Another thing to consider-if the "ERA limit" is a function of a certain percentage of BIP becoming hits even if they're not line drives, then the increase is\n K-rates over baseball history should facilitate better ERA+'s, since the best power pitchers have been assuming control over an increasingly large percentage of the outs in the game.

Re: Post #95.

I did say it was a cheap study, a quickie. I invite others to explore this issue - there should be several reasonable ways to access the question.

in 1968, Bob Gibson was never pulled from the game (he was pinch hit for 6 times).

Gibson wasn't having that good of an offensive year that year - no HR, .170 BA - so I guess it would make some sense to send up Johnny Edwards or Dick Simpson to bat for him. Most of those PH games happened early in the season. Gibson did average 8.95 innings per start. (Obviously, he did have games in which he went more than 9 innings.)

But there's a game I remember. It wasn't 1968, but one of the other nearby years - I'm not sure whether to look in '66-'67 or '69-'70 - in which (against the Giants, I think), Gibson gave up 9 runs in the first inning before the shocked manager recovered enough to figure out that maybe he should come out of the game.

Gibson gave up 9 runs in the first inning before the shocked manager recovered enough to figure out that maybe he should come out of the game.

June 29th 1967

GIANTS 1ST:
--Davenport singled to pitcher;
--Haller singled to right [Davenport to third];
--Mays singled to center [Davenport scored, Haller to second];
--Hart singled to left [Haller scored, Mays scored (error by Brock), Hart to second];
--McCovey was walked intentionally;
--Brown singled to center [Hart scored, McCovey to third];
--Lanier tripled [McCovey scored, Brown scored];
--Fuentes singled to center [Lanier scored];
--On a bunt Gibbon struck out;
--Davenport popped to second;
--Haller walked [Fuentes to second];
BRILES REPLACED GIBSON (PITCHING)

... then Briles gave up a single to Mays and a HR to Hart and then a single to McCovey before he got Brown to foul out.

Gibbon had a similarly brutal start in the bottom of the first but the Giants quickly pulled him.

Thanks, David. That's the game - and I knew that it was actually an 11-run inning with 9 charged to Gibson, only 7 of which had scored yet when he came out. Harry Carey was spinning it as Gibson's stuff being good, he was just getting unlucky with dinks and bloops. Of course, as you read through that, the spin gets gradually harder and harder to swallow.

I am not going to argue against Pedro having the best five year peak ever because I believe that he is a shoo-in as well. However, Koufax pitched mant more innings than Pedro (another reason for his lower ERA+?) and probably even pitched more compared to his peers as Pedro is not known as a workhorse.

Pedro has been relatively durable, he's topped 200 innings, the standard of durability for our time, numerous times and been awfully close in most others.

They say everything is relative: it's especially true with innings pitched.

Pedro Martinez 15-team league
8x top 10 in ERA+, 5x #1
6x top 10 in IP

Sandy Koufax 10-team league
6x top 10 in ERA+, 2x #1
4x top 10 in IP, 2x #1

Addie Joss 8-team league
8x top 10 in ERA+, 1x #1
2x top 10 in IP

Pedro was never the single-season workhorse Koufax was in a couple of his top seasons, but he is not weak on the durability side. And his performance on the per-inning quality side is definitely superior. He'll have to top the league in ERA+ several more times to catch Lefty Grove or Roger Clemens in times leading the league in this category, but a quick search doesn't find anybody else who managed the feat more than 5 times (others with five include Mathewson, W. Johnson, Maddux, and R. Johnson).

It would be good if the electorate would place less emphasis on ERA+ and more on DERA, which is a more meaningful measure of pitching value.

While I think I know where your going with this, I think it could use a bit more explanation....why is DERA "more meanigful"?

It isn't; it messes up knuckleballers and others with high-movement deliveries. Cicotte and Wakefield for the HOM!!

Really quick - OCF - I didn't mean to knock your study, it's very appreciated, and I appreciate you taking the time to do it. Hopefully my criticism didn't come across the wrong way . . . sorry if it did. I just wasn't 100% on board that it was the right way to look at the issue . . . I read the back and forth and now I'm more confused, I'll have to give it another pass to try to digest it all again.

***********

Does DERA do that? I think you are mistaken karl.

