Continuing my obsession with actual runs scored. I?ll probably lay off this once we get further into the 20th century. The explanation for R* is in the ?1905 Ballot Discussion? thread, starting at post #9 and ending at post #23. To summarize: the average of the number of runs of the #5 person in the league and 1/6 of an average team?s runs is set at 100,so 100 is a standard of excellence. Typical league-leading numbers come out in the 115-120 range. Billy Hamilton?s record of 192 runs was set in 1894 ? and 1894 is one of the all-time freak offense years. It comes out as an R* of 120 ? a good league-leading number , but not out of line.

I updated the standard of excellence for both leagues from 1901 to 1910. The most spectacular new single season R* totals both came in 1909: Leach at 138 in the NL and Cobb at 135 in the AL. There?s also a 129 (Beaumont, 1903 NL), and two 127?s: Dougherty, 1904 AL, and Burkett, 1901 NL.

Several caveats are in order.

Billy Hamilton's numbers indeed are inflated by the era he played in and the team he played for. But Larry Walker will be a serious candidate, and so is Mr. Hamilton.

10 straight OPS+'s over 125

Indeed, I am not convinced that Mr. Hamilton is a shoo-in. The experience of Mr. McPhee seems instructive. A close #4 in his first year of eligibility, but once the rush of excitement ebbs and everybody takes a deep breath and a fresh look he drops behind two incumbents and stays in fourth place, no closer to the top than before. I happened to be a supporter of Mr. Rusie in his first-year election, but I also don't doubt that if he had failed to gain election that year, he might well have dropped back the next time. I am learning to appreciate the urge toward caution on newly eligibles below the Ruth-Cobb-Johnson line.

Here's my ballot. Again, I use a combination of peak and career for the rankings. I also view each position on an equal basis. This doesn't mean that I have a quota to fill each position for my top ten. Sometimes a position will not have a viable candidate for a certain "year."

1) Ezra Sutton (1): Greatest nineteenth century third baseman. In fact, I think he's the best peakXcareer player at that position until at least Ray Dandridge (who I haven't analyzed yet). Baker was much better peak-wise, but wasn't nearly as durable (he also didn't play during 1915 and 1920). Not that far off from being the best NA third baseman over Meyerle. Best major league third baseman for 1872, 1873, 1875, 1883, 1884 and 1885. Almost the best first baseman behind McVey for 1876.

As has been stated before, third base at the time was more of a defensive position than second base. Offense at the "hot corner" has to be analyzed with that in mind. Third basemen tended to get beat up more than they do today so their career numbers seem truncated as compared to some of the other positions.

2) Bid McPhee (2): Greatest second baseman of the 19th century. If any AA guys should go in, he should be numero uno. Consistently near the top of the list for second baseman (and did it longer than any of them). Best major league second baseman for 1886.

3) Sliding Billy Hamilton (n/a): My pick for greatest outfielder of the '90s because of his amazing peak. Best major league leftfielder for 1890 and 1891. Best major league centerfielder for 1895, 1896, 1897 and 1898.

4) Cal McVey (4): Awesome player. I gave him credit for his pre-NA work, though I still decided not to give him any for post-NL. This might be unfair of me and I might decide later to include his career out west (does anyone have any info for this time of McVey's career?).

Never had an off year in the NA or NL. Best offensive catcher for the NA (possibly the best all-around). Best first baseman for 1876 (possibly 1879). Best catcher for 1877. Best third baseman for 1878.

5) Dickey Pearce (5): Really revolutionized the position of shortstop. All-around player at the position. Considered the best before George Wright. Caught many games as a catcher (even was an All-Star at the position one year). Even with my conservative evaluation, he has to rank near the top. He played for over twenty years in the best leagues or on the best teams of the 1850s and '60s. Even though his NA and NL was meager (he was 35 in '71), he still had the most value after 35 until Dahlen and Davis, FWIW.

If we are including pre-NA players, I can't see how anyone could leave him off their ballots, IMO.

I'm not giving him any credit here for the bunt, BTW.

6) Cupid Childs (n/a): Best second baseman of the '90s. Too short of a career to knock out McPhee for tops for the 19th century (but his stellar peak almost does it!). Best major league second baseman for 1890, (almost in 1891), 1892, 1893, 1894, 1895, 1896 and 1897.

7) Joe Start (6): Considered the best first baseman for the 1860s. Considering how old he was when he joined the NA and how well he did, that evaluation seems to hold water. Best first baseman for 1871, 1878 and 1879.

8) Charlie Bennett (7): Strictly as a catcher, extremely comparable to Buck Ewing value wise (though based more on career than peak value). Best major league catcher for 1881, 1882 and 1883. Most durable catcher up to that time (catchers absorbed much more abuse than they do today).

9) Billy Nash (8): The '90s had some terrific players at the "hot corner": McGraw, Collins, Joyce and Nash. Possibly the best defensive third baseman for the 19th century (and not too bad offensively).

Best major league third baseman for 1888, 1889, 1892, and 1893. Best PL third baseman for 1890.

10) Jack Clement (9): Very durable with a nice peak. Best major league catcher for 1891 and 1895.

11) Ed Williamson (10): Best third baseman for the 80s. Best major league third baseman for 1881. Best NL third baseman for 1882. Best NL shortstop for 1888.

12) Fred Dunlap (11): Most value as a second baseman for the 1880s (though McPhee and Richardson were still the better players career wise). Best major league second baseman for 1880, 1881 and 1884. Best NL second baseman for 1882 and 1886.

13) Lip Pike (12): Considered the fastest man of his time. Best centerfielder for 1874, 1875 and 1876. Best rightfielder for 1871. Star second and third baseman for half of the 1860s. He might deserve to move up.

14) Hugh Duffy: "Only" the third best centerfielder of the '90s, but that position was very strong for that decade. Best major league rightfielder for 1890 and 1891. Best major league centerfielder for 1892, 1893 and 1894.

15) Pud Galvin (13): Better than Old Hoss.

Best major league third baseman for 1872, 1873, 1875, 1883, 1884 and 1885.

It should be "only" 1875, 1883, 1884 and 1885. I forgot to correct it again.

It's hard to speak about placing Hamilton on the ballot with the 1906 election results not in yet, because he obviously needs to be placed among the top candidates, one of whom will be gone.

By fielding & season-adjusted WS, I see Hamilton as having the second-most documented career value of any player potentially on the ballot, trailing only Sutton, if Ezra is not elected. He could also trail Start or McVey (or Pike or Galvin, I suppose), depending on how much credit one gives for undocumented or non-major-league play.

By the same measure, I see Hamilton as having the highest documented peak value of any non-pitcher on the ballot. WARP concurs that Hamilton's record is tremendous.

If Sutton, my #1 in 1906, is not elected, I'm going to have to think hard about whether Sutton or Hamilton will be #1 in 1907. If Sutton is elected, Hamilton will be an easy #1 choice. Galvin, Spalding, McVey, Stovey, Start, & McPhee are all worthy HoMers, but I line them all up behind Hamilton without hesitation.

Whether I put Hamilton #1 or not, I think he's going to be a very strong candidate because his career value, his peak value, and his raw numbers are all impressive. Sure the raw numbers have to be considered appropriately in terms of context, but the context-adjusted metrics agree that Hamilton was a truly great player, one of the top 2 or 3 offensive forces of the 1890s.

