Many of those "top" scorers played for the 1876-CHI which scored 9.45 R/G, almost three runs ahead of second place.

Still, Start and White played for quite a while and should give you some decent LHB data.

baseball-reference lists RC data for the NA. I wonder how well that "adds up" given all the errors that were occuring. The 1876 NL score 5.90 R/G, but had an ERA of 2.31. The 1871 NA was 10.47 R/G and 4.22 ERA.

Notes on Pike, "reached on errors", and runs scored.

Where the batter hits in the batting order is a huge factor in the ratio of hits to runs scored in the early game, probably much larger than any differences in ability to reach on errors.

RC data on baseball-reference is based, as far as I can tell, on the basic runs-created formula. Whatever the formula, it misses runs scored by the sort of margins Brent is talking about.

Back in the early days of the HoM, when we were studying the NA players hard, I looked for evidence of meaningful differences in ability to reach base on errors, and I didn't see any. I'd be interested to see some of the results of the Woolner study, though. How much of an impact do differences in ROE ability make on player value, as he sees it?

How much of an impact do differences in ROE ability make on player value, as he sees it?

In the modern game, if ROE were included in OBP, it would raise the average level about 010 points, but varying (over a career) between 020 points for the best (Bob Horner in Woolner's 1978-2000 dataset) to 004 points for the worst (John Mayberry). Woolner compares ROE to HBP - worth knowing in extreme cases, but usually not worth bothering with.

Another comment. In the game of the 1870s, we're talking about ROE adding something like 100 points to OBP. (We can't calculate this figure exactly because presumably many of the errors did not result in ROE.) If the batter could affect that ROE rate, it's got to be huge and show up somewhere in the statistical record.

I agree that runs scored are a problem because they are heavily influenced by batting order and quality of the team. I think looking at runs scored, we're more likely to spot someone who is particularly successful at ROE (for example, if R > H+BB) than someone who is unsuccessful. But the default assumption seems to be that ROE is proportional to other hitting statistics, and that assumption is a problem too.

A digression back to midteens/twenties pitchers.

The best pitcher in baseball in 1917-20, on the other hand, was surely Cicotte.

I thought I would take a look at this so I looked at total win shares for the period, 20+/30+ win share seasons, all-star (top 4) and league bests, and WARP1. The following is a list every pitcher who was a 2 time league all-star between 1917-20.
I don't know whose BP numbers have adjusted for the new formula and whose has not.

American League
player     WS  20+ 30+ AS Lg#1 WARP1 WARP3
Coveleski  117  4   1   3  0    42.4  40.7
Bagby      108  3   2   2  1    34.9  32.3
Cicotte    106  3   2   2  1*   36.5  34.3
Mays       105  4   1   3  0    30.6  28.4
Johnson    104  3   1   2  1    43.9  42.8
Shawkey     70  2   0   2  0    23.7  22.4
  
* In 1919 he had 32 WS. If you discount because he threw the World Series, the second place pitchers had 27 WS. 

Ruth led the league in 1917 with 36 WS, one more than Cicotte, but he did not continue to pitch full time after 1917.
Shawkey had a great year in 1916, but that's not the time period.

National League
player     WS  20+ 30+ AS #1   WARP1 WARP3
Alexander  104  3   2   3  2 *  38.9  35.3
Vaughn     104  4   1   3  2    35.6  31.2
Cooper     101  4   1   3  0    30.6  26.1
Grimes      67  2   1   2  0    16.9  13.0
Adams       55  2   0   2  0    15.9  14.2

* (missed most of 1918 for WWI, 2 WS. 40 WS season in 1917)

The cut off sentence continued to state that the second place pitchers had 27 WS in case you discount Cicotte's 1919 year.

From the above numbers, I think Coveleski, Johnson, and Alexander are the best for the tight time period. Coveleski has the best WS, Johnson the WARP numbers, and Alexander would probably have the best if you give him credit for the missing WWI time. Cicotte has close numbers, but I heavily discount the 1919 season, and somewhat the 1920 season.

I also took a look at 1917-1923 to see what pitchers "dominated" the changeover to the lively ball/elimination of the spitball/grandfathering of spitball pitchers. Notice that the following lists have 4 of the granfathered spitball pitchers: Coveleski, Shocker, Faber, and Grimes.

