With Josh Gibson, that tandem easily has to be mentioned when discussing the greatest baseball duos that played on the same team.

1934-35 HOMers as Regular First Basemen:
Gehrig, Foxx, Greenberg, Terry, Charleston, JWilson, Leonard (ok, assumes)

In 1882 and 1885, we had:
Start, Anson, Brouthers, Stovey, Connor

Buck Leonard Data from Holway

Teams: 1933 Brooklyn Royal Giants, 1934-50 Homestead Grays

1933 No Data
1934 .330 for Homestead, 1b
1935 .332 for Homestead, 1b
1936 .260 for Homestead, 1b
1937 .356 for Homestead, 10 hr (3rd), 33 hr/550 (3rd), 8 2b (2nd), 2 3b (5th); 1b
1938 .396 for Homestead (2nd), 8 hr (2nd), 42 hr/550 (2nd), 1b, all-star
1939 .397 for Grays (3rd), 8 hr (3rd), 57 hr/550 (2nd); 1b, all-star
1940 .378 for Grays (3rd), 11 hr (1st), 27 hr/550 (4th), 21 2b (1st), 5 3b (1st); 1b, all-star MVP
1941 .294 for Grays, 9 hr (1st), 26 hr/550 (3rd), 6 2b (5th), 9 3b (1st); 1b, all-star
1942 .172 for Grays, 1b
1943 .321 for Grays, 6 hr (3rd), 11 hr/550 (4th), 20 2b (2nd), 11 3b (2nd); 1b
1944 .350 for Grays, 9 hr (2nd), 15 hr/550 (2nd), 23 2b (1st), 7 3b (2nd); 1b
1945 .365 for Grays (5th), 4 hr (3rd), 13 hr/550 (4th); 1b, all-star
1946 .265 for Grays, 19 hr/550 (5th), 4 3b (5th); 1b
1947 .419 for Grays, 36 hr/550 (2nd); 1b, all-star
1948 .395 for Grays (1st), 13 hr (1st), 45 hr/550 (1st); 1b
1949 no data
1950 no data

Career totals from Holway
779-2325, .335
79 hr, 15 hr/550 ab
10-24, .417 vs. major-league competition
Black Ink 22
Gray Ink 105

Career Totals from Macmillan 10th:
462 g, 1601 ab, 525 hits, 74 2b, 23 3b, 69 hr, .328 ba, .532 sa

I'll have MLEs in the next couple of days, estimated win shares within the week.

Whoa Nelly! Am I the only person surprised by these numbers?

In summary

Leonard .335-79 HR-1 per 37/22-105 black-gray ink
Beckwith .352-80-27.5
Suttles .341-237-14-56/145 black gray ink

I'm sure somebody can explain these!

Moore .355 (mean .365)
Wilson .354
Beckwith .352 (mean .366)
Gibson .351
Suttles .341
Charleston .340
Torriente .336
Leonard .335
Stearnes .332
Bell .328
Wells .328

Suttles 237 HR
Gibson 224
Charleston 197
Stearnes 197
Wells 138
Wilson 94
Beckwith 80
Leonard 79
Bell 73
Torriente 53
Moore 50

Gibson 1 HR/13 AB
Suttles 14
Stearnes 20
Charleston 24
Beckwith 27.5
Wells 27.5
Leonard 37
Moore 37
Wilson 46
Torriente 48
Bell 69

Suttles 56 black ink/145 gray ink
Charleston 54/174
Leonard 22/105

Torriente .436 vs. major league competition
Leonard .417 (10 for 24)
Stearnes .378
Gibson .375
Wells .353
Bell .342
Suttles .341
Charleston .330
Beckwith .319
Wilson .295

A big part of this is either underreporting or simply a lighter schedule. Notice that he was third in HR in 1945 with four. That's got to have something to do with the schedule he was playing or with the league context for power.

Also, was homepark were the Grays in during this period? I can never remember if they were in PIT or WAS, so....: If they were playing some or all of their schedule in Forbes Field, that might explain the seeming lack of HR power, and the same would be even more true if they were in Griffith Stadium.

As for park effects, that's what I meant when I said somebody could probably explain why Leonard doesn't look better. But Gibson was his teammate. In 1945 Leonard hit 4 HR and .365, Gibson hit 11 and .323.