I don't think many of us are going to lose much sleep trying to place Sliding Billy this week. The tougher rankings are going to be Duffy and Childs.

John, what, no Al Spalding?

John, what, no Al Spalding?

Oops! I thought the election was over. This discussion thread really shouldn't be up yet.

I would place Spalding #3 again, then everybody else moves down one (Galvin would temporarily be removed).

<i>John,

John, I don't follow. Are you multiplying WS times (WS per season)?

Yes. Why I do that is I believe that quantity (WS) is equal to quality (WS per season). Just my opinion and there is no correct way of doing it.

But, you aren't making a fielding adjustment to WS, then. That could affect how Bid McPhee does in your rankings. According to WARP1, he was the best 2b in baseball 5 times. Adjusting for fielding may not affect your numbers that much, but it may change the number of times McPhee finished first among 2b. And, among people who know more about WS than I do, it seems to be necessary to arrive at more accurate values for pre-1893 players.

Since I have McPhee second on my ballot, I don't think I'm seriously underrating him. :-)

Is WARP1 adjusting for league? I don't remember, to be honest with you.

You may have a valid point about the fielding adjustment. My problem with WARP is that the formula is not spelled out, so I remain wary of it until further notice.

Other voters look at your rating of McPhee (only once the best 2b in baseball) and downgrade McPhee based on a small peak. My argument is that by adjusting WS for fielding (and making WS more accurate), perhaps McPhee would more often show up as the best 2b in baseball for a given season.

I wouldn't worry about it, John. Anybody who doesn't agree with me about who the bests were for each position each year should 1) make their case as you have and 2) post their own bests. I'm always willing to listen to anyone's argument for or against.

John, WARP1 is not taking into account the difference in competition between the two leagues. That has to be factored in.

1907 Preliminary Ballot

A much more competitive year. Only one player leaves the ballot, and four have careers good enough to be considered for the 15 slots: Billy Hamilton, Hugh Duffy, Cupid Childs, and Elmer Smith. Smith is a sentimental favorite because he was a Pittsburgh native and a Pittsburgh star, but that's not enough to get him onto this ballot. The others are given preliminary positions as listed below. Only re-arrangement so far among the holdovers concerns the outfielders, whom I've been studying more carefully in preparation for the arrival of Hamilton, Duffy, Hoy, Delahanty, Ryan, and Van Haltren in the next few years. Charley Jones gets some credit for the years he was blacklisted, Sam Thompson gets more fielding credit, and Mike Tiernan loses some fielding credit.

I'm including numbers this time, for a change of pace. Career win shares are fielding-and-season adjusted, and adjusted for league difficulty when two leagues run simultaneously. Total peak is the sum of win shares above average in each season. I follow the numbers closely, but not exactly.

1) Ezra Sutton 470 CWS, 87 total peak -- Career value at difficult defensive position narrowly outweighs Hamilton's offensive peak.

John C, I know you didn't ask me, but... As of 1870-71, Joe Start had more career value than any player in the 25 year history of organized baseball. He was the star for the game's first dynasty, the New York Atlantics, who were the dominant team in the New York area. He is widely mentioned as one of the 3 best players at any given time (early on it was Creighton, later it was G. Wright, in between maybe H. Wright, while almost throughout the decade he was one of the best). I would assume that whatever value he had in the '70s, he had similar value ("normalized" to some kind of base) in the '60s. Of course there are NO numbers.

As for McVey, he played 3 years and was by 1870 perhaps one of the top 5, certainly a top 10, but vastly overshadowed by G. Wright and Spalding as a peak performer for those years. I think that 1868-70 were probably the weakest year's of McVey's career in that they were his "rookie" years, his formative years. By 1871-76 he was probably playing at a peak near Start's earlier (1865ish) peak.

I had McVey ahead of Start on my early ballots. Having explored the '60s, I now have to say that Start probably had a slightly higher peak and, of course, a 2X+ career.

At the tail end of the 1906 Ballot Discussion thread, Tom H brought up Hamilton's 1690 career R as an extraordinary accomplishment.

Now that I've got all this R* stuff, it's tempting to create career totals for it. I'd be skeptical of this, because I'm not sure what it really means, but you might as well know what I have. Of the players I've worked up, the top 4 career totals belong to Anson, O'Rourke, Gore, and Connor, with Kelly 6th and Brouthers 10th. I've got Anson, O'Rourke, and Gore that high even though my numbers start in 1876 (no NA).

Since none of that has much to do with our current task, I'll just list career totals for some players who are either on our current ballot, or who are not yet eligible but whose careers are over or nearly over. Since I left off some token years, these totals might in some cases be off by 5 or 10, but nothing big.

Career totals for R*:

Career votes pts leaders

John,

I'm giving Start 60 WS and McVey 50 WS for non-Major League play.

I don't have any absolute replacement levels calculated: all I have are the differences between WARP and WS numbers. For pitchers, I figure that at about 6 WS/ 162; for batters, 3.5-3.7 WS/162; I don't have a fielding replacement level.

For 19th-century catchers, to calculate peak, I set avg. performance at 60% of league avg. for position players. For career value, I boost catchers by 25%. I am considering a career boost of up to 40%, to make the career ratio more consistent with seasonal peak, but I haven't been able to run the numbers on enough cases yet to determine if this much of a boost is justified. In practical terms, should Charlie Bennett be near the end of the line of players I think merit eventual election, or should he be near the head? A 40% bonus puts him at 400 CWS, a 25% bonus puts him where he is now.

TomH - I'm not taking the career totals all that seriously myself. I mostly wanted to look at peak/extended peak. Of course, like any counting total, playing every day matters - and that's the problem with McGraw. He didn't play every day.

What about Ted Breitenstein? He should be eligible, no? Should he be given any consideration as a down-ballot candidate? He doesn't look like a HoMer to me, but perhaps someone who might merit mention at the bottom of a ballot. Anyone have any insights or thoughts?

Yes, Breitenstein is eligible. Since we consider his 3 games pitched in 1901 to be token, we ignored it and considered 1900 to be his last year. So he was also first eligible last year. That's why he doesn't show up on the list of new candidates.

Two of the years when McPhee was the best 2b in baseball were 1893 and 1895, when he was playing in the NL. In 1889, when he was the best 2b in baseball, some have argued that AA was equal to NL in difficulty. For other years, some discount may be applicable, but that discount wouldn't be on the level of the UA discount.

I'll take a look at him again. He's still not going to top Sutton, but a more thorough analysis of his career won't hurt.

A tale of Hugh Duffy, Frank Selee, Sam Thompson, and Pythagoras.

I've been bothered for some time by the fact that WS rates Hugh Duffy more highly than WARP does, on both offense and fielding. Usually the two systems evaluate offense, at least, pretty consistently -- there'll be small difference from year to year, but they even out over a career. Duffy is an unusual case in that WS likes him better, and Thompson is an unusual case in the WS likes him less. For those who care, I think I have the answer: Pythagorean record vs. actual record.

