I have him as a borderliner at this point. Don't know if he makes my ballot, but I may still be more bullish about him than others here.

Between Vern Stephens and Fregosi on my list of SS, not close to the ballot. Actually he's another one I have near Nellie Fox.

A while back I posted probably 40 times or something ungodly like that advocating Concepción's HoF/HoM credentials. Perhaps that thread should be re-posted or at least linked here? Anyways, I'll put up the essentials of my case over the next couple of days. I am hoping to return to voting after about a 60-"year" lapse to advocate for my favorites as the HoM nears completion.

Back when I was getting into sabermetrics, I did a little thing using WARP to rank the best Venezuelan players of all time. Imagine my surprise when Concepción, not Aparicio, ended up on top.

If I did that again sometime now, I would imagine Bob Abreu getting close to the top, and Johan and Miggy Cabrera getting close (if not already on top) on a peak-only basis.

No.

The diff between WARP's view of Davey, and Win Shares's take, is ... Huge.

At Cy Morong's page,
http://us.share.geocities.com/cyrilmorong@sbcglobal.net/WSperPA.htm

there are lists of players ranked by career win shares per PA (min of 7500 PA). Concepcion is 55th.

you want career instead? There are at least 30 SS with more career WS, if you adjust for schedule length, than D.C.

you want peak? Concepion's best season is 25 win shares. There must have been 70 MLB SS with a better season than that.

WARP, on the other hand, holds D.C. about as highly as Joe Morgan does :)

I'll be interested in the analysis, but right now I've got him even with half a dozen backlog shortstops.

>I am hoping to return to voting after about a 60-"year" lapse to advocate for my favorites as the HoM nears completion.

I hope that "fair to all eras" remains a part of the deal.

Dan, I don't know if you have everything worked out, but by your system, how do Fregosi and Campenaris rank? They weren't playing in the most wonderful time for shortstops, either.

Very good, higher than most would have them I think, but not close to Concepción or to electability. I'll start posting charts now.

Glossary

Pos = Position.
BWAA/Yr = batting wins above overall league (not positional) average per season.
BRWAA/Yr = baserunning wins above overall league (not positional) average per season.
FWAA/Yr = fielding wins above positional average per season.
WAA/Yr = total wins above overall positionless league average (BWAA/Yr + BRWAA/Yr + FWAA/Yr) per season.
Av-Pos = Difference in wins between the offense (hitting plus baserunning) of an average player at the given position and overall league-average offense per season. WAA/Yr - Av-Pos = wins above positional average per season.
Pos-Rep = Difference in wins between a league-average player at the given position and a replacement player at the given position per season. WAA/Yr - Av-Pos - Pos-Rep = wins above replacement player at the given position per season.
WARP/Yr = Wins above replacement player at the given position per season (see above).
&#xSe;as = Fraction of the league average plate appearances per lineup spot for the given year.
WARP = Wins above replacement player at the given position (WARP/Yr * &#xSe;as).

All seasons are straight-line adjusted to 162 games where necessary.

