Yest, it's working now.

testing

This is only going to show what Welch and Keefe did in 1880-1883 and 1885-1889 the years they were teammates for complete seasons.

------------------------Tim Keefe---Mickey Welch
wins--------------------------213---------221
losses------------------------136---------143
winning percentage------------610---------607
games-------------------------364---------385
games started-----------------361---------379
complete games----------------346---------370
shutouts----------------------27----------31
batters faced-----------------12836-------13987
innings pitched---------------3092--------3361.3
hits--------------------------2724--------3173
runs--------------------------1507--------1720
earned runs-------------------864---------982
home runs---------------------48----------71
walks-------------------------732---------862
strikeouts--------------------1535--------1217
hit by pitches----------------41----------29
times on base-----------------3497--------4064
wild pitches------------------166---------193
wins per game-----------------.585--------574
wins per games started--------590---------583
shutouts per win--------------.127--------.140
shutouts per game-------------.074--------.081
shutouts per games started----.075--------.082
wins per shutout--------------7.89--------7.13
starts per shutout------------13.37-------12.23
games per shutout-------------13.48-------12.42
hits per game-----------------7.48--------8.24
hits per start----------------7.55--------8.37
hits per 9 innings------------7.93--------8.50
runs per game-----------------4.14--------4.47
runs per start----------------4.17--------4.54
RA----------------------------4.39--------4.61
earned runs per game----------2.37--------2.55
earned runs per start---------2.39--------2.59
ERA---------------------------2.51--------2.63
walks per game----------------2.01--------2.24
walks per start---------------2.03--------2.27
walks per 9 innings-----------2.13--------2.31
strikeouts per game-----------4.22--------3.16
strikeouts per start----------4.25--------3.21
strikeouts per 9 innings------4.47--------3.26
times on base per game--------9.61--------10.56
times on base per start-------9.69--------10.72
times on base per 9 innings---10.18-------10.88
complete games per game-------.951--------.961
complete games per start------.958--------.976
starts per game---------------.992--------.984
innings per game--------------8.49--------8.73
innings per start-------------8.57--------8.87
strikeouts per walk-----------2.10--------1.41
batters faced per hit---------4.71--------4.41
opponents batting avg.--------226---------242
batters faced per run---------8.52--------8.13
batters faced per earned run--14.86-------14.24
batters faced per walk--------17.54-------16.23
batters faced per strikeout---8.36--------11.49
batters per time on base------3.67--------3.44
opponents on base percentage--272---------291

This is also only going to show what Welch and Keefe did in 1880-1883 and 1885-1889 the years they were teammates for complete seasons.
Everything that follows is based on the retrosheet so it only shows games that Mickey Welch or Tim Keefe started whether he got the decision or not. So Welch has 3 more wins and 1 more loss and Keefe has 1 more win and 1 more loss and Welch has 19 extra runs while Keefe has 59 extra runs. Keep in mind Welch started 379 games completing 370 of them and Keefe started 361 games completing 346 of them in this period.
So according to the retrosheet Mickey Welch has
224 Wins 144 Losses 11 Ties and a 609 Winning Percentage
1739 runs allowed 4.59 runs per start
And Tim Keefe has
214 Wins 137 Losses 10 Ties and a 610 Winning Percentage
1566 runs allowed 4.34 runs per start


in games that Mickey Welch lost he
gave up 24 runs 1 time
gave up 21 runs 1 time
gave up 18 runs 1 time
gave up 17 runs 3 times
gave up 16 runs 1 time
gave up 15 runs 1 time
gave up 14 runs 3 times
gave up 13 runs 4 times
gave up 12 runs 6 times
gave up 11 runs 7 times
gave up 10 runs 7 times
gave up 9 runs 6 times
gave up 8 runs 11 times
gave up 7 runs 16 times
gave up 6 runs 14 times
gave up 5 runs 23 times
gave up 4 runs 21 times
gave up 3 runs 12 times
gave up 2 runs 4 times
gave up 1 run 2 times

in games that Mickey Welch won he
gave up 12 runs 2 times
gave up 10 runs 1 time
gave up 9 runs 1 time
gave up 8 runs 7 times
gave up 7 runs 13 times
gave up 6 runs 11 times
gave up 5 runs 16 times
gave up 4 runs 19 times
gave up 3 runs 38 times
gave up 2 runs 34 times
gave up 1 run 51 times
and no runs 31 times

in games that Mickey Welch tied he
tied at 9 runs 1 time
tied at 7 runs 1 time
tied at 6 runs 2 times
tied at 5 runs 3 times
tied at 4 runs 1 time
tied at 3 runs 1 time
tied at 2 runs 1 time
tied at 1 run 1 time


And in games that Tim Keefe lost he
gave up 15 runs 3 times
gave up 13 runs 3 times
gave up 12 runs 6 times
gave up 11 runs 4 times
gave up 10 runs 12 times
gave up 9 runs 13 times
gave up 8 runs 14 times
gave up 7 runs 16 times
gave up 6 runs 9 times
gave up 5 runs 17 times
gave up 4 runs 18 times
gave up 3 runs 8 times
gave up 2 runs 8 times
gave up 1 run 6 times

in games that Tim Keefe won he
gave up 13 runs 1 time
gave up 9 runs 2 times
gave up 8 runs 9 times
gave up 7 runs 6 times
gave up 6 runs 9 times
gave up 5 runs 11 times
gave up 4 runs 38 times
gave up 3 runs 27 times
gave up 2 runs 45 times
gave up 1 run 39 times
and no runs 27 times

in games that Tim Keefe tied he
tied at 6 runs 1 time
tied at 5 runs 1 time
tied at 4 runs 2 times
tied at 3 runs 4 times
tied at 1 run 1 time
tied at no runs 1 time

Here it is one more time this time comparing the two

in games that they lost they gave up
---------Tim Keefe-----Mickey Welch
24 runs---no times----------1 time
21 runs---no times----------1 time
18 runs---no times----------1 time
17 runs---no times----------3 times
16 runs---no times----------1 time
15 runs---3 times-----------1 time
14 runs---no times----------3 times
13 runs---3 times-----------4 times
12 runs---6 times-----------6 times
11 runs---4 times-----------7 times
10 runs---12 times----------7 times
9 runs----13 times----------6 times
8 runs----14 times----------11 times
7 runs----16 times----------16 times
6 runs----9 times-----------14 times
5 runs----17 times----------23 times
4 runs----18 times----------21 times
3 runs----8 times-----------12 times
2 runs----8 times-----------4 times
1 run-----6 times-----------2 times

in games that they won they gave up
---------Tim Keefe-----Mickey Welch
13 runs---1 time------------no times
12 runs---no times----------2 times
10 runs---no times----------1 time
9 runs----2 times-----------1 time
8 runs----9 times-----------7 times
7 runs----6 times-----------13 times
6 runs----9 times-----------11 times
5 runs----11 times----------16 times
4 runs----38 times----------19 times
3 runs----27 times----------38 times
2 runs----45 times----------34 times
1 run-----39 times----------51 times
no runs---27 times----------31 times

in ties they tied at
---------Tim Keefe-----Mickey Welch
9 runs----no times----------1 time
7 runs----no times----------1 time
6 runs----1 time------------2 times
5 runs----1 time------------3 times
4 runs----2 times-----------1 time
3 runs----4 times-----------1 time
2 runs----no times----------1 time
1 run-----1 time------------1 time
no runs---1 time------------no times

