Sure it is. I think that there is always more than one right way to succeed, and that is perfectly reflected in that group. That was also true in the handful of pitchers NY Royal profiled. The various career paths indicate that there are a few historical precedents, though rare, for a pitcher performing this badly as a rookie to improve into an above-average #3 or better (even more rare).

Putting up a great second half and finishing with an above-average ERA+ would greatly increase his odds for success, barring injury. I think we all knew that, but I was trying to come close to quantifying that with this data. A great second half and an overall ERA+ of 97-100+ would give him a 72-75% of being at least that good again (with the possibility of being a good #2 someday being realistic) and better than a 50% chance of being at least that good relatively soon, ignoring injuries. Without any significant improvement this year, his odds of becoming anything more than a serviceable #3 who is most helpful in the fourth slot are less than 10%, and a significant improvement to an ERA+ in the low to mid nineties would give him a 20-25% chance instead.

Unfortunately, all data is rendered useless, and all comparison studies would be unfair.

Yes, right there.

Literally, I meant that comparisons to Tim Hudson or Brandon Webb or even Dan Cabrera or Willie Banks are unfair to them because they were better than this all along. To a lesser extent, it is also unfair to Luke because he’s been led down such an odd path and set up to fail. Sarcastically, I meant this as a criticism of the Royals organization. Comparing a bunch of sinkerballers who weren’t mishandled this way to dispell the myth that it is somehow unreasonable to expect above-average from any rookie pitcher - especially a 24-year-old one - is fair. A year from now, we’ll see how he’s doing. Maybe he’ll be around 97, and then we can base future expectations off of that as if 2009 were his rookie season.

if Hochevar continues to take his turn in the rotation, I would think it’s because the Royals think highly of his chances of progressing significantly, not because of a "tolerance for suck."

Who could the Royals have inserted into the rotation besides Hochevar and Davies? I don’t think they had any other contingent plans. That’s just speculation on my part, and I want to believe that you’re right (and I’m not suggesting that you’re wrong).

Yes, the 75-80% of rookie pitchers are above-average in year one seems to apply anecdotally beyond sinkerballers. The vast majority of highly successful major league pitchers were better than average by ERA+ during their rookie seasons and/or first years as starters. Maybe “lots of pitchers” could apply to 15-20% or less of rookie pitchers who struggled since nothing close to all of those who struggled went on to be #2s or better (more like a minority of 15-20% of rookie pitchers). I can say that it applies to sinkerballers from these eras definitively, and it’s more fair to look to those examples and their peripheral stats as being more like what Hochevar can/may produce in his career. Even if the trend of rookie success applies to all pitchers, the non-sinkerballers don’t provide meaningful insight with their peripherals, i.e., Hoch needs to get his K rate up to be more like Randy Johnson.

Oliver Perez ranked that high in NL ERA, and I would normally not think of him as an ace. I do think that if Hochevar has any hope of being ranked that highly in the league by ERA or ERA+ in any given year, it will be as an Oliver Perez sort of outlier which can’t be counted on to repeat.

For the record, I don’t see Luke as a top of the rotation guy, more like a #3 with the capacity to be dominant or downright bad depending on where the grounders go start-to-start. I also think that, given his minor league numbers, there’s room for real improvement in his K-rate. Do you disagree? I’d be interested to know.

I agree with the #3 assessment, yes. The best and least unfair comparison seems to be to Dan Cabrera right now, and since Hochevar could actually be in the fourth slot if Bannister would only get his act together (i.e., 4.25-4.75 ERA for the season), that would be fine. It wouldn’t necessarily be enough to make the Royals a winner, though.

I don’t know if there is room for improvement with his K rate or not. As you correctly pointed out, that is the only real encouraging data from his minor league history. I had thought that he could improve on that until I read that Luke wasn’t throwing quality strikes (John Sickels). That made me wonder if he’ll ever have the command to be less hittable and get more Ks. The other sinkerballers in the game today are definitely doing something (or more likely a few things) more effectively than Luke to avoid being hit like that, even without more Ks. Lowering his walk rate is almost as important as raising his K rate according to NY Royal’s post and one of Gopherballs comments here. I don’t discount that entirely since it would directly lower his WHIP and bring is walk rate more in line with Wang and others, but I think the crucial step for Hochevar will be to command and locate his pitches more effectively to produce fewer hits, which could result in more strikeouts, too. For whatever combination of reasons, Luke seems to require a higher K rate than Cook or Wang.

I don’t have data for this, but here goes another guess: all groundballs are not created equally, although they may appear that way statistically. Perhaps Hochevar has been generating gbs that are, say, 144% the velocity of Wang’s worst 20% or 160% of Hudson’s worst 20%. That’s the sort of information I’d love to see rather than BABIP, which tells us that he puts more in play than the average starters (nothing we didn’t already know) and that more are going in for hits (which we already knew by H/IP or WHIP). I doubt it’s all the defense’s fault or a fluke.

I don’t think you understand the point of the ERA+ comparison: to demonstrate that most rookie and first-year starters are above-average immediately. Very few aces and #2s started out significantly below average as rookies or first-year starters.