Dave Concepción

Year Pos BWAA/Yr BRWAA/Yr FWAA/Yr WAA/Yr Av-Pos Pos-Rep WARP/Yr &#xSe;as  WARP
1970   6    -1.5      0.2    -0.6   -1.9   -1.7    -2.0     1.9   .46   0.9
1971   6    -4.7      0.0    -0.5   -5.1   -2.3    -1.6    -1.2   .55  -0.6
1972   6    -3.1      0.5     0.6   -2.0   -1.8    -1.9     1.7   .70   1.2
1973   6     1.4      0.7     1.2    3.3   -2.4    -1.4     7.0   .55   3.9
1974   6     1.0      0.6     1.6    3.2   -1.7    -2.0     6.9  1.03   7.1
1975   6    -0.4      0.5     2.3    2.3   -1.9    -1.8     6.0   .88   5.3
1976   6     1.0     -0.1     2.0    2.9   -1.3    -2.4     6.6  1.01   6.6
1977   6    -0.8      0.3     2.1    1.6   -2.2    -1.5     5.3   .97   5.1
1978   6     1.6      0.0     0.7    2.3   -1.0    -2.6     5.9  1.01   6.0
1979   6     1.3      0.2     1.4    2.9   -1.2    -2.3     6.4  1.05   6.7
1980   6    -1.1      0.2     0.5   -0.4   -1.6    -1.9     3.1  1.05   3.2
1981   6     1.9     -0.3     1.2    2.8   -1.6    -2.0     6.4  1.11   7.1
1982   6     0.3     -0.5     1.5    1.2   -1.1    -2.4     4.7   .98   4.6
1983   6    -2.6      0.0    -0.1   -2.6   -1.0    -2.5     0.9   .93   0.9
1984   6    -1.6     -0.3    -1.0   -2.9   -1.8    -1.7     0.7   .93   0.6
1985   6    -1.3     -0.4    -0.8   -2.5   -1.2    -2.3     1.0   .99   1.0
1986   6    -1.2      0.5    -0.3   -1.0   -1.4    -2.0     2.4   .54   1.3
1987   4     0.9     -0.3     0.0    0.6    0.3    -1.3     1.7   .49   0.8
1988   4    -4.2      0.0     0.4   -3.7    0.1    -1.2    -2.6   .35  -0.9
TOTAL                                                                  60.6

Agh, I had a slight error in the replacement levels, plus for some reason it didn't want to make a percentage sign. I'll call (percentage of season) SFrac this time. Take two:

Dave Concepción

Year Pos BWAA/Yr BRWAA/Yr FWAA/Yr WAA/Yr Av-Pos Pos-Rep WARP/Yr SFrac  WARP
1970   6    -1.5      0.2    -0.6   -1.9   -1.7    -2.2     2.0   .46   0.9
1971   6    -4.7      0.0    -0.5   -5.1   -2.3    -1.8    -1.0   .55  -0.6
1972   6    -3.1      0.5     0.6   -2.0   -1.8    -2.0     1.8   .70   1.2
1973   6     1.4      0.7     1.2    3.3   -2.4    -1.5     7.1   .55   3.9
1974   6     1.0      0.6     1.6    3.2   -1.7    -2.0     6.9  1.03   7.1
1975   6    -0.4      0.5     2.3    2.3   -1.9    -1.8     6.0   .88   5.3
1976   6     1.0     -0.1     2.0    2.9   -1.3    -2.4     6.6  1.01   6.6
1977   6    -0.8      0.3     2.1    1.6   -2.2    -1.4     5.3   .97   5.1
1978   6     1.6      0.0     0.7    2.3   -1.0    -2.6     5.9  1.01   6.0
1979   6     1.3      0.2     1.4    2.9   -1.2    -2.2     6.4  1.05   6.7
1980   6    -1.1      0.2     0.5   -0.4   -1.6    -1.8     3.0  1.05   3.2
1981   6     1.9     -0.3     1.2    2.8   -1.6    -2.0     6.4  1.11   7.1
1982   6     0.3     -0.5     1.5    1.2   -1.1    -2.4     4.7  0.98   4.6
1983   6    -2.6      0.0    -0.1   -2.6   -1.0    -2.5     0.9   .93   0.9
1984   6    -1.6     -0.3    -1.0   -2.9   -1.8    -1.7     0.6   .93   0.6
1985   6    -1.3     -0.4    -0.8   -2.5   -1.2    -2.3     1.0   .99   1.0
1986   6    -1.2      0.5    -0.3   -1.0   -1.4    -2.0     2.4   .54   1.3
1987   4     0.9     -0.3     0.0    0.6    0.3    -1.3     1.7   .49   0.8
1988   4    -4.2      0.0     0.4   -3.7    0.1    -1.2    -2.6   .35  -0.9
TOTAL                                                                  60.7

Bert Campaneris

I don't have his '81 or '83, but they don't really matter. Note that shortstop was *much* weaker by the late 1970's than it was in the mid 1960's, measured either by positional average (Av-Pos) or replacement level (Av-Pos plus Pos-Rep). In fact, the 50s to mid-60s were the historical high-water mark of the shortstop position, where it was about as strong as 2B, while the decade centered around 1980 was its nadir, when season-ruiners like Mario Mendoza, Alfredo Griffin, Jackie Gutiérrez, and Rob Picciolo were handed starting jobs.