Now putting all decisions together they gave up
---------Tim Keefe-----Mickey Welch
24 runs---no times----------1 time
21 runs---no times----------1 time
18 runs---no times----------1 time
17 runs---no times----------3 times
16 runs---no times----------1 time
15 runs---3 times-----------1 time
14 runs---no times----------3 times
13 runs---4 times-----------4 times
12 runs---6 times-----------8 times
11 runs---4 times-----------7 times
10 runs---12 times----------8 times
9 runs----15 times----------7 times
8 runs----23 times----------18 times
7 runs----22 times----------29 times
6 runs----18 times----------25 times
5 runs----28 times----------39 times
4 runs----56 times----------40 times
3 runs----35 times----------50 times
2 runs----53 times----------38 times
1 run-----45 times----------53 times
no runs---27 times----------31 times

and in all games started (including ties) they gave up
---------Tim Keefe-----Mickey Welch
24 runs---no times----------1 time
21 runs---no times----------1 time
18 runs---no times----------1 time
17 runs---no times----------3 times
16 runs---no times----------1 time
15 runs---3 times-----------1 time
14 runs---no times----------3 times
13 runs---4 times-----------4 times
12 runs---6 times-----------8 times
11 runs---4 times-----------7 times
10 runs---12 times----------8 times
9 runs----15 times----------8 times
8 runs----23 times----------18 times
7 runs----22 times----------30 times
6 runs----19 times----------27 times
5 runs----29 times----------42 times
4 runs----58 times----------41 times
3 runs----39 times----------51 times
2 runs----53 times----------39 times
1 run-----46 times----------54 times
no runs---28 times----------31 times

I'm intrigued by how often Welch allowed at least 11 runs - 30 to 17 for Keefe.
Also, Welch allowed more than 15 runs SEVEN times, to none for Keefe. I'd suggest that those incremental runs may be meaningless. Is there a difference if the final score is 10-3 or 20-3?

I think this suggests, at least, that Welch's ERA+ (remember even the ERA stat didn't exist when he pitched) is 'off' by a few points. Maybe not enough to convince some voters, but does anyone think there may be something to this?

yest, great consolidation of info. I hope to make use of it in my 1939 ballot. Welch ain't getting in this week anyway, and his window won't run out any time soon.

There are two problems with the Keefe-Welch debate--i.e. with the hypothesis, basically, that Keefe and Welch are basically equivalent (hey, they both won 300 games and they were teammates for 9 years, so therefore I cannot tell them apart, they are equivalent.

1. All of the information, including yest's own info above, keeps showing that Keefe was better. Not by a lot, I'll grant, but better. And he was no Kid Nichols and not even a John Clarkson. I can't see where it is at all illogical that a line (HoM in/out) would be drawn between them.

OK, Keefe was a little more durable during the 9 years in question, (over those 9 years he threw about 1/2 years worth of extra innings) but c'mon, Keefe threw 5000 innings in his career, how non-durable is that? Besides, Welch threw 200 fewer innings for his entire career, which brings me to my second point.

2. The data above conveniently fails to note the following. Outside of their 9 years as teammates:

Keefe 129-89, ~2000 IP, 236 G, 233 GS, 208 CG, 12 ShO
Welch 86-67, ~1500 IP, 180 G, 170 GS, 155 CG, 10 ShO

So again, Keefe was better. OK, two years in the AA (during a time when the AA was improving rapidly; Keefe went 78-44, ERA+ 138 those two years, and then the very next year went 32-13, 170 in the NL).

And then in 1890, Keefe was 33, Welch 31. Keefe pitched in the better league, the PL, at age 33, and went 17-11, 138, while Welch at age 31 in the weaker league went 17-14, 117.

So in summary:

1. The hypothesis that Keefe and Welch were equivalent as teammates is not proven.

2. It ignores the fact that Keefe was clearly better when they were not teammates.

3. Keefe himself was no Kid Nichols, John Clarkson, Walter Johnson, Pete Alexander or even Ed Walsh, and drawing a line not too far below Tim Keefe (and therefore above Mickey Welch) is not unreasonable.

4. They were not actually teammates in 1883 either, but I have not tried to go any further with that.

5. Re. the idea that it is unfair to evaluate Keefe and Welch on the basis of ERA, all i can say to that is Win Shares and WARP. Maybe I can interest you in Ed Williamson's 27 HR.

Well said, Sunny Day 2. My guess is that, when this is all said and done, Keefe will rank as about the 175th best player in baseball history (maybe the 45th or 50th pitcher) and Welch will rank as about the 250th best (maybe the 60th or 65th pitcher). It just so happens that those 75 places are the difference between election and perpetual also-ran status.

Reposting some Welch comparison information of my own. Check the 3rd page of the Pitchers thread for info on more pitchers...

HoMers Against Other HoMers and Welch:

Radbourn         Clarkson        Keefe
Clarkson  9-8    Radbourn  8-9   Clarkson  9-10
Galvin   16-14   Galvin    7-5   Radbourn 14-10
Keefe    10-14   Keefe    10-9   Galvin    8-6
Ward      3-1    Rusie     3-4   Ward      3-7
Rusie     2-0    Caruthers 2-1   Rusie     5-2
Caruthers 0-1    Young     1-1   Young     0-2
Young     1-0    Nichols   0-1   Nichols   3-3
Nichols   0-1
total    41-39   total    31-30  total    42-40

Welch    10-17   Welch     7-13

Welch            Galvin  
Galvin   26-11   Ward      8-12
Radbourn 17-10   Radbourn 14-16
Ward      4-7    Keefe     6-10
Clarkson 13-7    Clarkson  5-7
Rusie     2-0    Caruthers 4-4
Nichols   0-1    Rusie     4-2
Caruthers 0-2    Young     1-1
total    62-38   total:   42-52
                 Welch    11-26 *
*Sometimes I only get 9 wins for Pud

Records by opponent position

finish  Clarkson Keefe Radbourn Welch Galvin
1st      26-20   30-38   27-35  22-30  29-58
         .565    .441    .435   .423   .333
2nd      32-32   29-44   25-29  27-29  34-55
         .500    .397    .463   .409   .382
3rd      24-23   33-21   36-27  48-23  36-29
         .511    .611    .571   .676   .554
4th      46-20   34-24   32-31  34-25  42-43
         .697    .586    .508   .576   .495
5th      42-17   39-24   34-18  32-30  44-36
         .712    .619    .654   .516   .550
6th      41-19   54-23   46-20  39-20  52-31
         .683    .701    .697   .661   .627
7th      41-22   52-20   52-17  51-21  49-24
         .651    .722    .754   .708   .671
8th      54-11   50-25   51-14  59-14  65-26
         .831    .667    .785   .808   .714
9-12     24-10   27-9                   6-5
         .706    .750                  .545  

The following numbers are from Chris J.'s site:
Run Support Index:
John Clarkson 109.46
Tim Keefe 107.16
Ol' Hoss Radbourn 106.83
Mickey Welch 102.79
Pud Galvin 102.11

Defensive Support Index:
John Clarkson +29.9
Radbourn +15.7
Tim Keefe +15.1
Mickey Welch +5.4
Pud Galvin -7.3
Clarkson, Radbourn, and Keefe are in the top 12 all-time, while Welch is #84. Galvin's is one of the worst all-time.

