Yes, but "All Star", I meant "would/should have made the All Star team," not "Was the best first baseman in the league." Otherwise the comparison to Baines is nonsensical. Was Baines ever the best RF in his league?

It's the hypothetical Top 25 in the league (or is it 30 or 32 now?) fairly evenly distributed across positions, and without the "every team represented" rule, not the Top 1 at the position. For many of the years jimd lists, Beckley was the #2 guy, and would therefore be a legitimate "All Star".

The relevant analogy may be Rafael Palmiero, who has made 4 All-Star teams despite rarely being the top first baseman.

Also, I should add that I don't see Harold Baines as obviously a horrible HoM selection. (Perhaps in the "Harry Hooper"range.) It'll depend on context and strength of competition, but I certainly see him at least making the ballot somewhere. (I'll have Hooper around 10 or 12). It's not like we're comparing him to Jose Cardenal.

OK. Beckley's an All-Star in that sense. I prefer the 1st-team All-Star meaning (which is what I used in the Dunlap resume).

That's quite a HOF class of '71. Paige ... (huge gap) ... Kelley, Hooper, Bancroft, Beckley, Hafey, Marquard.

Really, only one player of note here.

Players Passing Away in 1930

Candidates
Age Eligible
69 1898 Emmett Seery-LF
62 1905 Jack Stivetts-P
61 1903 George Hemming-P
60 1913 Frank Kitson-P
52 1923 Rube Foster-P/Mgr

Future Candidates
38 1936 Hal Carlson-P

Um, did anyone notice that BP's WARP3 calculations have changed? When did this happen? And does anyone know what was changed? Every player seems different (some a little, some a lot).

Browning's pitching WARP problem, FWIW, was finally corrected it appears.

I had just noticed that there's been reformatting going on, didn't know the numbers had changed as well. I'm rather excited that one can now get to team listings . . .

Could those who have old WARP data stored take a look at the new and see if they can deduce what the systemic changes are?

Just browsing the new listings it seemed like BP got rid of BRARP and replaced it with BRAA.

I just noticed it myself as I compile my lists of All-Stars, Gold Gloves, Silver Bats, and Bronze Arms.

Jennings can no longer claim 4 consecutive #1 seasons. George Davis now wins 1897 by a slim margin (it used to be vice-versa).

Joe will be ticked with the disappearance of BRARP. (I don't think Tango's system works anyway; about 1/4 of the regular position players appear to be below Tango-replacement in a spot check I recently ran.)

So, Ryan and VH are tied for career value under the new WARP metric. WS has VH something like 30 ahead, but that difference is almost entirely based on the fact that WARP gives him very little credit for his pitching while WS gives him a substantial amount.

Well, I guess we will see a shakeup with the ballots now...

Well, whatever the changes are, they still don't change my mind, which I made up five minutes ago, that BP's system is next to worthless. Here's why. Pick a team. How about the 1908 Washington Senators. That team went 67-85, with a Pythagorean record of 68-84. But the team is credited with 74.4 WARP1. How can a team be credited with more WARP than actual or Pythagorean wins? First of all, an all-replacement team wouldn't go 0-152; I'd peg it at about 46-106. But it sure as hell wouldn't go -6.5-158.5, that I know. How on earth do they account for this?

I'm not saying my WARP methodology is perfect by any means; it has tons of kinks to work out. But at least I am using an actual, consistent replacement level of a .300 winning percentage.

Dan,

Ouch! That bit of data doesn't give one much confidence in WARP. At least they're _showing_ team data like this, though, so we can check their findings against actual outcomes more readily . . .

Hmm. On third reading, that last post sounded like it might be making an implicit criticism of Dan's work and a defense of WARP. No criticism intended, and WARP's treatment of the 1908 Senators is obviously indefensible -- I am glad that WARP is showing us data at the team level, finally, so that we don't have to take its evaluations of individual players so much on faith.

Its not just their weird idea of a replacement level -- they seem to try and use (FRAR - FRAA) as a way of doing a positional adjustmentment -- their "AA" numbers are off too.

PRAA = -13
FRAA = -18
BRAA = +37

The Senators average game score was 3.09-3.48 in a 3.44 R/G league. Strangely its the pitching and defense which looks about right there, its the offense that's way off. They should be about 53 R below average. No CS data or other OOB, but 90 runs is a big swing and its usually offense that these systems can get quite right.

Its only one team, I suppose.

The Senators' park was a strong pitchers' park that year. That should explain part of the assessment. I'm always assume I am not handling adjustments for park factors properly, but it looks to me like the Senators' offense was still below average, but only by 5-15 runs.

I emailed BP yesterday about the changes to WARP, and they said that they haven't yet changed the WARP numbers for pitchers. That may be why the totals aren't adding up quite yet. They said they would be working on that over the next few days.

They also said they'd adjusted the fielding numbers to, among other things, better reflect infield-outfield ball distribution and to account for defensive park factors.

They indicated that an article will be forthcoming.

Well, maybe they're changing them again, but looking at it, the pitcher W3 numbers HAVE changed since we got the "new" WARP late last year. Here are the leading candidates for 1931-32 with the numbers pulled off their site today:

***1931 (August 1)—elect 1
WS W3 Rookie Name-Pos (Died)
321 95.7 1909 Harry Hooper-RF (1974)
265 79.7 1913 Bobby Veach-LF (1945)
290 75.5 1913 George J. Burns-LF (1966)
208 44.6 1909 Rube Marquard-P (1980)
160 45.7 1913 Dutch Leonard-P (1952)
188 42.8 1914 Milt Stock-3b (1977)
159 34.8 1913 Casey Stengel-RF (1975)

1932 (August 15)—elect 2
WS W3 Rookie Name-Pos (Died)
266 65.1 1913 Wilbur Cooper-P (1973)
215 70.4 1913 Hooks Dauss-P (1963)
243 60.1 1909 Babe Adams-P (1968)
227 51.1 1910 Stuffy McInnis-1B (1960)
206 59.6 1918 Ross Youngs-RF (1927)
142 38.5 1914 Everett Scott-SS (1960)
134 37.0 1912 Hank Severeid-C (1968)

The biggest thing I notice is the NL deadball players usually gain a lot with their new formula.

Oops. Stengel's new WARP3 is 40.5.

Is there any reason to put stock in BP's unexplained changes to their proprietary numbers? Maybe this is a misconception, but it seems to me that there's no solid basis for evaluating their system and their changes other than rumor.

Myself, I'd rather hold a seance to give Frankie Frisch a vote instead.

I tried checking some other boards to see what they thought of the new WARP but unfortunately, about half the internet seems to be down this morning...

No MyDoom for HOM though! Yeeha!

Trying a long shot here, but does anyone have an excel spreadsheet with 1871-2003 park factors that they wouldn't mind sharing with a newbie? If so, THANK YOU! I'll be much obliged if you could email to me at eric.chalek@heinemann.com.