During Hugh Duffy's prime years, he played for Frank Selee's Boston Beaneaters. They were the best team of the 1890s, and Duffy's prime years with the team, 1892-1899, they consistently beat their Pythagorean record. Their scores -- +8, +8, +5, 0, +1, +2, +4, 0. Total -- +28. Great team, great manager. Sam Thompson's Philadelphia club, during this same period, was a team on Pythagoras's bad side. During Thompson's big years from 1891 to 1896, here are the totals by which the team missed their Pythagorean record: -1, -5, -3, -6, -3 . Total -- 21. WARP is not adjusted to team wins. WS is adjusted to team wins. In most cases, I suspect, the year to year differences even out over a player's career, but in this particular case, they don't.

I haven't looked at Thompson's record carefully yet, but I can tell you that all the years where WS gives Duffy significant credit that he doesn't get in WARP are the years where the Beaneaters beat Pythagoras handily: 92, 93, 94, and 98. I also haven't figured out yet exactly how much total WS difference comes from the teams' out-performing or under-performing Pythagorean expectations, but I wouldn't be surprised in this case to see the difference be as large as 25 "raw" WS, +15 for Duffy, -10 for Thompson.

Having figured out the discrepancy, I now have to proceed to deciding how to account for it in the rankings. Which numbers better represent the players' merit?

Your thoughts?

Joe,

3-4 WS/162 games is the difference between replacement level in bRAR in the WARP system -- the zero point on its scale -- and the zero point for batting WS in win shares. I only use it for conversions between the systems -- I'm not using any formulas that depend on knowing exactly what true replacement level is, as you are.

Could Joe or someone please post this as my definitive 1907 ballot, as I'm on vacation till 25th (just in time for 1908 - some vacation!)

Joe:

I try to avoid comparing 70s to 90s with the new metric, but it does help with all the late 80s/90s guys, who are near contemporaries. I think it undervalues McVey/Pike/Meyerle compared with the 90s boppers, but the downgrading of Sutton compared to his contemporaries (and Start too, except I'm now convinced he was just about Numero Uno in the 60s) is I think valid, as is the upgrading of Thompson and the downgrading of McPhee, who are pretty well spot on contemporaries.)

I hope I've used it intelligently; the advantage of doing this every "year" is one can adjust as others find new evidence.

Dickey Pearce, -- Poor 1872, so even if you add 1871-2-3 together it?s unimpressive. Not convinced.

Karl, can you find a shortstop that accomplished more after 35 during the 19th century than Pearce? Until gloves were the norm, the average shortstop was had it by 30.

BTW, I keep reading here that Pearce was on a lower level to other players from pre-1871. Since I'm not an expert on that era, this might be so. However, I have yet to see any quote that documents this. Besides, it's one thing to say Wright or Creighton were better peak-wise than Dickey (which I wouldn't argue against), but Wright was about finished as a player by the age of 32, while Creighton was definitely finished as a person by the age of 21. Pearce kept playing longer than any other middle infielder of his time.

Marc:

How does Start have more career value than Pearce prior to 1871?

Andrew Siegel said: "Suggestion: how 'bout we all agree that if we leave any player off of our ballot who finished in the Top 10 in the last election, we will put a sentence at the end of the ballot explaining why?"

I have no problem with this rule. But, have you guys seen some evidence of strategic voting? I read the ballots but do not scrutinize them, so it's unlikely I would notice a pattern. But Joe, at least, is neck-deep in the ballots and patterns would emerge.

By the way my thinking to date has both Childs and Duffy scoring much higher than the posts so far.

As of right now, where would you place them?

For those curious, here's a list of the cumulative deviation from Pythagorean expectations of the team records of most of the leading position players on the ballot. NA years are not included because the estimated WS are not adjusted to team records. Note that these numbers are _not_ season adjusted, so high deviations in short seasons may weigh more heavily than they appear here. Also, major deviations in peak seasons may substantially affect peak WS even if career WS even out. But if you want to look further at this, these cumulative numbers may give some idea of what bears further scrutiny.

Sutton -15

Regardless of whether or not we adopt the "top 10" rule I will follow

here's a list of the cumulative deviation from Pythagorean expectations of the team records of most of the leading position players on the ballot.

Stovey -14

Most of Stovey's negatives come from his first three NL seasons.

As the only voter to put Stovey #1 last time, I've got to deal with this, and I'm going to need a little help understanding it. I've been taking those three seasons (with the Worcester Ruby Legs) as a large part of the reason why I have been supporting Stovey.

Some team context:

Year G R/G allowed League Record Pythag Deviaton

For TomH.

WS/162 (taken from NBJHBA)

Hamilton -- 34.31

Yes his career length was good, but as I observed in the other thread, he had fewer PA than Chet Lemon, who has the same lifetime OPS+.

Two things that separate Duffy from Lemon:

1) Stolen bases. I know they played in different times and under different rules, but Duffy appears to have been the superior baserunner.

2) Duffy's peak compared to Lemon's. Chet never had a season that he could be argued as the best player in the majors (you can make the case he was the best centerfielder over Dawson in '81). Duffy did.

For OCF:

<i>I don't have the right RC formula in front of me, nor do I have WS. Would someone check out these things for me:

Another thing about Duffy...

Duffy (along with Hamilton and Kid Nichols) was absolutely essential to the successes of the Boston dynasty of the 1890's (that translates into "actual pennants"). For me at least that militates in favor of Hugh (as well as Billy and Kid although they won't really need the help).

Also, Duffy (along with Curt Welch) was perhaps as good an outfielder as there was during his prime. And this was -- as supporters of McPhee hasten to remind -- an era when fielding was more important than it is today.

Generally I find the Chet Lemon comparisions rather distracting at this stage. For starters, all of our candidates played shorter seasons (some like Start or McVey much shorter) than Lemon did. At least until we have to start compring the Chet Lemons (or even the Harold Baines' or Roger Maris') with the best leftover 19th century greats I would counsel against it.

Very tentatively....

1. Joe Start

Very tentative first pass. May flip Duffy and Griffin.

Jason Koral wrote: "Duffy was not really near the offensive player that Hamilton, Delahanty, Thompson or Tiernan was. It is not clear to me that he was superior offensively to Griffin; I suspect not. It seems to me that a LOT of weight has to be given to his defense to put him in the top 10, much less than top 3."

I agree. I've got Duffy behind all the other "normal" outfielders on the ballot: Hamilton, Browning, Tiernan and Thompson, and a little ahead of Griffin (he contributed more because of career length, in my opinion). Duffy will be somewhere between 10 and 15 on my ballot.

Childs was a good second baseman, but not a HOMer in my opinion. He won't make my Top 15.

Ezra,

John C or Chris Cobb, do you have the post that introduced the 30% fielding WS increase? I can't seem to find it. I want to go over the reasons for the different approaches between the increase and James's 70/30 split.

Thanks!

I'm no good at finding stuff in the archives, but it's easy enough to re-post my original reasoning.

WS presupposes a 67% - 33% basic division between pitching and fielding for defensive WS, which is likely about right for the modern game. This ratio is adjusted, based on the pitching and fielding records, but it never moves teams substantially away from this ratio. If you check the ratio for every team in WS, you may find some as low as 60-40, but nothing lower than that. Given the much lower rate of defense-independent pitching outcomes in the nineteenth-century game, (the more so the farther back you go), it seems clearly that pitching is systematically overvalued and fielding systematically undervalued in WS. Based on discussions of the pitching-fielding ratio of defensive responsibility in the 19th-century game, it seems better to me to assume a baseline 50-50 ratio between pitching and fielding. This is still too high for the NA and early NL, but I think it's pretty good for the rest. (You'd have to dig into other threads to find the evidence for this. I'm relying on others' arguments and research here.) The purpose of the 30% increase in fWS is to bring them more into line with a 50-50 ratio between pitching and fielding.