In terms of over/under achieving based on the support they have been given:
Mickey Welch overachieved by 13 wins
John Clarkson underachieved by 2
Ol'Hoss Radbourn underachieved by 5
Tim Keefe underachieved by 18.

Interesting stuff, Kelly and yest. But once again, the meaning is hardly clear enough to offset the overall impression of ERA+ in ordering Keefe ahead of Welch (IMO).

E.g. Welch pitched more against the HoMers Clarkson, Radbourn and Galvin--37 times against Galvin, e.g., vs. Keefe's mere 14 times. I remember somebody once suggesting that the Giants manager regarded Welch as hit #1 guy because he matched him up against these pitchers more often.

But wait! Keefe started 141 times against 1st and 2nd place teams, Welch just 108. We don't know when these match-ups occurred, of course (nor the ones against Clarkson, Radbourn and Galvin). But we do know that Galvin pitched for some miserable teams. Maybe the Giants matched Welch against Galvin because Welch was NOT the #1 guy?

And Keefe's underachieving could just as easily be a result of those extra 33 starts against #1 and #2 teams, and chance, as Welch's alleged "pitching in a pinch."

And if Welch "rose to the occasion" against Pud, Hoss and Clarkson, why was his W-L 100 points lower than Keefe's against 5th place teams? This seems like the flip side of pitching in a pinch--i.e. losing your concentration in games you coulda/shoulda won.

So all the extra info. still fails to override the big picture, which was that Keefe was better. Add to that those 3.5 extra starts against 1st and 2nd place teams per year (for Keefe), and the extra 3 starts per year (for Welch) against Galvin and his lousy teams. Those two factors seem to reinforce that the Giants actually saw Keefe as the better of the two. Rather than Keefe underachieving his support, maybe it was Welch who underachieved considering his opposition.

The fact is that Keefe won 33 more games than Welch. If we were talking 50 years later and 100 fewer wins (242 wins versus 208) and if the guy with 242 had a better ERA+, there would be no questioning who was better. Baseball in the 1880s was still baseball. The basic logic seems to me to be the same.

Whether he was better than Keefe is irrelevant (he was, just a little, in my view, but it's close.) However, we elected Keefe almost 40 years ago, with very little delay, obviously regarding him as far above #175, more like #100.

The important thing to remember is that ERA+ is irrelevant for an 1880s pitcher, because they didn't calculate it and left starters in in blowouts. Thus Welch is a LOT better than these 20s bozos, including Covaleski, having almost 100 more wins (and indeed as good an ERA+ as several of them.) He's no Alex or Grove, but he and not Covaleski, let alone the overrated Vance, should be next in the HOM

And if Mullane's best year hadn't been in the 1882 AA, I'd argue for him too.

Frankly, I think that Keefe got somewhat of a free-pass in 1900-01. Many of the tools we use to evaluate pitchers now were not developed then; neither was the Retrosheet data available then on match-ups and game scores. I'm not saying Keefe wouldn't be elected if we did it over again, just that he would have had to undergo the same kind of intense scrutiny as Radbourn/McGinnity/etc.

Pitching Rank

      by Win Shares     by WARP
      Keefe   Welch     Keefe   Welch
1880    -     T 4th       -       6th
1881   11th   T 8th      12th     8th
1882   12th    20th       5th       !
1883    1st*   12th       4th     9th
1884   10th*   11th      13th     5th
1885  T 5th     2nd      15th     8th
1886    9th   T16th       8th    25th
1887    6th   T14th       5th    13th
1888  T 6th   T11th       7th    10th
1889  T14th     8th      14th    12th
1890  T35th   T35th      24th    25th

* in AA, unadjusted
! below replacement level

Neither were statistical standouts for a single seasons (no "Cy Young" awards, other than Keefe's 1883 AA, a league still not quite major in quality); they were both solid inning-eaters of some quality, 19th century Don Suttons, perhaps.

Lets calculate NHL stype All-Star teams (1st team, 2nd team), and use the following table to determine # pitchers for the team:

1880 1 pitcher, 1881-5 2 pitchers, 1886-97 3 pitchers

Each pitcher has 1 Win-Shares 1st team selection; Keefe also has an additional 2nd team selection. WARP says that Keefe 1883 was only good for 2nd team due to AA discount, and agrees with the 1887 2nd team selection. It bypasses Welch for 1885, presumably because the Giants had a great defense that year.

Neither do well by this standard, Keefe having only two selections, and Welch having only one controversial selection.

<i>I'm intrigued by how often Welch allowed at least 11 runs - 30 to 17 for Keefe.
Also, Welch allowed more than 15 runs SEVEN times, to none for Keefe. I'd suggest that those incremental runs may be meaningless. Is there a difference if the final score is 10-3 or 20-3?

I think this suggests, at least, that Welch's ERA+ (remember even the ERA stat didn't exist when he pitched) is 'off' by a few points. Maybe not enough to convince some voters, but does anyone think there may be something to this?<i>

If find it interesting that Welch's candidacy is being promoted because he gave up MORE runs!

Unless Keefe knew before the game started how many runs he team would score, how can giving up a lot of runs be considered meaningless?

Also, good pitchers generally DON'T give up huge numbers of runs in a game.

This argument is almost like arguing that if you remove all the losses of 3 or more runs, the Devil Rays would have almost the same record as the Red Sox...

How do their contemporaries do by this All-Star standard?

Spalding: WARP (5) 71B, 72B, 73A, 74A, 75A; WS (1) 76A (no NA 71-75)
Bond: WARP (4) 75B, 76B, 77B, 79A; WS (2) 77B, 78A (no NA 75)
McCormick: WARP (2) 81B, 82A; WS (3) 80A, 81B, 82B
Galvin: WARP (3) 81A, 83B, 84A; WS (2) 79A, 81A
Radbourn: WARP (4) 82A, 83A, 84B, 85B; WS (3) 82A, 83A, 84A
Whitney: WARP (4) 81A, 82B, 83A, 87B; WS (3) 81A, 82B, 83B
Buffinton: WARP (3) 84A, 88A, 89A; WS (2) 84B*, 88A
Clarkson: WARP (4) 85A, 87A, 89A, 91B; WS (4) 85A, 87A, 89A, 91A
Caruthers: WARP (4) 86A, 87A, 88B, 89B; WS (5!) 85B, 86A, 88A, 89A
Mullane: WARP (0); WS (3) 83B, 84B*, 87B

* ignoring UA pitchers
! WS selects Caruthers as 1st-team RF in 1887, Foutz as 2nd-team RF;
if reclassified as pitchers, PBC is the #1 pitcher; Keefe is bumped from #6 to #8, off the team

Note: I missed Keefe's selection to the WS second team in 1888, disputed by WARP.