Trying a long shot here, but does anyone have an excel spreadsheet with 1871-2003 park factors that they wouldn't mind sharing with a newbie? If so, THANK YOU! I'll be much obliged if you could email to me at eric.chalek@heinemann.com.

Oh, yes, I have park factors. What kind do you want, however, 3-year factors calculated like Total Baseball 8? 1-year factors calculated on simple home divided by away? Some othe combo???

In case others want them, I've posted the Total Baseball 8 3-year BPF and PPF numbers to the Hall of Merit egroup :

Total Baseball Park Factors

I was in Cooperstown Monday. I was unable to find the name and likeness of the following HoMers on display. Some may be in uncaptioned photographs; I may have missed others. In any case, they have to be considered among the most neglected great players in pre-1920 baseball.

Ross Barnes
Bob Caruthers
Bill Dahlen
Frank Grant
Pete Hill
Paul Hines
Home Run Johnson
Sherry Magee
Joe Start
Ezra Sutton
Deacon White (recognized him in a team photograph).

For him [Beckley] to make your ballot, let alone the HoM, one of three things need to be true:

(1) Consistently ranking among the top 30-40% of major league regulars for 18 years makes you worthy: Under some of your systems this may be true, but I'm never going to buy it.

This is exactly why I have Beckley high on my ballot. How many other players can make the same claim? Only a couple dozen or so eligible players have career lengths comparable to or longer than Beckley. All of them have been inducted except for the following, in order of what I see to be their relative level of consistency/"domination":

1. Jake Beckley
2. Harry Hooper
3. Deacon McGuire
4. Lave Cross
5. Kid Gleason

My ballot in/out line is between Hooper and McGuire. Sure, it's an arbitrary choice of where to draw the line on long careers without much peak, but it's a choice that's supported by more than 60 years of HoF voting patterns.

I haven't posted a ballot in about 15 "years" or so. It's been fun researching and reading the threads to get back on the horse.

this got me asking how many voters who once voted no longer vote?

I was keeping track of this up until a few elections ago (I'll catch it up to date when I get a chance). Through 1926, I have a total of 72 distinct voters (as far as I can tell; if someone changes their name and doesn't tell anybody, I won't know). Many of the missing only vote for a handful of elections or less and then disappear (e.g. Silver King, chocaholic, Dr. Feelgood, Shawn Weaver). Others "hit the wall" at about five/six months or so (David, thebigeasy, Jason Koral). The record was 16 elections before dropping (redsox12) though I think Yardape has broken that (20 votes through 1926, though I don't recall seeing his ballot lately).

I'd estimate there's 1 lost voter for every 2 active ones.

Apropos of nothing, we all know of the infamous Mathewson-Rusie trade.
Have any other HOMers so far been traded for each other?

Of the 29 voters who cast ballots in 1898, 27 (93%) are still active (David and thebigeasy are not, at least under those handles). 24 of the 46+ voters who've joined us since are still active, about 50%.

EricC #2-25
I was in Cooperstown Monday. I was unable to find the name and likeness of the following HoMers on display. Some may be in uncaptioned photographs; I may have missed others.

Did you find more than name and likeness (individual portrait, I guess) on display for the other HOMers who are not HOFers?

Did you find more than name and likeness (individual portrait, I guess) on display for the other HOMers who are not HOFers?

Most of the images of HoM/non HoFers that I recall were either on championship team photographs or on baseball cards, though Joe Jackson was part of the Black Sox display.

Recall the list of Keefe and Welch catchers (above, end page 2). Briefly:

Tim Keefe (594 starts, 39 shutouts)
GS ShO Catcher
185 13 *Bill Holbert (31%)
164 15 *Buck Ewing (28-%)

Mickey Welch (549 starts, 41 shutouts)
277 28 *Buck Ewing (50+%)
73 5 *Bill Holbert (13+%)

Buck Ewing worked 38.6% of Keefe & Welch career starts, 53.8% of their shutouts. His shutout rate with the two pitchers, almost 10%, was much higher than the rate for other catchers.

Thus Welch enjoyed some advantage over Keefe in his battery mates.

Catcher Records with Keefe & Welch in sum
GS ShO=complete game shutouts
1143 78 All Catchers (6.8% shutout rate)
441 43 Buck Ewing (9.8%)
702 35 Other than Ewing (5.0%)
258 18 Bill Holbert (7.0%)
444 17 Other than Ewing & Holbert (3.8%)
092 2 Jim O'Rourke
083 7 Pat Deasley (8.4%)
060 3 Willard Brown
235 5 Other than Top Five (2.1%)

Note: Catcher shutouts are complete games. Keefe & Welch each pitched one shutout with multiple catchers.

The differences in shutout rates are much greater than I would have guessed.

I just put up a query post on Mendez, eligible Monday, and it hasn't appeared on Hot Topics. We should probably make sure he's linked in the right place, because he will presumably be a very Hot Topic, being 1932's main uncertainty.

Thanks for posting the PFs, KJOK!

For those of you keeping track, how many of us have voted in every single election?

As of 1926 there were 20, and that has dropped by at least one. I'll have to update my list.

About the Keefe/Welch catcher stuff...

Holbert was in the AA with Keefe in 83-84 while Ewing was in the NL with Welch. That's 128 starts in the AA for Keefe where Ewing could not possibly have been his catcher. That makes up much of the battery mate difference.

Karl - which thread is your query on? The Mendez thread hasn't been posted to since early July.

since Welch’s record is based on what I posted before from the retrosheet he gains 7 wins and 3 losses on his record

assuming Welch’s team scored 5 runs every game he started he would have
294 wins 236 losses and a 555 winning percentage
assuming Welch’s team scored 6 runs every game he started he would have
349 wins 181 losses and a 658 winning percentage

This is how many run per game were averaged by the NL as a whole during Welch’s career
1880------------------4.7
1881------------------5.1
1882------------------5.4
1883------------------5.8
1884------------------5.5
1885------------------5.0
1886------------------5.2
1887------------------6.1
1888------------------4.5
1889------------------5.8
1890------------------5.6
1891------------------5.5
1892------------------5.1
1880-1892-------------5.3

This is exactly why I have Beckley high on my ballot. How many other players can make the same claim? Only a couple dozen or so eligible players have career lengths comparable to or longer than Beckley.

I doubt it's that many. If you normalize every season to the same number of games, here's the leaders among eligibles for most seasons worth of games played:


1. Jake Beckley 16.8
2. Lave Cross 16.0
3. Tommy Corcoran 15.3
4. Harry Hooper 15.2
5. Jimmy Ryan 14.6
6t. George Van Haltren 14.2
6t. Tommy Leach 14.2

I may have transposed GVH & Ryan. That's everyone over 14 that I know of. Beckley has a fairly decent career quantity lead right now.

My adjusted WARP (uses FRAA, as opposed to FRAR) career totals . . . using new WARP.