Ignoring that defense and ballpark have a great deal to do with how "hittable" a pitcher is, is exceptionally poor statistical analysis.

Well, the ERA+ is park-adjusted, but this wasn’t really an exercise in statistical analysis so much as an analysis of historical trends.

Here’s what your sort of analysis says: despite the facts that Tomko, Bale, and Fulchino have high G/F ratios and higher than expected BABIP, they are performing well, and their ERAs and WHIPs don’t matter.”

Mine: the facts that Tomko, Bale, and Fulchino have high G/F ratios and higher than expected BABIP only add to the information reflected in hits/innings and WHIP, that they are extremely HITTABLE pitchers! No wonder their ERAs are sky-high!”

BP projected Tomko would improve in 2008 also, FWIW, so I’m not just making this up.

The difference: during your analysis, you get attached to things which indicate a process over the results. The pitcher’s job is to get batters out and allow the fewest number of baserunners to cross the plate in the process. That is what ERA and ERA+ measure. It doesn’t matter that Carlos Zambrano relies a lot less on groundouts than Aaron Cook or Wang or anyone else you had mentioned while striking out a lot more compared to them. All three of them get the job done with a slightly different balance of Ks, control/command, and groundballs, but the objective is to have a low ERA, not to allow a hit parade of groundballs for the sake of the G/F ranking. If Zambrano tried to imitate Wang, he would probably fail.

BABIP and G/F are helpful tools to explain anomalies (is this former Dodgers starter’s WHIP really that high, or a function of BABIP? Well give him a $3M contract, then!) or for use as tiebreakers between two similarly performing starters (one had an average BABIP while the others was the lowest in the league). They were never designed to replace ERA.

I don’t have data for this, but here goes a guess: all groundballs are not created equally, although they may appear that way statistically. Perhaps Hochevar has been generating gbs that are, say, 144% the velocity of Wang’s worst 20% or 160% of Hudson’s worst 20%. That’s the sort of information I’d love to see rather than BABIP, which tells us that he puts more in play than the average starters (nothing we didn’t already know) and that more are going in for hits (which we already knew by H/IP or WHIP). I doubt it’s all the defense’s fault or a fluke.

To compare Luke Hochavar to Garland, Wang, or Hudson is unfair to them because they get the job done well. None of them have a 1.50 WHIP. (I didn’t look at Hudson since he’s in the NL.) Garland is close to Hoch in SLG allowed and BAA. Hoch has the worst OBA by 20 points. Hoch has the lowest K/BB of the bunch, and it isn’t really close to anyone. Luke throws the most pitches per inning, more than 1.5 more per inning than Wang. The fact that Hochevar has the highest K/9IP indicates that he should have an advantage over the others, but doesn’t because he’s too hittable.

That is utter nonsense. You think that after one half season we know within a pretty small range what kind of pitcher he’s going to be. Good lord that is ridiculous, even for you.

How is that more ridiculous than projecting a minor leaguer’s future? Or measuring the odds of a team winning a game before it starts? Or tracking the likelihood of winning a game while in progress…seems like you’ve posted that before. I’m not the only one doing this, after all, and I never said that anything was set in stone.

ERA is a lousy way to evaluate how individual pitchers will perform

We are in 100% agreement. I’ve never said that ERA or ERA+ should be used to determine how a pitcher WILL perform, but rather how well he did his job in any given season. ERA+ was used in the post above to track how often and how early a pitcher performed at league average or higher in his career. I understand that a few of you don’t like that, so feel free to track whatever data you want. It’s all compiled there.

The fact that Hochevar’s flyballs and line drives are mostly going for hits could be the D, or it could be much like Tomko or Bale—a result of simply not throwing quality strikes, not fooling anyone, and getting hit harder than most as a result. FIP is like a tiebreaker stat, especially if I’m a GM with a bigger park and a better D. I don’t like to rely much on FIP because I saw BP predict that Tomko would bounce back based on that. I’ll open my mind to FIP as soon as someone uses it to analyze Tomko’s 2007 Dodger numbers and Bale’s 2008 numbers without concluding that either pitcher is worth starting because they are not. I don’t need a stat to tell me that those guys suck, and I really don’t need one that suggests that they are any good when they are not.

Here’s a new possible comparison for Luke: Kameron Loe. Using whatever “better” stats you care to use, did the Rangers do the right thing by moving him to the bullpen?

Since I’m still (amazingly) interested in the initial thought of comparing Hochevar to past folks, any thoughts on what we should use to set up the test? What group of pitchers would you choose – obviously some sort of objective group of “good” pitchers that probably have had a chance to be not so good and then consistently good? What stats would you compare to judge?

Most starting pitchers take app. 3 years before they get into their breakout years…

Well, the main point of this post had been to survey similar pitchers in the past 30 or so years, and that wasn’t true in their cases. Some of them spent a year or two in the bullpen first, but the vast majority of them (75-85%) were at least as good as the league average from their first year as a starter, and none of them started out so badly.

If you have a link to some more comprehensive research which concluded that the breakouts mostly occurred around 3+ years, I’d love to read it.