If the ratio of pitching - fielding responsibility is taken to be about 50-50 rather than 67-33, then fielding WS ought to be increased by 50% while pitching WS are reduced by 25%. Given the roughness of across-the-board adjustments, I've felt a 50% increase was too high to be safe, so I fixed it at 30%, which still gives excellent fielders a significant boost in the rankings. It still provides an assessment of defensive value significantly below what WARP offers.

The across-the-board increase assumes that WS still does a good job of dividing defensive responsibility among positions and still does a good job of distinguishing good fielders from bad fielders. I'm pretty confident that the latter is true (though there's still a lot of room for argument over questions like "Was Charlie Bennett a greater defensive catcher than Buck Ewing?"). I'm less confident about the former; its possible that early outfielders are overrated and early first basemen are underrated significantly more than other positions, but I'm not sure enough about those possibilities to apply a systematic alternation in fielding WS for these positions.

I did something similar with my own analysis. I calculated the pitching/fielding WS split between every team from 1876-1893. As you mentioned, no team has more than 40% as fielding, and that's because James says in his introductory materials that he capped fielding at 40%. I don't have the data in front of me, but as I recall, there were about 35 teams with the 40% cap, all of them had losing records, and only three of them had a team winning pct. better than .350.

I calculated the correlation between the winning pcts of these teams and the percentage of defensive WS that was attributable to the fielding on those teams, and the correlation was about -.75. In other words, the better the team record, the lower percentage that fielding WS represented of all defensive WS.

The mean fielding/pitching split for these teams was very close to what Bill James said it was for all of baseball (and as Chris said, 67.5/32.5). Like Chris, I was afraid a 50% adjustment was too drastic (though it may be warranted), so I changed mine to a 60/40 split (or an approximate 25% increase in fielding WS and 10% decrease in pitching WS). When I used bigger numbers, pitchers couldn't compete with any of the hitters in my HOMer rating system, and that didn't seem quite right.

I then applied this increase/decrease to individual seasons, but it is inexact, because I do not have the exact batting/fielding/pitching WS breakdown for every player season. So I used the career batting/fielding/pitching WS splits listed in the Win Shares "decades" section and assumed a player maintained that split during every year evaluated. I broke total WS into the three categories, increased the fielding WS by 25%, decreased the pitching WS by 10%, and recalculated total WS. I then adjusted them for season length (and a few other things I do, like an AA discount).

The effect of my adjustments was to (i) drop all the genero-pitchers (Mullane, Welch, etc.) off of my ballot, (ii) drop Galvin a couple of places relative to the hitters (and behind Spalding), and (iii) give McPhee and Bennett a boost on the ballot because fielding was a large part of their value. Some other players got boosted too (Williamson, for example), but not to the Top 15.

I should probably mention that I only made those adjustments for seasons through 1892. I'm considering whether to extend them (or use smaller adjustments) for a few more years, but I'm not sure how far to go with it.

Chris, how far out do you make your adjustments? I stopped at 1892 because of the pitching rule change, but clearly there is a period after that where unearned runs still constituted a large percentage of runs scored (and thus, defense was still playing a major role vis a vis pitching).

Jeff,

I'm thinking about this issue myself. My tentative plan is to gradually scale down the fielding adjustment between 1893 and 1920 to 20%, then drop it to 10% then, then phase it out between 1920 and 1940. Right now I'm dropping the adjustment to 27% after 1892, but I don't have any great justification for that number.

I would find it _very_ useful to see a study that looks at 1) increases in defense-independent pitching contributions and 2) decreases in errors from 1890 10 1920 to get a better fix on what's happening to the value of fielding in the era we are entering.

Chris,

Carl,

I know you asked the question of Chris, but I think you have to decrease pitching WS if you increase fielding WS, just to balance things out. A team has a finite number of WS, so if you increase a team's fielding WS to account for the greater relative importance of fielding, they have to come from somewhere. I take them from the pitching WS, because I am not prepared to reduce the hitting WS (which, overall, would increase the value of pitching as a percentage of total WS).

However, I don't think it works if you increase fielding WS by a percentage and reduce pitching WS by the same percentage. That's too big a hit on the pitchers and doesn't produce the right math.

Assume a team won 81 games in a 162-game season (243 WS). Suppose there are 127 batting WS, 78 pitching WS and 38 fielding WS. This might represent a fairly typical team. If you increase fielding WS by 25% (to 48), then you have to subtract 10 WS from somewhere -- in my example, pitching (68), which is a 13% reduction. If you reduced pitching by the same pct as fielding, you'd have 59 pitching WS, making 234 WS total, or 9 WS too few.

A more extreme example: Assume a team won 56 games in a 162 game season (183 WS). Suppose there are 73 batting WS, 66 pitching WS and 44 fielding WS. If you increase fielding by 25% (55), you need to decrease pitching to 55 (17% discount). This example is so extreme (involving a really bad team with a very high percentage of fielding WS), I think this means you would never discount pitching by much more than 17% if your fielding increase was 25%.

The discounts for pitching would be different every season and for every team if you chose to go through that exercise on a team-by-team, season-by-season basis. And that would certainly be more accurate. What I've done is tried to move the mean 67.5/32.5 pitching/fielding split to 60/40, so that there's an approximate 25% increase in fielding and 10-11% decrease in pitching across the board.

Incidentally, b/c Bill James uses an upper limit of 40% for fielding WS as a percentage of defense, and a lower limit of 25%, doing an across the board increase to accomplish a higher mean also increases the upper and lower limits to 50% and 31.25%, respectively, which I think works better for this era.

Carl,

I know you asked the question of Chris, but I think you have to decrease pitching WS if you increase fielding WS, just to balance things out. A team has a finite number of WS, so if you increase a team's fielding WS to account for the greater relative importance of fielding, they have to come from somewhere. I take them from the pitching WS, because I am not prepared to reduce the hitting WS (which, overall, would increase the value of pitching as a percentage of total WS).

However, I don't think it works if you increase fielding WS by a percentage and reduce pitching WS by the same percentage. That's too big a hit on the pitchers and doesn't produce the right math.

Assume a team won 81 games in a 162-game season (243 WS). Suppose there are 127 batting WS, 78 pitching WS and 38 fielding WS. This might represent a fairly typical team. If you increase fielding WS by 25% (to 48), then you have to subtract 10 WS from somewhere -- in my example, pitching (68), which is a 13% reduction. If you reduced pitching by the same pct as fielding, you'd have 59 pitching WS, making 234 WS total, or 9 WS too few.

A more extreme example: Assume a team won 56 games in a 162 game season (183 WS). Suppose there are 73 batting WS, 66 pitching WS and 44 fielding WS. If you increase fielding by 25% (55), you need to decrease pitching to 55 (17% discount). This example is so extreme (involving a really bad team with a very high percentage of fielding WS), I think this means you would never discount pitching by much more than 17% if your fielding increase was 25%.