Pitcher          adjWARP2
Cy Young          208.7
Christy Mathewson 141.7
Kid Nichols       132.3
Eddie Plank       109.6
Amos Rusie         98.3
John Clarkson      93.8
Tim Keefe          86.4
Ed Walsh           83.7
Clark Griffith     82.7
Vic Willis         79.0
Mordecai Brown     78.6
Rube Waddell       78.3
Ed Cicotte         76.2
Charley Radbourn   76.1
Bob Caruthers      71.6
Pud Galvin         70.0
Joe McGinnity      68.4
Jim McCormick      64.0
Al Spalding        58.8
Addie Joss         58.2
Silver King        57.8
Tony Mullane       55.1
Rube Marquard      52.3
Mickey Welch       28.2
Tommy Bond         23.4

Batters WARP2, this is an incomplete list, I've only got HoMers through 1902 (Brouthers, not Ewing) and more recent electees in here so far, but I figure it doesn't hurt to list what I have.

Player           adjWARP2
Dan Brouthers      112.2
Jim O'Rourke        96.6
Sam Thompson        80.8
Jake Beckley        80.4
Paul Hines          79.8
George Gore         78.9
Jimmy Sheckard      78.0
Ross Barnes         74.5
Harry Hooper        72.5
King Kelly          71.5
Bobby Wallace       69.7
Deacon White        65.9
Fielder Jones       65.4
Hugh Duffy          64.9
Mike Griffin        64.9
Jimmy Ryan          64.2
Ed Konetchy         64.0
Mike Tiernan        62.7
Bobby Veach         62.5
George Van Haltren  60.8
Home Run Baker      58.3
Roy Thomas          57.7
Joe Start           56.7
George J. Burns     56.7
Joe Wood            56.0
John Ward           54.6
John McGraw         53.6
Jake Daubert        53.3
Hughie Jennings     53.0
Tommy Leach         53.0
George Wright       51.9
Lave Cross          50.7
Frank Chance        50.6
Cupid Childs        50.3
Charley Jones       48.1
Pete Browning       47.7
Lip Pike            47.4
Tom York            46.5
Miller Huggins      45.4
Jimmy Williams      44.7
Del Pratt           43.1
Gavy Cravath        41.3
Johnny Evers        40.4
Larry Gardner       39.1
Roger Bresnahan     35.7
Joe Tinker          35.4
Ed Williamson       34.6
Larry Doyle         34.1
Herman Long         32.7
Clyde Milan         31.4
Art Fletcher        31.0
Levi Meyerle        27.1
Donnie Bush         26.8
Ray Chapman         25.8
Harry Wright         3.4
Dickey Pearce       -9.7

Beckley will probably be my #1 this week. That's an enormous edge in career value over this pack, and I don't think his peak was non-existent. It was just Rafael Palmeiroesque.

Seeing Beckley and Konetchy rise makes me think WARP has fixed some of it's old-time defensive spectrum issues.

Home Run Baker seems a little low. IMO, he's overrated by this group (I had him 9th I think the year he went in); but I think he's still underrated by WARP.

I need to get the other HoMers done, I'd like to see how Sutton, and the teens OFs we elected stack up under the new system.

I've also got to post WS for these guys . . .

They've also fixed their issues with fielding in WARP2 for position players, so I'm not nearly as uncomfortable using that now.

The wholesale WARP3 changes have me nervous about using mainly it, so I'm looking to work WS in, too.

So my question: Does anyone have a list of "Seasonal" adjustments for Win Shares? Like, by what factor we need to multiply 1886 AAers to normalize their WS, etc. Just looking for season length adjustments, not a league talent level adjustment here.

Better yet, does anyone have a spreadsheet with all of the significant eligibles and their year by year adjusted WS already? :-)

WARNING --- no HoM content in this post
request for editorial/analytical review

I'm in the process of writing up the results of an extensive study on MVP voting; mathematical model, changes over time, how the superstars did, potential biases in race or big media markets. Those of you who are SABR Statistical Analysis Committee members already received part I many months ago, and now I'm working on the follow-up.

I'd like some feedback on my draft, which should be ready in a week or two. Any volunteers can email me at Han60Man@aol.com and I'll send the doc out as soon as it's ready.

I'll keep posting this damn request every day until someone gives me an answer:

Does anyone have a run estimator that is accurate for the 19th century and/or deadball era? The error in XR is too large.

Thanks,
Dan

Joe, the Mendez thread is closed too, like the Hot Topics thread was, so it didn't take my query. The query was simply a hope that some kind soul would do for Mendez what was done this "year" for Foster, so one can make a determination between the two -- my ranking of both becomes solider when I have 2 Negro league pitcher data points instead of one. My current belief is that Mendez was probably better than Foster, but a few "years" ago I expected to have Poles in my top 5 (as I currently do Santop) so more data shifts me a lot on these guys.

Also, thanks very much indeed for the new WARP numbers, since I'm not a member of BP. As a matter of now purely historical interest, does Caruthers' figure include his hitting and Ward's his pitching?

Glad we agree on Beckley. I know I tend to be stuck on old fashioned metrics, but every time I sneak a peak at the nearly-but-not-quite-obvious HOMers from the next few years like Wheat, Rice, Sisler, Beckley looks better (yes OK, he doesn't beat Ty Cobb, and probably doesn't beat Heilmann, but nevertheless safely the right side of dividing line.)

BP WARP numbers are available to all, I believe.

Harry Heilmann--10 hits away from hitting .400 four different seasons, with a lifetime OPS+ of 148--eats pieces of $h!t (stupid filter) like Jake Beckley for breakfast. And no, he doesn't eat pieces of $h!t for breakfast. (did I recycle that joke once already on BTF?)

Run estimator for 19th century/deadball

I would trust James's RC formulae, which are very different pre-1920. He put much into them to make sure they work for the non-HR speed-based era, and I believe published data to show how they work very well between 1900-1920. I'm not sure how well they do going backwards (James didn't seem to care much about 1880 era ball). They are available on Lee Sinins BB Encyclopedia. Don't know if Total Baseball publishes them. Tom's knowledge on this subject is hereby declared exhausted.

Joe--

I like your adj WARP-2 lists. Two questions, though: do you think your numbers are too kind to 1st basemen (Beckley a clear 1st, Konetchy top 10, Daubert in serious ballot contention)? If so, do you think the fault is in the WARP calculation or your modifications?

Dan R, the ungentlemanly tone of your riposte caused me to check again, and although I agree that, as I indicated initially, Heilmann was better than Beckley, it's really quite close. Heilmann had 2660 hits to Beckley's 2930; if you adjust Beckley to full length (154 game) seasons, he's on 3238. People talk about Beckley's triples as being caused by a "funny park." Are they suggesting they would have been singles in a normal field? Of course not, they'd have been homers, if playing on a normal field with non-soggy baseball. Heilmann has 334 triples plus homers (151 triples, 183 homers) Beckley has 329 (243/86). Yes, the 1890s were a hitters' era, but so were the early 1920s, with several .400 hitters including George Sisler, who will get into the HOM but isn't a slam dunk.