Year Pos BWAA/Yr BRWAA/Yr FWAA/Yr WAA/Yr Av-Pos Pos-Rep WARP/Yr SFrac  WARP
1964   6    -0.7      0.2    -1.0   -1.5    0.5    -2.7     0.6   .46   0.4
1965   6     0.3      0.0    -1.4   -1.1   -0.4    -1.7     1.0  1.00   1.0
1966   6    -0.3      0.6    -0.7   -0.5   -0.6    -1.8     2.0   .98   2.0
1967   6    -0.9      0.3    -0.9   -1.6   -1.0    -1.6     1.0  1.04   1.1
1968   6     1.7      0.2     0.2    2.0   -1.4    -1.3     4.7  1.14   5.4
1969   6    -1.8      0.9    -0.2   -1.1   -0.2    -2.8     2.0   .93   1.8
1970   6     1.5      0.3     0.7    2.5   -1.1    -2.0     5.5  1.02   5.7
1971   6    -1.7      0.3    -0.1   -1.4   -1.7    -1.7     2.0   .98   1.9
1972   6    -1.8      0.9     1.2    0.4   -1.8    -1.7     3.9  1.13   4.4
1973   6    -1.8      0.4     1.3   -0.2   -2.8    -0.9     3.5  1.04   3.7
1974   6     0.6      0.4     0.5    1.5   -1.5    -2.3     5.3   .92   4.8
1975   6    -0.5     -0.1    -0.5   -1.2   -2.3    -1.7     2.8   .90   2.5
1976   6    -0.7      0.5     0.4    0.2   -1.4    -2.6     4.3   .97   4.2
1977   6    -2.2     -0.1     1.3   -1.0   -2.3    -2.0     3.4   .95   3.2
1978   6    -5.6      0.6    -0.2   -5.1   -1.6    -2.7    -0.8   .47  -0.4
1979   6    -3.4     -0.2     0.6   -3.0   -2.3    -2.2     1.5   .43   0.6
1980   6    -1.8     -0.1    -1.0   -2.9   -1.6    -2.9     1.6   .36   0.6
TOTAL                                                                  44.1

Jim Fregosi

Fregosi I only have through 1971. After that he was moved off shortstop and never had another decently valuable season again; I don't think he had more than three career WARP after being traded for Ryan. I'm also missing his '62 which is probably worth a win and change. So add on five or so to this total.

Year Pos BWAA/Yr BRWAA/Yr FWAA/Yr WAA/Yr Av-Pos Pos-Rep WARP/Yr SFrac  WARP
1963   6     1.5      0.0    -0.8    0.7   -0.4    -1.7     2.7  1.01   2.8
1964   6     2.8      0.2     0.0    3.0    0.5    -2.7     5.2   .92   4.8
1965   6     1.8      0.3    -0.3    1.8   -0.4    -1.7     3.9  1.04   4.0
1966   6     1.3     -0.1     0.8    2.0   -0.6    -1.8     4.4  1.07   4.8
1967   6     2.5      0.2    -0.8    1.9   -1.0    -1.6     4.4  1.03   4.6
1968   6     1.2      0.2    -1.1    0.3   -1.4    -1.3     3.1  1.09   3.3
1969   6     1.9      0.2    -0.2    2.0   -0.2    -2.8     5.0  1.08   5.4
1970   6     2.9      0.0     0.1    2.9   -1.1    -2.0     6.0  1.06   6.3
1971   6    -0.5      0.1    -1.2   -1.6   -1.7    -1.7     1.8   .62   1.1
TOTAL                                                                  37.1

Plus five max makes 42.