American League
player     WS  20+ 30+ AS Lg#1 WARP1 WARP3
Coveleski  180  6   1   3  0    65.8  61.9
Johnson    165  5   1   2  1    65.9  62.8
Mays       158  5   2   4  0    44.6  40.5
Shocker    145  4   1   3  0    51.5  48.2
Faber      138  3   2   2  2    43.4  39.2
Shawkey    133  3   0   3  0    40.4  37.1
Bagby      118  3   2   2  1    37.1  32.9
Cicotte    106  3   2   2  1    36.5  34.3
Rommel      85  2   0   2  0    37.4  34.6

National League
player     WS  20+ 30+ AS #1   WARP1 WARP3
Cooper     178  7   1   5  1    54.5  44.5
Alexander  171  5   2   3  2    60.1  51.7
Grimes     128  4   1   3  1    38.8  29.2
Rixey      113  4   0   3  0*   36.0  27.0
Luque      109  2   1   2  1    37.0  30.2
Vaughn     104  4   1   3  2    35.6  31.2
Adams       97  2   0   2  0    29.5  24.9
Morrison    58  2   0   2  0    18.8  14.6
* 1918 no season because WWI service??

Rixey had a good year in 1916 and Faber is missing his good first three years.

I would call Alexander as the best pitcher because I would give credit of about 25 WS for the 1918 WWI year. Cooper, Coveleski, and Johnson would be in the next tier. I do agree that Johnson did not dominate as he had in the past. However, I think Alexander is still dominant if credit is given for WWI. YMMV.

I hope this helps with the discussion.

From the above, it seems that the pitchers who were successful in this period of change were either great control and hit prevention pitchers (Cooper, Alexander, and Mays), grandfathered spitballers, or Walter Johnson. Also, I noticed that Johnson and Cooper had reputations for easy motions, Alexander had a sidearm motion, Mays threw underhand/sidearm, and most spitballers didn't throw hard.

Back to your currently scheduled conversation about ROE and 1870s baseball. It is very enlightening and I am very interested.

Apologies -- I meant to post this a week ago, but it's been a strange and chaotic week (a gallstone scare, assorted snafus at work, and getting back into college after a layoff the length they make sitcoms out of).
I'd envisioned a rather drastic shift in this ballot, favoring the high-peak, short-career guys. But the aforementioned chaos prevented me from doing the work to back up the arguments I'd want to make in that direction, so I figure it's best if I take the next few years of "well, that went quick" candidates coming up to iron the kinks out of my thinking, and in the meantime play it relatively conservative.
The biggest jumps here are mostly players with significant time in the 1880s, explained under Charley Jones.

1. Luis Santop (n/e; PHoM 1932). So much for my wariness about the possibility of overrating Negro Leaguers ... still, it looks like I'm in fairly safe company here.
2. Tommy Leach (1; PHoM 1930). Willl probably drop down further once I get the peak/career thing worked out in my head ... but he's gonna divebomb like everybody else pretty soon anyway, so it probably won't be noticeable.
3. Rube Foster (4; PHoM 1932). Mmmm ... Kool-Aid ... seriously, the more discussion of Foster I read, the more I think his post-play accomplishments are *hurting* his HoM chances, instead of merely being irrelevant, which is kind of sad.
4. Clark Griffith (2; PHoM 1931). The wacky theory lasted one week. I'd still have no problems with his getting a plaque, though.
5. George van Haltren (3; PHoM 1926). Lost a little faith in the credit I was giving him for the pitching.
6. Hughie Jennings (5; PHoM 1919). Nothing much to add.
7. Hugh Duffy (11). Whack-A-PHoM bumps up a little more.
8. Addie Joss (6; PHoM 1930). Stopped myself from much higher ranking and a "FoAJ" rant this week, because I want to do a little more number-crunching and I'm not comfortable with the resources or time I have at the moment ... but the quickie version is, I don't believe in "hang-on" time and I'm having stronger qualms lately about the "he wasn't a horse" argument.
9. Spotswood Poles (OB). Broke him away from the pack of question-mark guys from last week after re-reading some of the discussion.
10. Dobie Moore (N/A). Will probably spend the next month or so bouncing around between Jennings and Childs.
11. Pete Browning (15). Gets help from the Jones Effect described below.
12. George Burns (7). A little surprised he ended up dropping down this far--would probably be top-5 on a "guys I just like for no particular reason" ballot.
13. Charley Jones (OB). The talk about Jones's missed seasons led me to take another look back at him ... at which point I realized, a couple of the ways I'd set up to look at peaks had screwed up the season lengths, and thus screwed over the earlier players. I think the word I'm looking for here is "mortified".
14. Cupid Childs (9; PHoM 1921). Durnit. Dropped because ... well, because of Santop, Moore, and Poles, mostly.
15. Rube Waddell (10): Personally, I'd have lived with the headaches. Then again, I'm a dog person, too.


