hot topics

"Am I in heaven?"

"No, Minneapolis."

Like Satan, I come when my name is invoked.

Perhaps some weren't paying attention in 1925 when I began pushing the great Cravath juggernaut that now consists of me and three other voters:

Gavvy Cravath's Minor League Numbers

First, PCL stats. Remember that the PCL was a major "pitchers league." I have at my disposal currently team offensive stats (batting average) from every PCL team in 1905, which should give a taste for what 'league average' was. (Cravath played in L.A. from 1903-1907)

Seattle: .238
Los Angeles: .236
Portland: .232
San Francisco: .228
Tacoma: .226
Oakland: .215

1903 (age 22): Rookie 22 year old Cravath hits .274 in 209 games, with a team leading 7 home runs. Los Angeles wins the PCL pennant.

1904: No stats available (to me at least)

1905 (age 24): Cravath hits .259 (the team average is .236, see above). He had 33 doubles, 9 triples and 9 homers and 44 stolen bases. He was third on his team in BA. Just below him in BA is George van Haltren, who hit .255, two years removed from hitting .257 with the Giants. Angels win the pennant.

1906 (age 25): Cravath hits .270.

1907 (age 26): Cravath hits .303 with 10 homers and 50 stolen bases. Angels win the pennant. He is purchased by the Red Sox.

1908 (age 27): Cravath has a 136 OPS+ for the Red Sox. Great power player in a deadball period that didn't appreciate it.

1909 (age 28): Bizarrely sent to Chicago after a successful 1908. He drops off to a 108 OPS+, but in a very small sample size (70 plate appearances) and then is shipped off to Washington for 7 games and is released. He goes to play for the Minneapolis Miners in the AA. He hit .290 – second best on the team -- with 4 homers. This was significantly better than player/manager and HoMer Jimmy Collins (.273), who had just retired from Philadelphia.

1910 (age 29) – Cravath hits .326 with 14 homers for Minneapolis. Both of these numbers lead the league. Miners win the pennant.

1911 (age 30) – Cravath hits .363 with 29 homers for Minneapolis. Both of these numbers lead the league. Miners win the pennant.

1912 – 1920 (age 31-39) – the stats are readily available. He earned about 200 win shares in those 9 years, which included being the best hitter in the game from 1913-1915. There can be no doubt that his 3 year peak, or 5 or 7 year prime in these years is HoM-worthy.

The only issue is whether the preceding nine years is sufficient to provide the “career bulk” to this peak. I think it clearly does.

Come up with whatever bizarrely conservative translations you want. 10 Win Shares a year for being one of the stars of the PCL? 15 a year for being THE star of the AA?

I think those numbers are hideously low, but granting that, 5 PCL years plus 3 AA years give 95 extra win shares. Add that his 202 Major League win shares, and you've got 297. Any doubt that a player with 297 win shares and Cravath's major league peak (leaving aside that his actual peak included some AA years) would be a first ballot HoMer?

What realistic credit can you award to Gavvy and his five minor league pennants that could induce you to leave him off of your ballots? Inquiring voters want to know.

If credible numbers could be created for Gavvy, I think that would be helpful to the electorate deciding if Cravath belongs. They sure helped me out with the Negro Leaguers.

I give Gavy credit at his 1912-20 level for 1910-1911, due to MLB's unappreciation of his talent. For 1909 I give him full-season playing time, with the missing time at his established level

I think it's tough to give him credit pre-1908, though I could see giving him credit for 1907.

I think this additional credit is enough to have put him 5th or 6th (can't remember) on the most recent ballot. The man has a 150 OPS+, most of it from age 31 on. It's a Joe Start type of argument if you ask me.

'with the missing time at his established level'

Read that as at 130 or so OPS+, he had a 138 OPS+ in 1908.

PhillyB, I confess I hdan't spent much time looking at Gavvy, and he probably belongs in my top 30.

I can't put him near Bill Terry, tho, for example. He had poorer rate stats by either Win Shares per 162 G, or EqA (plus Terry was a better fielder by FRAA). I also think he took unusual advantage of the Baker Bowl (1914, led the NL with 19 homers; all at home!), which knocks him a bit in my book. Still, a fine hitter. Close to Hack Wilson maybe?

I can't put him near Bill Terry, tho, for example. He had poorer rate stats by either Win Shares per 162 G, or EqA (plus Terry was a better fielder by FRAA).

Pardon my contradiction, but I think this is wrong.




Bill Terry: 278 Win Shares in 1721 Games = 26.2 WS/162.

Gavvy Cravath: 202 Win Shares in 1220 Games = 26.8 WS/162.


Further, it appears to me that -- unless I am missing some era adjustment that I shouldn't be -- Cravath had fewer Plate Appearances per Game (perhaps Cravath was used as a pinch hitter more?), and the difference between them expands in you look at WS/600 PA instead of 162 games.

Gavvy Cravath: 26.1 WS/600 PA
Bill Terry: 23.5 WS/600 PA

Cravath had the better rate by Win Shares.

EQA (from Baseball Prospectus Pages:
Gavvy Cravath: .312 (timeline adjusted to .303)
Bill Terry: .308 (timeline adjusted to .304)

I mark this "essentially identical".

The question is only whether Terry's 76 extra win shares (in the majors) make up for Cravath's better rates, and unrecorded 8 years in the PCL and AA (i.e., did he average at least 9-10 WS equivalents per season). I think the answer is clearly yes.

Both Cravath and Terry will be in my Top 10 this years, but Cravath will be higher.

I would like to give Cravvath credit for 1910-1911 and the partial 1909. Unfortuantely, I am not in a position to guesstimate on how much those seasons would have been worth on the major league level.

Can someone smarter than I take

1) His fielding ratings from the majors
and
2) His recorded minor league numbers from those years

And give a Win Shares estimate for those 3 seasons?

Gavvy Cravath: 202 Win Shares in 1220 Games = 26.8 WS/162.

James has him at 21.51 WS/162 in the NBJHA, so somebody is wrong here.

Perhaps some weren't paying attention in 1925 when I began pushing the great Cravath juggernaut that now consists of me and three other voters.

LOL.

Keep fighting the good fight, PhillyBooster.

James has him at 21.51 WS/162 in the NBJHA, so somebody is wrong here.

202 Win Shares is the Number listed in the BJHBA and in the 1926 discussion thread in this site.

The BJHBA lists 1221 games, and baseball reference lists 1220. So both Bill and I agree on the raw imputs.

202 divided by 1220, times 162 definitely equals 26.82295. So the problem is EITHER both of our inputs, or Bill James's output.

If I can reverse engineer Bill James's bug, 21.51 WS/162 over 1221 games would lead to a career total of . . . 162 Win Shares. I'd say either the book is wrong and Cravath REALLY earned 162 career win sharesm not 202, or -- mor likely -- somebody stuck the Games per Season figure into the "Career Win Shares" column on somebody's database.

While I an open to someone checking on the former, I would put money on the latter.

The WS electronic supplement clearly has Cravath at 202 career win shares.

Phillybooster's deduction about the source of the error in Cravath's published rate sounds plausible to me.

While I an open to someone checking on the former, I would put money on the latter.

I would, too. After seeing Christy Mathewson as a starting pitcher during the fifties and other errors, I wouldn't be surprised if they just screwed up on their end.

I'm happy to give Cravath some credit (particularly for the years between his major league stints), but I think you need to be careful about giving credit for every season someone was capable of playing major league baseball. In a perfect world, you'd give that credit and you'd also give equivalent credit to everyone else for similar such seasons (thereby raising the bar on what it takes to have an HoM career). I always thought that we didn't give credit for pre-and post-major league seasons in most cases not b/c/ they aren't worthy, but b/c/ for most players they balance out.

Therefore, for someone like Cravath, he should get credit for a few of his minor league seasons b/c/ he played MORE good seasons in the minors than anyone else, but you can't give him credit for all of them b/c/ 2 or 3 very good minor league seasons are assumed to be part of most players careers.

Incidentally, if a 24-year-old Cravath only outhit a 39-year-old Van Haltren by 4 points, doesn't George deserve some extra credit as well?

If we're giving extra credit to Cravath, we should surely do so to Leever; both missed some of their prime seasons because of structural/economic problems in the game.

"James has him at 21.51 WS/162 in the NBJHA, so somebody is wrong here."

wow, good catch. I wonder if the presumed database error affected Bill's ranking of Cravath.

Okay, let's assume by WS that Cravath has a slight rate edge, which might go away if you timeline. By BP, we see their offense rate is eveen, but I still maintain that Terry was more valuable defensively. So it seems that Cravath's case over Terry boils down to making up the -40% career shortfall with assumed minor league time.

If we're giving extra credit to Cravath, we should surely do so to Leever; both missed some of their prime seasons because of structural/economic problems in the game.

Yes, but when Cravath missed them, he was still playing high level professional baseball, and was among the best in his top-minors league each year.

I'm not planning to give Bo Jackson credit for his NFL Pro-Bowl seasons either.

Leever was not wanted by the majors after he got out high school because he did not have a fastball and he did not develop his curve until later. I could see giving the credit if he dominated out of high school, but he was playing semipro Sunday ball - not even low minors.

And there was a depression during the 1890s so baseball was a very viable alternative. Fielder Jones actually went into baseball because he could not get a job with an engineering degree.

And it was not difficult for him to be scouted. One of his teammates on the semipro Norwood Maroons, Kid Elberfeld, made the majors as well.

In 1890, his age 18 baseball season, there were 3 major leagues. He was not wanted.
1891 - 2 major leagues - not wanted.
1892-1896 - one major league with some BAD teams and pitchers - he was not wanted.
He had to develop another pitch. If he was happy as a schoolteacher so he didn't take the time to develop a curve ball, why should he get credit for that? He had the opportunity to make a $1500-$3000 a year in the majors or several hundred for a half year's work in the minors OR he could make less.