Beckley is not Cobb, but, because of his exceptional career length and perfectly decent OPS+, not far short of Heilmann. He looks to me significantly better than Wheat, Carey, Sisler or Rice, at least some of whom we will presumably elect fairly easily.

People talk about Beckley's triples as being caused by a "funny park." Are they suggesting they would have been singles in a normal field? Of course not, they'd have been homers, if playing on a normal field

This is perhaps supported by the fact that Beckley hit only 34 of his 87 career homers in home games. Seems his parks were turning some homers into triples. Or maybe it was his lack of speed turning some IPHR into triples(?) Hmm.

Andrew Siegel wrote:

I like your adj WARP-2 lists. Two questions, though: do you think your numbers are too kind to 1st basemen (Beckley a clear 1st, Konetchy top 10, Daubert in serious ballot contention)? If so, do you think the fault is in the WARP calculation or your modifications?

Joe may have a different analysis, but it looks to me like it is a combination of the two. Here's my intepretation. WARP's FRAA values are position-specific, so that the average fielder at every position has a FRAA of 0. WARP adjusts for the differences in defensive value between positions in turning FRAA into FRAR. If you drop out that modification, as Joe does, then you have a system that will privilege players at less valuable defensive positions The average first baseman gets just as much credit as the average shortstop. That seems to me to be an incorrect evaluation, and it leads to a ranking that is top-heavy with first-basemen and corner outfielders and bottom-heavy in middle-infielders.

I don't know enough theory to dispute with TangoTiger's premise that replacement level for fielding is league average, but I would suggest that even within that theory, one would need to have a mean for all fielders, not a set of position-specific means for use of fielding replacement level as league average to yield accurate results. I'm also not convinced that, in the early game, defensive replacement level was league average. That seems likely to be true of baseball since 1930, but unlikely to be true of baseball before 1900.

Karlmagnus--I hope you understood I was being ungentlemanly towards Beckley, not towards his supporters. There is a big, big difference between 148 OPS+ and 125 OPS+. You seem to be saying that Beckley and Heilmann had similar career value. Unless you are a 100% career, 0% peak voter--someone who thinks somebody who played for 30 years at 5 WARP a year is a slam dunk HOFer--the fact that Beckley took so much longer to acquire the same career value as Heilmann suggests a major difference in their HOM credentials.

He looks to me significantly better than Wheat, Carey, Sisler or Rice, at least some of whom we will presumably elect fairly easily.

Good point, karlmagnus. Of course, I'm not too enamored with the others either. :-)

I'll keep posting this damn request every day until someone gives me an answer:

Dan, try decaf. :-)

Those of you who are SABR Statistical Analysis Committee members already received part I many months ago, and now I'm working on the follow-up.

Tom, I'm on the committee, but I don't remember receiving it. Could you send me a copy? My e-mail setting on my BTF profile allows e-mails. Thanks!

TomH--thanks so much. I don't have a copy of the Sinins encyclopedia, would you mind emailing the formula/ae to danrosenheck@economist.com? Thanks a lot.

My first reaction was somewhat similar to Chris's. I've always thought that one of the most difficult things to figure in evaluating players was what to do with guys who are great hitters and play key defensive positions but play them poorly or only adequately(Piazza, Hornsby, Lance Berkman in CF, McVey at C, Bresnahan, Doyle, etc.). Every holistic rating system that I've ever run into seems either to overrate or underrate these guys.

I think I am going to adopt a new rating system next year where I use all the available quantitiative evidence to evaluate players' hitting and fielding, but put those two elements together subjectively, kind of like what KJOK does.

You seem to be saying that Beckley and Heilmann had similar career value.

Since first base during Beckley's time was more defense-oriented than it was post-1920, karlmagnus may not be too off the mark here. I also don't consider Becley's long career a minus for him either.

Unrelated: Neyer writes today that Collins was the greatest 2B ever "until Joe Morgan set a new standard in the '70s." Unless you're a hard-core timeliner, I'd take Collins (574 career WS with 8 35-WS seasons; 182.8 WARP3 with 10 10-WARP seasons) over Morgan (512 WS with 5 35-WARP seasons; 155.1 WARP3 with 6 10-WARP seasons) by a small but not insignificant margin.

I'd take Collins (574 career WS with 8 35-WS seasons; 182.8 WARP3 with 10 10-WARP seasons) over Morgan (512 WS with 5 35-WARP seasons; 155.1 WARP3 with 6 10-WARP seasons) by a small but not insignificant margin.

I think I would also, but I'd take Morgan over Hornsby.

So would I. I'd take him over Speaker, too; be interesting to see what the HOM voters as a whole do. Presumably the Collins/Speaker winner gets in in '34, the other in '35.

So would I. I'd take him over Speaker, too

Tentatively, I agree with you. I haven't fully analyzed them yet. I'll be shocked if Collins is not my number #1 that year though.

So would I. I'd take him over Speaker, too; be interesting to see what the HOM voters as a whole do. Presumably the Collins/Speaker winner gets in in '34, the other in '35.

I was going to put Pop Lloyd at #2 in 1934 which would make both Collins and Speaker wait a year to get into my PHOM. (No big deal for me, they are all getting in eventually).

It might be time to request the Lloyd, SJWilliams and Torriente threads. Similar projects who only induct a token amount of Negro Leaguer's induct both Lloyd and Williams.

The average first baseman gets just as much credit as the average shortstop. That seems to me to be an incorrect evaluation, and it leads to a ranking that is top-heavy with first-basemen and corner outfielders and bottom-heavy in middle-infielders.


Positional adjustments are certainly tricky. Most top players of all time lists are overloaded with CF-SS. Win Shares solves the problem of making the numbers "add up" very nicely, but I'm not overly thrilled with the positional balance on an all-time Win Shares list. Obtaining the appropriate level of positional balance in the HOM is going to be an interesting task.

I was going to put Pop Lloyd at #2 in 1934

Forgot about him. Fun year that '34.

Win Shares solves the problem of making the numbers "add up" very nicely, but I'm not overly thrilled with the positional balance on an all-time Win Shares list.

I'm not either. They (and most other systems) also don't take into account each position's attrition level. There's a huge differecne playing twenty years as a modern-day first baseman than as a third baseman.

My adjusted WARP (uses FRAA, as opposed to FRAR) career totals . . . using new WARP.

Joe Dimino, where are you getting your values for BRARP? It is not on the new WARP pages.

If you are using BRAA instead (same column that BRARP used to be), then you might as well use PRAA, join the Palmer school, and call your new statistic WAA (Wins Above Average).

If you are using BRAR instead, then glove positions are hurt badly by your new statistic. An average fielding 1B-man and an average fielding SS will each have 0 FRAA. They will get the same rating from the same level of offense, even though we all know that the shortstop is much rarer, and hence more valuable.