Ozzie Smith



So yes, Concepción was basically Ozzie, just for much less time. His peak is extremely similar, a tad worse (Ozzie 7.6, 7.2, 6.5, 6.4, 6.4; Concepción 7.1, 7.1, 6.7, 6.6, 6.0), but he only had two or three seasons outside of that as an All-Star level player and then practically none as a solid regular, while Ozzie has the entire first half of his career. Saying that Concepción's worse than Ozzie doesn't mean he's not Meritorious, of course, since Smith is a no-brainer. As I've said repeatedly, Concepción is right on the borderline. He had eight years playing at a strong All-Star level (but not really an MVP candidate), and nothing else.

Dan, could you give us Cal Ripken, Alan Trammell, and Robin Yount for comparative purposes (SS only if you want). Just to get a sense of the scale we're operating on.

I forgot to mention, what an incredible baserunner Ozzie Smith was! It's a very substantial chunk of his value--nearly 10 career wins above average on the basepaths. Hard to do.

Barry Larkin

Concepción's successor, and another no-brainer.



A well above-average Hall of Famer. Also an outstanding baserunner (8.2 career baserunning wins above average). He was the second-best player in the NL in 1988 after deserving MVP Gibson, the second-best player in the NL in 1990 after deserving MVP Bonds, should have won the MVP in 1991 (besting Bonds and Terry Pendleton), and probably took his 1995 MVP from Bonds but he was the 2nd-best player (or maybe 3rd after Piazza, I haven't done catchers yet). Talk about underrated.

SS is the only position I have AL data for, and only post-1960. So I don't have Ripken's 3B or Yount's OF years.

Let's just be clear that the early 1980's AL was probably the all-time high for standard deviation at the SS position. You had three absolute no-brainer Hall of Famers, one of whom, Ripken, is inner circle (2nd or 3rd greatest ever at the position), and then NOTHING--Houston Jiménez, etc. All three were, thus, extraordinarily valuable commodities. Yount was the MVP in '82, Ripken in '83, it should have been Ripken in '84 by a landslide, and Trammell was robbed in '87. It's fitting that their teams won the pennants in 82-84. Of course Concepción will look like nothing compared to that Holy Trinity, but all three are waaaaaaaay over the in/out line. As you may recall from the debates I got into the last time Concepción was discussed, I don't believe in comparing people either to the top or to the middle of the distribution (Jeter/Nomar are no less great because ARod is in the league). I believe in comparing them to the bottom--that is, the replacement level--which reflects their actual value to a team. It may very well be that the enormous spread in talent among early-80s AL shortstops was nothing more than random bunching, but that has major effects on wins and losses. My argument for Concepción is not that he was "the best of a bad lot"--I wouldn't elect Gil Hodges for being the "best full-time 1B of the 1950s." My argument is that had he gone down, Cincinnati would have lost more marginal games than, say, San Francisco would have if Willie McCovey had gone down in many years.

Here goes.

Cal Ripken

This is actually probably conservative on his defense, since Win Shares only thinks he was good-not-great in the field once you factor in he was playing 162 games a year. I'm sure he had some +20 run defensive seasons in there that I am not giving him credit for.

Robin Yount

Alan Trammell

Ernie Banks

An interesting one. He is a HoM'er, based mainly on the strength of his peak. But he's a lot closer to the borderline than you'd think--not only because he played just barely above replacement level at 1B for most of the 1960's, but because the 1950's NL was a golden age for SS--the position was just as deep as 2B at that time.

I don't have positional averages or non-SB baserunning runs as we go back further in time. So instead of Av-Pos and Pos-Rep, there is just one Av-Rep column--the difference in wins, per season, between a positionless player hitting and fielding at the league average and a replacement player at the given position.

whoops, forgot to put the career total for Banks! It's 62.6.

Good stuff Dan.

Please respond to comment #7 when you get a chance :-) Thanks!