Gibson's big years were 1937 (21 HR, .462) and 1943 (22, .449). Leonard hit a total of 21 HR and .321 and .356. Gibson hit .351 to Leonard's .335 and 3X as many HR/AB. So at a glance I don't know about park effects.

But of course nobody was Josh Gibson except Josh Gibson. The real question is whether Leonard was really as good as Mule Suttles?

But of course nobody was Josh Gibson except Josh Gibson. The real question is whether Leonard was really as good as Mule Suttles?

We'll have to wait for the MLE's. The 1920's guys have better raw numbers because offense levels were much higher in the 1920s.

I wait patiently for Chris' wisdom, but Leonard once again appears to bear out my thesis that the HOF grossly overrated a number of players in the 1930s/40s NEL, at the expense of earlier eras. Thus the HOM/HOF ratio for the NEL, which has been 2 or 3 to 1 for the early eras, may turn out to be well under 1 to 1 for the period immediately prior to integration. Even at the very top level, I expect to find that Smokey Joe was better than Paige, for example.

Chris' numbers may disabuse me of this thought, of course...

David Foss is right about waiting for the MLEs, but I don't think it's likely that Buck Leonard is going to show up as a better hitter than Suttles or Beckwith. He does, however, have a top-notch defensive reputation that they lack, so his overall value may be similar.

Chris Cobb-

Actually Leonard, like Scales and the great Jud, may end up looking a lot better than you think because his OBP is going to be very VERY high. Buck Leonard walked a lot, a real lot.

After the collapse of the Homestead Grays at the end of the 1950 season, Leonard played in the Mexican League from 1951 to 1953. In those three years, he played 227 games and walked 235 times (with only 35 strikeouts).

In other words, he walked more than once a game. The published stats from his Negro League days support the fact that he was walking a ton then too (very very low AB per game ratios).

Basically Leonard was an awesome fastball hitter. His basic batting strategy seems to have been to work the count, laying off breaking stuff out of the zone, while waiting for a fastball to crush.

I think he'll end up looking pretty good.

Mexico League 1951 to 1953, reduced to 154 Games:
G-AB-R-H-2B-3B-HR-RBI-BA-SA-SB-BB-SO
154 514 104 168 37 3 18 117 .326 .515 14 159 24

The walk data I have for Leonard from Cuba is very similar, though small sample size.

gadfly,

Thank you! I was just going to post asking if anyone had walk data from the Mexican League or the NeL for Leonard. The Riley biography mentions his high walk rates in Mexico, but it's good to have that confirmed by numbers and other anecdotal evidence.

Buck Leonard MLEs

Notes

1. For walks, I have used Mel Ott and Dolf Camilli as models. The MeL data provided by gadfly shows that Leonard had great plate discipline, but I don’t know how to translate the rates of early 1950s MeL into the NL of the 1930s and 1940s. I assume, therefore, that Leonard was at or near the top of his league in walks but does not exceed the best actual major-league players.

2. For playing time, I looked at Gehrig, Foxx, Greenberg, Terry, and Mickey Vernon as models. It’s pretty clear that first basemen at the top of their game played 150+ games a year unless they were injured. Leonard suffered a broken hand in 1942 which caused him to miss the East-West game and hindered him in the World Series, so I posit considerable missed playing time that season, but high durability elsewhere. All of Leonard’s slugging contemporaries left the game at a much younger age than he did. I’ve modeled his decline in playing time loosely on Mickey Vernon, who did play regularly past the age of 35 and who was of lighter build than the big sluggers. Leonard had more power than Vernon, but he was not a slugger on the Gehrig/Foxx/Greenberg/Mize model.

3. You may notice that Leonard’s slugging average and batting average rise in his final two seasons, at age 40 and 41. This is partly an artifact of regression: Leonard’s off-year in 1942 exercises an unusual downward pull 1940-1944, which I am tempted to excise by leaving it out of the regression formulas. The late years outside this pull balance out Leonard’s record somewhat. Also, offensive levels were rising at this time, so OPS+ may diminish the jump somewhat. There is a third factor surely at play here, however, which is the decline in quality in the Negro Leagues with the beginning of integration. Bill Byrd shows marked improvement from 1946 to 1948 while he was in the over-40 age bracket, just as Leonard does. The standard conversion factor probably overestimates these seasons somewhat. However, it should be noted that since these seasons (though not Leonard’s stats in particular) are part of the set of seasons used in calculating the conversion factor, this is another bit of evidence suggesting that the standard conversion factor may be slightly low for the rest of NeL history.

