Kid Nichols - How Good Was He?

[Bothrops Atrox]

Note, I have implicitly defined the "present heroes" by their debuts beginning 1984, selected for the debut of Roger Clemens, the earliest of the recent quartet. That selection of the starting point, often called "cherry picking" around here, overstates the concentration of great pitchers from any "debutspan"


If that concentration of debuts and careers is not a mistake then there is no bias in favor of either deadball era pitchers or present heroes to be explained. I think "everyone knows" there is no bias in favor of Ted Lyons' or Warren Spahn's generations. The only other time period plausibly the object of favorable bias is Tom Seaver's generation.


That is a good general point, increasing numbers in recent years, but it does not explain the concentration of "greatest pitchers".

"All pitchers in baseball history" are not relevant in discussion of the greatest. The debuts beginning 1984 cover perhaps 35% of more than 8000 men who have pitched in the major leagues and 30% of more than 4000 who have pitched 100 innings, but only 16.4% of the 688 pitchers with 1500 innings entering this season and 12.7% of the 407 pitchers with 2000 innings. (The latest debut of the 2000-inning pitchers was Tim Hudson 1999.)

Those shares 16.4% and 12.7%, and 11.5% of the 234 pitchers with 2500 career innings, do not represent much recent concentration of long-career pitchers measured by innings pitched. (Why not? Regular starting pitchers work fewer innings each season. In the career sums at this level, that roughly balances the greater number of teams.) Nevertheless the recent period and the deadball era produced heavy concentrations of long-career pitchers with high ERA+. Consider the thresholds 2500 career innings and career ERA+ at least 120 or 130.

		all 2500 IP	ERA+ >= 120	ERA+ >= 130
debut		count		count	pct	count	pct

before 1890	30 (12.8%)	8	26%	2	7%

1890 to 1911	43 (18.3%)	16	37%	8	19%

between		134 (57.2%)	19	14%	4	3%

1984 to dddd	27 (11.5%)	12	44%	4	15%

--------	---		---		---
all-time	234 (100%)	55	24%	18	8%

Blue marks column percentages. Other percentages ("pct") measure the high-ERA+ subset among the high-innings pitchers: row percentages.

Within the latest period, all twelve of the 2500-inning pitchers with ERA+ 120 or greater debuted 1984 to 1992, demarcated by Roger Clemens and Pedro Martinez.
: the big four, Clemens, Maddux, Johnson, and Martinez
: the little four, Brown, Glavine, Schilling, and Smoltz
: four more, Key, Saberhagen, Appier, and Mussina
They are twelve among only 55 in major league history. Sixteen more of the 55 debuted 1890 to 1911, demarcated by Cy Young and Pete Alexander.

Merely nineteen debuted during the long intervening period 1912 to 1983, which produced much more than half of the 2500-inning pitchers.

As far as perception, you are right - there is probably an ERA+ bias. Of course a huge majority of guys from the 1984-on group would be under a 120 ERA+ if they threw 4,500 to 5,000 IP like in the dark area, instead of the 2,500-3,500 most of them pitched. That is why ERA+ always has to be considered with a time-played element, as you have done. I wonder how many of the Steve Cartons, Warren Spahns, Phil Niekros, Perrys, Blylevens, or even Johns, Koosmans, Tiants, and Kaats etc. had a 120 or higher ERA+ at 2,500-3,500 innings or roughly the number of IP by most of these 1984-on guys. I would bet most of them. The lack of IP seems to be the single driving factor in higher ERA+ numbers (of the recent era at least.) When you look at the ERA+ years of 200 or more in the past 15 years, virtually all of them are either in seasons of fewer than 220 IP or connected to steroid users. Hopefully those who rank players will keep into consideration IP (BF) just as much as ERA+ by itself. This is why I have Niekro, Perry, Blyleven, etc, ranked much higher than most.

That is what is nice about WAR and WSAB, WPA/LI, etc. - they all look at runs allowed based on baeRuns (and defense/park neutralizes them) considers era run environment, IP (BF) compared to league, etc. And the results show a pretty evenly distributed group of pitchers, even though there doesn't necessarily have to be - it is possible for a better crop at one position to play at one time. And yes, the big four and little four score very well on those lists as well.

Just some thoughts - not disagreeing with anything you said.