"Unless Keefe knew before the game started how many runs he team would score, how can giving up a lot of runs be considered meaningless?"

KJOK, if it's 15-1 in the 6th inning, is it fair to say that Welch (or Keefe) know who's going to win?
Once a pitcher gives up 10 runs in a lopsided game, what difference does it make (other than in ERA, which hadn't been invented) how many more he allows? These guys pitched the whole game regardless.
Welch obviously gets no benefit from allowing extra runs - check the annual balloting. The question is, are a decent number of those runs irrelevant?

In the late 1880's, incomplete games were as much of a novelty as complete games are now. Starters were relieved about 4-5% of the time. This was partly a function of the rule book (no substitutions without the opponent's permission, except for obvious injury), and partly a function of strategy. In the 1870's, many teams played a second pitcher in RF as a backup. This could cost you a bat from the lineup, so some teams would forego the backup pitcher and play a hitter; the Giants became such a team with Tiernan. The overall league relief rate depended on how many teams went with the pitcher in RF vs how many had a slugger; this rate was actually higher in the 1870's (8-9%) due to more teams using the backup pitcher strategy then. (Note that the ideal was to have a pitcher that could also slug; Caruthers & Whitney were a manager's dream come true.)

This all changed when the substitution rule was changed in 1890; the relief rate peaked at 19% in 1895, subsided, and then went back up and up - 19% again in 1905, 48% in 1913, 51% in 1922, 60% in 1936, 69% in 1954, 80% in 1965, 96% in 2002.

Welch had 24 incomplete games compared to an expected 37. The differences occur in his rookie season of 1880 (7), and from 1885-1889 (7), when he was relieved a less than typical amount.

If your team didn't use the backup pitcher strategy (and the other team had no mercy), then, as a pitcher, you took your lumps when you didn't have it that day. Apparently, Welch took more than most.

KJOK, if it's 15-1 in the 6th inning, is it fair to say that Welch (or Keefe) know who's going to win?

Well, if he's the one who gave up the 15 runs, then he's not demonstrating he's a good pitcher, regardless of if his team scored 1 or 14 runs. Yes, we know who's going to win, but it's because of the pitchers lack of success that the team is in that "no win" situation in the first place.

KJOK, now we're getting somewhere!
RIGHT that Welch deserves the loss.
And right that Welch's record should reflect his poor pitching that day.

But the key is: I'm not saying AT ALL to throw out those games where he gave up lots of runs. I'm talking about the law of diminishing returns here, and saying that we don't want to overemphasize irrelevant numbers, either.

Roughly speaking, there is a certain point where the incremental additional runs allowed are utterly irrelevant. The game is decided, and someone is mopping up.
For the Giants, that guy is Welch, not a substitute.
Is it such a stretch in that circumstance - 15-1 in the 6th - to imagine that he's not caring whether the opponents score more runs? Why strain your arm in a lost cause? And there's no ERA in existence; all that matters is wins and losses.

It's like doing Pythagorean runs scored, and having the Yankees lose a 22-0 game (as they did this year). That mucks up their seasonal record slightly.
Imagine if they lost three such games.
You would then have serious people claiming that the Yankees were "overachievers" this year, simply because the 66-0 differential skewed things dramatically more than three 6-0 losses.
But those people would be missing something. I think something similar is happening with Welch.

(This is the THIRD time I've attempted to post this....computer frustration mounts....)

Using yest's last set of #s in post 5, I calculated the diffference between runs allowed and expected winning percentage, given the different distributions of runs theat Keefe and Welch allowed as teammates. Keefe allowed 4.34 runs/g in 361 starts. Welch allowed 4.59 r/g in 379 starts. By pythag, one would expect Welch's wpct to be .027 lower (4.5^2/(4.5^2+[4.5-.25]^2). But using a reasonable-man estimate of win pct by runs allowed per game, I find Welch's expected wpct to be only .008 lower - a gain of .019. This might equate to 10 wins over his career, IF we assume he 'pitched to the score' in such a manner compared to a typical pitcher. This is the upper end of credit I give Mickey above his ERA+/WS/WARP numebrs, which does move him onto my ballot.

I also believe those Welch fans who credit him more on his Wins than his ERA should be voting for Clark Griffith - it's hard for me to fathom how Griffith can be behind Welch, given that we have both anecdotal AND empirical evidence that Griffith indeed 'pitched to the score'.

Judging by complete games and W-L decisions alone, yest correctly included only 1880-82, 85-89 (not 1883).

--
[as teammates]
--------------------Tim Keefe---Mickey Welch
W-L decision-----------349-----364
CG/decision----------0.991---1.016

--
[career]
Catching Tim Keefe and Mickey Welch - Starts & Shutouts
Walt Wilson, Baseball Records Cmte newsletter, Feb 2002

TIM KEEFE
GS ShO
594 39 career
185 13 BILL HOLBERT
164 15 BUCK EWING
_60 _1 JIM O'ROURKE
_50 _2 Jack Clements
_41 _5 Pat Deasley
_40 _2 WILLARD BROWN
_14 _0 Lave Cross
__9 _0 Charlie Reipschlager
__7 _0 PAT MURPHY
_24 _1 ten others (inclg one shared shutout)

MICKEY WELCH
GS ShO
549 41 career
277 28 BUCK EWING
_73 _5 BILL HOLBERT
_42 _2 PAT DEASLEY
_32 _1 JIM O'ROURKE
_20 _1 WILLARD BROWN
_20 _0 PAT MURPHY
_18 _1 Artie Clarke
_16 _0 John Humphries
_14 _0 Dick Buckley
_37 _2 fifteen others (inclg one shared shutout)

UPPERCASE = catchers who appear on both lists

On the Keefe list, "Pat Deasley" should be uppercase

HoMey for #23, great post.

This seems like the clincher in the Keefe/Welch debate. Buck was da MAN. Holbert was strictly a defensive catcher (47 OPS+), yet Buck looks better defensively, judging by the shutout percentages. (O'Rourke, ugh, not a catcher) And Welch enjoyed Buck behind the plate a LOT more than Keefe. This enabled Welch to appear almost as good as Keefe, having Ewing's superior defense and bat in the lineup.

HoMey for #23, great post.

I have to agree. Paul definitely knocked some of the steam out of the Welch/Keefe comparisons.