From a quick scan of google - teacher salary 1890, 1895.
Sudbury MA - top schoolmistresses $13 a week. 1889
Sioux City - avg male schoolteacher $43 a month. 1890
Kansas City, KS - about $50 a month depending on how you tested. - 1890.
Since they worked 9 months a year - that is about 350-450 a year.
I'll post some 1895 figures later.

Thank you, Kelly. Can we copy and paste that post every time karl brings up Leever? :)

Thank you, Kelly. Can we copy and paste that post every time karl brings up Leever? :)

How about on the Vic Willis/Sam Leever thread?

Kelly's precisely in reverse on the depression, as I wrote on the Fielder Jones thread. Leever had a steady job teaching school, so stayed there even though he could have joined the disntegrating and poorly paying NL. Jones couldn't get a job as an engineer, so became a ballplayer. In both cases, the depression increased risk aversion, but the outcomes were different because their circumstances were different.

There is no evidence I know of that Leever made any attempt to get an MLB job before he actually did so. He didn't try out, or anything.

Well, if the Depression increased Leever's risk aversion, it didn't do so for long, considering he was playing in the minors in 1897. Historians have described the 1890s Depression as 1893-1896, but I am not aware there was a 19th century version of Alan Greenspan telling everyone there was nothing to worry about, that the Depression was over, 1897 is the year to try new things. It takes a while to get out of a Depression.

I think he developed a reputation in the Sunday leagues in Ohio for having a great curveball and the Pirates took a chance on him and signed him.

Oh, salaries for schoolteachers in 1895:
In Wood County, Ohio, in 1890 a male high school teacher made $52 a month. In 1894, $60 a month.

In Kansas City, KS, schoolteachers were often not paid because of the financial condition of the city. Businesses were sponsoring schools so that teachers could get paid or would donate unused space in their buildings for schoolrooms. If they got paid, it was $40-$60 a month.

Ohio isn't Kansas, which suffered badly from depressed farm prices. The depression was already lifting by mid 96 (true nadir was early 95, during the gold crisis), and was ended by the election of William McKinley and the Full Dinner Pail, which was followed by rapid upturn. Unlike FDR, McKinley put in place policies that ended the depression rather than prolonging it. Spring 97 is a full year after things had clearly turned up. Also, by then it was clear the NL wasn't going to disappear, which wasn't at all clear in '92-93.

I doubt that anybody had good breaking stuff in the mid-90's after they moved the pitching distance from 50' to 60'6". Your existing curve ball (for example) broke prematurely, so you had to redevelop it to the new distance. It was Nirvana for fastball hitters, with only a change-of-pace to worry about.

I also think he took unusual advantage of the Baker Bowl (1914, led the NL with 19 homers; all at home!), which knocks him a bit in my book.

Maybe it's just me, but isn't the ability to take advantage of your home park above and beyond what others can do a reason for extra credit, rather than demerits?

"Maybe it's just me, but isn't the ability to take advantage of your home park above and beyond what others can do a reason for extra credit, rather than demerits?"

No, in the discussion of the immortals, I think it merits demerits (say that really fast 3 times!).

Basically a unique ability to take advantage of a park shows that in most other times, in most other places the player wouldn't have been as valuable. Sure real wins result from that, and he gets credit, but I think a slight demerit is reasonable, because he probably wouldn't have been as good in other parks.

I'm not a big fan of adaptibility of skills in theory. I think skills are what they are, and some parks just suit some players better than others.

The Cravath Catechism
(with apologies to the Catholic Church, and Ulysses, Chapter 17.)

Part I

How should we approach Cravath's minor league career?

With an open mind.

Couldn't Cravath have just been a late bloomer?

It is not impossible.

You don't sound conviced.

I am not.

Why?

Because once he allegedly "bloomed", his career was historic.

Define historic.

Inner circle Hall of Fame historic.

Please explain.

Beginning at age 31 . . .

Wait. Aren't you concerned about multiple endpoints?

Age 31 isn't cherry picking. It is when he first became a regular with the Phillies. I will cherrypick later.

Okay.

From age 31 until the end of his career, Cravath had an OPS+ of 152 in 4241 plate appearances.

Is that good?

Yes.

Can you list every person in baseball history to ever have an OPS+ of greater than 152 from age 31 onward, with a minimum of 4000 plate appearances?

Barry Bonds, Babe Ruth, Ted Williams, Willie Stargell, Frank Robinson, Willie Mays, Hank Aaron.

Anyone else?

No.

Is Cravath helped by having a relatively short decline phase?

No. In each of Cravath's last two seasons, he was the oldest player in the National League (6th and 3rd oldest in all of the majors.)

Wow. And that's without cherry picking a start date?

Yes. Were I to cherry pick a start date, I would omit Cravath's age 31 season, where his OPS+ was only 119, lower than it was in his first abbreviated stint in the majors.

What happens then?

Cravath's OPS+ from age 32 onward is 157 in 3739 plate appearances.

Who had a higher OPS+ in more plate appearances from age 32 onward?

Babe Ruth, Barry Bonds, Ted Williams.

Anyone else?

No.

What if I lower the plate appearances threshold from age 32 on to 3000?

Throw in Mark McGwire.

So, you don't think Cravath was a late bloomer?

Cravath was almost certainly a late bloomer, if you define late blooming as a late career improvement. The question is really what base he improved from. When considering Cravath's pre-age-31 stats, it is important to have the right mindset of what his major league hitting level actually was. That way, one can view the reasonableness of whether he was below average, average, or above average in the years leading up to the period when he was one of the five to ten best age-adjusted hitters ever.

(Note, relatively complete Gavvy Cravath minor league stats will be forthcoming in the next day or two, time permitting. Feel free to let the anticipation grow amongst yourselves.)

Okay. Here's the raw numbers. Now that I have them, I'm not 100% sure what to do with them. I bolded all the places where he led his league. Maybe someone well versed in MLEs or Negro League stats could do some sort of translation here? I have lots of raw data, which I think indicates that the PCL was generally a pitchers league, aside from the fact that it was in the deadball era, but it's hard to tell how much. Also, while Cravath never led the PCL in homers, we was second 3 times, third place once, and fourth place once in his five years there. And while he led the PCL twice in doubles, he also had two third place finishes (1903 and 1904).

Sorry in advance for the formatting:

Maybe someone well versed in MLEs or Negro League stats could do some sort of translation here?

That would be awesome if that would happen, Matt.

Thanks for digging up the numbers!

Hmmm. The bold worked on preview. Let me try again . . .

I guess not. What's the good of "preview," if it doesn't look like it does on the preview screen when you post it?

Anyway, besides what you can get from bb-ref, he led the AA in homers and batting average in 1910 and 1911, and he led the PCL in doubles in 1906 and 1907. The other near-leads I mentioned above, except that he was also third in the PCL in stolen bases in 1907.


Thanks for digging up the numbers!

You're welcome. I actually now have every PCL stat available from 1903-1957. If you suspect any of your other favorite candidates have hidden PCL value, I'd be happy to look those up too.

Here are more complete stats for his AA career with Minneapolis:

Yr---G--AB---H--D--T-HR---R--AVE-SLG
09-125-413-120-23-07-04-060--290-409
10-164-612-200-41-13-14-106--326-505
11-167-608-221-53-13-29-147--363-637

(Walks are unfortunately not in the official stats.)

League averages:
Yr-AVE-SLG
09-237-300
10-243-310
11-268-357

Black Ink:
1910: hits, doubles, triples, home runs, AVE, probably SLG
1911: hits, doubles, home runs, AVE, probably SLG

He did play in a very good hitters' park.

I'll drop in here what I posted way back in the 1926 ballot discussion. That was before Gavvy had a thread of his own...

I'd like to join PhillyBooster on the Gavvy Cravath discussion. I did some looking into players who generally played around the time of Cravath to see if we could place some expected results for his 3 missing years of 1909-1911. (I'm counting 1909 as missing since he played so little ML ball before finishing the year in Minneapolis) Cravath's age 27 and 31 seasons are at the bottom. Above him is a set of OFs who played fairly significant amounts of time for ages 27-31. The list is not exhaustive since I limited it to players with a minimum OPS+ of around 100 at the endpoints of age 27 and 31. Plus I missed some players I'm sure. It isn't intentional - sorry about that. The table displays OPS+:

                                 Age
Name                27     28     29     30     31  

Sam Crawford       159    159    153    130    163
Birdee Cree        137    151    140    100    140
George Stone       145    192    151    131    121
Johnny Bates       132    130    126    122    127
Mike Mitchell      122     89    152    125    120
John Titus         114    135    152    124    129
Sherry Magee       119    137    156    128    103
Chief Wilson       126    134    101    107    110
Bob Bescher        115    115    108    111    109
Dode Paskert        95    126     96    122    106
Frank Schulte      137    156    105    113     95
Jimmy Sheckard     114    113    102    110    114
Tommy Leach         92    107    136    125    112
Zach Wheat         107    150    134    130    124
Harry Hooper       103    114    116    142    112
Bobby Veach        142    136    159    121    158
George Burns       146    128    142    120    107
Jack Gravey        102    111    110    106    100
Burt Shotten       116    131    107    134    126
Jimmy Walsh        101    105     77     94    109
Red Murray         115     97     85     74     98
Max Carey          126    114    120    104    116
Rube Oldring       104    101    113    108     99

Gavvy Cravath      136    ???    ???    ???    119
Median Values 
Excluding Cravath  116    128    120    121    112

The math is maybe a little rough but the general idea is we're talking a comparable player to the rest of the set. Was Cravath likely to be a 120 OPS+ hitter in those years? I think so. Three reasons:

1.He's above the median OPS+ values on the above list, a list including many players who didn't have a peak resembling Cravath's. Unless Gavvy was one of those few players who peaked ages 32-36 or learned to hit at that late age his expected level of performance was in that ballpark.