[Paul Wendt]

high-innings, high-ERA+ careers --concentrated in time

Note, I have implicitly defined the "present heroes" by their debuts beginning 1984, selected for the debut of Roger Clemens, the earliest of the recent quartet. That selection of the starting point, often called "cherry picking" around here, overstates the concentration of great pitchers from any "debutspan"

This may be true. But what if that concentration isn't a mistake?

If that concentration of debuts and careers is not a mistake then there is no bias in favor of either deadball era pitchers or present heroes to be explained. I think "everyone knows" there is no bias in favor of Ted Lyons' or Warren Spahn's generations. The only other time period plausibly the object of favorable bias is Tom Seaver's generation.

6/16 from the 1890's-1910's is a little surprising. 4/16 from the mid 80's on is a little less so, as probably more than 25% of all pitchers in baseball history have played in that time.

That is a good general point, increasing numbers in recent years, but it does not explain the concentration of "greatest pitchers".

"All pitchers in baseball history" are not relevant in discussion of the greatest. The debuts beginning 1984 cover perhaps 35% of more than 8000 men who have pitched in the major leagues and 30% of more than 4000 who have pitched 100 innings, but only 16.4% of the 688 pitchers with 1500 innings entering this season and 12.7% of the 407 pitchers with 2000 innings. (The latest debut of the 2000-inning pitchers was Tim Hudson 1999.)

Those shares 16.4% and 12.7%, and 11.5% of the 234 pitchers with 2500 career innings, do not represent much recent concentration of long-career pitchers measured by innings pitched. (Why not? Regular starting pitchers work fewer innings each season. In the career sums at this level, that roughly balances the greater number of teams.) Nevertheless the recent period and the deadball era produced heavy concentrations of long-career pitchers with high ERA+. Consider the thresholds 2500 career innings and career ERA+ at least 120 or 130.

		all 2500 IP	ERA+ >= 120	ERA+ >= 130
debut		count		count	pct	count	pct

before 1890	30 (12.8%)	8	26%	2	7%

1890 to 1911	43 (18.3%)	16	37%	8	19%

between		134 (57.2%)	19	14%	4	3%

1984 to dddd	27 (11.5%)	12	44%	4	15%

--------	---		---		---
all-time	234 (100%)	55	24%	18	8%

Blue marks column percentages. Other percentages ("pct") measure the high-ERA+ subset among the high-innings pitchers: row percentages. For example (big), beginning with 1984 debuts 44% of 2500-innings pitchers have achieved career ERA+ 120.

Within the latest period, all twelve of the 2500-inning pitchers with ERA+ 120 or greater debuted 1984 to 1992, demarcated by Roger Clemens and Pedro Martinez.
: the big four, Clemens, Maddux, Johnson, and Martinez
: the little four, Brown, Glavine, Schilling, and Smoltz
: four more, Key, Saberhagen, Appier, and Mussina
They are twelve among only 55 in major league history. Sixteen more of the 55 debuted 1890 to 1911, demarcated by Cy Young and Pete Alexander.

Merely nineteen debuted during the long intervening period 1912 to 1983, which produced much more than half of the 2500-inning pitchers.

[Bill Burgess]

It's been my experience that to be ranked well, one normally needs some energetic supporters to push their case. That is often what it took to get candidates elected to the Hall of Fame.

A few good men to write letter, lobby voters, etc.

That is what I have done here for Cobb, Ewing and some others. The Kid will need someone here to push his case, with his case logically presented, and presented regularly and consistently. Sometimes, good credentials are not enough.

[Bothrops Atrox]

--

If that concentration is a mistake, I'm not sure that "idolization" of present heroes and those from a century ago is the cause. There may be a bias of another kind. Perhaps participants have relied on the ERA+ measure but some general conditions rather than individual greatness generated more high ratings from those two time periods.

This may be true. But what if that concentration isn't a mistake?

6/16 from the 1890's-1910's is a little suprising. 4/16 from the mid 80's on is a little less so, as probably more than 25% of all pitchers in baseball history have played in that time.

I bet we will also see the next 30 pitchers in that poll absolutely littered and inidated with guys from the late 50's through the early 80's. They may be a perceived as a little below the top 10-15, but they will be well represented soon. Not sure what that all means - just an observation.

[Paul Wendt]

In this forum this month,

About twenty readers voted Kid Nichols "15th Greatest Pitcher". Numbers one to fifteen have been named by a continuing series of five-man rank-order ballots; the "16th" is now underway.