Regarding presentation of any work of this kind, I prefer to see the column sums, the career (or career-at-position) totals for the components of the grand measure.
And which components are by Dan Rosenheck? (whose 'h' I just notice)

What would suggest I wouldn't be fair to all eras?

I've just completed all non-catcher NL starters from 1893-2005, and developed a (fairly sophisticated, I think) adjustment for the standard deviation of performance in each era (which further strengthens Concepción's case). How do I get a thread to post the spreadsheet on the HoM? Much easier than typing in 8,500 player-seasons!

To get column sums, the easiest thing to do is just multiply each column by the Seasonal Fraction for that year and then add up. I could do it myself but clearly the simplest thing is to just post the spreadsheet.

All components are by me, what do you mean? My methodology is outlined in detail on the old Concepción-for-the-Hall thread (I'll repost it when I get a new thread to post all my data).

The diff between WARP's view of Davey, and Win Shares's take, is ... Huge.

Win Shares doesn't really say anything bad about him either though.

Here are is MLB-SS WS rankings from 73-82:

73-4
74-1
75-3
76-3
77-3
78-1
79-2
80-8
81-1
82-8

Win Shares just doesn't give much of a bonus (if any) for being the best of a weak lot of players.

What would suggest I wouldn't be fair to all eras?

I think it was this:

I am hoping to return to voting after about a 60-"year" lapse to advocate for my favorites as the HoM nears completion.

I think it's a misunderstanding, but this could be read to mean "I'm coming back now because my favorites are only now eligible, and I'm going to come back and push for them, and only them, now."

I think (correct me if I'm wrong) that's what Mark and Joe were getting at--they were worried that the "favorites" you speak of all debuted after 1965 or so.

That is correct Mark . . . you aren't wrong :-)

Oh no. As a matter of fact I stopped voting precisely because I didn't feel I could be fair to all eras without determining the value of every player-season in MLB history. Which is now what I have decided to do! <gulp>...

All components are by me, what do you mean?

Then you should come up with a name other than "WARP".

Actually, my defensive FRAA are a heavily modified mix of Fielding WS, BP FRAA, and Dial, UZR, and Fielding Bible for when they are available, so those components are not by me.

That is a fair point. Nonetheless, I *am* measuring Wins Above Replacement Player...do you have a suggestion for a new name?

Chalek, we need an ackronym, and it can't suck!

OK, boss, how about?

-WiCER: Wins Contributed in Excess of Replacement [pronounced Wicker]
-WinCER: Same, but pronounced Wincer
-WinGER: Wins Gained in Excess of Replacement

Chalek, you numbskull! WinGER is too close to Wing Man, and Coors has that one locked down.

OK, boss, how about this?

-Total CARBs: Total Contribution Above Replacement Baseline.

Better, but the nutrition angle might turn off the overweight.

Uh, how about this one, then?

-PALACE: Position and League Adjusted Contribution Estimation.

Too regal, we're not trying to appeal to Zsa Zsa Gabor here. You'd better come up with something on the double because I'm meeting with the big boss in ten minutes!

No, wait I've got it:

-POWER: Player's Optimized Wins in Excess of Replacement

Hey, I really like that! It works, still, we might do better. Consider it a fall back.

OK, boss, I've got the perfect one for us:

-MR. ACE: Multiple Regression Adjusted Contribution Estimation.

What? My name? Well, frankly, I like it, it's a keeper, Chalek. Tell the boys down in the copy room to get cranking on our next campaign, and I'll get it okayed by the suits upstairs.

the Reds annoced they'll be retiring Concepcion's number
<url=http://news.cincypost.com/apps/pbcs.dll/article?AID=/20070501/SPT05/705010329/1027>see here</url>

Move over Hrbek and Horton ;-)

Tomorrow is when number 13 is officially retired:

mlb.com

I've been threatening to look at Davey Concepcion to see if other methods corroborate that he delivers the kind of positional advantage for his team that DanR's methods find. And for one thing, DanR is correct: the SS position in the NL during Concepcion's day was a freakin' leper colony.