2.During his time with the Minneapolis Millers, especially 1910 and 1911, he played on a team of many players who performed in the majors. Otis Clymer, Dave Altizer, Claude Rossman, Hobe Ferris, Jimmy Williams are all on that team and had at least average major league seasons before going to the minors. In 1910 and 1911 Catcus led the team in both batting average and home runs above all these players.

3.He led the American Association in batting average and home runs in 1910 and 1911. Runs batted in were not kept for the league those years but its been written Cravath probably led the league both years. Would a two-time Triple Crown winner from a high minor league be overwhelmed at the major league level, especially after showing he was capable of hitting just a couple years prior? I lean toward no.

So what's it all mean? I'll probably keep him close to the bottom of the ballot. Man could hit the long ball at a time others could not. Had he been around a decade later his skills would have fit the environment better.

Reposting of a lot of info about Cravath and the PCL in 1905 from the 1906 Spalding Guide:

Here are team by team batting totals:

Team Games AtBats Runs Hits Avg. R/game
Sea   208   6851  681  1632  .238  3.27
LA    218   7089  766  1672  .236  3.51
Por   207   6676  667  1548  .232  3.22
SF    230   7290  838  1660  .228  3.64
Tac   219   7107  718  1604  .226  3.28
Oak   228   7337  687  1580  .215  3.01
Lg Avg: 218 7058  726  1616  .229  3.33

The Guide lists extra base hits / stolen bases / sacrifices for all players in 15 or more games played so approx totals of extra base hits are possible (Seattle, SF, and Oak had a lot of players split teams so I’ll divide their totals in half by team. Best I can do without any games by team.) I’ll also give the approx at bat totals for those players and the official team total for runs.

Team ABs Hits   2B   3B  HR  SBs  SAC  Runs SLG
Sea 6634  1563  222  49  11  360  270  681  .289
LA  6890  1635  280  43  27  376  256  766  .302
Por 6497  1547  278  41  16  300  261  667  .301
SF  7187  1637  255  50  10  415  315  838  .281
Tac 7021  1599  290  36  36  326  290  718  .295
Oak 7290  1585  255  52  11  363  201  687  .271

Oakland is not even that good. They shared 3 players with Seattle and I split them equally, but I think Seattle should get more of their stats, but I have no idea how to split them. But the idea is there.

Fielding Totals:

Team PO     A      E     FPct.
Tac  5815   2842   384  .958
LA   5931   2996   439  .953
SF   6284   3066   473  .952
Por  5482   2734   466  .946
Oak  6144   3146   538  .945
Sea  5525   2476   467  .945
Lg Avg: 5864 2877  461  .950

I don’t know if those are good defensive numbers or not. Only Tacoma had fewer than 2 errors per game.

Only two hitters played in more than half their teams games and hit over .300. Blinkenship of Sea in 106 games hit .311 and Brashear of LA hit .303 in 189 games.

Cravath:
Gms ABs Hit 2b 3b HR SB SH Avg. Slg.
204 703 182 33 09 09 44 44 .259 .370

Fielding:
Gms POs As Es Pct.
204 227 33 14 .957

He had the 8th best fielding percentage of an outfielder who played in at least half his team games in the outfield. I assume a corner spot from his PO totals compared to other full time outfielders (a few with 350-450).

Among players with at least 150 games, his average was 12th. He finished one spot ahead of Van Haltren, who played for Oakland.
Van Haltren
Gms ABs Hit 2b 3b HR SB SH Avg. Slg.
220 860 220 18 10 02 47 17 .255 .307

Cravath’s place in various totals:
Home runs: 2nd in league with 9 to Eagan of Tacoma who hit 21. Only 2 other players were over 5.
Triples: 7th with 9. League leaders were 2 with 11, Nealon and Hilderbrand for SF, then 4 with 10.
Doubles: 13th tied with 33. League leader was Nordyke with 57. Totals down to Cravath: 57, 50, 49, 49, 44, 41, 40, 39, 39, 39, 35, 35
Stolen Bases: 20th or 21st with 44.
Batting Average: 12th among players with 150 games played. If half of team games is sufficient, he finished 13th.
Slugging: I figured the slugging for hitters with BA over .250.

Player Team  Gms Slug    BA Isolated Power
Eagan    Tac 210 .430  .278  .152
Nealon    SF 207 .405  .287  .118
Brashear  LA 189 .397  .303  .094
Blnknshp Sea 106 .394  .311  .083
Nordyke  Tac 219 .381  .271  .110
Dunleavy Oak 227 .375  .264  .111
Cravath   LA 204 .370  .259  .111
Hoshldr P/SF 210 .358  .267  .091
McLean   Por 180 .355  .280  .075
Hildrbrnd SF 225 .345  .264  .081
Dillon    LA 216 .340  .272  .068
Waldron   SF 196 .332  .279  .053
VanHltrn Oak 220 .307  .255  .052
Irwin     SF 226 .304  .266  .038

Cravath was 6th or 7th depending on the minimum appearances. Only one person looks to have appreciably more power.

Can you list every person in baseball history to ever have an OPS+ of greater than 152 from age 31 onward, with a minimum of 4000 plate appearances?

The classic "we-can-make-a-group argument". Tweak the cutoffs slightly, and you can make a similar argument for Ken Williams or Jack Fournier. A HoM case for Cravath begins by researching the full story of Cravath vs. K. Williams, Fournier and other true comparables, and not with the the notion that he was some uniquely overlooked inner-circle type.

There's nothing really wrong with Ken Williams or Jack Fournier. They were almost as good as Cravath in lots of ways, but they never really dominated.

Black Ink (average HoFer=27):
Williams: 11
Fornier: 10
Cravath: 46

And that's without giving any extra credit to Cravath. Actually, I don't really think you could create a group that included inner circlers, Cravath, Williams and Fournier. Honestly, if you expanded the group any, the first person you'd be likely to pick up who isn't a HoFer (yet, or maybe ever) is Edgar Martinez.

But, if you want, I will give you that Cravath's missing AA (age 28-30) years would have been exactly as good as the average of Fornier's 28 & 30 (he missed age 29) or Williams's 29 & 30 (he missed age 28). You don't have to assume that Cravath would have dominated in his missing years -- he's got he "peak" argument locked up with his time in Philadelphia. All he really needs from Minneapolis and Los Angeles is enough "bulk" to add on to his "grey ink" type credentials.

(Walks are unfortunately not in the official stats.)

Oddly, in the PCL walks were recorded for pitchers, but not for hitters. It seems to me that, if you know how many walks there are to go around, you should be able to estimate how many walks each player had by comparing his at bats to games played (fewer ABs would likely mean more walks).

I'm not sure how to go about figuring that, though. Cravath was top 5 in walks every year from 1914-1918, so it seems like he wasn't a hacker. A reasonable estimate seems like it should help his case.

By the way, talks to Gary A for the AA stats. Based on what we have now, here's a complete raw "career line", including NL, PCL, and AA stats.

I recognize that adjustments need to be made, but they would be relatively small compared to, say, Negro leaguers since half of the stats were actually earned in the majors. Anyway, just for comparisons sake, I put a couple of other stats in parentheses.

G: 2666
AB: 9107
R: 1394
H: 2642
2B: 566 (Cobb was the active ML leader in 1920 with 429)
3B: 156
HR: 211 (Roger Connor held the ML only lead with 138)
SB: 297 (excluding AA)
BB: 551 (majors only)
AVG: .290
SLG: .456

talks to Gary A

Er, that's "thanks."

I've been traveling this last week, so I'm just now catching up on the discussion. Before I left I had managed to collect team data for most of Cravath's minor league teams, so together with the career data posted by PhillyBooster in # 29 and 31, I should be able to estimate his MLEs. I'll try to compile them this week.

I've been traveling this last week, so I'm just now catching up on the discussion. Before I left I had managed to collect team data for most of Cravath's minor league teams, so together with the career data posted by PhillyBooster in # 29 and 31, I should be able to estimate his MLEs. I'll try to compile them this week.

Brent will do a much better job with Cravath's MLEs than I ever could, but I attempted some back-of-the-envelope math to get a basic and conservative sense of what Cravath's MLEs might have looked like for each of his PCL and AA seasons.

I melded the very most basic portions of Brent's and Chris Cobb's approaches, using Brent's conversion ratios for R, AVG, and SLG, then trying to locate season-by-season WS comps for Cravath among MLBs in each year I was MLEing.

I also made some small assumptions
1) That Cravath would play 85% of his team's games each season, so 131 games in a 154 schedule. During his MLB prime, Cravath played in about 90% of his teams' games, but I wanted to be conservative.

2) That his walk rate would be the same as his walk rate in the majors (about one walk for every 7 or so ABs).

3) That his AA parks (described as good hitting parks earlier in the thread) would have a PF of 105. This could be too low, surely, but 105 seemed like a good starting point.

I didn't do anything to contextualize him within his league's or team's run-scoring environment, nor do I have any idea of the PF for his PCL years, so take these with a grain of salt and for only what they are meant to be: a conservative, unshaped notion of what Cravath's PCL and AA stats might look like in translation:

YEAR_AGE_ABs_R___H___2B_3B_HR_BB_RC__AVG/OBP/SLG
1903_22__504_62__127_28_7__3__66_60__252/339/353
1904_23__477_62__119_27_1__7__63_57__249/337/354
1905_24__451_48__108_17_4__5__60_49__239/328/328
1906_25__468_69__116_23_5__4__61_54__248/335/344
1907_26__442_70__123_28_2__6__58_63__278/362/391
1909_28__433_57__113_20_6__3__56_53__261/345/356
1910_29__489_75__141_27_8__9__62_78__288/368/431
1911_30__477_103_155_34_8__19_61_105_325/401/549

TOTALS
AB__3742
R___547
H___1002
2B__204
3B__41
HR__56
BB__487
AVG_268
OBP_352
SLG_389

Notice that these MLE total rate stats look almost identical to Cravath's 1908 AL season line: 256/354/383.