The discounts for pitching would be different every season and for every team if you chose to go through that exercise on a team-by-team, season-by-season basis. And that would certainly be more accurate. What I've done is tried to move the mean 67.5/32.5 pitching/fielding split to 60/40, so that there's an approximate 25% increase in fielding and 10-11% decrease in pitching across the board.

Incidentally, b/c Bill James uses an upper limit of 40% for fielding WS as a percentage of defense, and a lower limit of 25%, doing an across the board increase to accomplish a higher mean also increases the upper and lower limits to 50% and 31.25%, respectively, which I think works better for this era.

Carl,

Freak years enable fluke performances.

What comes to mind first are performance in the year's direction, which tend to stand out in absolute terms: Hack Wilson's RBI in 1930, Bob Gibson's ERA in 1968. HR took a spike in 1994, and Matt Williams might have broken Maris's record had the season continued, and when Maris's record was broken, it was surpassed 6 times in 4 years. We all know of lists of players who hit unusual (for them) numbers of HR in 1987. But when we get into the habit of measuring performance relative to league, we can see something else: fluke years can enable unusual performances in the opposite direction. Gibson, going with the flow, had an ERA+ of 258 in the "Year of the Pitcher" - but in the 1994 freak HR year, Greg Maddux had an ERA+ of 273 (in far fewer innings).

This comes up because 1894 was one of the all-time freakiest offensive years. The team runs/game tower over the neighboring years. Hugh Duffy went .440/.502/.694 as part of a .324/.384/.449 career. Of course, we know that Gibson's 1968 is way out of line with the rest of his career - but we still aren't going to have any trouble electing Gibson. That thought isn't much help; Duffy's going to be hard for us to deal with.

There's a great "against the flow" outlier performance in 1894: Amos Rusie's ERA+ of 189 in 444 innings. That one year did weigh very heavily on the 1st ballot vote for Rusie, didn't it?

Billy Hamilton set the single-season run record in 1894. I read that as just another Billy Hamilton year - he had quite a few others of nearly the same quality. Since scoring runs was what he did, he rode the wave of 1894 to the record. In my mind, the idea that this year wasn't out of line for Hamilton is a good reason to rate him very highly on extended peak considerations.

Having read Jeff's post, let me add that if one were to go through and redistribute pitching and fielding WS on each team based on a reassessment of the pitching-fielding ratio, then I'd have much more confidence in pitching WS, but I don't have time to do that!

>Posted 10:03 a.m., August 6, 2003 (#62) - Ezra Sutton 'I have no problem with this rule. But, have you guys seen some evidence of strategic voting?' Somebody left me off their ballot entirely. I'm not accusing, but an expalnation should be warranted.

I have said before that I build a consideration set of players with the highest peaks (about 20-25 players per ballot) and then rank them finally about 50% peak and 50% career. When this method produces weak candidates at the bottom of the ballot I sometimes insert a player of high career value down there (McPhee, Sutton, etc.), other times I throw in a high peak/low career player, but hey the #10-15 isn't a HoMer anyway. I am in the process of recalculating my peak numbers (which include a 100% fielding bonus pre-'93) though I don't expect any FOES to thank me for that.

As to strategic voting, you gotta be kidding. Ezra has been bouncing around 12-20 on my ballot from day one. And who the he** could possibly have forecast that Spalding and Sutton would leapfrog Galvin and McPhee from #3 and #4 among the holdovers from 1905 and end up head to head.

But all of this has been posted before. Ezra, how many times is the same explanation warranted?

For the record, I wasn't Ezra in #62. I did laugh though :-)

Wasn't me, either.

ed, TomH has just posted what looks to me like a pretty good case for ranking McPhee ahead of Childs. What do you see in Childs that puts him #2 for you?

ed, TomH has just posted what looks to me like a pretty good case for ranking McPhee ahead of Childs. What do you see in Childs that puts him #2 for you?

Joe wrote: "I would just advise everyone to take those spots just as seriously as #1-9, and not just 'throw in' players."

I agree wholeheartedly. And while I'm not sure it has much impact on the final tallies, I'm don't like half votes. I see them in our final tallies, but what is the HOM rule on that?

Ran across this post #40 in the 1/6/02 - 1/12/02 archive, posted by jimd (who, by the way, has some very good posts in that particular archive):

"The following link may be of some interest to the HOM voters:

http://www.baseballprospectus.com/news/20020129davenport.html

The article by Clay Davenport is about the relative strength of the Japanese Leagues vis-a-vis MLB, but about 2/3 of the way down, he mentions similar ratings of the Federal League, the American Association, the Union Association, and the Players League, all of which have interest to us here."

Also in that particular archive are some discussions of strategic voting and how to prevent them, which might be interesting to those who weren't around back then. See post #13 from jimd and post #s109-116 from others, including me and John Murphy.

Here is my preliminary ballot.

As I've written in my ballot,I've been having alot of trouble reconciling WS and WARP. I like Chris's post about adjusting fielding WS, since I also believe that they seriously undervalue defense and overvalue pitching and it may put WS more in line with what I believe to be the players true value. Here is my prelim. ballot that I wrote from before I saw his post. I may not have time to use his adjustment before the next ballot (Work and all), but I will have it done by the ballot after that. (Hamilton's probably a lock this year anyway)

HOMer?s

% Breakdown of put-outs by half decade:

----- SO's C-SO Infd Outf -- SO's C-SO Infd Outf

Forgot to mention: Data source is the Lahman Database. (Errors in calculations are most likely mine, though I'll try to blame the DB software ;-)

So 1890's outfielders caught significantly more flies then their predecessors. One could speculate on a couple of possible reasons for that.

The questions that matter: what happened to the difference in value between a bad outfielder and a good outfielder? Did the increase in plays make outfielders more valuable, or was that mostly just an increase in routine plays?

We are considering quite a few 1890's outfielders now and in the next few years.

>Billy had a longer career, was likely better defensively

"Likely" or not, this needs to be checked out. Based on other skills, I would think it "likely" that Rickey Henderson is a better fielder than Don Mattingly. Not so.

First cut, career defensive WS: Hamilton 50.55 Thompson 28.3. Round one to Billy.

<I> To add to what Marc said (but to consult a different source) Baseball Prospectus has Hamilton as a below average fielder.

Prelim ballot:

1. Billy Hamilton (NA)

this has been said before but needs repeating: the fielding runs for warp methodology is published in bp 2002 (with possibly some minor adjustments - I havent been able to make it match up exactly but it is very close).

I know more about WARP methodologies than I do win shares methodologies. This is completely do to lack of effort on my part, since I havent bought any win shares books. Also, I know alot about TB fielding runs, and that only makes me not want to even consider using them. I would trust BP warp fielding numbers even if i had no knowledge about them before I would trust TB fielding numbers.

I think there is probably enough descriptions floating around the web/the bp site/bp books/usenet to reverse engineer warp if someone wanted to put the time into it. I've considered it, but dont have the time at the present.

On Hamilton vs. Thompson: I'm convinced by the raw fielding numbers that Thompson was a good right fielder: fairly sure-handed, strong arm, at least decent range.

However, I think the arguments running down Hamilton's defense are less convincing. Clearly, he did not have a strong arm -- I imagine that's why he played right field only one year. However, he did have enough arm and range to play centerfield effectively for many years. It doesn't make sense to dismiss his putout totals because he was playing center field -- he was getting those putouts because he could play centerfield, so he should get credit for them.