I'll still have Welch in the top ten though because I wasn't really buying the arguments here anyway.

I don't know why we need to compare Welch to Keefe, who went in easily, rather than against the 20s bozos, who are clearly inferior. If you knock out blowouts (more than 10 run differentials) I bet Welch's ERA+ is as good as Coveleski's, and he won a lot more games.

However, on Welch/Keefe, the fairly small differences in catcher are NOT the end of the story -- for one thing, Ewing caught less than half his games, which suggests he WASN'T a great defensive catcher, and may well have been in the lineup for Keefe starts when he wasn't catching.

In any case, you also need to look at hitting, if we're being picky; Keefe had 390 hits, 134 RBI and an OPS+ of 58, whereas Welch had 492 hits, 202 RBI and an OPS+ of 68. Clear advantage to Welch, who IMHO was marginally better than Keefe (better peak, for one thing, in 1885), is significantly better than Griffith and Cicotte, a LOT better than Coveleski and the 20s mob, and should have been in the HOM 30 "years" ago.

There seems to be some confusion here.
Welch doesn't have to be as good as Keefe to get in, not at all.

Also, I used to be just about Griffith's top supporters, and I continue to vote for him every year. He and Welch figure to be near the top of this weak ballot.

Griffith vs Covaleski is an interesting matchup, no?

I don't know why we need to compare Welch to Keefe, who went in easily, rather than against the 20s bozos, who are clearly inferior.

karlmagnus, would you say milk is better today because it's more expensive than it was during the Great Depression? I'm appealing to the economist in you that there are reasons why Welch (who I support) could do what he could do during the 1880s and why Coveleski could do what he could do during the twenties. You're allowing statistical illusions to cloud your judgment on this subject, IMO.

Milk was MUCH better in the Great Depression; it had no disgusting chemicals and preservatives in it, and there were no health faddists making you drink the skim stuff instead of cream you could stand a spoon up in!

I agree we should avoid statistical illusions, which is why one adjusts for changes in league ERA, for example. 1880s pitchers wore their arms down less, so could pitch more innings per annum; on the other hand they were more overused, so their arms tended to fall off. Indeed, ERA+ in an era that never heard of the ERA stat is precisely such a statistical illusion, because of there being few or no relief pitchers in blowouts.

However, we must also guard against ignoring the possibility that random flcutuations in pitcher quality from period to period are very substantial, because at the top HOM end of the quality scale there are no large numbers to balance out. The fact that NONE of the 20s pitchers got anywhere close to 300 wins, while the late careers of Johnson and Alex were going well beyond that mark, and Grove managed it in the 30s with a career truncated 5 years at the beginning, and that their ERA+ figures are uniformly uninspiring, compared to either earlier or later pitchers (and ERA+ corrects for the hitting increase) suggests that the crop of great pitchers was unusually thin in the 1920s.

We do not therefore need to elect half a dozen 1920s pitchers to the HOM; if we are worried, as we should be, that we don't have enough pitchers, we should go back and elect Welch, Griffith and possibly Cicotte, or reach across to the Negro Leagues and elect Bullet Joe Rogan. IMHO there are a subsantial number of better candidiates we could be electing in 1938 than Coveleski, yet we seem poised to push him in in spite of his numbers being significantly inferior to those of other ballot-eligible pitchers. Coveleski may well be the best of a mediocre 20s bunch, but we don't need a 20s pitcher quota in the HOM, and his inclusion may block either Welch from the 1880s, or, via a long causal chain, say Blyleven from the 1970s/80s, both of whom were much better pitchers.

Both WS and WARP agree that Ewing was the best defensive catcher in the majors in 1889 (most FWS and most FRAR). He could also hit, very well. Cumulative WARP fielding rating of 120 at catcher.

Both WS and WARP agree that Holbert was the best defensive catcher in the majors in 1883 (played with Keefe in AA). WARP also gives him that title in 1882 and 1880 (in Troy with both pitchers). He couldn't hit a lick. Cumulative WARP fielding rating of 128 at catcher.

The difference defensively is probably minimal, maybe slightly in Holbert's favor, maybe dependent on the distribution of starts over the years. There's no disputing that having Ewing catch helped the run support.

I'm intrigued by how often Welch allowed at least 11 runs - 30 to 17 for Keefe.
Also, Welch allowed more than 15 runs SEVEN times, to none for Keefe. I'd suggest that those incremental runs may be meaningless. Is there a difference if the final score is 10-3 or 20-3?


Here are scores from all their losses during those years when giving up 13 runs
Welch
1-24
9-21
3-18
1-17
2-17
6-17
8-16
10-15
7-14
4-14
7-14
3-13
1-13
7-13
8-13
Keefe
3-15
6-15
9-15
11-13
3-13
9-13

4. They were not actually teammates in 1883 either, but I have not tried to go any further with that.
that was a typo it should read 1882

IMHO there are a subsantial number of better candidiates we could be electing in 1938 than Coveleski, yet we seem poised to push him in in spite of his numbers being significantly inferior to those of other ballot-eligible pitchers.

Since Coveleski destroys Welch in black and grey ink and does much better than Griffith and Cicotte in that regard, I don't know why you don't think that Coveleski is at least in the ballpark with those three.

Milk was MUCH better in the Great Depression; it had no disgusting chemicals and preservatives in it, and there were no health faddists making you drink the skim stuff instead of cream you could stand a spoon up in!

I think you knew what I meant. :-)

If you have a generation where there are no great pitchers, the remaining mediocrities will rack up black and grey ink. Doesn't prove anything -- you have to look at W/L, ERA+ and by all means strikeouts etc.

If you have a generation where there are no great pitchers, the remaining mediocrities will rack up black and grey ink. Doesn't prove anything -- you have to look at W/L, ERA+ and by all means strikeouts etc.

You're going to hard pressed trying to persuade people here or elsewhere that the talent pool for pitchers was greater during the 1880s than it was during the twenties.

Some thoughts on Mickey Welch...

[Please note that all figures used in this message are my adjusted ones. IP, PRAR1 and WS are all adjusted for games played (to the 2/3 power based on a 162 game season), and WS are adjusted for league quality at the rate of 1/2 BP's penalty. When in doubt, split the difference, right? As well, all ratios are based on white major leagues only; I wasn't going through an entire process of recreating decades of blackball statistics for this.]

Our HoM, in current form, contains a very large chunk of the pitchers from the first 20 years of truly organanised ball, and an even greater proportion of the quality pitchers. This proportion drops steadily (with one exception, which I'll get back to) through the mid-'10s, when it falls off the cliff.

Hall of Meriters as a % of Total…
% of.....IP....PRAR1....PWS

71-75...10.3....93.6...12.5 *Created WS based on PRAR, actual team wins
76-80...13.2....75.7...17.1
81-85...14.8....53.0...20.8
86-90...11.1....21.1...13.8
91-95...12.2....23.4...12.7
96-00...07.1....13.7...06.7
01-05...08.5....16.7...09.3
06-10...08.8....16.0...10.9
11-15...06.4....12.0...09.5
16-20...03.2....06.7...04.9

It seems that we are very well established with pitchers (and good ones) from the early part of the 1880s. Let's add Mr. Welch to the above data, for kicks.