Off-Ballot Mentions:
Lip Pike: Ranked at 16. Would have edged out Rube, but I want to digest what Brent brought up before I could post the first draft of this.
Mickey Welch (14): Ranked at 17. Nudged down by Santop etc.
Jose Mendez: Ranked at 18 ... and that's totally unfair to him. Thinking of him just for the pitching would probably bump him down 3-4 spots; the projections and anecdotal evidence isn't as compelling as I thought it would be. But I know I'm pretty much just paying lip service to his hitting contributions by putting him here, and he deserves better than that.
Jimmy Ryan: Ranked at 21. Red Queen of the ballot -- "moved up" some spots before adding the new names to the list, ended up in about the same place.
Roger Bresnahan (13): Ranked at 25. Took a tumble from a smoothing-out of a positional bias I'd worked in, and from a jump Ned Williamson took.
Jake Beckley: Ranked at 29. Really about where he was last time, just bumped down by the new faces.
Ross Youngs: Ranked at 30. If he moves higher, it's probably proof I'm a sucker for a sob story.

I'll post the "legit" ballot tomorrow, to give myself time to make sure I was awake when I finished writing this. :)

Thanks Kelly!

Sinin's SABR encyclopedia agrees. Using RSAA, Coveleski is the "stud" of the changeover. (Although 1917-1923 does bracket his off years of 1916 & 1924).

Cicotte scores best if you go back to include his fine 1913-14 seasons. He passes Coveleski and edges out Vaughn for 3rd place if you do this.

Trying to figure out how to get Sinin's numbers league/context adjusted, so I won't post a table. RCAA could be done by just using adjusted RC, but doesn't look like it can be done for RSAA.

Tanketra, is that your ballot (post #309) or are you only submitting a preliminary one this time?

But the default assumption seems to be that ROE is proportional to other hitting statistics, and that assumption is a problem too.

I would agree, but I'm not convinced it is a significant one. I think a useful analogy for thinking about this matter is hits allowed on balls in play for pitchers post 1880. The idea that pitchers don't have an influence on their hit rate has been proven wrong, but study of DIPS is still going to show you most of the best pitchers. Reached on Errors is something that is mostly, but not entirely, out of the batter's control, so even if ROE makes up 100 points of a players OBP (just as hits on balls in play makes up most of a pitcher's ERA), I think that our deduction in the absence of clear evidence must be that the range of variation within that 100 points that is susceptible to the hitter's influence is going to be small.

Another avenue for study of this issue might be to compare the number of actual runs scored by NA teams to the number of runs predicted by the basic RC formula. Pike was one of the few power hitters in the early game; does his presence on a team significantly affect their scoring profile? Another thing to look at at the team level would be the ratio of runs scored to RBI. There again, significant variation between teams might show something about a team's hitters' ability to force opponent errors. If Pike's teams were to show up consistently on the low side of measures like these, that would be a piece of evidence that he was not as productive an offensive player as his hitting numbers suggest, since he looks like he would be the major offensive force on most of his teams.

That was just a prelim ballot, yes. I didn't post it here by mistake ... I'll post it in the ballot thread later today.

Since I work graveyard shift, and work on my ballot there, I try to give myself a chance to see it 'in print', and catch anything really screwy, like "I voted for Ryne Sandberg 75 years too early" or something.

Sorry for any confusion.

Reached on Errors is something that is mostly, but not entirely, out of the batter's control

This may be the best assumption we can make in the absence of other information, but making this assumption would still change our picture of 1870s baseball. For example, let's assume that ROE during Pike's career is equivalent to adding 100 points to the league average OBP and SLG, and that the ROE are spread out among batters randomly in proportion to the number of hitting outs. If I've done the math correctly, Pike's unadjusted career OBP and SLG would increase to .431 and .558, while the lgOBP* and lgSLG* would increase to .387 and .438, causing Pike's OPS+ to fall to 139.