I have updated the linked summary "who ranks Nichols where?" (#8) to include this latest result.


--

Oh come on, there are just as many if not more guys around here who glamorize/idolize old players and come up with any reason possible to make current players look bad. Many completely ignore LQ all together including integration, etc. which is way more laughable than putting Nichols down a few slots.

The biases run both ways.

(my emphasis)
It seems likely that the series linked above will name Joe Williams the 16th greatest pitcher. If so then the sixteen will include six pitchers from the 1890s to 1910s (debut 1890-1911), four from the 1990s and 2000s (debut 1984-1992), and only six from the long intervening period (debut 1912-1983; none pre-1890). If that concentration is a mistake, I'm not sure that "idolization" of present heroes and those from a century ago is the cause. There may be a bias of another kind. Perhaps participants have relied on the ERA+ measure but some general conditions rather than individual greatness generated more high ratings from those two time periods.

[Paul Wendt]

Boston 1897

Five times the Boston pitcher started on one day's rest (*): Nichols against the Baltimore champions ...

June 22 and 24, first and last of a three-game series versus Baltimore. That followed his pitching two of three versus Chicago with two days rest, then skipping for a three-game series at Brooklyn. Immediately following the Baltimore series Lewis pitched first and last of a three-game series vs Brooklyn with one day rest.

Jun 16 - Jun 30
vs Chi: Nich-Klob-Nich (two days rest)
at Bro: Klob-Stiv-Lew
vs Bal: Nich-Klob-Nich (one day)
vs Bro: Lew-Klob-Lew (one day)

Klobedanz worked one in each series with three days rest each time.

Nichols worked two of three in another Baltimore series at the close of the season.

Sep 16 - Oct 2
vs NYk: Klob-Nich-Klob (one day rest)
vs Bro: Nich-Lew-Stiv
at Bal: Nich-Klob-Nich (two days)
at Bro: Lew-Klob-Lew (two days)

[Paul Wendt]

Boston 1897

Next season Boston certainly used better pitchers led by Kid Nichols a little more against stronger rivals. Indeed, Nichols started half of the 12 games against Baltimore, including one on two days rest. He also started one-third of the 36 games against other strong teams and one-quarter of the 87 games against seven mediocre and weak teams!

Meanwhile Fred Klobedanz started almost one-third of the games against Baltimore and the other strong teams and one-quarter of the games against mediocre and weak teams. Ted Lewis, in his first full season, worked more against the mediocre and weak teams. Jack Stivetts and Jim Sullivan, who helped eat innings in 1895-96, worked much less in 1987. (Stivetts had been a star on the 1892-93 champions but he was over the hill by this time.)

Pitcher starts, Boston NL 1897

	all	Nichols	Klobed.	Lewis	Stivet.	Sullivan
all	135	40	37	34	15	9

by opposing team, ordered and grouped according to full-season wins
Bal	12	6 *	4	1	1	0

NYk -	12	3	4 *	2	2	1
Cin	12	4	3	3	2	0
Cle	12	5	4	2	0	1

Bro	12	1	3	6 *	2	0
Was	13	3	2	3	1	4
Pit	12	4	4	4	0	0
Chi +	13	5	4	3	1	0
Phi	13	4	4	3	2	0
Lou -	12	3	3	3 *	1	2

StL	12	2	2	4 *	3	1

by opposing team grouped according to full-season wins
Bal	12	6	4	1	1	0
next3	36	12	11	7	4	2
next6	75	20	20	22	7	6
StL	12	2	2	4	3	1

in doubleheaders
dh	26	5	6	7	3	4

Reference: Retrosheet, The 1897 Boston Beaneaters Game Log

Grouped by the previous season's final standings, New York (-) and Chicago (+) change places and the pattern in favor of Nichols is two games stronger.

Five times the Boston pitcher started on one day's rest (*): Nichols against the Baltimore champions, Klobedanz against New York, and Lewis three times against mediocre teams (*).

The team played 20% of its games in doubleheaders and relied more on its lesser pitchers in those games.

[Beady]

In order to assess how great these guys were we need their records against the rest of the league. I'd bet Kid faced a weaker opposition than Young.

I have counted up the numbers for 1896, maybe not the best year to pick because Boston was in their mid-decade trough and actually finished behind Cleveland.