Here's my proposal for a FRAR system:

Use whatever stat you want to get FRAA--BP FRAA, UZR, Defensive WS minus league average WS at position times 3.33, etc. Then, divide positional league average fielding WS times # of teams divided by total league fielding WS to get a positional fraction of defensive responsibility (eg 1B 5%, SS 17%, etc.). Somehow determine a pitching/fielding split, and pick a replacement level winning percentage. For that league's run scoring environment, determine RS/RA for that winning percentage, take that team's RA minus league average RS, and multiply by the fielding percentage and then by the positional percentage. This would give you the # of runs an average fielder at that position saved above replacement level. Add this to FRAA, and voila. Note I don't have positional average WS data.

Example: 1896 Hughie Jennings. BP gives him 31 FRAA. Judging from the WS leaderboards that year, I'm going to guesstimate the average full-season catcher had about 5.6 fielding WS, 1B 1.3, 2B 3.7, SS 5.5, 3B 3.8, and OF 3. Thus, C is responsible for 19% of fielding WS, 1B 4%, 2B 13%, SS 19%, 3B 13%, and each OF position 10%. I'm going to arbitrarily set a pitching/fielding split at 60/40. The league scored 6.03 runs per game in a 130-game season, for an average of 784 runs per team. Also arbitrarily setting replacement level at a .300 winning percentage, an all-replacement team in 1896 would score 626 runs and allow 942. Of the 158 runs below average on the pitching/defense side of the equation, we'll give 63 to the fielders. 63*.19 for SS = 12 FRAR for a league average SS. Add that to Jennings' 31 FRAA and you get 43 FRAR.

yest #2-41 (#762717)
since Welch’s record is based on what I posted before from the retrosheet he gains 7 wins and 3 losses on his record

yest,
Please explain.
I can't interpret your position on the official record and the record derived from Retrosheet.

All,
Don't be too quick to toss WARP in the dustbin or to take the Retrosheet game log as gospel.

An example of why BRARP+FRAA does not work.

This is the bottom 25% of starting REGULARS (at least 66 games GP in 132 game schedule) or multi-position players from 1895, as ranked by BRARP1+FRAA1. (I've ordered them by WARP-1; these are pre-July WARP-1 numbers.)

According to TangoTiger/JoeDimino they are ALL allegedly below replacement level. Bottom at each position I can accept, not 3 or 4 per position. Also note every team is carrying at least one "below-replacement" player, the champion Orioles are carrying two.

QY TT                AdjGm  BRARP1 FRAA1 PRAR1 WARP1
            DUMMYHOY 108.7       5    -6     0   3.8 -  Cin-LF-8/CF
        JIMMYMCALEER 131.0     -24    10     0   3.7 -  Cle-CF-8
        CHARLIEABBEY 138.1      -1   -11     0   3.3 -  Was-CF-9
       TOMMYMCCARTHY 118.2       2   -10     0   3.3 -  Bos-LF-9
          WALTWILMOT 108.0      -7     1     0   3.2 -  Chi-LF-10
     GENERALSTAFFORD 122.1       9   -17     0   3.1 -  NYo-2B-9
             TOMDALY 120.0      19   -28     0   2.9 -  Bro-2B-10
        JOHNANDERSON 102.0       4    -8     0   2.9 -  Bro-LF-11
          KIDGLEASON 105.9      16   -11    -7   2.8 -  BAL-2B-11
          MONTECROSS 107.6      12   -23     0   2.5 -  Pit-SS-10
          EDDIEBURKE  95.5      -8    -1     0   2.5 -  2Tm-LF-12
        DOGGIEMILLER 123.4       6   -17     0   2.1 -  StL-3B/C
        CHIPPYMCGARR 108.7      -8    -2     0   2.0 -  Cle-3B-11
            TOMBROWN 119.0     -16    -5     0   1.9 -  StL-CF-10
         BILLMERRITT  86.0      -4    -7     0   1.5 -  2Tm-C-15
          HARRYSPIES  85.9       1    -6     0   1.4 -  Lou-1B/C
         FRANKGENINS  71.9      -4    -1     0   1.4 -  Pit-Util
        GEORGEDECKER  77.3      -5     0     0   1.2 -  Chi-Util
        FRANKSHUGART 114.3       4   -28     0   1.1 -  Lou-SS-11
      GEORGETREADWAY  86.4      -1    -8     0   0.9 -  Bro-RF-10

       BILLHASSAMAER 110.1     -11   -14     0  -0.6 -  Was-RF
         SCOOPSCAREY 123.8     -11   -11     0  -0.7 -  BAL-1B
           JACKBOYLE 133.0     -16   -24     0  -2.5 -  Phi-1B

Note: Jimmy McAleer loses 1895 WARP-1 CF Gold-Glove to Bill Lange 42-40 FRAR1.

Adventures in Time-Traveling SABRmetrics: Joe Dimino tells Patsy Tebeau that Loafer McAleer is no good and should be replaced. Patsy spits tobacco juice in Joe's face and decks him.

Sorry, Joe. I couldn't resist. ;-)

Dan,

I agree that's a good way to do it. It's just a SMALL (not) matter of gathering the necessary data and making the necessary calculations.

There is one shortcut you could take.

Instead of calculating a season-by-season positional fractions of defensive responsibly, you could simply use the "intrinsic weights" built into the WS system, with appropriate period modifications to account for changing conditions that lead to changes in the responsibilities of the defensive positions.

These are

Post-1930
C 19%
1B 6%
2B 15%
3B 13%
SS 18%
OF 29%

Pre-1930
C 19%
1B 6%
2B 14%
3B 14%
SS 18%
OF 29%

The seasonal calculations would get you back pretty close to this formula, anyway.

The accuracy of the system depends on how you set the pitching/fielding split, of course, and how you set the positional percentages.

Both of those settings could be the subject of extensive discussion and analysis . . .

Does anyone have a run estimator that is accurate for the 19th century and/or deadball era? The error in XR is too large.


BASE RUNS would be your best bet. (Just use GOOGLE to find it)


The 1900-1907 for runs created is:
A factor = H + W - CS + HBP - GDP

B Factor = (TB + (.26 x (TBB - IBB + HBP)) + (.52 x (SH + SF + SB)))

C Factor = AB + W + HBP + SH + SF

Then A Factor X B Factor / C Factor

1938 Hall of Fame Election (thought this was interesting..)

ELECTED VOTES
George Sisler 235
Eddie Collins 213
Willie Keeler 207

OTHERS VOTES
Rube Waddell 179
Rogers Horsnby 176
Frank Chance 158
Ed Delahanty 145
Ed Walsh 132
John Evers 107
Miller Huggins 97
Walter Maranville 82
Jimmy Collins 72
Roger Bresnahan 67
Fred Clarke 59
Mordecai Brown 54
Wilbert Robinson 46
Chief Bender 40
Herb Pennock 40
Ray Schalk 35
Hugh Duffy 34
Ross Youngs 34
Hugh Jennings 33
Joe McGinnity 32
Frank Baker 30
Addie Joss 28
Eddie Plank 28
Mickey Cochrane 28
Frank Frisch 26

Hey ... I've been lurking over the Hall voting for several "years" now, and the back-threads have proven more interesting reading than a lot of the books on my shelf. I decided I was finally comfortable jumping in. Only, the one thing I can't seem to find on any of the threads is an indication of if the voting's open. (Or if there's a process for jumping in. Or if there's a jumping in.) I sent some e-mail asking for clarification, but I think it got lost in the cosmic shuffle.