Hall of Meriters (plus Mickey Welch) as a % of Total…
% of.....IP....PRAR1....PWS

71-75...10.3....93.6...12.5 * Created WS based on PRAR, actual team wins
76-80...13.2....75.7...17.1
81-85...18.1....59.8...25.0
86-90...13.0....24.1...15.8
91-95...12.2....23.4...12.7
96-00...07.1....13.7...06.7
01-05...08.5....16.7...09.3
06-10...08.8....16.0...10.9
11-15...06.4....12.0...09.5
16-20...03.2....06.7...04.9

25%. One quarter of all Win Shares earned by pitchers between 1881 and 1885 would be enshrined. Honestly, now.

From that decade, we have Charles Radbourn and Pud Galvin, who were pretty effective for the whole decade, plus the front end had a dominant Johnny Ward. The back half of the ‘80s saw the emergence of the best pitcher yet in John Clarkson, and the red-hot peak of Bob Caruthers. In addition, Tim Keefe was around. As has been mentioned, he’ll not be listed in the greatest pitchers of all-time, but seems to be by all accounts better than Welch.

To add some more fun to this list, let’s add Jim McCormick, who during Welch’s best 6 years (his first 6) pitched more innings (4000 to 3700) at a better rate (400 PRAR to 250) and collected more Win Shares (147 to 133). What we can say is at no time during Welch’s career could he be considered the 2nd or 3rd or maybe even 4th or 5th best pitcher in a 16 team environment. If you believe the above paragraph, he was the 6th best pitcher at any one time (time meaning 3 –5 year stretch) during the 1880s. That might make him more qualified than the 4th best of another decade in some people’s minds, but certainly not mine.

I’m hoping to draw your attention to the 2 potential talent lapses in the pitching wing of our little club. First, the back half of the 1910s. We currently have pitchers who represent 3% of the innings, 7% of the PRAR and 5% of the Win Shares collected from 1916-1920. Your response might be that, “A lot of those pitchers are still going, so their not eligible. And the ones that are weren’t really that good, anyways. It’ll look a lot better after we elect Coveleski.”

Things will look a little more balanced after Coveleski’s enshrined, no doubt. However, if we were to attempt to match the 25% of Win Shares earned during this period, we’d need some more. For example, Eppa Rixey will probably be a strong candidate, as will at least a couple of Dolf Luque, Red Faber and Burleigh Grimes. All of those guys were active during the teens, and so was that Boston lefty who should sneak by.

To get to 25% of all Win Shares we’d need to elect all of them. Not some, All.
And Eddie Cicotte.

And Urban Shocker and Carl Mays.

And Hippo Vaughn, Jim Bagby and Wilbur Cooper.



And Bob Shawkey.

That leaves us a handful of WS short, but we’ll call it even.

As for our other lull, in the late ‘90s, we’ve elected the best three. Young, Nichols and to some extent Rusie are no-brainer, inner-circle guys. The 4th best has been sitting patiently for 35 years now. Clark Griffith has a much better case for election than Mickey Welch. As mentioned above, he even has a better case for “pitching to the score” than Welch, and Welch is getting the benefit of the doubt on that one. Hell, if Jim McCormick was better than Welch at his best, how can this be a question?


I by no means think that these ratios that I’m referring to need to be constant. The game changes, talent ebbs and flows, what have you. Let’s just realize that there’s at least Coveleski and Griffith, and possibly Cicotte or Shocker who should have the kind of mobilized support that Welch is getting.

Re McCormick vs Welch:
I'll never understand why anyone would give credit for the 21-3 and other ridiculous stats of McCormick in 1884 with the UA. It's a minor league, and not nearly the best one of the 19th century. Sure, it's part of McCormick's career, but why would we ignore the fact that it came against guys (aside from Dunlap and a couple of other examples) who hardly knew which end of the bat to use, relative to real major leaguers? Either give proportional credit for other players' minor-league standout seasons, or drop it. And if that seems extreme, it's no moreso than pretending McCormick accomplished those things in the NL or even UA.

As for percentage of innings, it's simply the case that pitchers had a major, major influence on the game back then. The best guys pitched damn near all the innings. Seems silly to me to compare that to an era 40 years later, when the best guys pitched far fewer innings. How much influence can you have pitching 220 good innings, compared to 400?
Also, for those of us who would not have elected Caruthers, are we less allowed to vote for Welch because of it? Seems like a double whammy.

No offense intended; just making the other case.

The 25% figure is a silly comparison. Top pitchers pitched a much greater percentage of their team's games than in the 1920s, so naturally an even weighting of top pitchers between the two eras will have a higher percentage of innings pitched in the 1880s.

As far as quality is concerned, the gap between Welch and his HOM contemporaries is very small (possibly negative in the case of Keefe and indeed Galvin); the gap between Welch and the lesser picthers of the 1880s is much more substantial, when looked at closely. There is a clearly defined set of "top 1880s pitchers" which does include Welch and doesn't include Mullane or McCormick.

That's not true in the 1920s (or rather it is, but the set contains only Johnson, Alexander and Grove.) The next group of 1920s pitchers form an amorphous mass, no one standing out, and no clear line possible to be drawn between HOM-worthy and not HOM-worthy. If the top pitchers stand out so little from their contemporaries, it's an indication that they weren't all that good.

I agree with Max

Howie,

I'm sorry, I should have clarified. There's no UA data in my system. McCormick's results are AA and NL only.

MP

A big part of the difference in Keefe & Welch workloads with Holbert & Ewing behind the plate is 1883-84, not included in the "as teammates" data presented by yest far above. Keefe & Holbert then played for the Mets, Welch & Ewing for the not-yet-Giants.

Catching Tim Keefe and Mickey Welch - Starts & Shutouts
(This table is not transcribed as in #23 but derived from
Walt Wilson, Baseball Records Cmte newsletter, Feb 2002.)

GS ShO - for pitcher starts by Keefe & Welch

441 43 BUCK EWING
258 18 BILL HOLBERT
_92 _2 JIM O'ROURKE
_83 _7 PAT DEASLEY
_60 _3 WILLARD BROWN
_50 _2 Jack Clements ( all Keefe)
_27 _0 PAT MURPHY
_18 _1 Artie Clarke ( all Welch)
_16 _0 John Humphries ( Welch)
_14 _1 Dick Buckley ( Welch - ShO 0 is a clerical error in #23)
_14 _0 Lave Cross ( Keefe)
__9 _0 Charlie Reipschlager ( Keefe)
_61 _3 ten others ( inclg two shared shutouts)


Catcher Games Played - starts and substitutions

538 BILL HOLBERT
_93 1876-79
159 1880-82 Troy
127 1883-84 NY Mets
159 1885-88

636 BUCK EWING
_79 1880-82 Troy
143 1883-84 NY Giants
296 1885-89 NY Giants
117 1890-92 NY PL/NL
__1 1893-97 Cle, Cin (OF 164, 1B 176, 2B 1, C 1)

(Cliff Blau http://mysite.verizon.net/brak2.0/ has written a few articles on the Mets. As Holbert's Baseball-Reference sponsor he writes, "One of the best defenseive catchers of his time. Holds the ML record for most ABs in a career without a homer." Tom Hanrahan is Buck Ewing's sponsor.)