Young started two games against Brooklyn, six each against Baltimore and Cincinnati and three to five against everybody else. Nichols started between three and five times against every opposing team. In more detail:

Young, 3 starts: Philadelphia, St. Louis, Washington
Young, 4 starts: NY, Pittsburgh
Young, 5 starts: Boston, Chicago, Louisville

Nichols, 3 starts: Cincinnati, Louisville, Pittsburgh
Nichols, 4 starts: Baltimore, Chicago, St. Louis, Washington, Brooklyn, Cleveland
Nichols , 5 starts: NY, Philadelphia

I don't see much of a pattern to be extracted from Nichols' record, but Young does seem to have been spotted against the tougher teams (although why Louisville?). If Tebeau did that with him in 1896, then quite likely he would have done it every year.

[brett]

Top 15 inferred from Top 100 Pitchers, tier 1
Young, Mathewson, Johnson, Alexander
Grove, Paige, Feller, Spahn
Gibson, Seaver, Carlton
Clemens, Maddux, Johnson, Martinez

I think Nichols is in the Gibson, Carlton, Pedro territory. I tended to slip Nichols ahead of Feller and Spahn in the bast, but I have separated Feller and Spahn from those 4 over the years, and moved them into my 9 and 10 slots (not including Paige who I could place anywhere in the top 10, but most safely in Feller/Spahn territory).

[Paul Wendt]

According to baseball prospectus, Nichols' defenses were significantly better than Young's. I don't know of any other measures of team defense. Does anyone else?

Win Shares. Batting/running, fielding, and pitching win shares for every major league team are published in the book.

For example, the 1892 Boston and Cleveland teams met in the National League championship series.

Win Shares, 1892 full season
Bat ; Field ; Pitch : Sum (= 3*wins)
122.4 49.2 134.4 : 306 Boston
122.7 49.5 106.7 : 279 Cleveland

Batting Win Shares follow simply from the allocation of wins between Batting/running and Fielding/pitching, that is between run-scoring and run-prevention, which should be uncontroversial. In the example, the shares allocated to batting/running are 40% of the Boston wins and 44% of the Cleveland wins, which come to the same 41 wins or ~123 win shares.

The allocation between fielding and pitching, within run-prevention, is one of the sharply criticized elements of the rating system. "Everyone" agrees that James gives too much credit to pitching, too little credit to fielding, for some early period that certainly extends through the careers of Nichols and Young.
(If the misjudgment is proportional for all contemporary teams, then the Boston fielders should be credited with some part of the 134.4 pitching win shares, and Cleveland fielders credited with the same fraction of 106.7 pitching win shares.)

[Bothrops Atrox]

Nichols had much better teams behind him. In that era, I gotta believe that the syndicate that got all the great players had both better offensive and better defensive teams. Clearly, the offensive part is true.

According to baseball prospectus, Nichols' defenses were significantly better than Young's. I don't know of any other measures of team defense. Does anyone else?

In order to assess how great these guys were we need their records against the rest of the league. I'd bet Kid faced a weaker opposition than Young. This reminds me of the discussion we had not long ago about the ridiculously bad record of the 1922-33 Red Sox, which strongly called into question the records of the superstars of that era, but more importantly, the quality of the league itself.

When I said defensive support being "in favor of Young" I meant that the poor defense behind him favors his ranking, not that he was favored by a better defense. Poor wording on my part.

[csh19792001]

In terms of comparing the importance of defensive support (in favor of Young) vs. offensive support (favoring Nichols).

Nichols had much better teams behind him. In that era, I gotta believe that the syndicate that got all the great players had both better offensive and better defensive teams. Clearly, the offensive part is true.

According to baseball prospectus, Nichols' defenses were significantly better than Young's. I don't know of any other measures of team defense. Does anyone else?

In order to assess how great these guys were we need their records against the rest of the league. I'd bet Kid faced a weaker opposition than Young. This reminds me of the discussion we had not long ago about the ridiculously bad record of the 1922-33 Red Sox, which strongly called into question the records of the superstars of that era, but more importantly, the quality of the league itself.

[Bothrops Atrox]

Have you looked at Nichols' run support compared with Young's?

RMB pointed out earlier in this thread that Nichols only got 94% of the league average run support. That ain't good.

How did Young fare in the 90s?