In a moment, WS estimates.

To give a rough estimation of Cravath's WS based on the above translation, I located players who had similar rate stats to Cravath's in the same season he had them. I usually came up with several who were in the ballpark, some higher, some lower, and I tried to mentally adjust for ballparks and positions. Then I just averaged the comps together and when appropriate spat out a number that seemed a little lower or higher as appropriate.

So the following list is Cactus's estimated WS, plus the comps I looked at, including their Win Shares
YR___WS: COMPS(ws)
1903_15: McGann(12), J Farrell(16), Ganzell(18)

1904_18: Wallace(23), Hartsel(21), Browne(20), Odwell(19), Dolan(16), Ritchey(22), Beaumont(24), Smoot(19), J Delahanty(20)

1905_18: Dahlen(24), Ritchey(17), Courtney(17), F Jones(29), Mcintyre(20)

1906_18: Isbell(26), G Davis(29), Wallace(23), Mcintyre(19), Scheckard(25), Nealon(18), T Leach (19)

1907_22: F Clarke(29), T Leach(29), Titus(22), Seymour(20), J Delahanty(17)

1909_20: H Davis(19), H Lord(21), Engle(23), Demmit(18), T Leach (26), Chance(14), Hoffman (27), Lennox(15)

1910_23: Merkle(20), Doyle(25), Byrne(27), Bates(24), Knight(23), Cree(22)

1911_30: Doyle(28), Schulte(31), Crawford(32) [Really, no one in 1911 has a truly comparable line, so this is one is really just a rounded off guesstimate.]

Total: 164

Add 164 to his 202, and you've got someone in the Keeler/Goslin/By Williams range.

Again, I was trying to make commonsense, conservative judgements, so feel free to disagree with my assessments. It should be noted, however, that if you look at the comparable players by year, they lend credence to the notion that Cravath was a late bloomer. He starts off with solid veteran types like McGann, jumps up to the borderline star area (McIntyre), then into the occasional All-Star area (H Davis, F Jones, T Leach), then finally in 1911 hits paydirt and starts the run of outstanding seasons that would continue in the NL.

Also, the persistent presence of first McIntyre, then Leach, then Doyle suggests the kind of arc his career was taking.

OK, so this is my conservative, back-of-the-envelope look at him, I can't wait to see if Brent's translations agree and can provide a more specific sense of the shape of Cravath's career.

Minneapolis Millers (not Miners)

Fitting, Gavy Cravath and Chuck Klein on the agenda together

Awesome stuff guys and way to go getting this starting again Philly Booster!

I am one of Gavvy's best friends, thanks for getting this back on the radar. I just don't see how he could have been that good from age 31 on and not be a HoMer. Worst case scenario is the McGwire comparison.

It should be noted, however, that if you look at the comparable players by year, they lend credence to the notion that Cravath was a late bloomer.

It's equally consistent with the conclustion that the National League during Cravath's peak was relatively weak and that his stats (as well as his black ink total) were especially inflated by his home park. Then his late peak is not out of context with the rest of his career.

Having said that, there is a case for Cravath as a HoMer. As I said above, however, for those making the case, it would be worthwhile to spend equal effort on the minor league years of Ken Williams and Fournier (and military service of Williams.

Two hurdles Cravath has to overcome to get on my ballot:

(1) I have significantly downgraded the other NL stars of his era for league quality reasons. (Doyle and Roush are off ballot and Carey was just off ballot when elected.)

(2) There are other A list hitters with lousy defense who are still in the pool but not in my top 25 (Browning, Tiernan, Fournier).

I need to be convinced that even giving Cravath credit for his minor league years (or an appropriate portion of them) that he is better than Roush and Doyle. I then need to be convinced that he is better than Browning, Tiernan, and Fournier. Neither task is impossible, but I'm not convinced yet.

It's equally consistent with the conclustion that the National League during Cravath's peak was relatively weak and that his stats (as well as his black ink total) were especially inflated by his home park. Then his late peak is not out of context with the rest of his career.

I don't think anyone is trying to argue that Cravath did not have a late peak. The object for reasonable disagreement, I think, is where he peaked from? If you think he was essentially a replacement level player before become one of the best players ever from age 31/32 on, then you won't give him any credit.

If you think, as I do, that he peaked from a base that was somewhat above average, then I don't see how you can possibly withhold credit from those "sub-peak", "abve average" years.

Seeing as how high his peak was, I just don't find a below-average pre-age-31 career reasonable. Besides, this isn't like Sam Leever, where we are debating what could have been. Cravath actually played the games in those years.

Dr. Chaleeko's rough estimate of 164 extra win shares is actually higher than I had been expecting. My uninformed estimate was 100-150 extra win shares, giving his 300-350 career. In my mind, 300+ Win Shares with a huge peak is an easy call. I'll be waiting expectantly for more complete MLEs, but I don't expect huge changes.

In short, saying he "peaked late" isn't enough, I don't think. It's where he peaked from that is more relevant, and is what the MLEs are starting to show.

EricC,

I agree that it would be appropriate to get the minor league figures for Fournier and Williams, though I suspect their cases would rely somewhat less than Cravath's on their minor league records.

As for Cravath, himself... Andrew Siegel, how significant is your deadball-NL discount? While I think we all agree that the AL was a stronger league, the NL was still a stronger league than the PCL, AA, or IL. Are you talking 10%? 5%? 2.5? I too apply a discount, but it's more a mental note than anything else, a reason to rank a player lower than someone else who appears otherwise equal but played in the AL of the same period.

In terms of Tiernan, Fournier, Browning. If you give Cravath MLE credit only for his AA years (1909-1911), he's essentially a clone for Tiernan, and Fournier (with no MLE credit) is third. Browning I'm not sure about because it really depends on what level of difficulty you see the AA as having.

Thing is, however, I'm not convinced that only giving Cravath credit for his AA years is appropriate. My rough and less-scientifc-than-Brent estimates above suggest that Gavy was a solid regular (15-18 WS) from 1903 through 1906. I estimate he'd have broken 20 in 1907. In 1908, he posted 12 WS in 94 games in the AL as a 27-year-old, a 20-win-share pace for a 162 season, which is entirely consistent with the growth pattern he was already showing.

Therefore, I think it's not too wild to suggest that his 1907 season should also be credited to him. I would, myself, take it a little further however. He debuted in the PCL at age 22. He was performing at the level of an MLB regular that year and just a year later was Matty McIntyre. I don't think it's unreasonable to suggest, therefore that after two PCL seasons, he was a major-league-caliber player.

I think giving him credit for 1905-1907 (ages 24-26) does not seem unreasonable.

So now he looks like this if you add in the 1905-1911 MLEs, by age:

AGE___WS (prorated to 162)
24____19
25____19
26____22*
27____13 (94 g in AL)
28____23* (2 WS in AL)
29____24*
30____32*
31____16
32____30
33____29
34____37
35____27
36____27
37____12
38____17
39_____2
========
TOTAL_350
*estimated

BEST ANY 3: 99
BEST ANY 5: 155
BEST ANY 10: 270
BEST ANY 15: 345

Again, I have no idea what his PCL PFs were like, and I only used a 105 PF for his AA years based on something said in this thread that said he was in a good hitter's park. They are estimates.

But what I'm driving at here is that Cravath was clearly MLB-ready by age 24, and if we accept that premise and accord him MLE credit, he looks an awful lot like Goose Goslin or Willie Keeler or Billy Williams do, valuewise: Good career total, decent peak, good prime/extended prime.

The career certainly puts him over Tiernan. And it would put him over Browning too, but again, his peak is a matter of discount structure. Finally, I do agree that we need more a more systematic study of Fournier's MLEs to see if we've been overlooking him. Ditto Williams.

I'm struggling with this one and reserving judgment, but I do have one point to make:

If we decide to give every player credit for all the seasons they were major league ready but playing in some other league, then the bar for what constitutes an average HoM career will go up. We need to compare Cravath, et al., to the new higher standard (say 350 or 360 WS) rather than the older standard (say 310 or 320 WS).

Is this true, though?

Other right fielders in the HoM (just running down the Bill James Top 100) -- Babe Ruth, Sam Crawford, Harry Heilmann, Elmer Flick, King Kelly, Willie Keeler, Sam Thompson.

Top 7 eligible candidates for RF(also BJ Ranking): Gavy Cravath, Sam Rice, Chuck Klein, Harry Hooper, Ross Youngs, Mike Tiernan, Wildfire Schulte.

Did any of these 14 (other than Cravath) have any substantial minor league service at major league quality? It is possible that one or two of them do -- but if so I haven't heard of it (and would like to!) Most played almost all of their assumedly prime years (say, 23-38) in the majors. For the few who didn't (e.g., Youngs, who died), I don't know of any high-minor league play. Perhaps, from that group, you could (maybe) point to one or two player who I missed who had an extra 40 MLE minor league win shares, but that's a far cry from a baseline that is 40 points higher.

It is my impression that when there is a credible case for minor league credit, someone brings it up.

a number of voters have/are giving Earl Averill minor league credit for one or two seasons. It doesn't raise the bar for CFs, but it situates Averill more appropriately within the context of his peers whose careers were not stalled. Cravath's case feels much more cut/dried than Averill's.

If you are giving credit for all seasons where the guy played somewhere and was good enough to play in the majors, just about every HoMer should get some credit, whether it be 6 months, a year-and-a-half, three years, or eight years.

Moreover, the bar gets raised for 2 reasons:

(1) You give credit for the additional WS earned by the guys already in the HoM.

(2) Guys like Cravath get a lot of credit and vault over other guys, knocking them to the other side of the in-out line.