Paging thru my archive . . . here is a list of pitchers covered in the series by Walt Wilson, Baseball Records Cmte newsletter, Oct 2001 - Apr 2004, bimonthly.

Hoyt, Pennock
?? missing
Keefe, Welch
Vance, Grimes
-- none
Nichols, Willis
Brown, Walsh
-- none
Waddell, Joss
Ryan, Sutton
Johnson, Mathewson
Lemon, Wynn
Clarkson
-- none
Spahn, Roberts
Faber, Rixey

Among those, a few worked more than halftime with a single catcher: Faber & Schalk, Lemon & Hegan, Welch & Ewing, Vance & DeBerry, Walsh & Sullivan, Waddell & Schreckengost - led by three above 60%
Lemon-Hegan 266/350
Vance-DeBerry 245/349
Faber-Schalk 307/483

"Howie,
I'm sorry, I should have clarified. There's no UA data in my system. McCormick's results are AA and NL only."

Aw, way to take a wind out of my sails! ;)

And with a polite response to make me feel even worse, lol...

karl,

Referring to a pretty basic thumbnail analysis as silly isn't exactly pushing the discussion forward. Although the tenor of my entire post may have disguised this small piece, allow me to quote myself:

"I by no means think that these ratios that I’m referring to need to be constant. The game changes, talent ebbs and flows, what have you."

Howie (and to a lesser extent karl),

I think that you've got the flip side of this debate framed perfectly. Well-structured and factual arguments are what will make me change my mind - both in this project, as well as in general.

To your point re: Amount of IP by key pitchers (i.e. the type we'd elect)...

For the two 5-year periods, 1881-85 and 1916-20, I compared the "quality" of ace pitchers, or more properly #1 starters. Specifically, I measured the PWS earned by the "best" pitcher on each team in a given season as a percentage of all PWS earned that year. For instance, in 1881 Jim Whitney had 28.1 PWS for Boston, Galvin had 27 for Buffalo, McCormick had 31.3 for Cleveland, etc. The 1881 NL had 270 PWS, and the sum of the team leaders (my proxy for Aces, or HoM candidates, whathaveyou...) was 180.6, or 67% of the total.

1881 67%
1882 64%
1883 61%
1884 61%
1885 54%
Total 60%


1916 28%
1917 28%
1918 31%
1919 29%
1920 28%
Total 29%

In words, Top pitchers were accumulating half the ratio of WS in the 1910s as they were in the '80s. The numbers prove your contention.

Based on that result, if we can infer straight-line value transfer (and I'm not convinced that we can, but I'll roll with it for now), for lack of a better term, the better comparison between the '80s and the '10s isn't the "silly" 25% (Total % of WS by HoMers) to 25%, but 25% to 29/60*25%, or 12.5%.

Recall that that figure is currently 4.9%, so there's still some addition to do, but under these assumptions, the "gap" in HoM pitchers from the late teens is closed simply with the more popular candidates plus Cicotte. A much more reasonable comparison, however do any of us feel that all of those pitchers will be elected?

I continue to stand by my basic point. That point is not that Welch should never be elected, which may have gotten lost amid my soapboxing. Rather, it is that Griffith and Coveleski should get the attention before Welch. Let's save the energy of a "Best X not already enshrined" campaign for Welch until that time when he actually is said "Best".

I think both of Max's posts are well-argued and supported.

I hope (and expect, really) that the electorate will choose a sufficient number of pitchers for the HoM in the next 5 elections to bring the late teens up to the 12.5% level that Max suggests as the neighborhood for reasonable representation of pitchers.

Myself, I support Coveleski, Griffith, and Welch in that order. I see Welch as a HoMer, but I agree 100% that Coveleski and Griffith and some of the other pitchers who’ll be becoming eligible in the next 3 years should go in first.

Sorry for "silly" -- I usually try to be more diplomatic but don't always succeed. :-((
Incidentally, whoever said "Why is there no claque for Cicotte? -- trust me, there will be; I think I'm the best FOEC, or up there anyway :-))

Karl,

No offense taken whatsoever. In hindsight, I probably shouldn't have even mentioned it. Already gone.

MP

Max Parkinson #44
posted the league portion of Pitching Win Shares reaped by the team-best pitchers, for five season early in the careers of Mickey Welch and Stan Coveleski (quote):

1881 67%
1882 64%
1883 61%
1884 61%
1885 54%
Total 60%

1916 28%
1917 28%
1918 31%
1919 29%
1920 28%
Total 29%

Good.

Fundamental regulation of pitching occurred throughout Welch's career and most of the increase in number of games scheduled also occurred then. Length of Early MLB Schedules

Since the number of games increased mainly by playing more frequently rather than earlier in Spring and later in Fall, I suppose that schedule length was an important cause of proportionate workload (the most important cause, I guess).

Griffith worked almost entirely after the fundamental regulation and entirely after the main increase in number of games, although a significant spike (from 132 to 154 to 140) occurred during his prime.

What is the Max P statistic in the 1890s? Maybe too costly to calculate. Does anyone have Win Shares in a manipulable database? (Well, anyone but Bill James and Jim Henzler?)

Paul,

Year....PWS...#1 WS..Rate

1894...937.1..374.8..0.40
1895...929.4..354.2..0.38
1896...937.7..340.5..0.36
1897...935.6..316.5..0.34
1898...971.5..328.8..0.34
1899..1025.8..312.4..0.30
1900...632.0..189.7..0.30
1901..1242.6..412.2..0.33

Total.7611.7.2629.1..0.35


Sorry that it took this long to respond, busy weekend and such...

Thanks, Max.

1894-97 shows a strong decrease with 60'6" pitching distance and 132g schedule constant. Management learning, I guess.

1898 shows no immediate decrease in #1 pitchers' share of workload with a 17% (154/132) increase in number of games.

Mickey Welch's regular teammates 1889: HOMers in bold

Ca Ewing (teammate 81-89)
1B Connor (teammate 80-89)
2B D.Richardson (tm 84-89)(good defensive 2B, 2 WARP gold gloves)
3B A.Whitney
SS Ward (teammate 83-89)
LF O'Rourke (teammate 85-89)
CF Gore (teammate 87-89)
RF Tiernan (teammate 87-90)(OUAT, a serious HOM candidate)

They also may have something to do with his winning record.

FYI, there's more Welch-ing on the 1939 ballot discussion right now...