The discussion of who was playing CF instead of Hamilton early in his career is interesting, but it's not the whole story. When Hamilton goes to Boston, he stays in center field, while Hugh Duffy, whom most everyone seems to agree was a great outfielder, moved to left field. Either this decision calls into question the notion that managers had a particularly good grasp of outfield defense, or it suggests that Hamilton was a quite effective centerfielder, or both.

It's obvious that Hamilton was not a great defensive outfielder: he was slightly below average _for center field_, but there he's being compared, overall, to a group of very good defensive outfielders, and he was good enough to play center for an outstanding defensive team: the Boston Beaneaters of the late 1890s. I conclude, therefore, that he was somewhat more valuable than Thompson defensively, though the difference is not large.

Among the 90s outfield group now on ballot, I put them in this order based on defensive value:

Mike Griffin

Following up on the posts regarding fielding WS and pitching WS adjustments, both Chris and I said we were applying our adjustments through 1892 but were not sure where to draw the lines to phase out the adjustments.

With that in mind, I evaluated the percentage of unearned runs to total runs for each season. I used the unearned runs analysis because an upward adjustment to fielding WS means a downward adjustment to pitching WS, if you want to stay within the basic WS methodology.

The following are the noticeable shifts in percentage of unearned runs. I used a 3-year historical moving average for smoothing, and I refer to the shifts as noticeable because they represent the biggest year-to-year moving average shifts. I also excluded the AA from the analysis because of quality-of-play issues, and excluded UA and PL from the analysis because the one year data doesn't add much.

1876 60% from 63%

I purchased the WS Digital Update, so I now have batting/fielding/pitching WS splits for every player season. Using the criteria in post #97 (modified to reflect comments), I plan to do the following:

For the NA years, increase a player's fielding WS by about 54% to reflect the 50/50 pitching/fielding split (that is, moving the fielding from 32.5% of defense to 50% of defense). Since Bill James didn't do team pitching/fielding WS splits for the NA, I'll then just reduce a player's pitching WS by about 26% (that is, moving the pitching from 67.5% of defense to 50% of defense). For hitters, I'll use the WS conversion formulas from WARP that Chris developed. For pitchers, I'll examine their post-NA years to determine a ratio between actual wins and WS, and apply that same ratio to the NA wins to get their NA WS. For determining how those WS are split b/w hitting/fielding/pitching, I'll use the same ratios as published for their entire careers in WS.

Beginning in 1876, I will increase a player's actual fielding WS by the appropriate percentage (e.g, about 54% through 1880, 32% through 1886, 23% through 1893 and 14% through 1909). Then, instead of applying a blanket percentage decrease to pitching, I'll do it on a year-by-year basis, as follows:

I'll examine the fielding WS and pitching WS for a player's team during a particular season. I will increase the team's fielding WS by the same percentage as I increased the player's fielding WS. However many fielding WS that increase produces, I will subtract from the team's pitching WS (so that total WS still balance out), and determine the percentage by which I have decreased the team's pitching WS. I'll then decrease the player's pitching WS by that same percentage. It will be different for each year and for each player, because the pitching/fielding splits for each team are different -- though they produce a mean of 67.5/32.5.

I'll apply all this to unadjusted WS.

This will involve a lot of work, so let me know what you think about the process. If we can get a process nailed down, I can then publish here the cumulative changes in total unadjusted WS for the significant players, and everyone can make their own subsequent adjustments for season-length, league quality, etc.

Chris, I don't disagree with anything you said. My point was not that Sam Thompson was a better OF than Billy Hamilton but that the "image" of Hamilton as a great OF and of Thompson as a lousy fielder need to be examined. Hamilton did make the plays that a CF is supposed to make--the ones Jim Burns, Eddie Burke, Billy Sunday and Ed Delahanty made. I can't say Sam would have made them or that he wouldn't.

Good pt. about Duffy and Boston. I missed that.

All in all, without further analysis I'm sure Ham and Sam are both top 5, maybe higher. Billy gets on base, Sam drives him home. Think Rickey and Donnie.

this has been said before but needs repeating: the fielding runs for warp methodology is published in bp 2002.

They did? I have to search for my copy to see if the metric makes sense to me via their explanation.

I would trust BP warp fielding numbers even if i had no knowledge about them before I would trust TB fielding numbers.

I'm in definite agreement with you there. Without comparing formulas, I trust BP to a much greater degree than TB.

Jeff, great work. I have been recalc. WS fielding numbers (and obviously total WS for position players, having already finished reducing the pitching WS) through '92. About halfway through I realized it would have made sense to use a sliding scale rather than the same ratio throughout and then coming to a dead stop. I'm going to complete my current method, however, since I well into it, but I like your method a lot. Can't wait for the results!

BTW, thanks to Chris and Jeff for answering my question concerning the pitching/defense split. I'll be studyng it the rest of the week and make some followup posts later on.

Interesting excerpt here re the genesis of the NL, with a reference to Ezra Sutton that supports the notion he was a very highly regarded player:

"An incident that occurred in 1874 provided the spark to Hulbert's National League of Professional Base Ball Clubs. This was a time of rampant "revolving," or contract jumping, and the White Stockings were nervous that their diminutive shortstop, Davy Force, would desert them at season's end, as he had done with several other clubs. In September 1874 they signed him to a renewal contract for 1875; then they learned that, because the season was still in progress, the contract was invalid by Association rules. Chicago signed Force to another contract in November, but the organization blundered by backdating the contract to September, thus voiding it once again. In December the Philadelphia Athletics offered Force a contract, and he signed it. The Association Council, led by a Philadelphia official, upheld Force's deal with the Athletics.

Albert Goodwill Spalding, the great pitcher of the Boston Red Stockings, met with a distressed Hulbert in mid-1875. Years later Spalding would write of the wounded civic pride and, by modern standards, paranoia that afflicted Hulbert: "It was borne to him one day that the reason why Chicago, whose phenomenal achievements on other lines were attracting the wonder of all the world, could make no better showing on the diamond was because the East was in league against her; that certain Base Ball magnates in the Atlantic States were in control of the game; were manipulating things to the detriment of Chicago and all Western cities; that if the Chicago Club signed an exceptionally strong player he was sure to be stolen from her; that contracts had no force, because the fellows down East would and did offer players increased salaries and date new contracts back to suit their own ends."

Hulbert reminded Spalding that he had made his name as a pitcher in the 1860s through his exploits with the Forest City club of Rockford, Illinois. "Spalding, you've no business playing in Boston; you're a Western boy, and you belong right here." Within a few months, Hulbert proceeded to give the Easterners, who had, in his view, rustled his prize shortstop, a taste of their own medicine. He not only raided Boston for Spalding but also snatched three other Red Stocking heroes, Ross Barnes, Deacon White, and Cal McVey. From the hated Philadelphia Athletics he took Cap Anson and Ezra Sutton, although Sutton later reneged on his commitment and returned to Philly.

When word leaked in the summer of 1875 that Chicago had denuded Boston of its stars for the following season, a columnist for the Worcester Spy wrote of Boston's loss: "Like Rachel weeping for her children, she refuses to be comforted because the famous baseball nine, the perennial champion, the city's most cherished possession, has been captured by Chicago."