I give credit for Averill for one year, Cravath for 2+ years, and Fournier for 2+ years.
My rationale for Averill is he played in a league that did not have to sell its players to the majors at set prices, so he could be held until a ML paid the right price. Also, Averill was playing at a consistent level over the 2nd and 3rd years in the PCL and continued at a similar level in the majors in his first year with Cleveland. This leads me to believe he was playing at a major league level the year before he got to the majors, he was just stuck in the PCL.
Cravath and Fournier get mid-career credit because they had established a certain level of better-than-average performance, but their teams did not recognize it. Then, they went to the minors and continued to perform at that level or better, returned to the majors and performed even better. They had not failed, their teams just failed to recognize their ability, which other teams later did. I don't know about pre-major league credit for either of them yet.

I don't know of any other players who were sent down after establishing (to our eyes, at least) that they were average or above average players who were released to the minors and returned to the majors.
Nichols went to pitch for his own minor league team (and didn't need the credit anyway.)
Vance was never healthy in his many trials.
Joe Jackson did not establish himself as a regular in 1908 with Phil.
That's all I can think of off the top of my head (I am at school now). I just don't remember any similar cases to Fournier and Cravath.

Andrew,

Whether or not you feel the bar is raised by offering war, minor league, NgL, or positional-bonus credit is really a personal call. I don't believe that's an accurate assessment, but I understand why we disagree.

But addressing your second point. Once you've determined someone was denied access to the majors or was of major-league quality, and once you determine the amount of MLE credit to give them, the whole point of it is, in fact, to see whether it makes a difference in how you assess the rankings that drive our balloting process. The very question being asked about Cravath is "Should he be getting more votes?" The discussion of MLE credits is a way to answer that question.

But let me also address one other point since you brought it up:

If you are giving credit for all seasons where the guy played somewhere and was good enough to play in the majors, just about every HoMer should get some credit, whether it be 6 months, a year-and-a-half, three years, or eight years.

I generally don't agree with this statement. For instance, no one is suggesting GVH should be getting any credit for his PCL seasons in the 20th century. When he retired as an MLB player, that's that, even though his PCL stats probably indicate he could have played on a couple more years.

What we're really dealing with here (and in most cases, except, say, Fournier or Charley Jones) is the start of a player's career, the ladder he climbs to reach the big leagues. Talent seems to have been more haphazardly dispersed among the minor leagues in Cravath's time, but there also seems to be a consenus that the AA and the PCL were probably the two highest-caliber leagues underneath the majors or among the top three. So just like when we examine pre-MLB service time time today, we should take dominance into account relative to age and the level of competition.

When we say that Averill or Cravath was posting major-league caliber numbers in the PCL, this is a far, far different thing than saying he was posting them in the (I'm making up a name here) Backwoods League.

So when at 24, in a typical growth year for a player, in his third year in the league, Cravath hits .259 and slugs .368, the league hits .229 and slugs .289(!!!), his average is 13% higher than the league average, his SLG is 27% higher than the league, and on his own team, he creates 18% of the runs that the RC formula can account for without knowing any walks totals, that's not really something to just overlook. That tells us a lot.

To post similar numbers in some lower-level league would not suggest major-league quality, doing it in the PCL very clearly does.

I don't believe Cravath should get full credit for every minor league season. I think it's appropriate to exclude his first two to err on the side of caution. But year three (1905) seems like a logical place to start administering credit because he repeated the level of dominance over the league he displayed in 1904 (i'm assuming 1905 was a particularly difficult one for hitters), suggesting he had matured as a player.

I don't really follow your argument, Doc.

If the point of this exercise is to determine who the best baseball players were period, then it doesn't matter whether the guy was playing in the NL, the Negro Leagues, the PCL, the Backwoods League, or the Army. If he was good enough to play in the majors he should get credit. (Obviously you need to put up much better numbers in the Backwoods League to establish that you were major league quality than in the PCL, but we can account for that.)

I think it is fairly obvious that many superstar players had a couple of years where they could have played in the majors at either end of their career. What I assume people are doing is saying that those seasons even out and we don't need to worry about them, only about the EXTRA big league quality seasons some players have. I have no problem with that assumption, we just need to be careful that we apply those rules fairly across eras and leagues.

Most every HOM quality player had a season or so of Major League quality in the minors prior to their career. This is what attracted the scouts attention. I dont' think any credit should be given for this simply because it is a lot of work for no result since it is evely dispersed. The only people we should be looking at are those who spent 3-4 years as an MLB caliber player in the minors.

jschmeagol,

You've said very concisely one thing that I've been unable articulate very well. Cravath's first couple of years look like typical minor-league ladder-climbing. It's thereafter that I belive he should be given MLE credit for producing at a major-league level at the highest non-major classifications.

In # 42 I promised to compile MLEs for Cravath. It's gone more slowly than expected - little things like Form 1040 and the home opener have taken precedence. (Note - thank you, Dr. C, for filling in your MLEs; mine shouldn't be that much different.) I'll try to finish up the MLEs soon, but first I thought I'd report on a couple of issues that I see in his minor league record, and solicit your comments on how to deal with them.

PCL league quality

When looking through PCL rosters during the seasons that Cravath spent there (1903-07), I noticed that compared to the 1920s, it seemed that fewer players went on the the majors and those who went generally seemed to have had less successful careers. This raised a concern that the league quality may not have been as high as in the 1920s.

I also noticed that the Guides during those years did not give the PCL the same prominence that was afforded to the Eastern League (as the International League was then known) and the American Association. (For example, in the table of contents to the 1906 Spalding Guide the Eastern League is the first minor league listed, followed by the American Association. The PCL appears much further down, between the Kansas State League and the Northwestern League.)

Logically, it certainly seemed possible that the quality of the league improved between the aughts and the twenties. The population of the major west coast cities other than San Francisco was still quite small in 1903, though they grew very rapidly during the next two decades. For example, from 1900 to 1920 the population of Los Angeles grew from 102 thousand to 577 thousand, while that of Seattle grew from 81 thousand to 315 thousand. Also, the league, which was founded in 1903, went through growing pains as it competed with other leagues (the "Pacific National League" and an "outlaw" California League that didn't honor the reserve clause). In 1907 the PCL contracted from six teams to four.

The clincher, however, came when I looked in the Reach Guides to try to find when the PCL was classified as one of the top 3 minor leagues (at that time known as the "Class AA" classification) along with the Eastern League and the American Association. It's an interesting story that is told in the 1909 Reach Guide, which began with a rebellion by the Eastern League and the American Association against the lower-ranked minor league teams, which however dominated their governing association, the National Association of Professional Base Ball Leagues. These two leagues withdrew from the National Association and attempted to get the major leagues to recognize them as a separate organization.

Eventually a deal was negotiated and the two teams were induced to rejoin the National Association, but one of the conditions was the formation of a "Class AA" classification that brought with it special priveleges. The Guide includes the following explanation of why the PCL also joined in the new classification:

The Pacific Coast League was granted the privelege of entering the new classification because of a promise made to it when it came into organized base ball that it would never be asked to accept a classification more than one point below the major leagues. When this was decided, Cal Ewing, the president of the Coast League, said of course he would go into the new organization because he wanted his league to advance as rapidly as possible, not because he wanted priveleges. He said he would not be doing his duty to his league unless he took everything that was handed him.

It appears that at the time of the agreement everyone involved, including the PCL, acknowledged that it was lower in quality than the Eastern League and the American Association, but the PCL expressed a commitment to raise its quality to a matching level.

Unfortunately, I don't really have any solid information on how much lower its quality was. Two possibilities are (a) that it was equivalent to a modern double-AA league, that is roughly twice as far from the majors as the triple-AAA level that I've used for 1920s and 30s PCL translations, or (b) that it was halfway between modern double-AA and triple-AAA in quality.

Does anyone have an opinion? It's hard to judge, since I don't have any solid information.

I'll also mention that the effects of this on Cravath's MLEs will be partly offset by the fact that the Los Angeles ballpark of that period was a pitchers' park. My other issue has to do with park effects and run environment, but it's late now so I'll wait to write that up tomorrow.

I meant to mention another major disruption to the PCL during Cravath's tenure there - the 1906 San Francisco earthquake. See this article for more information.

Here’s some more information on and questions about Cravath.

Batting record

To what PhillyBooster and Gary A posted in # 31 and 33, I’ve added two seasons at the end of his career—his 1921 season as player-manager for Salt Lake City in the PCL, and a brief encore season with Minneapolis in 1922. I’ve also filled in things like SH and HBP where available and rounded the averages using the computer, which sometimes changed them by .001. I believe these statistics represent his complete major and minor league batting record, unless there was another minor league season we may have missed (the NBJHBA reports that from 1899-1902 Cravath was a telegraph operator, playing semi-pro ball in his spare time):

Year  Lg   G  AB   R   H 2B 3B HR RBI SB CS BB SO  AVG  OBA  SLG  TB SH HBP

1903 PCL 209 804 108 220 51 13  7  -- 34 -- -- -- .274   -- .396 318 --  --
1904 PCL 211 771 107 207 39  4 13  -- 45 -- -- -- .268   --  380 293 22  --
1905 PCL 204 703  81 182 32  9  9  -- 45 -- -- -- .259   -- .368 259 20  --
1906 PCL 177 633 102 171 39  9  6  -- 34 -- -- -- .270   -- .389 246 --  --
1907 PCL 182 614 106 186 45  5 10  -- 50 -- -- -- .303   -- .441 271 27  --
1908  AL  94 277  43  71 10 11  1  34  6 -- 38 -- .256 .354 .383 106  8   4
1909  AL  23  56   7   9  0  0  1   9  3 -- 20 -- .161 .382 .214  12  1   0
1909  AA 125 413  60 120 23  7  4  -- 21 -- -- -- .291   -- .409 169 16  --
1910  AA 164 612 106 200 41 13 14  -- 25 -- -- -- .327   -- .505 309 41  --
1911  AA 167 608 147 221 53 13 29  -- 33 -- -- -- .363   -- .637 387 29  --
1912  NL 130 436  63 124 30  9 11  70 15 -- 47 77 .284 .358 .470 205 16   3
1913  NL 147 525  78 179 34 14 19 128 10 -- 55 63 .341 .407 .568 298 11   3
1914  NL 149 499  76 149 27  8 19 100 14 -- 83 72 .299 .402 .499 249 19   3
1915  NL 150 522  89 149 31  7 24 115 11  9 86 77 .285 .393 .510 266  7   6
1916  NL 137 448  70 127 21  8 11  70  9 -- 64 89 .283 .379 .440 197 15   5
1917  NL 140 503  70 141 29 16 12  83  6 -- 70 57 .280 .369 .473 238 16   1
1918  NL 121 426  43  99 27  5  8  54  7 -- 54 46 .232 .320 .376 160  8   1
1919  NL  83 214  34  73 18  5 12  45  8 -- 35 21 .341 .438 .640 137  4   2
1920  NL  46  45   2  13  5  0  1  11  0  0  9 12 .289 .407 .467  21  0   0
1921 PCL 112 341  62 111 22  0 18  --  3 -- -- -- .326   -- .548 187  7  --
1922  AA  52  90  14  25  3  0  4  19  1  0 10 14 .278 .363 .444  40  2   2

If there are any errors or omissions in these statistics, I welcome your input.