"DERA

Defense-adjusted ERA. <i>Not to be confused with Voros McCracken's Defense-Neutral ERA. Based on the PRAA, DERA is intended to be a defense-independent version of the NRA. As with that statistic, 4.50 is average. Note that if DERA is higher than NRA, you can safely assume he pitched in front of an above-average defense."</i>

emphasis mine up there . . .

All DERA does is adjust for the quality of the defense behind pitchers, it isn't a DiPS thing.

Cicotte 3.83 NERA, 3.94 DERA, means his defense helped him some. I think that'd be obvious, he pitched for some great teams.

Wakefield's DERA is better than his NERA, 4.33 vs. 4.38, meaning his defenses were a little worse than average. He has a career ERA+ of 109, since a 4.50 DERA or NRA is equivalent to a 100 ERA+ (in theory) I'd say they are quite similar, but one adjusts for quality of defense and the other does not. Cicotte had a 123 ERA+.

Wakefield had a 116 ERA+ in 2001 - league average was 4.53, and he was 3.90. So Cicotte's 3.83 being 123 is pretty close.

I agree that while there may be issues, I put more weight on DERA and Translated IP than ERA+ and IP.

I have much less faith in Translated ERA though:

"Translated Pitching Statistics

Converts all pitching statistics into a standard context. Pitchers are translated to a league where the top five pitchers (in innings) pitch an average of 275 innings. An average pitcher will have rates, per nine innings, of 9.00 hits, 1.00 home run, 3.00 walks, 6.00 strikeouts, and 4.50 earned runs. In the standard context, a replacement level pitcher has a 6.00; the translation is set up to conserve runs above replacement (alltime PRAR). Wins and losses are set using the pythagorean formula with average run support, with the pitcher's actual deviation from his real expected win percentage added back in."

I don't really like the way they convert the components on this one. We know how many runs he gave up, there isn't a need to adjust the component numbers, IMO.

While I think I know where your going with this, I think it could use a bit more explanation....why is DERA "more meanigful"?

To elaborate slightly on what Joe has said, I view DERA+ as more meaningful than ERA+ for two reasons:

1) it takes better account of fielding support than does ERA+. ERA+ measures fielding support only in terms of distinguishing "earned runs" from "unearned runs" on the basis of errors. We know that this is not a particularly accurate measure of fielders' contributions to run prevention. DERA measures fielding support in terms of team defensive efficiency, which more accurately reflects the fielders' contributions to run prevention.

2) DERA is normalized to a standard run-context. This removes one major factor that may lead to ERA+ scores being higher in some run-scoring environments than in others.

ERA+ is not a bad statistic, but if we are to use a rate stat for pitching effectiveness, DERA is better.

In a nutshell, DERA is the ERA a pitcher would have had in a standard or neutral or "average" run context, and in front of an average defense, right?

And where do we get DERA numbers? Do us peakers have access to seasonal DERAs somewhere?

In a nutshell, DERA is the ERA a pitcher would have had in a standard or neutral or "average" run context, and in front of an average defense, right?

Yes.

DERA numbers are part of WARP, so they are available on the bp site along with all other things WARP. There are seasonal numbers as well as career numbers.

I put it to you that if you've got defense-adjsuetd ERA and defense neutral ERA, both abbreviated as DERA, us simple souls have the right to be confused. DERA drove me mad this season, because nutters on Sox Therapy kept trying to prove that Wakefield was the worst starter in the Red Sox rotation, when with 16 wins and the lowest ERA he clearly wasn't, he was about the best, in a year without a Pedro or a 2004 Schilling.

This DERA seems much more sensible, because it doesn't make silly assumptions about every ball in play being exactly equivalent, which if you think about it has to be nonsense when you have pitchers like knuckleballers (and presumably Hubbell's screwball) who rely on deception and preventing the hitter from clouting it square on. It certainly makes sense that Cicotte had an above-average defense behind him, and from observation Wakefield has always had a pretty terrible one, because Red Sox GM's tend to regard fielding as being for little people.

Karl,

I think that the defense nuetral ERA is generally called DIPS. Though I made the same mistake when I began the project and was corrected by, I believe, Chris.

I don't think that DIPS is silly, it just has to be taken with a grain of salt.

j, DERA and DIPS are definitely different things. Just don't ask me to explain why.

>I put it to you that if you've got defense-adjsuetd ERA and defense neutral ERA, both abbreviated as DERA, us simple souls have the right to be confused.