Normally, I'd just say "heck with it, post and be damned" ... but I realize this is a one-guy-only year, and it's fairly late in the voting week, and I could see where people would be irritated if a Johnny-come-lately threw off a close election.

So, I figured, I'll post what *would* have been my 1931 ballot if I'd gotten off my tuckus sooner, here. And if there's no reason why it can't be an official ballot, yay and so forth. If I didn't fill out the proper forms, or if it's really too late to jump in for this "year", or if my theories are the worst kind of popular tripe, my methods are sloppy, and my conclusions are highly questionable, then just have fun blasting me and I'll slink off to a corner somewhere. (Why, yes, I *am* neurotic and self-conscious, now that you mention it ...)

(FWIW: I'm more peak than career; I like Win Shares like pork chops--more than is healthy for me, but I'm trying to cut down; I haven't gotten a handle on the metrics over at BPro quite yet, but neither has BPro; and when push comes to shove, I'm rather susceptible to peer pressure.)

1- Tommy Leach (PHoM 1930). Reading between the lines of the NBJHBA seems to suggest that a 300 WS career with 3 "Great Seasons" (30+ WS) would be the definition of a no-brainer Hall of Fame player. Of the folks eligible, Leach, Griffith, and Burns are the players closest to that standard, depending on how you view "closest". Not that some silly extrapolation of James is warping my thinking ... nah, I'd wager instead that my thinking is warped by a paranoid tendency to overrate third basemen to counter the underrating of third basemen (call it the Santo effect).
2- Clark Griffith (PHoM 1930). There's only so many ways I can look at things and try to figure out why it's telling me Griffith's the best pitcher available before I have to try the wacky theory that hey, Griffith's the best pitcher available.
3- George van Haltren (PHoM 1926). Dr. Career's hopeless attempt to bring balance to Mr. Peake's ballot before the cops break down the door to the lab.
4- Rube Foster (PHoM 1931). I have a tendency to drink the Kool-Aid when it comes to Negro Leaguers, and I'd like to play it a little conservative here. But I really am impressed by the arguments that have been made in his favor.
5- Hughie Jennings (PHoM 1919). I am perhaps too easily swayed by people able to put "best player, period" on their resume, even if it's for just a year or two. I'm seeking treatment for it, there are some new drugs available, but right now Ee-yah still makes my top 10.
6- Addie Joss (PHoM 1929). About comparable to half a season of an average Vic Willis ... in his worst season, pitching while being scouted by T.G. Reaper. OK, so I like me some peak, but there's no unpeak here. It's turtles all the way down. If he'd been alive to serve up utterly lousy pitching for 4 more years, I'd still be arguing for him here. And I don't even like Cleveland.
7- George Burns.
8- Bobby Veach.
I feel like I should have one of these guys up around 5, and one of them down around 11, but I can't figure out which is which. Burns looks a little better to me WS-wise, Veach looked a little better in the BB-ref vision-quest. Short of trying the latter again with actual controlled substances involved, I'll link the two together here for now.

9- Cupid Childs (PHoM 1921). I didn't really like Bid McPhee, for some reason, and I'd feel better about him if Childs had a seat next to him in Plaqueville.
10- Rube Waddell. Sort of the Georgia-in-"1776" slot: My gut sneers at simple W-L issues and says put him higher, other people seem to be saying "not so fast", I'm saying "not so fast" until I have the guts to sneer at them too.
11- Hugh Duffy. The PHoM Whack-a-Mole. It seems like every time I come up with some sort of new system of fiddling with the numbers available to me, it churns out Duffy near the top of the list. Short of coming up with a huge discount for being Hugh Duffy, I can't put him much *lower* than this.
12- Dickey Pearce (PHoM 1913). On the one hand, the greater amount of snide in an argument, the less substance it eventually seems to have ... and Pearce has been the target of an awful lot of snideness. On the other hand ... I probably jumped the gun in my PHoM, and I'm mature enough to admit that--when they go to choose up sides for baseball in Riverworld, I don't have any *firm* confidence in how high he goes.
13- Roger Bresnahan. Probably overrated. Might drop down quite a bit when the baseball groundhog sees Santop's shadow. But I'm not sure I buy the "he played too much outfield" arguments, and he *was* a pretty damn good catcher.
14- Mickey Welch. A big thank-you to the people who left him off their ballot entirely, so I don't have to stay awake at night feeling guilty about if I've discounted him too much.
15- Pete Browning. I'm currently convinced he didn't make my PHoM yet just because I'm slamming AA years *too* reflexively. (OTOH, not sure if Greg Luzinski being one of my early faves has me giving Browning bonus points for his "legendary" fielding ...)

Off-Ballot Disclosures:
Spot Poles: 16th. Playing a game of flip-flop with Browning, Ryan, Pike, and Cravath on pretty much a daily basis as I try to figure out which set of question marks bother me the least.
Lip Pike: 17th. See Poles.
Jimmy Ryan: About 20th. No, I'm not real comfortable with the distance between Ryan and the other CFs (coughhiGeorgecough) that made my ballot, either. But there's nobody ahead of him I'm real comfortable bumping down to tighten the gap, either.
Harry Hooper: 22nd. My caveman brain lumps him in with Beckley, but likes him better.
Jake Beckley: 24nd at the moment. He moves up about 1-2 spots every time I look at this. One season that makes the evil voice in my head stop whispering "Doug Mientkiewicz with a full career" would help a lot.

Welcome, Tanketra (as a fan of Creature Features as a kid, I don't mind saying that your screen name rocks!)

Your ballot looks fine to me, so repost it any time that you want on the 1931 Ballot thread

Figures. I tell Bryce that I like his screen name and then he changes it. :-0

Welcome Bryce. Hope you stay & enjoy it. Few comments:

1- Tommy Leach (PHoM 1930). Reading between the lines of the NBJHBA seems to suggest that a 300 WS career with 3 "Great Seasons" (30+ WS) would be the definition of a no-brainer Hall of Fame player.

Careful about using the boldfaced phrase around here. It's usage has sparked debate before.

Re: Rube Waddell. I iniitally had him on my ballot because of his ERA & ERA+ - but there's a problem with those numbers. Compared the % of runs he allowed that were unearned with the % of unearned runs on the teams he played for & it turns out he allowed about 50 more unearned runs over the course of his career than one would expect. Adjust for that & his ERA goes up to about 2.30 & his ERA+ down to about the mid-120s. Not sure how much importance you give to ERA+, but the more you give to it, the more damning the above inforation should be to Rube's case.