4. In all other respects, these MLEs follow the earlier ones. I look forward to seeing OPS+ numbers for Leonard, at which point I will begin work on batting win shares.



Year Team EqG PA BB Hits TB BA OBP SA
1934 Grays 126 529 77 148 210 .326 .424 .463
1935 154 647 96 166 223 .301 .404 .405
1936 153 643 98 165 270 .303 .409 .496
1937 151 634 98 170 264 .318 .424 .493
1938 152 638 99 185 305 .343 .445 .565
1939 154 647 103 183 312 .337 .443 .575
1940 154 647 105 177 291 .326 .436 .537
1941 151 634 105 145 244 .274 .394 .460
1942 85 357 58 65 92 .218 .346 .308
1943 154 647 107 146 217 .269 .390 .401
1944 152 638 108 157 222 .295 .414 .418
1945 140 588 92 157 207 .316 .422 .418
1946 135 540 81 131 197 .285 .392 .429
1947 120 480 72 143 221 .349 .447 .543
1948 100 400 60 118 201 .347 .445 .592
2081 8669 1358 2255 3477 .308 .417 .476

</pre>

-First you have Year, Team(s), PA.
-Second you have Chris's MLE's
-Third, in parentheses, you have pitchers-removed offense context. MLB for the 20s, then NL
-Fourth, you have AVG+/OBP+/SLG+
-Lastly, is the OPS+


1934 Grays 529 0.326/0.425/0.463 (0.287/0.342/0.408) 114/124/113 138
1935 Grays 647 0.301/0.405/0.405 (0.286/0.341/0.407) 105/119/100 118
1936 Grays 643 0.303/0.409/0.496 (0.286/0.345/0.400) 106/119/124 143
1937 Grays 634 0.318/0.423/0.493 (0.280/0.342/0.397) 114/124/124 148
1938 Grays 638 0.343/0.445/0.565 (0.275/0.339/0.391) 125/131/145 176
1939 Grays 647 0.337/0.442/0.575 (0.280/0.346/0.401) 120/128/143 171
1940 Grays 647 0.326/0.436/0.537 (0.272/0.337/0.391) 120/129/137 167
1941 Grays 634 0.274/0.394/0.460 (0.266/0.337/0.375) 103/117/123 140
1942 Grays 357 0.218/0.345/0.308 (0.256/0.328/0.356) 85/105/ 87 92
1943 Grays 647 0.269/0.391/0.401 (0.265/0.334/0.360) 102/117/111 128
1944 Grays 638 0.295/0.415/0.418 (0.268/0.335/0.377) 110/124/111 135
1945 Grays 588 0.316/0.423/0.418 (0.273/0.343/0.377) 116/123/111 134
1946 Grays 540 0.285/0.393/0.429 (0.263/0.338/0.368) 108/116/117 133
1947 Grays 480 0.349/0.448/0.543 (0.274/0.349/0.407) 127/128/133 162
1948 Grays 400 0.347/0.445/0.592 (0.269/0.343/0.398) 129/130/149 178
</pre>

Using 1934-48 data for Buck Leonard


Counting stats (+/- 2 for rounding?)
8669 PA
7318 AB
2256 H
3478 TB

Percentages
Leonard -- 0.308/0.417/0.475
Context -- (0.274/0.340/0.388)
Plusses -- 113/123/122

OPS+ = 145

Marginal, IMHO, but like Wells, probably just the right side of the line. Probably the first HOMer who made it primarily because of his walks -- the HOM Kevin Youkilis.

Among the list of OPS+ career leaders, only Edgar is coming across as a likely comp. Here's everyone within three OPS+ points of Leonard's MLEs:


Rank Player Adjusted OPS+
40. Lance Berkman 148
40. Harry Heilmann 148
40. Willie McCovey 148
43. Vladimir Guerrero 147
43. Edgar Martinez 147
43. Mike Schmidt 147
43. Gary Sheffield 147
43. Willie Stargell 147
48. Brian Giles 146
48. Todd Helton 146
48. Sam Thompson 146
51. Sam Crawford 144
51. Mike Donlin 144
51. Ken Griffey 144
51. Bill Joyce 144
51. Hack Wilson 144
56. Albert Belle 143
56. Harmon Killebrew 143
56. Eddie Mathews 143
56. Lefty O'Doul 143
56. Alex Rodriguez 143
56. George Stone 143
56. Harry Stovey 143
63. Jack Fournier 142
63. Frank Howard 142
63. Kevin Mitchell 142
63. Tip O'Neill 142
</pre>

Is he Edgar Martinez or Jaques Fournier with a glove?