Post of the week.

Well, I thought it was Defensive-Efficiency (adjusted) Run Average

Chris Cobb #112
2) DERA is normalized to a standard run-context. This removes one major factor that may lead to ERA+ scores being higher in some run-scoring environments than in others.

I think this depends entirely on how the data is normalized to 4.50 runs per game. Simplistic normalization preserves differences between low-scoring and high-scoring environments debated by Joe Dimino and OCF this week.

In a nutshell, DERA is the ERA a pitcher would have had in a standard or neutral or "average" run context, and in front of an average defense, right?

A nutshell isn't going to answer the question about whether DERA is really better than ERA+.

How is it determined what is an "average" defense vs. a "good" defense, especially considering we mostly talking about seasons pre-zone data? What is the specific adjustment - it is in runs, or is it applying a ratio to the ERA? Is the defensive measure park-adjusted? If so, how? I think the group seems to be blindly believing that DERA is somehow superior...

I think the group seems to be blindly believing that DERA is somehow superior

Most of the information needed to see that DERA is superior to ERA+ is available in the BP glossary, but for the sake of convenience, I'll run through the defense of DERA here. This is what they do:

1) Find raw runs allowed average in the pitcher's innings.

2) Adjust raw runs allowed average to NRA: normalized run average. BP describes the adjustment as follows:

"Normalized runs" have the same win value, against a league average of 4.5 and a pythagorean exponent of 2, as the player's actual runs allowed did when measured against his league average.

This is the method of adjustment that I proposed as being appropriate in post 97 above, and it should answer the concerns that Paul Wendt raises in the preceding post. By normalizing the run context based on the win value of runs saved, the process of generating DERA removes one potential shortcoming of ERA+, namely the influence of changing run environments on its values.

3) From NRA, calculate Runs saved above average: RAA. BP describes this process as follows:

Runs above average. At its simplest, this would be the league runs per inning, times individual innings, minus individual runs allowed. However, we have gone one step beyond that, because being 50 runs above average in 1930, in the Baker Bowl, doesn't have the same win impact as being +50 in the 1968 Astrodome. The league runs per inning need to be adjusted for park and team hitting . . . and then you can multiply by individual innings and subtract individual runs. Finally, that quantity needs to be win-adjusted. See win-adjustment.

Note -- league runs per inning is park-adjusted and team-hitting adjusted, so it is calculated off of an appropriate baseline. bbref's ERA+ is calculated off of the same baseline, so DERA matches ERA+ here, and both are good.

4) From RAA, calculate PRAA: pitcher-only runs above average. BP describes this process as follows:

Pitcher-only runs above average. The difference between this and RAA is that RAA is really a total defense statistic, and PRAA tries to isolate the pitching component from the fielding portion. It relies on the pitching/fielding breakdown being run for the team, league, and individual. The individual pitching + defense total is compared to a league average pitcher + team average defense, and the difference is win-adjusted.

To see exactly how PRAA is derived, we would need to see exactly how team average defense is derived and that information is not available in the bp glossary, but we can see from this description that PRAA is ultimately calculated by subtracting runs. A ratio is not used.

5) From PRAA, calculate DERA. BP defines DERA as follows:

Defense-adjusted ERA. Not to be confused with Voros McCracken's Defense-Neutral ERA. Based on the PRAA, DERA is intended to be a defense-independent version of the NRA. As with that statistic, 4.50 is average. Note that if DERA is higher than NRA, you can safely assume he pitched in front of an above-average defense.

This is less helpful than some of their other definitions, though it's good to be reminded of the run environment. However, a little calculating shows that the formula for calculating DERA from PRAA differs little from calculating good old earned-run average. Multiply Adjusted Innings Pitched by 4.5 (the run environment), and divide by 9. This gives you league average runs allowed in the pitcher's adjusted innings. From this number, subtract PRAA to get pitcher's normalized runs allowed. Divide this total by XIP (adjusted innings pitched), multiply by 9, and you have DERA.


I hope that account will give any voters who have not already examined DERA to make an informed decision about whether or not to prefer it to ERA+.