Re Leach: Yeah, probably should have been more careful there. For the record, I wasn't saying *I* thought that made for a no-brainer, just that it seemed to be one definition.

I don't believe in a no-brainer Hall of Famer, personally. I like to sit up at night and come up with reasons why Babe Ruth doesn't belong in the Hall of Fame. (Well, not so much "like" as I work graveyard so I *have* to sit up at night, and it's more interesting than infomercials.) The best I've come up with so far is the trick-bat thing. :)

I like ERA+, but I'm not married to it. And I did catch the problem with the unearned runs earlier. I still like his case better than quite a few of the other available pitchers (though, obviously, not better than at least 3 others).

Quickly so my "official" ballot will count.

1931 BALLOT
(1) Dickey Pearce - most impressive "best player at his position" of returning candidates.
(2) Lip Pike - last of the great NA/NL players.
(3) Eddie Cicotte - best pitcher; lost 2nd half of career to Landis' decision.
(4) Rube Foster - jumped way up the list to this spot over OF glut.
(5) Cupid Childs - dominated 1890s second basemen.
(6) Mickey Welch - the vs. opponents data has put him in a strong position on my ballot.
(7) Larry Doyle - best second baseman of the deadball era (after Collins/Lajoie).
(8) Ed Konetchy - best 1Bman of the 1910s.
(9) Rube Waddell - up over Addie Joss on this list now.
(10) George Van Haltren - overlooked star of 19th century.
(11) Tommy Leach - re-thinking my third basemen has put him on the ballot after years off it.
(12) Jake Beckley - career too strong not to be this high.
(13) Addie Joss - dominated while he played.
(14) Spotswood Poles - probably next best OF (over Jimmy Ryan now).
(15) Hughie Jennings - why I've never taken to this high-peak icon before is beyond me.

yest,
Please explain.
I can't interpret your position on the official record and the record derived from Retrosheet.

I'm not sure I understand your question correctly so I hope this answers it.

since the retrosheet only says the pitcher who started the games and not the pitcher who got the decision, the chart I made in post 127 saying how many runs Welch gave up has the same flaw.

post 241 is showing how many wins Welch would have had if his team would have scored the exact same number of runs every game he pitched in. this is based on the chart in post 127.

"If you drop out that modification, as Joe does, then you have a system that will privilege players at less valuable defensive positions The average first baseman gets just as much credit as the average shortstop. That seems to me to be an incorrect evaluation, and it leads to a ranking that is top-heavy with first-basemen and corner outfielders and bottom-heavy in middle-infielders."

I don't think so Chris.

I'm using BRAR which is batting runs above replacement position. So positional value is taken into account on the offensive side of the ledger. I then add in FRAA (fielding runs above average) because replacement level for fielding is average. I'm fairly certain that Tangotiger is right on that statement. I've checked with him and I'm very sure that I'm doing this properly.

Jim - I believe the bottom 15% of regulars in the majors are the definition of replacement level, so your numbers don't surprise me.

Good players will have bad years (happens every year) certainly they don't lose their job the second they go below the level where their production could be replaced by a similar player.

Also, I think you've got to use the new WARP numbers, since that's what I've used. Your numbers are great, but they don't lead me to believe I'm doing anything wrong. If anything, you are saying I'm setting my replacement level too high, which would hurt, not help Beckley. Anything that lowers replacement level helps Jake Beckley (and any player with above average career length).

Back to new WARP. It's quite possible that Prospectus fixed what I thought was an issue with regard to the value of 1B in this era. WARP2 is an 'all-time' adjustment, so it was treating all positions at there historical average in terms of value - which is wrong. The guys that played 3B in 1892 would have played 2B had they been born in 1952 and in all-time adjustments they need to be treated as such - what I mean is that all-time adjustments should really just take into account qaulity of the era/year; and quality of the league relative to other leagues that year.

The all-time positional adjustments were docking old 1B too much, treating them like they had the same defensive value as they do post 1930.

It appears to me that they've corrected this, but without a glossary, I have no idea.

DanR - WS is wrong in it's valuing of 1B pre-1930. James missed that defensive spectrum shift, but 1B clearly had more defensive responsibility the further back you go. Before 1930 it was certainly more important than either LF or RF, before 1900 I'd say it was more valuable than CF even. As the gloves got better, the position became much less abusive for one.

Also Dan, I'd use the James formulas, though they aren't perfect. For ANY RC system pre-1920, at this point, you must adjust for the team total runs and credit/debit the difference to the players on the team. Short of something like BaseRuns (which I haven't worked with, I'm just guessing), there simply isn't anything accurate enough to use without reconciling the numbers with the team runs total.

"Also note every team is carrying at least one "below-replacement" player, the champion Orioles are carrying two."

The Yankees are starting Enrguel Cailson right now. They won despite the last couple of years of Tino and O'Neill in Bomber pinstripes, despite the fact that they were basically replacement level at the end, not to mention the presence of Scott Brosius.

The Cardinals start Mike Matheny and have the best record in baseball. The Astros play Brad Ausmus and Craig Biggio was basically replacement level last year, when you figure in the atrocious defense.

Joe Dugan on the 1927 Yankees. I could go on and on. Replacement level is higher than most people think, I believe. Once you get to where the curve expands the players are interchangable. This happens around the level of the bottom 15% of major leaguers who are basically only slightly better than (and in some cases worse) than top minor leaguers.

Welcome Bryce/From the Depths of Lurk rises ... Tanketra!

Why don't you call yourself The Necromancer or something, we need a good Rush reference here. Bytor would be good too.

As for the 300 thing, I wouldn't call 300 a slam dunk. I think there are roughly 200 players with 300 career WS (if you adjust for season length), so only in a strict career vote would that be a magic number. Many pitchers will go in with less than 300 WS too.

I'd say 350 is closer to the 'magic number', below that, you are in a grey area from 250-350. 175-250 for pitchers is grey over 250 is in pretty much.

since it was unclear what I meant I'll try one more time.
This is based on the retrosheet so it only shows games that Mickey Welch started whether he got the decision or not. So Welch has 7 more wins and 3 more losses. Keep in mind Welch started 549 games completing 525 of them

in games that Mickey Welch lost
gave up 24 runs 1 time
gave up 21 runs 1 time
gave up 18 runs 2 times
gave up 17 runs 3 times
gave up 16 runs 1 time
gave up 15 runs 3 times
gave up 14 runs 4 times
gave up 13 runs 8 times
gave up 12 runs 9 times
gave up 11 runs 12 times
gave up 10 runs 12 times
gave up 9 runs 11 times
gave up 8 runs 17 times
gave up 7 runs 22 times
gave up 6 runs 19 times
gave up 5 runs 32 times
gave up 4 runs 28 times
gave up 3 runs 15 times
gave up 2 runs 9 times including 1 forfeit when the game was tied
gave up 1 run 4 times

in games that Welch won he
gave up 12 runs 2 times
gave up 10 runs 3 times
gave up 9 runs 2 times
gave up 8 runs 12 times
gave up 7 runs 20 times
gave up 6 runs 17 times
gave up 5 runs 23 times
gave up 4 runs 28 times
gave up 3 runs 55 times
gave up 2 runs 52 times
gave up 1 run 62 times
and no runs 41 times

in games that he tied
tied at 9 runs 1 time
tied at 8 runs 1 time
tied at 7 runs 1 times
tied at 6 runs 2 times
tied at 5 runs 3 times
tied at 4 runs 2 times
tied at 3 runs 2 times
tied at 2 runs 5 times
tied at 1 run 2 times

according to the retrosheet Welch has
317 wins
213 losses
19 ties
and a 598 winning percentage
for his career