Is he as good a hitter as Jud Wilson?

I'd say Edgar with a (superior) glove fits my image of Buck Leonard pretty well. They both seem to approach the maximum batting value for a player who is more of a line-drive, doubles hitter than a home-run hitter. Leonard's developing this still of hitting may have been influenced by his home ballparks -- Forbes Field and Griffith Stadium -- which were not homer-friendly but which gave good value for doubles and triples.

One could also say that he is like Goose Goslin, who also lost his home-run power to Griffith stadium, with a bit more plate discipline.

The Negro League summary table with Leonard added:

Player Pos Begin End BA OBP SA OPS+
Gibson C 1931 1946 0.327 0.431 0.595 180
Leonard 1B 1934 1948 0.308 0.417 0.476 145
Suttles 1B-OF 1923 1941 0.302 0.366 0.538 137
Beckwith 3B-SS 1919 1935 0.333 0.387 0.522 137
Wilson 3B-1B 1922 1938 0.336 0.431 0.447 132
Scales H 2B-IF 1923 1938 0.292 0.392 0.440 118
Wells SS 1926 1946 0.294 0.371 0.425 114
Scales M 2B-IF 1923 1938 0.277 0.382 0.413 109
Bell OF 1924 1946 0.297 0.365 0.382 100
Mackey C 1920 1941 0.301 0.359 0.393 98
Lundy SS 1919 1935 0.299 0.333 0.391 92


Player Pos Begin End G PA BB H TB
Gibson C 1931 1946 1930 7837 1210 2165 3941
Leonard 1B 1934 1948 2081 8669 1358 2255 3477
Suttles 1B-OF 1923 1941 2420 10163 924 2791 4967
Beckwith 3B-SS 1919 1935 1905 8010 648 2451 3847
Wilson 3B-1B 1922 1938 2352 9879 1413 2845 3789
Scales H 2B-IF 1923 1938 1986 8136 1153 2039 3072
Wells SS 1926 1946 2682 10836 1181 2839 4502
Scales M 2B-IF 1923 1938 1986 8136 1174 1936 2876
Bell OF 1924 1946 3230 13637 1371 3710 4665
Mackey C 1920 1941 2255 9020 745 2493 3249
Lundy SS 1919 1935 2212 9160 451 2600 3408

</pre>

Looks HOM material to me. He'll probably be #1 on my ballot in 1955.

I would like to see WS estimates but right now I do like him more than Suttles (I have a soft spot for high OBP guys) and Mule will be #3 on my ballot in 1954. Likely #1 for 1955 right now.

I also am waiting for WS estimates, but he looks like a very strong #1 candidate for 1955, with better hitting than Suttles, plus an actual glove.

His best competition may be Ray Brown, who also looks awfully impressive on a quick glance.

Well, I kinda knocked Leonard based on his raw numbers (above), which didn't include any BB data that I could see. Assuming the BB analysis is accurate, he now looks damn good--maybe not quite as dominant as I had expected. I had expected Josh Gibson. But short of Arky Vaughan failing to get elected in 1954, I don't see any reason why my 1955 ballot won't be:

1. Leonard
2. R. Brown

But, yes, I suppose I should shut up until the WS are in....and yes, Suttles is not that far behind.

Bump.

BUCK LEONARD MLE WIN SHARES

This is just my best guesstimate for Leonard. I used David Foss’s OPS+ numbers and the usual method of comparing Chris's MLE OPS+ to MLB OPS+es. Given Leonard's strong defensive reputation, I used Bill Terry as a general proxy with a rate of 3 fws per 154 games.


year bws fws ws
======================
1934 19.6 2.9 22.5
1935 17.5 3.5 21.0
1936 24.4 3.5 27.9
1937 26.9 3.5 30.4
1938 28.7 3.5 32.2
1939 29.1 3.5 32.6
1940 29.1 3.5 32.6
1941 25.4 3.5 28.9
1942 5.4 2.0 7.4
1943 19.4 3.5 22.9
1944 23.0 3.5 26.5
1945 20.6 3.2 23.8
1946 17.8 3.0 20.8
1947 20.4 2.6 23.0
1948 18.4 2.2 20.6
----------------------
TOTAL 325.7 47.4 373.1
</pre>

I’d love some feedback from our translation experts. Someone let me know if these look wrong. Thanks!