The single "black-box" element we face with DERA is that it uses WARP's calculations of fielding runs above average. Is it better to use that fielding adjustment, whose derivation we are not shown in detail, to determine the fielders' share of responsibility for runs, or to use the earned run/unearned run distinction? Although no fielding metrics are as reliable as our batting metrics, experience with WARP's fielding numbers strongly suggests to me that they are as good as anything available, and that we get more accurate results by using them than by not using them. I've done a lot of my own calculations of basic team defensive efficiency and applied them to pitchers' totals in the ways that seem most reasonable to me, and I have found that my results and BP's DERA track one another well.

Someone who has BP print publications that go into more detail about the fielding system might be able to give us more information here.

experience with WARP's fielding numbers strongly suggests to me that they are as good as anything available

Really? Remeber this is the group that tells us that Barry Bonds gets 20% of his total value above replacement from his fielding. I understand that could be an error of application and not of calculation on their part, but it doesn't inspire much confidence that they've got it together on fielding.

On the other hand, neither does ERA+.

But why must we prefer one to another. I think it's probably likely that we should simply look at both of them. When they agree, great! When they don't, what about a pitcher's circumstances would lead one measure or the other astray.

On the other hand, neither does ERA+.

If you believe that the difference defensively between Andrew Jones (2001) and Bernie Williams (2005) is Errors only, then ERA+ is the stat for you. The utility of ERA+ is based on the proposition that the range of the fielders behind him have negligible effect on the runs surrendered by a pitcher.

Remeber this is the group that tells us that Barry Bonds gets 20% of his total value above replacement from his fielding.

Read carefully through Chris Cobb's excellent description of DERA. All calculations are relative to AVERAGE. Your beef is with BP's definitions of replacement level, which is a concept that is applied AFTER the calculations relative to average, and which is a separate debate.

Thanks Chris, that's a good explanation, and I feel a little better about using DERA, although as Dr. C says, I'll be using ERA+ too.

I like Jim's point about the difference between replacement level and average for WARP. I personally think that WARP woudl be a little more accurate if it either a)used FRAA instead of FRAR or b) had a much higher replacement level.

Let's throw my buddy Rixey into the mix, playing off the previous charts:

Top 10 ERA+ AND Top 10 IP

BoLemon (3-4-4-5-6).... (1-1-1-1-2-3-4-5-10)

Ferrell (2-5-5-6-6-7-7) (1-1-1-2-3-4)

EpRixey (2-2-3-4-5-6-8) (1-3-3-3-4-7-8-8-9-9)


ERA+ seasons (top 7, minimum 180 innings)

BoLemon 144 39 36 34 33 13 12

Ferrell 146 35 33 30 26 24 24

EpRixey 143 42 39 36 29 24 13

But this is awfully convenient for Ferrell, as he has a 133 in 181 IP.

At 210 IP minimum, it's:
ERA+ seasons (top 7)

BoLemon 144 39 36 34 33 13 12

Ferrell 146 35 30 26 24 24 (92)

EpRixey 143 42 39 36 29 24 13

Also, Rixey has a 144 in 162 IP, and a 115 in 177 IP (we'll forget about his 144 season in 115 IP, which has no match in the other two careers).

So at 162 IP minimum (a familiar number), listing all qualifiers of at least 100:
ERA+ seasons

BoLemon 144 39 36 34 33 13 12 08 03 01

Ferrell 146 35 33 30 26 24 24 06

EpRixey 144 43 42 39 36 29 24 15 15 13 10 09 09


I'm often amused at how crappy-hitting catchers always seem to be described as "good fielders," maybe because the announcers figure that must be how they got to the majors. But some of them stink on both ends.

We have something a bit different here: Rixey pitched for so long that he tends to get compared to the all-time greats. And of course he's a bit short there.
But then you have Ferrells and Lemons come along, and they're seen as "big peak guys" because that's almost all they have. And Rixey - not much of a peak guy compared to his career-IP peers, but a damn good one compared to the next tier - gets a little lost in the shuffle on some ballots.
Rixey also missed his age 27 season to WW I, and a shortened age 28 season upon returning.
I like Eppa better than Lemon as a pitcher, for sure.
He was a dreadful hitter, though (22 career OPS+).

He was a dreadful hitter, though (22 career OPS+).

dreadful or merely below average?
Coveleski 9, Vance 10

By the way, it appears that Rixey improved as a batter. Coveleski and Vance declined, which is the norm for a pitcher, I guess.
Also btw, Rixey was the tallest major leaguer for a while.