Run environment

The Angels played in Chutes Park, which appears to have been a pitchers’ park. As I’ve done with other PCL teams, I calculate the run environment based on the team’s offensive R/G, then use their W-L record and the Pythagorean formula to derive an estimate of opponents’ runs per game (OR/G). Note that the run environment is not the same as the park factor, which would need to be calculated from data on runs scored and allowed in home and away games (unavailable for these leagues). The run environment as I’ve calculated it is affected by the team’s offensive and defensive performance and by deviations from the team’s Pythagorean projection; however, I believe that most of the variation is attributable to ballpark characteristics.

Unfortunately, the Guides did not provide data on team runs for some seasons; I have data for 1904 from the 1905 Reach Guide and for 1905 from the 1906 Spalding Guide. For 1903 I have data for the Angels only from an article on minorleaguebaseball.com, but my microfilm was missing the 1907 volume of the Reach Guide that would have carried the 1906 stats, and the 1908 volume did not include PCL team batting data for 1907, though it did include the individual stats. So for 1906-07 I will use the average of the Angel’s major league factors for the prior 3 years.

For the 1909-11 American Association and the 1921 PCL, I used the same method described above; for the 1922 American Association I am able to use the actual data on opponents’ runs. (These data come from the relevant Reach Guides.) The Minneapolis Millers played in Nicollet Park, which was generally known as a hitters’ park, and Salt Lake City appears to have had a Coors Field-type phenomenon going on.

In the next table the run environment is calculated as (1/2)*((R+OR)/G), which is followed by its ratio (times 100) to the minor league and major league average run environments for the same season, which I call the minor league and major league (run environment) factors:

    
Year  Lg Team REnv MinLFct MLBFct
1903 PCL   LA 4.35      --   97.9
1904 PCL   LA 3.60    90.0   96.7
1905 PCL   LA 3.31    99.6   85.0
1906 PCL   LA   --      --   93.2*
1907 PCL   LA   --      --   93.2*
1909  AA  Min 3.18    96.2   89.7
1910  AA  Min 4.21   112.5  109.9
1911  AA  Min 5.27   113.9  116.8
1921 PCL  SLC 6.32   128.1  130.1
1922  AA  Min 5.82   110.1  119.4
* = assumed

As you can see, the Millers’ run environment factors for 1909 are notably low compared to the other Minneapolis seasons. My question is whether I should smooth out these factors by taking an average across the 3 seasons 1909-11 (MLBFct = 105.5, which is essentially what Dr. Chaleeko used), or alternatively, I should just use the values for each season. Because the average is close to Dr. Chaleeko’s, resolving this question is likely to affect the distribution of Cravath’s estimated MLE performance across those 3 seasons, but not his overall MLE for the 3-year period.

Arguments for using an average:
- For calculating park factors for the modern era, the consensus among the sabermetrics community is to take averages over 3 to 5 seasons to eliminate some of the year-to-year noise.
- Chris Cobb’s park factors for the Negro Leagues are based on several years data when available (though those park factors are based on much smaller samples).

Arguments for using each year’s factor:
- Bill James and Stats, Inc. used one-year park factors for early years when little information was available about changes in ballpark conditions.
- The introduction of the cork-centered ball in 1910-11 may have changed the run environment of Nicollet Park relative to other parks in the league.

I’m leaning toward using the factors for the individual seasons, which is what I’ve done for my other minor league MLEs. Any comments or opinions?

Brent #60
It appears that at the time of the agreement everyone involved, including the PCL, acknowledged that it was lower in quality than the Eastern League and the American Association, but the PCL expressed a commitment to raise its quality to a matching level.

Do you now mean the 1909 agreement rather than the earlier agreement between PCL and Organized BB?

--
The very low rank of the PCL in the table of contents to the 1906 Spalding Guide probably reflects publisher interest and perceived reader interest rather than perceived quality. The New England, Connecticut, and New York State Leagues are covered more thoroughly than the PCL in Sporting Life 1905-1913, and there may be others below class A that received more coverage --including the "outlaw" Tri-State League in 1909. But I doubt anyone considered the PCL weaker than fifth in player quality below the other As (American, Eastern, Southern, Western).

Have you looked at tables of contents much later than 1906?

--
By the way, what is your source for the Guides?
Are there multiple microfilm editions?
(Example of what can happen: The Library of Congress edition of Baseball Magazine is missing #1.1, although it includes the December 1907 prospectus that preceded #1.1.)

jschmeagol #58
Most every HOM quality player had a season or so of Major League quality in the minors prior to their career. This is what attracted the scouts attention. I dont' think any credit should be given for this simply because it is a lot of work for no result since it is evely dispersed. The only people we should be looking at are those who spent 3-4 years as an MLB caliber player in the minors.

andrew siegel #57 and Dr. Chaleeko #59 evidently agree.

But I'm doubtful, with Matt Philly #52.

Do you now mean the 1909 agreement rather than the earlier agreement between PCL and Organized BB?

Yes, I am interpreting the PCL paragraph from the article on the 1909 agreement as saying that the PCL was then generally considered to be somewhat inferior to the Eastern League and the American Association.

The very low rank of the PCL in the table of contents to the 1906 Spalding Guide probably reflects publisher interest and perceived reader interest rather than perceived quality.

Ok. I guess those illiterate westerners didn't read the guides. :-)

In later guides the PCL starts appearing with the International League and the American Association, though I can't tell you exactly when the transition takes place. (I've generally just photocopied the pages I was interested in.)

I'm using the Library of Congress microfilm edition of the Reach Guide. I'm new historical research, so I'm not really familiar with other editions that might be available.

Great stuff, Brent! I can now work on the numbers to see where he now falls on my ballot in '50.

Brent,

Are the numbers above MLE's or Cravath actual numbers? Right now I think that 1907 is the only PCL year that he would be deserving aof any credit, and of course the years in Minny should give him some credit.

Jschmeagol,

The numbers above are Cravath's actual numbers. (Sorry if I wasn't clear about that.) I'm planning to finish the MLEs soon, but I expect them to look fairly similar to those Dr. Chaleeko posted in # 43 and 44.

Brent #60:
For 1903 I have data for the Angels only from an article on minorleaguebaseball.com, but my microfilm was missing the 1907 volume of the Reach Guide that would have carried the 1906 stats, and the 1908 volume did not include PCL team batting data for 1907, though it did include the individual stats. So for 1906-07 I will use the average of the Angel’s major league factors for the prior 3 years.

What would be useful for this project? --Cravath project?
PCL team batting data 1903 06 07 21?
AA team batting data 1909 10 11?
Is there league-season for which individual batting stats would be useful?

I would say be specific by email, Brent, since I will need your email address if not street address in order to send anything.

But someone else in this forum may be able to photocopy and mail crucial batting tables from the PCL book by Carlos Bauer(?) or the AA book by Marshall Wright(?). I have Wright's Texas League, which does include a team batting table for every season. (For pitching, only individuals grouped by team, only earned runs and that by inference from innings and ERA.)

Brent #65:
Brent #65:
I'm using the Library of Congress microfilm edition of the Reach Guide. I'm new historical research, so I'm not really familiar with other editions that might be available.

The knowledge is specific to each publication that has been microfilmed. I don't know that info about any publ's has been coordinated. SABR should do that.

As a start, I now have from Brent: 1907 missing from LOC edition of Reach Guides. (Many years on one reel, I presume. Do you know the scope of the reel, Brent?)
and from my own research yesterday: 1908 May #1.1 missing from LOC edition of Baseball Magazine. (Vols 1-11 otherwise complete, ie thru 1913 Oct.)

Back to the substance of your research,

The numbers above are Cravath's actual numbers. (Sorry if I wasn't clear about that.)

I misunderstood your post, too. Cravath will have to wait to be reevaluated.

PCL History (Historical Society) with information about the encyclopedia by Carlos Bauer.

Paul Wendt wrote:

What would be useful for this project? --Cravath project?

The main gap right now is PCL team batting data for 1906-07 (especially team runs scored). Maybe PhillyBooster has them. (I believe he has the PCL History that you linked to in # 72.)

Even without these data, I will go ahead and finish the MLEs, assuming the Angels' run environment for 1906-07 is equal to the average for 1903-05.

Thanks.

great research, and I await more results. Gavy hasn't made my ballot yet, but he could. Interesting comps look like P Browning and C Jones and H Wilson. I can't see Cravath above Averill though.

The main gap right now is PCL team batting data for 1906-07 (especially team runs scored). Maybe PhillyBooster has them. (I believe he has the PCL History that you linked to in # 72.)

I have the book, and while it does not contain team data, it does contain data for each individual player on the team. It shouldn't be too hard to put the data into a spread sheet to come up with team data. Just time consuming.