More generally, though, I disagree with this assumption. In modern baseball, it's not unusual to observe batter G/F ratios smaller than .5 or larger than 1.5. Assuming similar variations existed in 19th century baseball, they would have to have affected the probability of ROE, though we may never know by how much.

As another check on my perception that Pike didn't score as many runs as expected, I looked at the BR career leader board for OPS+ to see how many other hitters with career OPS+ close to 155 failed to show up among the season leaders in runs scored. Looking at the top 55 players with OPS+ >= 144, there were few, if any, leadoff hitters. Most of the 55 players spent many seasons among the league leaders in R - in fact, most of them led the league for at least one season - but there were a few exceptions. Five players were in the top 10 in R only 2 times (OPS+ in parentheses):

Mark McGwire (163)
Manny Ramirez (157)
Jim Thome (151)
Brian Giles (150)
Willie Stargell (147)

Three were in the top 10 in R only 1 time:

Mike Piazza (153)
Gavvy Cravath (150) (the one time, though, he was 1st in the league)
Vladimir Guerrero (146)

And two 19th century players never were in the top 10:

Dave Orr (161)
Bill Joyce (144)

Most of these players either had short careers or had trouble staying in the lineup. By contrast, if I've counted correctly, Lip Pike missed only 6 games during the period 1871-77.

So the bottom line is that Pike's failure to show up among the season leaders in runs scored is unusual, but not unprecedented among players with high OPS+.

First impression: the people on that list ran like freight trains. Pike was reputedly the fastest player in the game.

In any case, until you adjust for the team's offense, which varies widely in Pike's time, I don't think you're going to approach a good answer.

the people on that list ran like freight trains.

It's interesting how that's also true of many of the 45 names that I didn't list because they did spend several seasons among the leaders in runs scored, guys like Frank Thomas, Ralph Kiner, and so forth.

Pike's teams were above average in R/G for 1871-73 and 1876, below average in R/G for 1874-75 and 1877. He never played for the best or worst offense in the league.

Regarding Brent's point about G/F ratio..., hopefully this hasn't been touched already.

Does anyone know when the one-bounce rule went out or whether the NA had it at any point? If so the incentive to swing for power or to hit the ball in the air is diminished considerably.

Another consideration for ROE in the 1860s, NA, and early NL is the fair/foul hit. This play in particular would seem like a prime opportunity to create ROE since a third baseman or catcher would need to make an even further and more hurried throw. Obviously we'll never know the answer, but it would be interesting to find out whether hitters like Ross Barnes generated higher ROE as a result of their use of this tactic.

(Speaking of which: [Shamelss plug alert!!!] For Boston-area residents interested to see basball played in the 1860s fashion or to play it, The Essex Base Ball Club of Danvers, Mass practices on Weds nights and plays games most Sundays, April through October. Come out and see/join/support the EBBC!)

I made a quick-n-dirty team runs created formula for the NA.
It is (OBA+K)*(SLG+K)/(1-AVG)*26. The factor K was .173 for 1872-73, and .137 for 1874-75; it was chosen to make the actual avg r/g equal the prediction.
The r squared between actual and predicted r/g is .99. I only used teams who played >30 games from 1872-1875.
Pike's teams
1872 scored 1.29 r/g more than predicted
1873 scored 0.28 r/g more than predicted
1874 scored 0.14 r/g less than predicted
1875 scored 0.13 r/g less than predicted

Does anyone know when the one-bounce rule went out or whether the NA had it at any point? If so the incentive to swing for power or to hit the ball in the air is diminished considerably.

Another consideration for ROE in the 1860s, NA, and early NL is the fair/foul hit. This play in particular would seem like a prime opportunity to create ROE since a third baseman or catcher would need to make an even further and more hurried throw. Obviously we'll never know the answer, but it would be interesting to find out whether hitters like Ross Barnes generated higher ROE as a result of their use of this tactic.


One bounce -- that one went away in 1864 or 1865. Elite players/teams had been trying to get rid of that rule since 1859 or so. One bounce rules were for "muffins" they said.

Fair Foul -- Ross Barnes did score a ton of runs, he played for high scoring teams, too, though.

Other factors -- I have no idea how lifely the ball was back in those days. There was the Mahn Ball and then by 1879 the Spalding Ball. I have no idea how far a ball would go if a strong player hit it solidly in say 1873.

lively... not lifely...