Chris Cobb
2) Adjust raw runs allowed average to NRA: normalized run average. BP describes the adjustment as follows:

"Normalized runs" have the same win value, against a league average of 4.5 and a pythagorean exponent of 2, as the player's actual runs allowed did when measured against his league average.

This is the method of adjustment that I proposed as being appropriate in post 97 above, and it should answer the concerns that Paul Wendt raises in the preceding post. By normalizing the run context based on the win value of runs saved, the process of generating DERA removes one potential shortcoming of ERA+, namely the influence of changing run environments on its values.

Yes, where "its values" means the the different win values, in different run environments, for ERA+ of any given magnitude.

Not the influence of run environment on ERA+ magnitudes, which is the focus of recent debate by OCF, Joe Dimino, et al. Is it "easier" or "more difficult" to achieve ERA+ of any given magnitude in high or low run environment?

Howie, Rixey certainly rates ahead of Ferrell and Lemon on my ballot. Someday I expect him to make my top 15.

Paul,

Since most of "merit" turns on the win-value of player's actions, I think that establishing win-value consistency addresses the truly consequential issue in the variability of ERA+.

The question of "easier or more difficult" to achieve an ERA+ of a given magnitude is an interesting question, but one of less moment to us.

Nevertheless, on that subject, what do you think of the argument that it ERA+ values will range higher in a low-scoring environment because as the denominator gets closer to zero, the same differences between RS and RA will produce larger ratios between RS and RA?

Well, Paul, dreadful but not as much so as his compatriots.
I think Rixey was better than Covaleski or Vance, but I've already lost those battles. I do have to concede that Lemon and Ferrell have that bonus that Rixey doesn't, even relatively speaking.

Remeber this is the group that tells us that Barry Bonds gets 20% of his total value above replacement from his fielding.

17% to be precise. But which Barry Bonds are we talking about? The HR-masher of the past 5 years with the .400+ EQA has 4% of his value from his fielding. Make him a DH and you don't lose much. The young 30-30 Bonds (who was a CF playing LF) in Pittsburgh was 30%. Making him a DH would have been a crime. The career is a composite.

Chris Cobb #129:
The question of "easier or more difficult" to achieve an ERA+ of a given magnitude is an interesting question, but one of less moment to us.

I agree.

Nevertheless, on that subject, what do you think of the argument that it ERA+ values will range higher in a low-scoring environment because as the denominator gets closer to zero, the same differences between RS and RA will produce larger ratios between RS and RA?

I doubt it because I don't think the "same difference" is a compelling focus of attention here. Because runs must be an integer for each team in each game, the "same differences" 0 and 1 do turn up frequently (eg, the man who pitches better commonly yields 1 more run than the one who pitchers worse). For that reasons, discrete mathematics may be attractive. But I think it is significant only regarding single games and small numbers of games, especially where most games pitched are complete games. I think that a continuous model is adequate for full seasons.

IIRC, this explains why I disagree with two recent explanations, by Chris Cob and Dr. Chaleeko. I think they both turn on the point that runs must be an integer.

Great post on Rixey Howie. I'm one of his best friends, and I hadn't even realized that until I saw it in your post.

Chris I think your theory on ERA+ is pretty interesting, maybe OCF's study was on to something afterall . . .

Ugh, there I go posting before reading everything . . . Paul your explanation makes sense too, my mind is turning to mush as this goes round and round.

My mind has long been mush on this. As karl said, we mere mortals have a right to be confused on DERA. It is among the more complex measures, and we have seen confusion (above) among DERA I (defense-adjusted), DERA II (defense neutral) and DIPS. I will continue to use ERA+ and make my adjustments in the ######## dump.

does anyone have team splits for Lemon

#33: Who the hell is Rixey Howie? Another player I'd never heard of before this project! ;-)

The Virtual Cy Young thread turned to Lefty Gomez v Lemon and Walters.
Joe Dimino wrote in #19:
Gomez was such a bad hitter, that he never, not once produced offense [runs] at even 1/2 the league average. That's right, Gomez was such a bad hitter, that he never, not once produced offense at even 1/2 the league average. His career high OPS+ was 15.

That isn't unusual for a pitcher (but Gomez was a bad hitting pitcher).

He never produced runs at 50% of league-average rate, estimated by OPS. His career batting record is a bit worse, estimated by OPS, than half of league onbase average and half of league slugging average, which would be OPS+ = 0.