I’m sorry I’ve kept many of you waiting. Over the weekend I did a translation of Cravath’s batting statistics to major league equivalents, but I didn’t finish (and still haven’t finished) the final step, which is calculating the win shares that are associated with the statistics. I think Dr. Chaleeko’s approach in # 44 to calculating win shares is as good or better than what I’ve done, so if Dr. C would be interested in recalculating them he is certainly welcome to. Otherwise, I’ll try to finish them in the next two or three days.

The main new information used in these MLEs compared to what Dr. C did in # 43 is that I have actual data on run environment for each minor league season except 1906-07, for which I have filled in an assumption based on the average for 1903-05 (see post # 62 – I am using each year’s run environment factor rather than an average across several years). Also, I’ve added Gavy’s end-of-career 1921-22 seasons with Salt Lake City and Minneapolis (though the 1922 season with Minneapolis is probably below major league replacement level).

There is one assumption on which I wish to differ with Dr. C. He assumed:

That Cravath would play 85% of his team's games each season, so 131 games in a 154 schedule. During his MLB prime, Cravath played in about 90% of his teams' games, but I wanted to be conservative.

While I agree with Dr. C that playing time is scarcer in the major leagues, so the playing time should probably be trimmed down, I disagree with just applying a flat rate of 131 games per year.

It seems to me that playing time depends on two factors: (a) injuries or the player’s need for rest (which are already reflected in the minor league playing time), and (b) the extent to which the manager chooses to use other players. Unfortunately the second factor is very difficult to predict, since managers and team are very heterogeneous. Some managers like to platoon, while others prefer a fixed lineup. Some teams have depth on their bench (or in their minor league system), so that many players are competing for playing time, while other teams have little depth.

Although it is probably difficult or impossible to accurately predict playing time, it seems reasonable that for a player who is projected to an all-star level (as Cravath is for his Minneapolis seasons of 1909-11) we should assume that he would have played except when he was injured or needed rest. So for these seasons, I took his actual playing time and adjusted it proportionally to a 154-game schedule. For other “good” seasons (1906-07, 1921), I reduced his playing time by 10 percent, and for “marginal” seasons (1903-05, 1922), I reduced it by 30 percent. Of course, if you disagree you are free to apply your own playing time adjustments. Also, some of you will choose to discard some seasons. (I will probably discard 1903-04 and 1922.)

Here are Cravath’s MLEs:

Year  Lg Age   G  AB   R   H 2B 3B HR BB  AVG  OBP  SLG OPS+
1903 PCL  22 104 410  42  98 23  4  3 30 .239 .291 .337   84
1904 PCL  23 101 380  40  90 17  1  5 29 .237 .291 .326   93
1905 PCL  24  98 354  35  84 15  3  4 32 .237 .301 .331   95
1906 PCL  25 133 493  63 119 28  5  4 47 .241 .307 .343  104
1907 PCL  26 124 436  60 118 29  2  6 42 .271 .335 .388  129
1908-09 AL
1909  AA  28 115 391  53 109 21  5  4 42 .279 .349 .389  132
1910  AA  29 150 563  77 164 34  8 10 70 .291 .370 .433  142
1911  AA  30 153 553 100 177 43  8 20 73 .320 .399 .535  161
1912-20 NL
1921 PCL  40  83 246  30  67 14  0  9 29 .272 .349 .439  103
1922  AA  41  33  58   6  14  2  0  2  7 .241 .323 .379   81

I also have some data on his minor league fielding, which I will post later. Basically, it looks like Cravath was a pretty good right fielder while in the PCL 1903-07, but his performance seems to have been below average while with Minneapolis 1909-11. Did he experience any kind of injury during 1908-09?

I didn't mention, for league quality of PCL in 1903-07 I decided to use a factor (in units of runs) of .71, which places it slightly above modern double AA minor leagues. (The factor for modern triple AAA, which I also use for the American Association and the 1921 PCL is 0.82). The reasons for using a lower league quality are discussed in post # 60 above.

(For comparison with Chris Cobb's factors, a .82 for runs translates to about a .92 factor for batting average, which is a little higher than the .90 he uses for Negro Leagues. A .71 for runs translates to about a .86 factor for batting average.)

Great stuff Brent.

I probably would not give him any credit for 1921-22, and I'm a huge fan. At this point, he was an established major leaguer, and they just decided he was done. Or he decided he was done. Or whatever, he was 40 years old at that point. He only played 89 games in 1919 and was basically a pinch-hitter in 1920. I'd say his ML value was done at that point.

On Cravath's MLES; clearly in 1907;1909-1911, he was a star player and would have been in the majors. I'm inclined to not give him 03 credit, but what about 04-05 when he was close to but still below average(at least in terms of OPS+)? I'm thinking my credit for him might only extend back to 1906 or 1907, the point when he was clearly and above average player in the bigs. I realize that, had those MLEs been actual ML stats, he would have had some value to his team(and thus add to his career value) as long as he was better than replacement. But those stats weren't real, and someone who hit at those rates might have been bumped for playing time or stuck in a platoon(did managers platoon back then?). My gut says to go back to '07, but I could see '06. I don't know if I'm arguing with someone here or if everyone agrees with me. Those who are going to give Cravath some ML credit, how far back are you going and why?

Brent,

I'm feeling like the walk rate might be a little low for Cravath's PCL years. His MLB walk rate was 1 BB per every 7 AB for his career, and specifically in 1908 in his first 90 some games, he was walking at that rate (actaully about one per every 7.3 AB). Then in 1909 in just 56 ABs he drew 20 walks (wow!).

The MLE walk rates are about half that for his PCL seasons (particularly the early ones, they gradually increase from there), which doesn't appear consistent with the MLB walk rates he established in his earliest seasons in the AL.

I'm a FOGC, but I hope you won't think I'm picking at your translations just to help his case. I want to get him right, not enshrined, and I just think the walks were a big part of his game all along (as demonstrated by his early AL walk rates), and so the MLEs are, therefore, undervaluing his probable offensive contribution---perhaps considerably.

But, then again, you may have had specific walk data for the league or team that suggest otherwise. If so, please share!

I agree that generally there is no good reason to award credit for post-age-40 post-career minor-league decline.

As to 1903-1907, I think that credit should be awarded for any year following an above-major-league-replacement-level year. And certainly credit should be awarded for an above-average-year that follows an above-replacement-year.

I would credit 1906-1907 at the full MLE value. Perhaps some deductions would be warranted for 1904 and 1905 (unless a re-examination of Cravath's walk rate brings those years up above average too).

Philley,
1st off, I am crediting him for '07 and probably '06. My problem is with the prior years when he was both below average and unestablished as a big-time player. I agree that those MLEs would had had value to his team had he put up those major league numbers in 03-05. The problem with using those number with a below average player is playing time. To give credit for the same PT, you have to assume he would wind up on a team where a below average/above replacement OF is an improvement so that he gets the PT. With an above average hitter, thats a pretty good assumption. Once a player is below average, that becomes a much greater leap of faith. Are you confident that 84 OPS+ would have gotten him into 104 ML games in 1903? I'm not even sure 84 OPS+ is above replacement for an OF.
Concerning a possible walk rate upgrade; that will help him in our minds, but I'm not certain it would have convinced a majorleague manager in 1905 to give him more PT.
Conclusion:
1903-No way
1904-05 I might give half credit to whatever WS estimates are created for those MLEs. There's no way I'm going to assume that those stats would have gotten Gavvy into 199 gms those 2 years. I'm still inclined to completely throw out those 2 years and start with '06.

I am not even sure I will give him 1906 credit, in fact I probably won't.

What assurances do we have that above repalcement level players were usually playing in the major leagues in the 1900's? I would assume very little.

The first season in which it even looks like he would have been productive in in MLB is 1906, and a 104 OPS+ is hardly impressive for an outfielder. How far above replacement level is that? And if he wasn't terribly impressive in 1906, why would teams give him a chance for 1907? Most every major league player had to have at least one good season in the minors to attract the majors attention. So I think that 1907 is even up for debate.

Also, why exactly did Cravath lose an MLB job after 1908? I honestly dont' know but if it was because he failed to play up to a standard where teams simply weren't interested in him (despite the potential that obviously existed) then should he receive credit for 1909? If his MLB exile was for reasons well beyond his control (i.e. not becuase of performance) than he deserves full 1909 credit. I think he deserves 1910 nad 1911 credit.

So to recap I dont' think he really deserves credit for 1903-1906. I think that 1907 and 1909 are in play and could go one way or the other but I am not agaisnt giving him credit for these years. If he was exiled in 1908 for all the wrong reasons or if his 104 OPS+ actually is impressive, then he gest both years.He desrves full MLB credit for 1910 and 1911.

Remember if we give Cravath credit for all or most of these years we must also do the same for everyone else who spent a few seasons in the minor leagues.

On credit pre 1907. Below average hitters need not apply. Here are the starting ML RF'ers of 1905 (by GP), with BP's EQA and Fielding rates (rank-ordered by me).

 EQA Rate
.330 102 Flick
.310 110 Crawford
.292 114 Seybold
.318 102 Titus
.287 109 Keeler
.274 106 Browne
.270 108 Maloney
.283  97 Green
.289  93 Selbach
.270 106 Clymer
.289  93 Frisk
.265 105 Dolan
.297  88 Lumley
.264 104 Dunleavy
.281  90 Anderson
(.270  87 Freeman)

Cravath's 94 OPS+ is similar to Dolan (97) and Dunleavy (91). If he can field as well as they can, then he's a marginal major leaguer. If not, then he's probably below replacement until he can hit better. (The other guys on the margin are butchers with better bats and/or reputations.)