Now Hulbert had real cause for worry. Because his club's contracts had been signed, yet again, in midseason, the Association Council could invalidate them and, perhaps, expel Chicago for gross misconduct. Then he came up with a brainstorm. "Spalding," he said to his ally in revolution, "I have a new scheme. Let us anticipate the Eastern cusses and organize a new association before the March (1876) meeting, and then see who does the expelling."

On February 2, 1876, at the Grand Central Hotel in New York City, Hulbert met with other large-city magnates and persuaded even the Eastern faction that a new circuit should be founded: the National League of Professional Base Ball Clubs."

Dan, interesting stuff. Especially Spalding's own characterization of Hulbert as suffering from "paranoia." It seems from this summary that the NABBP was right in voiding Hulbert's contract with Force, but it also appears that Hulbert had no real basis for thinking the NA would void his own VALID contracts. The contracts that were voided and/or that he feared would be voided were in clear violation of the NA's rules. It is the easiest dodge in the world to conjure up imaginary wrongs in order to justify one's own means.

I'm starting to think that one of the reasons that Sutton doesn't get mentioned that often as a contemporary star is that he didn't PLAY with the guys who got asked that question. Check out the forgettable list of teammates in the mid-1880s during his 'second wind.' I doubt sportswriters were asking those guys who impressed them most; they weren't important enough.

I apologize if I missed this somewhere, but I see a lot of posting about adjusting Win Shares for the difference in defensive responsibility between pitching and fielding (which I agree with), but I don't see much discussion about adjusting for the difference in responsibility between catching, infielders, outfielders, and the defensive spectrum.

In particular, adjusting fielding win shares upwards for outfielders due to the difference in pitching vs. fielding responsibility in the 19th century and NOT adjusting fielding win shares downward for outfielders due to the difference in outfield vs. infield/catcher responsibility will result in outfielders being overvalued and catchers and infielders being undervalued.

Dan Greenia,

Is the estimation of Win Shares for NA1871-75 by Chris Cobb explained in http://www.baseballprimer.com/hom/archives/00000084.shtml#28 and the immediately following exchange with JohnC (#29-30)? That is, I think, simply 2.75*WARP1 for single season and career measures; more complicated for peak measure.

The adjustment of fielding and pitching shares, discussed above, seems to be a separate matter.

Is estimation for 1871-75 obsolete, given the "Update" by Bill James?

Paul Wendt

This is in response to Chris Cobb's information in post #57.

Win shares are constrained to team wins, so a team that underperforms its Pythag squeezes down the WS of its players - maybe fairly, maybe unfairly. Of course with a truly awful team like Worcester '82, the Pythagorean estimate may not be all that accurate. On the other hand, maybe the Pythag is as "good" as it was because the team was overproducing runs - but Chris didn't find any evidence that they were. On the other hand ... that's too many hands. Too much noise, maybe no signal at all.

The WS listing had Richmond (the primary pitcher) at 13 WS, Stovey at 10 (10.3). Per 162 games, that would be 25 for Richmond (except you can't do a linear extrapolation for a pitcher!) and Stovey 20. A 20 WS season isn't much of a HoM argument. So is that fair?

One thing I found is that the pitchers are probably getting too much credit. Richmond should get a fair amount of credit for his hitting, since he was the second best hitter on the team. But if you look at runs allowed rather than earned runs, something unusual emerges. In the 1882 NL, 87 unearned runs scored for every 100 earned runs - but Richmond allowed 172 unearned runs to 171 earned runs. His ERA+ is 82 but his RA+ would be about 76. Given average offensive support, he could have had an 18-31 record; given the support he had, his 14-33 record is in line. It turns out that Richmond allowed the lowest proportion of unearned runs; the other pitchers allowed significantly more unearned runs than earned. Both Mountain and Corey had ERA's better than Richmond but RA's that were worse than Richmond, and Clarkson in his 3 games had an RA+ of 50. Corey had an RA+ of about 67 or 68, which is pretty close to the 0 WS line. Most of the underperformance of the team Pythag came with Mountain, Corey, and Clarkson, who were a collective 4-31.

That brings us back to Stovey. He's getting very little defensive credit. He played half 1B and half OF, and committed more than his share of outfield errors (of course OF errors are usually less costly than IF errors). His line of .289/.330/.422 just isn't all that impressive. The STATS encylopedia lists that as 69 RC - but he did score 90 runs. In all of his other years, his R are pretty much in line with his RC, but in that year they're way out there. He scored 23.7% of his team's runs that year, and that stands as the all-time record. Chris said he's being given 27% of the team's batting WS. Maybe that's enough, but maybe it isn't. Take a little credit away from the pitchers, shift a little more batting credit Stovey's way for the runs scored ... what can you get the WS to? 12? 13? 15?

I'm still holding that Stovey's 1882 was an extraordinary performance, but I'm not sure how strong I can make that argument.

Paul:

The Hulbert info is from the linked homepage above.

<i>Runs scored per out (at bats minus hits), career leaders, minimum 1500 runs scored

In 4th place ..Ted Williams ....356

Paul wrote: "Is estimation for 1871-75 obsolete, given the "Update" by Bill James?"

You might know of something out there that I don't, but the "Digital Win Shares Update" that I mentioned is a bit of a misnomer. It doesn't update to 2002 or add the NA years. It has three sections: (1) a breakdown for every player year of the batting/fielding/pitching WS splits, (2) a seasonal breakdown of Top 12 players and Top 5 fielders at each position and (3) year of birth charts (which don't seem that interesting to me).

Thinking about Stovey, Thompson, and Hamilton:

They all had 14-year careers, although Thompson got so little at the end that you might as well call his a 12-year career. Thompson played 1407 games, Stovey 1486, Hamilton 1591. Normalize that for season length, and Stovey moves slightly ahead of Hamilton in games played. We will hold Thompson's shorter career against him, even if it is mostly a matter of getting off to a late start.

Stovey aged the best, with the evidence including leading the PL in SB at the age of 33 and the NL in slugging at the age of 34.

Stovey played about 1/3 of his games at 1B, mostly early in his career, and divided time at all three outfield positions. Early in his career, he committed more than his share of errors in the outfield; that cleared up later. I don't sense that he was really a liability anywhere, but he also wasn't going to win any Gold Gloves. Thompson was a right fielder and a good one in the classic strong-arm mold. Hamilton started his career in LF and later moved to CF. We've seen that career pattern in our own time - think of Willie Wilson and Brett Butler. It suggests speed, it hints against much arm, and it really doesn't say anything about quality but doesn't rule out being an outstanding CF. I would have the defensive order as Hamilton > Thompson > Stovey, but the differences may not be huge.

Raw data:

KJOK wrote "I apologize if I missed this somewhere, but I see a lot of posting about adjusting Win Shares for the difference in defensive responsibility between pitching and fielding (which I agree with), but I don't see much discussion about adjusting for the difference in responsibility between catching, infielders, outfielders, and the defensive spectrum."

Some adjustment to fielding WS may be required for the different defensive spectrum -- I don't know -- but Bill James was certainly aware of the different defensive spectrum and he has separate fielding WS formulas to attempt to take some of those things into account. I have not evaluated whether he did a good job, but he took a crack at it. There's an essay in the WS book about the defensive spectrum shifting around 1930 (particularly for 2b and 3b).