Lemon's career OPS+ = 82 (not close to uniform) means a bit better than 90% of league onbase and 90% of league slugging averages.

Although the sharp drop in Lemon's batting productivity happened after only three of his nine prime seasons as a pitcher, his batting during those nine seasons was significantly better than during his entire career, about 86.5 to 82. If you call it ten prime seasons, about 92, but his 10-yr pitching record is not as good.

Bob Lemon
full career : pitch 119, bat 82
1947-1956 : pitch 122, bat 92 (2613.3 ip, 1162 pa)
1948-1956 : pitch 123, bat 86.5

Check out Kevin Appier and Bob Lemon.

Even after adjusting for era and Lemon's hitting, I have Appier considerably ahead.

Appier's 1990-1997 is one outstanding prime. 142 DRA+ over 1889 tIP.

I always liked Appier, but I think my impression of him was colored by the 1998-2004 Appier more than it should have been.

Compared to Lemon, year by year, including Lemon's hitting:

Appier 8.4, 7.5, 6.6, 6.4, 5.9, 5.8, 4.6
Lemon 8.2, 7.9, 6.4, 5.9, 5.7, 4.7, 4.4

top 3 consecutive - Lemon 21.8, Appier 21.7

top 5 overall - Lemon 34.1, Appier 34.8

Career WAR - Lemon 54.3, Appier 61.2

Career PA - Lemon .824, Appier .930

Food for thought.

That's funny, Joe, I was going to email you to check out Appier as a candidate.

Paul Derringer is also looking interesting to me recently. And you're old pick, Jim Whitney.

Actually, Appier, Saberhagen, Hershiser and Gooden all are pretty close to the bottom level of what the Hall of Merit has established.

Pennants Added:

Bret Saberhagen .994
Orel Hershiser .932
Kevin Appier .930
Dave Stieb .891
Dwight Gooden .879
Dennis Martinez .873

All are quite peaky too, except Martinez.

Using the scoring system of NHBA - they currently rank #27, #34, #35, #39 and #43 (Martinez is #121). The fit in pretty well with the group of Wes Ferrell, Stan Coveleski, Urban Shocker (don't forget war credit for 1918), Tommy Bridges, Billy Pierce, Don Newcombe, Burleigh Grimes, Bucky Walters, Luis Tiant, Virgil Trucks and Wait Hoyt.

That group ranges from .947 (Ferrell) to .860 (Hoyt) Pennants Added. I italicized the HoMers.

I've got Derringer much lower Jim Sp. .787, which puts him basically tied with Jimmy Key, Schoolboy Rowe and Vida Blue.

Derringer's peak is similar to Mike Garcia's, except that Derringer has like 60% more career.

I've got Derringer similar to Waite Hoyt too, just with a worse back end of the career. Peak wise you could compare him to Tommy Bridges too I suppose.

I'm not sure that floor is "established" yet. Maybe Stieb garnered some support based on the coincidence of his career with the 1980s. Dennis Martinez may be a true contemporary; that seems reasonable to me for this purpose. Appier is ten years later and shares environment with Clemens, Maddux, et al. The others are between.

I understand that Lemon and maybe some other HOM pitchers are below the floor in Joe's terms.

Oh man, I almost left out the best of the bunch.

1.070 Pennants Added. Tied with Dazzy Vance.

#27 peak rank (moves the rest of the group down a notch).

122 DRA+ in 3300 tIP. The guy most hurt by the 1994 strike (his best season, 8.8 WAR).

A 3-year peak that's in the same general class with Fergie Jenkins, Jim Bunning, Frank Viola (.827 PA himself) and Wilbur Wood.

I'm talking about David Cone. Who I will support pretty strongly come 2007.

Yeah Paul, I wasn't saying I endorse Appier yet - just that he's closer to the fringe than I realized.

Lemon shared the environment with Spahn, Feller, Roberts, Wynn, Newhouser, Ford and Pierce.

Don't forget, we've got 60% more teams in the 1980s-1990s than we had back in the day - it hold that the group of electees should also be that much larger.

I'm starting to think maybe the original play of electing 4 players every 4th year is something we should have stuck with. Perhaps we should revisit the idea, post 2008.

BTW, Cone's peak from 1993-95 was so good, there's a non-zero case for his going in as a Royal, especially considering it's his hometown team.