Look, I admit it. I did overrate Tim Keefe. Bought the whole ERA-is-everything deal when in 1885 it wasn't. Still, Keefe is still at least a slightly-above-the-line HoMer, instaed of a sure-fire HoMer.

And I overdid Pud Galvin too. Bought the pitched-in-front-of-bad-defense WARP view, which, while somewhat true, might be overdrawn. So Pud might be fighting for an elect-me spot on my ballot if he were around today.

But in the end, making an earlier mistake doesn't pertain to Welch. He's gonna have to stand up against Griffith, Faber, et al to get in.

Howie Menckel #34
Re McCormick vs Welch:
I'll never understand why anyone would give credit for the 21-3 and other ridiculous stats of McCormick in 1884 with the UA.

The statistical justification is that those whom the UA signed from the 1883-1884 Majors and those who subsequently played in the 1884-1885 Majors didn't generally garner ridiculous stats in the UA. Pete Palmer identified multiplier .87 for UA pitching; that is, only 15% increase in ERA+ above MLB performance.

(multiplier .76 for UA batting, or OPS+ up 32%)

Thanks, Paul!
I still think a similar multiplier sometimes would apply to players in top minor leagues, but most of us give little or no credit for those..

Howie, well, I still don't get it. I mean, basically, think of this as a "campaign" to "elect" Mickey Welch (which it is). By analogy, Mickey Welch for president. And your fundamental campaign platform is--he gave up more than 15 runs a game on several occasions, so therefore he's a better pitcher than you thought.

A problematic campaign from the git-go.

But as to the details. You also said we were just talking about those games in which he gave up more than 15 runs and lost by 10 runs or more, so that 16-8 game doesn't count. I count 24 runs in 6 games.

24 runs in 6 games, out of a career of 4800 IP with 2550 RA in 500 games. I just can't see how this changes his profile.

Besides, I could argue that this is not evidence of Welch "pitching in a pinch," but of the converse--not pitching in a non-pinch. Meanwhile, when Keefe gave up 15 runs and/or fell behind by 10 or more, he "pitched with pride." That's my campaign slogan: Welch failed to "pitch with pride" when he got down.

Seriously, I don't know that Welch didn't pitch with pride. More likely, his arm was tired. but just as seriously, I don't know that he pitched in a pinch or anything like that. All's I know is he gave up 200 more WHIP and 75 more runs in 200 fewer IP. The configuration of those runs raises as many questions as it answers.

And as for Welch's comp to Keefe not being relevant, see posts #3-4-5. I don't think I made this into a discussion about Welch and Keefe.

And just for the record, I never had Pud Galvin on my ballot. I for one believe Welch was better.

Well, my approach is for "Everything points to Welch except ERA+" voters.
I agree that those who didn't vote for Galvin - or who have buyer's remorse - don't have to 'compound the mistake' and go for Welch now.

Basically the premise is that ERA+ isn't quite as compelling then as it is now. You'll disagree, I think, which is fine.

For all I know, I may wind up putting Rixey and/or Faber ahead of Welch next week. At the moment, I'm just trying to clarify what it is that Welch actually did....

My big question about Welch is how much of his success in 1884 and especially 1885 should be attributed to the fielding abilities of his teammates. (Without those two years, there really isn't a good HOM case for Welch.) My understanding is that WS and WARP come out somewhat differently on the question, with WARP (especially for 1885) attributing quite a bit of his success to the Giants defense. How strong do you think their defense really was?

> Pete Palmer identified multiplier .87 for UA pitching;
> that is, only 15% increase in ERA+ above MLB performance.

(multiplier .76 for UA batting, or OPS+ up 32%)

Howie Menckel #55
Thanks, Paul!
I still think a similar multiplier sometimes would apply to players in top minor leagues, but most of us give little or no credit for those..

I agree, the multiplier for some minor leagues must be smaller than .87 /15%, not to mention .76 /32%. I don't know which leagues. Clay Davenport calculated smaller multipliers AAA, I recall, but I don't know the time period.

Looking at 1880s pitchers, it occurrred to me to check unearned runs, which were as follows (I have no idea how to make tables line up.)

xxxx ER R UER UER Ratio
Keefe 1472 2468 996 40.36%
Radbourn 1348 2275 927 40.75%
Clarkson 1417 2376 959 40.36%
Galvin 1910 3355 1445 43.07%
Caruthers 891 1393 502 36.04%
Welch 1447 2556 1109 43.39%

This shows that Welch was about 5% (3/60) worse at preventing UER than all his HOM contemporaries except Galvin. Given that Welch's ERA is his main weakness, I think I've decided he was an innings eater, and will downgrade him next year. Alas, poor Mickey....

Nice to have my prejudices against Galvin reinforced. He and Welch from the above list are not PHoM (and never will be). Thanks karl.

Not to defend Welch and Galvin necessarily, but it should be kept in mind that error rates were dropping rapidly during this era, so it stands to reason that pitchers who threw more innings early the era would have higher ratios of UER to ER.

I would suggest that a check of seasonal league ratios of UER to ER be done before too much is concluded from this data, since, for the pitchers listed here, a later start to the career seems generally to correspond to a lower ratio of UER to ER. I'd also be interested to see data for other notable early pitchers, say Tommy Bond, Jim McCormick and Tony Mullane, to put these numbers more into perspective.

>I'd also be interested to see data for other notable early pitchers, say Tommy Bond, Jim McCormick and Tony Mullane

I bet they average about 40 percent ;-)

OK, the things I do for you guys. :-) I've taken keefe, Radbourn, Clarkson, Galvin, Caruthers (HOMers all) plus Welch, Buffinton, Bond, McCormick, Mullane,Matthews, King and Weyhing, i.e. all pichers who won more than 200 games and pitched more than 3 years between 1880 and 1891 (Welch's 12 year career, ignoring the last tiny year in '92.)(Rusie and Stivets started in 1889, I left them out.) This therefore has a mix of older, contemporary and younger pitchers.

Totals:
Welch 43.39% UER rate (Unearned runs divided by total)

entire group 41.39% UER rate. (9667/23357)

Year by year

Year Group Welch Welch excess
1880 49.18% 49.53% +0.35%
1881 47.02% 49.87% +2.85%
1882 47.09% 51.13% +4.04%
1883 46.98% 52.40% +5.44%
1884 42.92% 43.64% +0.72%
1885 46.12% 46.47% +0.35%
1886 38.77% 40.50% +1.73%
1887 34.21% 32.46% -1.75%
1888 39.93% 41.67% +1.74%
1889 34.54% 35.71% +1.17%
1890 38.00% 33.10% -4.90%
1891 39.44% 44.12% +4.68%

In other words, it bounces about, and Welch clearly improves relative to the league, but he's worse than the league in 10 years out of 12. Note the interesting developments in 1889-1891 -- clealry going to 3 leagues in 1890 played hell with fielding!

that's impoves relative to the GROUP of 200-game winners. No mediocre pitchers in this lot!