This estimate looks pretty good to me, but a little on the high side.

The batting win shares are plausible. If one prorates Harmon Killebrew's (OPS+ 143) batting win shares to Leonard's PAs, one gets 298.7. If one prorates Eddie Mathews' (OPS+ 143) batting win shares to Leonard's PAs, one gets 337.6, so I'd guess that Leonard's total would fall somewhere between 295 and 340. Leonard and Mathews are higher OBP players, so that would suggest that Leonard should project more like Mathews. I'd have to go season-by-season to get a more precise estimate.

The fielding win shares look a bit high. I project 8.7 defensive innings per game played, which gives Leonard 18,104.7 defensive innings. If his fws rate were as good as Bill Terry's (2.28/1000, and A+ rate, among the best of all time), that would give him 41.3 fws. Given that Terry retired young and most of the players with similar rates by win shares also did so and that Leonard played past 40, I would project him more around 2.10/1000, which would give a career total of 38.0 fws.

Thanks Chris,

Given that information, everyone should probably nick about .5 FWS per season off of my estimates.

Actually, just as a nit from a Twins fan, Mathews hit .271 and Killebrew .256, but Harmon walked more than 100 more times than Eddie and both have OBP .376.

Both hit over 500 HR, of course, with Harmon hitting an extra 60 HR. But Mathews hit more than 100 extra 2B and 3B, and amazingly both also have SA of .509. Mathews' .509 is incrementally higher, however, and Jeff Heath of all people squeezes between them on the all-time list. Mathews is 66th, Heath 67th and Harmon 68th (or was, as of the end of the 2003 season); there are a number of active players in the .510 range--Shawn Green, Bobby Abreu, Scott Rolen, Moises Alou--so the ranking could have changed a bit, but all 3 will still be at a nominal .509).

Big fan of Buck, but his peak isn't quite high enough, I think, but me to put him ahead of the two long-time peak candidates which lurk near the top of my ballot (or Brown, for that matter, who is the 2nd best NL pitcher we've seen, and easily, obviously the best pitcher currently on the ballot, IMO).

So my ballot will look like:

1. Brown
2. Ferrell
3. Jennings
4. Leonard

FWIW, if I were still doing my PHOM, Leonard would be inducted this year, as Wes and Hughie would have long ago been in (Hughie almost 50 years ago now!)

Actually, just as a nit from a Twins fan, Mathews hit .271 and Killebrew .256, but Harmon walked more than 100 more times than Eddie and both have OBP .376.

Not that this is a huge difference, but the context OBP was 4 points lower for Matthews while the SLG was 7 points higher. So while they amazingly enough have the exact same OBP and SLG, Matthews' OPS+ is slightly more OBP heavy than Killebrew's.

Big fan of Buck, but his peak isn't quite high enough...

Guess it depends on what you consider peak. Brown beats Buck on 3-year and 5-year peaks if you ignore consistency. But keep in mind that Brown was erratic with his best seasons. Buck was not.

Brown's three best seasons were 108.2. Buck's were 95.9. But Brown's three best consecutive seasons were 92.0. Buck's were 95.9. Brown's five best consecutive seasons were 140.6. Buck's were 154.2.

IMO, Buck had the better peak (and Buck has the better 7 and 9 year peaks no matter how you measure).

As for Jennings, it will be hard to find someone with a better 3 and 5 year peak, but take it to 7 years and Buck is right there with him. Brown lags behind a tad. (Of course Hughie basically stopped right there, while Buck and Browning kept going for another 100 or so WS).

All of Ferrell's peaks and career measures are behind Buck and Brown.

They are by WS...they are not by WARP (Buck at least; I have Brown #1 on my ballot. Of course, I'm sort of having to fudge a bit from WS to an imaginary WARP conversion as a primarily WARP voter, but you get the picture). As you know, WARP has a significantly more favorable view of both Hughie and Ferrell than does WS.