The adjustments I've been advocating have been to correct a blind spot in fielding WS generally, because defensive WS are geared towards a 67.5%/32.5% pitching/fielding split, and most of us agree that doesn't work during the early years. If James' formulas are "right" with respect to the different defensive responsibilities of individual positions, my adjustments ought to keep everything in the same proportion. If he's wrong, I don't feel qualified to quantify how.

Are the Digital Win Shares Updates online? Maybe I missed a post, but where?

Sam Thompson is a classic slugger with more SA than OB? Perhaps. But c'mon, his OBA is .384. The top 100 of all time bottoms out at .388. I'm really not clear re. exactly what weaknesses Sam really had. Classic slugger, plenty of power, but not in the Dave Kingman mode, he hit .331 with .388 OBA. In his only World Series he hit 2 HR, 7 RBI and .362. Good RF in the strong-arm mode. He "only" played 14 years...same as Hamilton and Stovey, yet that seems to be a complaint re. Sam and not the others. I'm confused.

Are the Digital Win Shares Updates online? Maybe I missed a post, but where?

Here you are, Marc:

Digital Win Shares Update

Good RF in the strong-arm mode.

Are there any contemporary references to this? As I have posted before, the only reference that I have seen regarding Thompson's fielding was a negative one from his obituary in 1922.

Marc,

Thompson had appearances in 15 seasons, but if you add together 1897, 1898, and 1906, you get 25 G, 107 AB. That's why I was calling it really a 12-year career, a year or two short of Hamilton and Stovey.

On his career he exceeded league BA by .055, OBP by .041, and SLG by .129. He appeared on league leader lists at some point in nearly everything, but his OBP gray ink list is the shortest and least impressive of the lot. That's why I said more SLG than OBP. And I wasn't thinking of Kingman - I was looking higher than that, but there are many good cleanup hitters in history.

At that, I am considering moving him up from where I had him.

Joe, I don't disagree with anything you say. My point was to counter the "buzz" I've been reading here which seemed to be:

Hamilton was a really, really good CF because he was so fast!

I agree with KJOK that the 19thC fielding position weights for Win Shares are probably markedly different.

The PO breakdowns that I published above seem to indicate that Catchers played a much larger fielding role than they do today. I suppose one can argue that game-calling has become more important (due to larger pitch repertoires) as their number of actual fielding plays has dropped, therefore little has changed in terms of value. But it's also true that 19th century catchers were much worse hitters than any of the other fielders (excluding pitchers), unlike today when they are competitive with the middle infielders. I think this adds up to pre-1920 catchers being underrated by Win Shares.

I'm not sure what to make of the OF play of the 1880's. They are not making many putouts but the assist rates are phenomenally high; the assists are either due to very aggressive baserunning (possibly even more so than the deadball era), or a different style of OF play (much closer to the infield with very few batters capable of muscling the ball over the OF head), or some of both, of course. Anybody have any knowledge to share on this? We did seem to establish on that thread that RF'ers (particulary in Chicago) appear to have experimented with a "short-fielder" style of play. In the 1890's PO's are up, A's are down; OCF has a good point asking whether the increase in total plays made reflects an increase in value or just an increase in routine plays. That I don't know, though we do know that they became better hitters relative to the league, FWTW.

I don't have any data either but re:

>whether the increase in total plays made reflects an increase in value

I suppose a high fly with time for an outfielder to get there - to us a routine play - was a little more of an adventure before they wore gloves.

JeffM wrote

One things I do know about OF Defense - although we don't have much info on park dimensions, the info we do have suggests that for some reason many parks thru the 1880's had deeper LF's than RF's, and it appears that teams often played their worst fielders in RF, probably because the ball wouldn't be able to roll very far away?!

These parks all had significantly closer RF fences than LF fences:

KJOK, I think you are right about the separate formulas applying only to 2b, 3b and C. In my opinion, that's where most of the adjustments should be made in terms of adjusting for the spectrum, but maybe there are some other adjustments that can be made. We don't have separate fielding data on the three outfield positions, so I'm not sure how we would adjust among the outfielders.

Maybe, as you say, there should be increased importance of infielders over outfielders. Seems like James has some "intrinsic ratings" he gives to the various positions, so maybe that's where the tweaking must be done.

James messes with the fielding weights of 2B and 3B pre-1930. He messes with how Catchers shares are calculated at different times, but that's because the available fielding data changes (like RC keeps changing depending on the batting stats kept); the weight of the catcher within the overall fielding schema stays the same though.

I do recall that for 19thC Catchers, James cut the effect of passed-balls in half, saying something like "otherwise they swamp the system". That struck me as very odd (and also is a possible source of why BP and WS disagree about Ewing and Bennett in fielding).

jimd wrote: "I think this adds up to pre-1920 catchers being underrated by Win Shares."

That may be true, and as my response to KJOK says, we may have to look at James' intrinsic values...because otherwise, I believe he has taken into account the greater role in fielding bunts, etc. and given the catchers some benefit of the doubt where the numbers are concerned. He also acknowledges that his weights for the various catcher plays are certainly arguable, because we all know that a large part of what catchers do cannot be reflected in the statistics.

As I've said before, I'm interested in how many runs a guy saves at a particular position, whether through tangible things like fielding bunts and catching base stealers, or intangible things like calling a good game. By the way, although there are more types of pitches now, not many catchers are actually calling games today -- that happens in the dugout, so I would argue the old catchers had much more value there. Current catchers are just relaying signals.

I'm willing to give nods to old catchers for not having the benefit of shinguards, gloves, etc., and to all catchers for the general wear and tear. I give catchers an offensive bump on the assumption that they might be too tired or beat up to hit as well as some other guys. But I don't give them a DEFENSIVE bump simply because they are willing to take a greater beating. They get defensive credit for intangibles (hard to measure), and the plays they make.

Since the Charlie Bennett thread started a few weeks ago, I've paid close attention to the activities of current catchers (primarily Javy Lopez and Braves' opponents), and frankly, I don't see catchers as more valuable today than most middle infielders or center fielders. I know it isn't a popular thing to denigrate the defensive value of catchers, but: they aren't calling games anymore, they field maybe one or two bunts per game (plays which virtually all of them make) and they throw to second every once in a while (without great success because of slow pitcher deliveries). There's rarely an actual play at the plate, and there are few pop-ups that aren't fielded by someone else with a more flexible glove. I'm beginning to think that most of a catcher's value in the modern game is ability to prevent wild pitches (which isn't reflected in statistics) and their willingness to sit back there doing a job no one else on the field wants.

I guess what I'm saying is that the old catchers were probably much more valuable defensively because they were involved in more plays. If we see evidence that James doesn't take into account that they were making more plays, an adjustment to fielding WS is warranted. If we simply want to boost catchers defensive ratings because they took a beating and were poor hitters, then an adjustment to fielding WS isn't warranted in my opinion.

Not only that, but hasn't DIPS shown also that the value of the catcher even if he DOES call the game and "handle" the pitcher is also virtually NIL--i.e. that there are only insignificant differences among catchers in the DIPS results that they get. (And of course everything else is dumb luck anyway ;-)

Chris Cobb - I am just now getting a chance to dig into your NA WS.