Class of 1898

1898 players who have been elected to the HOM
Player/1898 votes/Career HOM votes
1. Deac White 657 657 Elected 1898, 1st try
2. Paul Hines 654 654 Elected 1898, 1st try
3. Georg Gore 553 553 Elected 1898, 1st try
4. Ros Barnes 476 476 Elected 1898, 1st try
5. C Radbourn 427 4558 Elected 1905, 8th try
6. Geo Wright 420 2145 Elected 1901, 4th try
7. Ezra Sutton 380 6070 Elected 1908, 11th try
8. Richardson 366 4077 Elected 1905, 8th try
9. A Spalding 339 4449 Elected 1906, 9th try
11. Jo Start 297 8378.5 Elected 1912, 15th try
12. Pud Galvin 209 6585 Elected 1910, 13th try
13. C McVey 198 7989.5 Elected 1914, 17th try
31. Dickey Pearce 6 8073 Elected 1931, 34th try

Not Elected
Player/1898 votes/1931 votes/career votes
10. Williamson 328/127/3491
14. Tip O'Neill 132/0/791
15. Lip Pike 123/533/9480
16. Ch Jones 96/220/2443
17. M Welch 96/407/4386
18. Fred Dunlap 87/19/828
19. McCormick 70/57/2954
20. Dave Orr 49/0/86
21. Dalrymple 31/0/31
22. Whitney 24/6/864
23. T York 24/16/369
24. T Bond 21/46/433
25. H Wright 16/6/878
26. Creighton 12/0/29
27. L Meyerle 7/7/287
28. Mathews 7/8/110
29. Joe Clapp 6/0/6
30. Har Nichol 6/0/6
32. B Sunday 4/0/4
33. Cummings 1/0/1

Comments: Williamson lost a mano-a-mano 3B battle with Ezra Sutton and never recovered... O'Neill, Orr, and Dalrymple's patrons didn't have as much staying power as many others... Welch and McCormick seem to have split the 1880s leftover P vote... H Wright had a brief renaissance before Pearce gained full steam for pioneer votes.

Class of 1899 HOMers
Player/1899 votes/Career HOM votes
1. J O'Rourke 699 699 Elected 1899, 1st try
2. King Kelly 625 625 Elected 1899, 1st try
4. Tim Keefe 480 1719 Elected 1901, 3rd try
5. Ha Stovey 410 9576 Elected 1916, 18th try
11. Caruthers 279 10704 Elected 1930, 32nd try
12. Bennett 272 11503 Elected 1921, 23rd try

Class of 1899 Not Elected
Player/1899 votes/1931 votes/career votes
13. Browning 271/361/7066.5
26. Harry Larkin 10/0/10


Class of 1900 HOMers
Player/1900 votes/Career HOM votes
1. John Clarkson 756 756 Elected 1900, 1st try
2. John M. Ward 617 617 Elected 1900, 1st try

Class of 1900 Not Elected
Player/1900 votes/1931 votes/career votes
18. Tony Mullane 83/ 38/ 1148
32. Curt Welch 3/0/3*
* - actually first eligible 1899


Class of 1901 HOMers
Player/1901 votes/Career HOM votes
3. Ja Glasscock 619 2878 Elected 1904, 4th try

Class of 1901 Not Elected
Player/1901 votes/1931 votes/career votes
23. Arlie Latham 36/0/36
25. Dave Foutz 28/0/136
28. Hutchinson 13/9/141
33. Oyster Burns 6/0/6

Good stuff, Howie!

If I can extrapolate from a very small sample - it looks like if a player makes it into the top 9 on his first ballot, his chances of eventually making it in are pretty good. If he starts out in places 10 to 15, he may or may not eventually make it. If he starts out any lower, his name needs to be Dickey Pearce.

Happy birthday Cupid Childs.

Getting a little ahead - just updated Urban Shocker on my site. I'll hopefully have more info on his matchups with the Yankees this week, but I haven't had much time lately, so who knows.

Just tossed the extra info I wanted on the Urban Shocker thing linked in the above post (hmm, that's probably a little awkwardly put, ah well).

Fun fact: despite starting 1/3 of his games with the Yanks, he faced them more than any team except the Senators.

stupid #### #### 6000 character limit