Although as much as it pains me, his most value was probably with the Mets. Depending on how you count post-season value of course, which could tip the tide to the Yankees.

Although as much as it pains me, his most value was probably with the Mets.

Yep, and I will cut out the heart of anybody who says differently.

:-D

I think his three-year peak is the Toronto-KC-Toronto stints, mid-1992 to mid-1995. He helped Toronto a little more in September-October 1992 than he helped New York in August-October 1995. Also, he was a 250-inning pitcher only in 1992-95.

But, no, he can no more wear Toronto or KC than Roger Clemens can wear Toronto. In his eight best seasons by ERA+, leading the league each time, he ranked 56169(Boston) 13(Toronto) _(Houston) in innings pitched.

Joe Dimino, does your system have the entire second tier of 90s pitchers (Brown, Schilling, Mussina, Glavine, Smoltz, Cone) making it in? I definitely see Cone on the wrong side of the line (and Brown as pretty clearly the cream of that crop). Not to mention a potential vote for Appier...it seems to me that you may be making a mistake by adjusting for usage (by translating IP) but not for standard deviation. A 128 ERA+ which was good for second in the 1955 NL would have been 10th in the 2003 AL (and yes, I know the 2003 AL had more teams, but still), and there are very good reasons why: the presence of the 2003 Tigers, much higher strikeout and home run rates giving pitchers greater control over outcomes, etc....if you correct for IP norms but not standard deviation, you'll wind up with a HoM chock full of third-tier 1990s hurlers. I've done research into stdev for pitchers and can send you my regression results if you're interested in incorporating that factor into your system.

I would think it would Dan - and I'd probably agree with that.

I'm on vacation this week, only checking in intermittently.

But I would think the 1988-2007 time-frame should have around 1.5x-1.75x the number of pitchers as MLB had from say 1943-1962, for example - just because of the increase in the number of teams.

I haven't actually sorted all of that out yet.

But all of the names above that you mentioned (Brown, Schilling, Mussina, Glavine, Smoltz, Cone) sound intuitively like Hall of Famers to me - again, without having looked into it that deeply.

My point is that part of the reason all of them sound like Hall of Famers is that they played in an era when it was easier to put up high ERA+ scores than in previous periods of the game's history. Again, on the other hand, it was harder to put up high IP totals. But that doesn't mean those two factors cancel out exactly--as has been noted, the late 1980s in particular are devoid of the monster seasons Cy Young winners rang up in the late 1990s or early 1970s. But you have to correct for both factors, otherwise pitchers from high-stdev, low-IP eras (eg the modern one) will be grossly overrepresented.

Dan R,

So what pitchers from, say, the 1980s should get support that we are currently overlooking? We've elected Stieb and Eckersley (choices I agree with) and Blyleven and Ryan were both still active and effective through the 1980s. Saberhagen is about to become eligible, and I at least think we should elect him before we catch up.

Who else is there? Do you think Hershiser (who's not yet eligible) should get support, say, more than Cone? What about Reuschel and Guidry, who are the other serious candidates from the 1975-90 era? Or should we be looking at pitchers from other eras? If you think SD needs to be factored in to Lemon's case, that might suggest that Don Newcombe (who also lacks truly gaudy ERA+ numbers, but who was 2nd or 3rd in the league with ERA+ totals like 128 and 130) ought to be garnering much more support.

I'm trying to get a more specific sense of what to do in practical terms with the SD issue you have raised. If 1990s pitchers drop, who should rise?

I'm trying to get a more specific sense of what to do in practical terms with the SD issue you have raised. If 1990s pitchers drop, who should rise?

This would seem an opportune moment for a Happy Jack Chesbro post. You out there, Jack?

You out there, Jack?

He's dead, Jim...er, I mean, Mark!

Chris Cobb,

I'm afraid I won't have a real answer for that until I have calculated WARP for pitchers, which I definitely won't finish before we're done with the HoM. I will say that the projected SD for the mid-1950s NL is *very* low--Newcombe's low ERA+ totals are DEFINITELY caused in part by adverse league factors. Combined with his hitting and his being blocked, I certainly think he deserves a good long look. As of now, the best thing I can do is make my research so far (on IP, career length, and standard deviation) available to the group and let everyone make of it what they will. Let me know if you're interested.

152-158, 109-135
13, 22, ? lost