I also do not compare pitcher WS or WARP directly to hitter; the pitcher will lose out every time. There has to be some adjustment upward, especially for career, when comparing pitchers to hitters, or candidly I wouldn't have a single non-Ferrell/Brown/Mendez pitcher on my ballot.

I also do not count consecutive peak at all, so that's not affecting my rankings.

I also do not count consecutive peak at all, so that's not affecting my rankings.

Which is better:

Player 1:

WS:
15
22
22
32
32
32
15

Player 2:

32
15
22
32
15
32
22

They have the same career totals (170), and the same three year peak (96) and five year peak (140). Player 1 has a three year consecutive peak of 96 and Player 2 has a three year consecutive peak of 69. Player 1 has a five year consecutive peak of 140. Player 2 has a five year consecutive peak of 123.

IMO, Player 1 is better because consistency matters in baseball and you can count on player 1. Also, now that we know what an incredible amount of randomness is involved in pitching stats (see McCracken's DERA formula), and how little control the pitcher actually has, if Player 2's numbers are somewhat inconsistent from year to year, it seems like those random factors are a significant part of his numbers.

Seems to me Player 1 and Player 2 won exactly the same amount of games for their team, so they are equal. The unexpected bad years are balanced out by unexpected good years.

As for DERA, pretty sure WARP incorporates that to a degree, at least moreso than WS (which does so on the team pitching level, but not the individual level, and thus swings with luck among pitchers on a staff). Yet another reason WARP is superior.

On this business of consistency: if you're using "consistency" as a reason to have Leonard #1 and Brown #2 on your ballot this year, ok, Leonard probably was more consistent than Brown, but it's crucial to remember that "consistency" in the MLEs for Negro-League players is partly an artifact of the translation system.

Negro-League position players are generally going to look quite consistent in comparison to major-league players because their seasonal totals are regressed in the translations to try to get the general height of their peak correct.

Negro-League pitche3rs are generally going to look quite inconsistent in comparison to major-league players because their seasonal totals are _not_ regressed in the translations. I haven't regressed pitcher MLEs because (1) pitchers tend to have greater variance season-to-season than position players due to smaller sample sizes and greater frequency of injuries affecting performance and (2) I don't have any good ideas for how the regression should be handled.

But if you look at Leonard's regressed MLEs and say "consistent" and Brown's MLEs and say "inconsistent," much of what you are judging is what the translation system has determined that you will find.

Good to know, Chris, thanks.

Of course, I'm sort of having to fudge a bit from WS to an imaginary WARP conversion as a primarily WARP voter, but you get the picture).

I'd be interested in knowing more about how you are converting WS to WARP -- not for purposes of evaluating your player rankings, but because I am interested in the relationship between the two. Seems like if you derive WARP from WS you still have WS, just in another form (i.e., WS in WARP's clothing).

There is a straightforward relationship between Batting Win Shares and the hitting portion of WARP1. It seems to me that for position players, WARP differs primarily in three ways: (1) treatment of fielding, (2) season-length adjustment, which isn't significant during this era and (3) the timeline, which to my knowledge no one outside of the BP insiders knows the formula for.

By contrast, there seems to be very little consistency between pitching WS and the pitching portion of WARP1.

Anyway, just curious.

Definately agree on the pitching point...my rankings of Mendez/Redding/Brown et al are not based on any sort of number conversion, more a look at the WS, the DERA estimates, and the career shape, and comparing to the white players in my rankings. My Mendez comp is Ed Walsh (slightly shorter peak, but more value outside the peak) as an example.

I do something similar for batters. Using our OPS+ estimates, fielding rep, etc., I try to find someone similar and adjust from there. For Buck, it, like most people in the thread, was Bill Terry. Somewhat better hitter, longer career, so he's above where Terry would be, but I still try to find a benchmark white player from our MLEs, and then move up or down.

(FWIW, Mackey's comp is Schalk; higher OPS+, much higher peak, shorter productive career; Schalk is in my top 50, Mackey is just off the ballot. Bell is, of course, Carey. Some players defy comparison, unfortunately...their rankings tend to be more fluid.)

Good question though...something I hadn't really fully thought about. To answer it as best I can, I use batting WS, and then WS for a basic career shape (Does he have a peak? How long's the productive career?), but lean just as much on the MLEs for thought in a WARP world.