This is how many run per game were averaged by the NL as a whole during Welch’s career
1880------------------4.7
1881------------------5.1
1882------------------5.4
1883------------------5.8
1884------------------5.5
1885------------------5.0
1886------------------5.2
1887------------------6.1
1888------------------4.5
1889------------------5.8
1890------------------5.6
1891------------------5.5
1892------------------5.1
1880-1892------------5.3

Now assuming his team scored 5.3 runs every single game he started he would have
363 wins 186 losses and a 661 winning percentage

and if you don’t want to include tie games
363 wins 186 losses and a 658 winning percentage

Yest, you'd have to weigh those R/G totals based on Welch's innings. Then you'd have to adjust each year based on his park. But you are definitely on the right track.

Chris J has done this for Welch - using the retrosheet data and he gets his career Run Support Index at 103. Link here:

http://runsupportindex.blogspot.com/2004/06/mickey-welch.html

I'm trying to find out how he would do if he the average run support

and if you don’t want to include tie games
363 wins 186 losses and a 658 winning percentage

make that
and if you don’t want to include tie games
349 wins 181 losses and a 658 winning percentage

Excellent! I've figured out how to use BaseRuns as a run estimator for every season. Updated WARP data will be forthcoming...

Does anyone have a run estimator that is accurate for the 19th century and/or deadball era? The error in XR is too large.

....

The 1900-1907 for runs created is:
A factor = H + W - CS + HBP - GDP

B Factor = (TB + (.26 x (TBB - IBB + HBP)) + (.52 x (SH + SF + SB)))

C Factor = AB + W + HBP + SH + SF

Then A Factor X B Factor / C Factor

This seems to be my week to be an idiot - I posted the NEWEST formula instead of the OLDEST - you'd think caught stealing would have given it away...

Here is the Runs Created Formula for 1907 and earlier:

A Factor = H+W+HBP
B Factor = [1.025x(TB+SB)]+(.75xSH)
C Factor = AB+W+HBP+SH

Then A X B / C

Joe re BRAR is BRARP:

I'm using BRAR which is batting runs above replacement position. So positional value is taken into account on the offensive side of the ledger.

It may be too soon to say for certain what the new BRAR means, since it hasn't been introduced for all players, so we're judging from incomplete information. But _all_ the evidence I can find indicates that the new BRAR is _not_ BRARP. Here are two pretty strong indicators.

1) Take a look at George Davis's or Bobby Wallace's or Buck Ewing's player cards (all of their cards are the new ones), players who changed position in their careers. When you read down their fielding columns, you can see the shift from position to positioned registered in the changing gap between FRAR and FRAA: when Davis and Wallace they hit shortstop after 3B, their FRAR jump in relation to their FRAA, when Ewing plays first or right field, his FRAR drop in relation to his FRAA. All this is as we would expect, since we know that this is what WARP does. When you read down their batting columns, there is _no_ corresponding shift in the difference between BRAR and BRAA, which we ought to see as they shift from a better-hitting position to a worse-hitting one.

2) Compare Joe Sewell's 1922 season to Harry Hooper's 1923 season (both have new cards):

SS Sewell, 422 outs, .273 EQA, 9 BRAA, 28 BRAR
RF Hooper 422 outs, .272 EQA, 9 BRAA, 29 BRAR

There's no way BRAR here is adjusted for position.

It's possible that your system would be valid using the old WARP that supplies BRARP, but I don't think it can be adapted to the new unless you independently calculate BRARP.

Santop thread doesn't accept posts either -- another we're going to need come Monday. Joe, your magic is needed again!

Reinforcing what Chris Cobb said on old BRAR vs new BRAR.

I went through the old and new team sheets for the 1895 Orioles. Differences: Brodie (36,35), Carey (-8,-7), Esper (-9,-8), Jennings (62,61), Keeler (57,56), Kelley (72,71), McGraw (44,43), Robinson (-5,-4); the other 12 players are identical. The only difference appears to be a change in the rounding method, or something else relatively trivial.

Conclusion: new BRAR is old BRAR; it still contains no positional information, and represents how many runs the player generated above .230 EQA. Using BRAR+FRAA will come to the conclusion that an average fielding 1B-man and an average fielding SS that hit the same stats will have the same value. Sorry.

I have been working on doing workups for Clarkson and Radbourn similar to what I have done for Keefe and Welch. I should be able to post the results later today. A little tidbit so far. Clarkson's record against HoMers:

Keefe     10-9
Radbourn   8-9
Galvin     7-5
Rusie      3-4
Caruthers  2-1
Young      1-1
Nichols    0-1
totals    31-30

Welch 6 or 7-13

In comparison, Welch went 62-35 and Keefe went 42-40.
The record against Welch is different than the 12-5 or 5-12 (depending on you look at it) I have previously posted because I forgot the 1891 season when I was handcounting all Welch's decisions. In 1891, they split 2 decisions. But, comparing game logs for both players 13-6 in favor of Welch is correct. I didn't know how to account for a 2-2 tie game that was forfeited to Clarkson's team in 1890.

yest, Thanks for the recap.

official W-L record of Mickey Welch
307-210
W-L-T record of his team in Mickey Welch's starts
317-213-19

If I now understand the "gain" correctly, Welch gains 10 wins and 3 losses,
not 7 wins and 3 losses [#241]. Right?

--
Regarding his record (ie, the record of his team) given 5, 5.3, or 6 runs uniformly scored all of his starts, you have:
294-236 @ 5 runs [#241]
349-181 @ 5.3 or 6 runs [#241, #292]

I think you have a clerical error, or a problematic treatment of ties or fractions or both.

If I now understand the "gain" correctly, Welch gains 10 wins and 3 losses,
not 7 wins and 3 losses [#241]. Right?

right I'm noy sure how I made that mistake
---------------------------------------------
Regarding his record (ie, the record of his team) given 5, 5.3, or 6 runs uniformly scored all of his starts, you have:
294-236 @ 5 runs [#241]
349-181 @ 5.3 or 6 runs [#241, #292]

I think you have a clerical error, or a problematic treatment of ties or fractions or both.

I'm not sure what you mean by this