I just went to baseball-reference.com and found out that Cravath had a 136 OPS+ in 1908 with Boston. It seems odd that they would then not retain him for 1909. He was well above league average in SLG and OBP (two stats that I dont' think were kept back then) but barely above league average in BA. A corner OFer that to many baseball men (i.e. those that only look at BA)might be barely above average isn't one to hold on to I guess. He wasn't particularly good in 1909 with Chicago but he only had a few AB's. After looking at this I think that he deserves credit for his play in 1909 with Minnesota (Minneapolis?).

However, I am still not sure about 1907. Sure he was obviously performing at an MLB level, even a star level, but was his performance prior to that godo enough to get him noticed in most conditions. If not then MLB teams may not have had any reason to sign him for 1907. Remember, your performance needs to garner the attention of MLB scouts for you to get signed by them.

Put this another way, should DiMaggio receive 1934 or 1935 credit? Ted Williams 1938 credit? David Wright 2004 credit? I don't think they do.

For all of the reasons JS states, I can't see how you can give Cravath credit for anything prior to 1907. There must be scores of young players who played their age 22-25 seasons at equivalent MLE rates. They finally get noticed or a spot opens up for them in their age 26 season or later. Only in 1906 does he finally have an MLE that makes him above average and therefore hard to keep down any more. I'd start him 1907.

It wouldn't be hard to find a dozen examples of players in the past 10 years with AAA MLEs at 104 and probably at 129. If I'm right about the 129, then I'd hesitate to even give 1907.

I have no problem with 1909-10-11 credit -- and appreciate the numbers to use. But with only those three years added, Gavvy still falls somewhat short of my ballot. Even with 1907, he's hard pressed to pass my outfield glut in the mid-20s of GVH, Ryan, Duffy, Roush, Averill.

I just read posts 45 to 60, which I'd missed, and address the group's views on some of what I raised in the post above. But I also think it is important to look at the number of MLers who went on to have non-HOM careers that also spent 3 or so years in the minors at 100+ OPS+ MLE level before being noticed. If they exist, and I think they do, that would demonstrate that there is nothing particularly unusual or unfair by giving no credit to Gavvy's age 22-25 seasons. No matter the era, his MLE batting averages for those 4 years never exceeded .241. Is it really a gross unfairness that he wasn't in the majors?

I agree with everything Daryn just said except for the scores of minor-leaguers with 129 MLE OPS+. 104 I agree, 129 I seriously doubt there's scores of guys that good and the minors not getting the call-up. 1907 credit is a no-brainer to me. I can go either way on 1906(but am now leaning towards 'no') and am officially refusing any credit prior to that.

For all of the reasons JS states, I can't see how you can give Cravath credit for anything prior to 1907.

I have to say that I'm leaning that way myself. 1906 appears to be his "hey, I'm about ready now, guys" year that almost every major leaguer has prior to his debut.

However, I am still not sure about 1907. Sure he was obviously performing at an MLB level, even a star level,

Quibbling with the "even a star level". A 110-115 OPS+ is average for RF during this decade. He's promising but he's not Elmer Flick or Sam Crawford who are routinely putting up 150+.

His competition in RF in Boston was Doc Gessler who had a career year in 1908 (162 OPS+), and an 8 year career with a 127 OPS+. In LF, it was Jack Thoney, who put up the same average as Cravath, but without the walks and slugging, but with double the SB's, less errors, and more range. It was a bad move, but may have made sense to the managers, Deacon McGuire or Fred Lake, a backup IF during the 1890's. In any case, Speaker and Hooper are starters the next year, joined by Lewis in 1910, all 22 years young that year.

Agreed with the forming consensus...1909-1911 were all years I hope we were giving him anyway. 1907 is the only year I'm even considering major league credit for before his Boston days, and even it is iffy.

1906 appears to be his "hey, I'm about ready now, guys" year that almost every major leaguer has prior to his debut.

1905 PCL  24  98 354  35  84 15  3  4 32 .237 .301 .331   95
1906 PCL  25 133 493  63 119 28  5  4 47 .241 .307 .343  104
1907 PCL  26 124 436  60 118 29  2  6 42 .271 .335 .388  129

Sorry. I think 1907 is that year. There's little difference between .237 and .241, or 95 OPS+ and 104 OPS+. He's kinda slow, and not that good a fielder (error-prone without good range). He has to HIT. 1907 announced he could, and he got the tryout.

The real question to me is why he didn't get that much interest in 1910 and/or 1911.

'A 110-115 OPS+ is average for RF during this decade. He's promising but he's not Elmer Flick or Sam Crawford who are routinely putting up 150+.'

Hitting 14-19 OPS+ points higher the the average at your position is pretty darn good. In our altered assessment, 1907 would be Cravath's rookie year. Is it fair to compare his 'rookie' year with the prime years of 2 HoMers?

Carl,

I think that Jim was just calling me on calling Cravath a star in 1907. a 104 OPS+ isn't terribly impressive for a corner outfielder, espeically one as good defensively as Cravath. How common was it for to be bought mid season?

Also, might Cravath have been decent defensively while in the PCL? 22-25 is usually the time when a player's defensive aiblity is at its peak according to MGL.

Sorry. I think 1907 is that year. There's little difference between .237 and .241, or 95 OPS+ and 104 OPS+. He's kinda slow, and not that good a fielder (error-prone without good range). He has to HIT. 1907 announced he could, and he got the tryout.

I agree that, since his defense is suspect, he needed more pop from his bat. In this case, you might be right that 1907 should be his breakout minor league season instead of '06.

Agreed on no credit after his NL career, but...

At the risk of sounding too apologetic for Cactus, and in the interest of just making sure, I think we should wait until Brent responds regarding the walk rates.

I think it's imperative to reserve judgement until then because it could mean a difference of 20 walks a year:

Here's Brent's numbers combined with Cravath's MLB numbers:

Year Lg Age AB BB OBP OPS+ AB:BB
1903 PCL 22 410 30 .291 84 13.7
1904 PCL 23 380 29 .291 93 13.1
1905 PCL 24 354 32 .301 95 11.1
1906 PCL 25 493 47 .307 104 10.5
1907 PCL 26 436 42 .335 129 10.4
1908 AL 27 277 38 .354 136 7.3
1909 AL 28 56 20 .382 92 2.8
1909 AA 28 391 42 .349 132 9.3
1910 AA 29 563 70 .370 142 8.0
1911 AA 30 553 73 .399 161 7.6
1912 NL 31 436 77 .358 119 5.7
1913 NL 32 525 63 .407 172 8.3
1914 NL 33 499 72 .402 160 6.9
1915 NL 34 522 77 .393 171 6.8
1916 NL 35 448 89 .379 147 5.0
1917 NL 36 503 57 .369 153 8.8
1918 NL 37 426 36 .320 106 11.8
1919 NL 38 214 21 .438 213 10.2
1920 NL 39 45 12 .407 145 3.8


Cravath's got a 7.04:1 AB:BB ratio in his MLB career and until he's done, he's not varying much from it. I think it's probably reasonable to suggest that he starts out with a 10:1 AB:BB ratio or maybe 9:1 in his first year and improves incrementally from there, but I think it's rather unlikely that he goes from a hacker in 1903 to a guy drawing a walk every 7.3 ABs in 1908. More likely he could walk all along.

So the difference could be as many 20-30 walks a year for some of those early MLEs. Add 30 walks to a guy's OPS+ and we're suddenly talking about a VERY different player, one who's much further from replacement than the current MLEs let on.

I think we should let Brent speak to this issue before consensus forms on Gavy's early years to judgement.

I think that Jim was just calling me on calling Cravath a star in 1907.

Exactly. His 1908 season is even better rate-wise, and he's a middle of the pack LF when BP includes defense. (McIntyre, Clarke, and Magee are the elite.) And Cravath's not a kid; he's 26 in 1907 so he may already be in his prime or maybe he gets better. It's not like he's doing this at 22.

Crawford is only a year older, and Magee is 3.5 years younger.

I think it's rather unlikely that he goes from a hacker in 1903 to a guy drawing a walk every 7.3 ABs in 1908. More likely he could walk all along.

This must be some horrible breach of etiquette, but I'm quoting myself here. Actually I'm calling myself out on this one: that's a pretty generalized statement and I should try to substantiate it instead of just tossing it out there.

I used the SBE to find players with similar MLB AB:BB ratios to Cravath's. I found every player between 3500 and 4500 big league ABs who walked between 500 and 600 times. [i'm using an old SBE that ends at 2001 BTW]. These forty players amassed 162681 AB and walked 21901 times for a collective 7.4:1 AB:BB ratio.

Then I spreadsheeted their age 22 through age 26 MLB seasons to see whether they grew into walking or whether they just always walked; in addition I compared their age 22-26 seasons to the rest of their collective career to see if there was any notable change between phases of their careers.

Here's how it all breaks out (there's rounding which is why 8.2-7.4 = .7...)

Career avg: 7.4

Age__AB:BB__diff from total
===========================
22___8.2____+0.7 AB
23___7.2____-0.2 AB
24___8.1____+0.7 AB
25___7.4____+0.0 AB
26___7.5____+0.1 AB

AVG 22-26: 7.6, diff: +0.2 AB

Rest of career: 7.3, diff: -0.1 AB

In nine of the forty cases a player's AB:BB average from ages 22-26 vary by more than one AB versus his career average. In most of those cases, the outlying variation is explained by the player having few or no ABs before age 27, leading to errors associated with smaller samples. [Minor league numbers to augment this study would be helpful, but...]

CONCLUSION
This little ministudy suggests that among players characterized by this career length and career walk total, by the time a player hits 22, there's not too too much untapped upside in his walk rate, probably within one AB. Therefore, while walk rate can and does improve, it's likely that someone who walks at age 27 onward has probably always walked.

Happy to share the data with anyone if they want it.

"his MLE batting averages for those 4 years never exceeded .241. Is it really a gross unfairness that he wasn't in the majors?"

What were the league averages at that time? Pretty low IIRC, no time to look it up now. I'd expect the league hit less than .250 during the early 1900s.