Run support certainly has a major impact on W-L record, but there are no studies that show any correlation between the amount of run support and amount of run prevention. Sure, if some guys have huge leads, they may relax and pitch better, but you always hear about the best "pitching their best" when the games are close/lots of pressure/low scoring, etc. In reality, very few pitchers have statistically significant differences in OPSA or ERA+, etc. between margin less than 4 games and margin greater than 4 games. Young and Mathewson may have gotten lots more run support, but if their ERA+ ( or FIP or whatever you want to use) is different (higher or lower), then run support is pretty much irrelevant. If one guy wins 8-0 and another wins 1-0, it is nearly impossible to prove that the 1-0 winner has pitched better.

This brings us to Nichols, Young, Mathewson, etc. I would only worry about run support if you are looking at W-L records. If not, why bother?

In terms of comparing the importance of defensive support (in favor of Young) vs. offensive support (favoring Nichols): if you are looking at W-L to compare the pitchers, then both made a huge impact. If you are looking to wards run prevention data for comparison, defensive support is monstrously more important. Especially in a low strikeout, high contact era. A horrible or great defense could completely make or break a pitcher.

[Buzzaldrin]

Have you looked at Nichols' run support compared with Young's?

RMB pointed out earlier in this thread that Nichols only got 94% of the league average run support. That ain't good.

How did Young fare in the 90s?

[csh19792001]

Behind only Young from 1890 to 1905.

The matchups between Cy Young and Nichols had really were legendary. Accordingly, Young had a ton of respect for Nichols, and many speculated that Kid's fastball was the equal of Amos Rusie, who by all accounts was Nolan Ryan pitching 80 years earlier.

My stance on Nichols changed drastically when I started to read more about syndicate ball of the 1890's. I was always one of his more ardent supporters back in the days when I used to vote/participate in polls and rankings.

In his 2000 Abstract Bill James asserted that Nichols was superior to Young during the 1890's. In coming to this conclusion, though, he neglected to take into account the incredible lack of parity in syndicate baseball. I wonder if Nichols was ACTUALLY better than Cy Young while they were in the same league at the same time.

Being on a dynasty...not only did he have superstars and some of the best defense teams in the league behind him year in and year out, but Nichols was also pitching a large percentage of his games against at 2-3 "second division" teams stock full of scrubs who weren't even remotely close to what we would consider major league caliber today.

The top teams won nearly 70% of the games in that decade, and the worst teams lost 75%. There were at least five teams in the 1890's that finished 50, 60, and 80 games out of first place. Dreyfuss and Robison-owners of two franchises each, put all their stars on one team and let the other atrophy into bankruptcy. The competitive balance was akin to little league.

Nichols' teams:

YEAR  PLACE   W    L   PCT   GB    TITLE
1890  5th     76   57  .571   12   
1891  1st     87   51  .630   +3.5   NL CHAMPIONS
1892  1st     52   22  .703   +2.5   NL CHAMPIONS
1892  2nd     50   26  .658    3   
1893  1st     86   43  .667   +5     NL CHAMPIONS
1894  3rd     83   49  .629    8   
1895  5th     71   60  .542   16.5 
1896  4th     74   57  .565   17   
1897  1st     93   39  .705   +2   
1898  1st    102   47  .685   +6     NL CHAMPIONS
1899  2nd     95   57  .625    8   
1900  4th     66   72  .478   17

7 times in a decade Kid's teams either led a 12 team league or finished second. Only one losing season and never on a bad team. Was he a huge part of this? Of course. But no more than Young contributed to his team's successes. They had extremely similar IP and ERA totals during these years.

Compare that with Young's teams:

YEAR  PLACE   W    L   PCT   GB    TITLE
1890  7th     44   88  .333   43.5 
1891  5th     65   74  .468   22.5 
1892  5th     40   33  .548   11.5 
1892  1st     53   23  .697   +3   
1893  3rd     73   55  .570   12.5 
1894  6th     68   61  .527   21.5 
1895  2nd     84   46  .646    3     WORLD CHAMPIONS
1896  2nd     80   48  .625    9.5 
1897  5th     69   62  .527   23.5 
1898  5th     81   68  .544   21
1899  5th     84   67  .556   18.5 
1900  5th     65   75  .464   19

Significantly different, I'd say.

The stats from baseball prospectus also suggest that Nichols' teams had significantly better defenses than Young's. With a much higher percentage of balls put in play then (BB and K were rare) and the incredibly high error rate and percentage of unearned runs, I think defense was significantly more important then than it is today.

I think Cy was actually a better pitcher during the 1890's.

Thoughts?