Larry Bowa had full time work for the first 12 years of Concepcion's career with OPS+ 57-66-72-38-75-94-65-73-91-71-70-87, and THEN the Cubs swapped Ivan DeJesus to get him and he rewarded them with--well, what did you think?--68-77-49. All the while providing a C glove. Yikes.

DeJesus went 75-88-89-77-44 (after which the Phillies acquired him)-73-85-79, at which point the Phillies went to Steve Jeltz 41-60-66-54. At this same time Chris Speier had himself a pretty nice career at 80-116!-87-91-111-72-79, at which point the Giants decided they needed a little less offense at SS and traded Speier for Tim Foli who at that point had gone 59-55-63-58-80-49. The Giants immediately moved Foli to the Mets and installed Johnnie LeMaster (78-77-59-70-50-77-56). With the Mets Foli continued 71-83 and got traded for Frank Tavaras (63-52-95-69-82-75 with Pittsburgh, and 75-80-58 with the Mets). Foli finished up with the Pirates at 74-64 and was replaced by Dale Berra. But back to Speier: with Montreal he went 79-87-78-91-70-88-88, and was replaced by Angel Salazar 7 (yes, that's right, 7). Bill Russell was 99, 81, 97, 52, 85, 78, 92, 80, 79, 64, 99, 64. You don't even want to know about Bud Harrelson, but Templeton and Ozzie were a little better.

Atlanta had the following succession: S. Jackson 71 M. Perez 60-50-79 C. Robinson 52 Blanks 61 Chaney 82 Rockett 64 Royster 78 Frias 62 L. Gomez 26 R. Ramirez 63. Ramirez, of course, stuck, but the Braves had had a different SS every year from 1973 to 1981 and never an OPS+ above 82.

Long story short. During Concepcion's 16 years as the Reds' SS there were a total of 17 seasons of OPS+ ?100 by NL SSs. Concepcion had 5, Templeton and Thon 3, Speier and Brooks 2, and Menke and Uribe 1.

Concepcion was better than the median NL SS in OPS+ 12 times and below the median 4 times. In the median year, he was 8.5 OPS+ points above the median NL SS. (One might say that, despite this incredibly low standard, Concepcion was only 8.5 OPS+ points above the median NL SS in his median year on that measure.)

His most typical years in this respect were 1979 at +7 and 1980 at +10. Let's look at 1980.

Concepcion OPS+ 84 and 17 WS
Templeton 108 and 17
Speier 91 and 12
Tavaras 80 and 10
Russell 79 and 12
DeJesus 77 and 14
Foli 74 and 12
O. Smith 70 and 17
Bowa 70 and 9
C. Reynolds 63 and 8
LeMaster 59 and 4
Gomez 26 and 4

Foli was pretty much the median at 74 OPS+ and 12 WS (!) and Concepcion's advantage over Foli is 10 OPS+ points and 5 WS. Davey earned 9 offensive WS and 8 defensive, Foli 4 and 8. Russell and Speier also earned 12 WS, both at 7 on offense and 5 on defense. Against those 3 more or less median SS, Concepcion's advantage averaged out to 3 on offense and 2 on defense.

I don't like 1982 as well because the Reds sucked (61-101) and I wouldn't use an "extreme" year for his team as a "typical" year for him. But that year he was +14 on OPS+, and +1 offensive WS and +2 on defense. In his +7 year his advantage would be expected to be a little less, and those are his 3 most typical seasons.

So +3 WS on offense and +2 on defense versus his positional cohort seems to be a fair summation of how WS sees him.

So far:

Concepcion +3 offense + 2 defense = +5 WS
Doyle +10 offense -1 defense = +9 WS
Rizzuto +1 offense +4 defense = +5 WS
Nettles +5 offense +2 defense = +7 WS

This at least is what WS says.

1. Didn't Concepción have six seasons above 100 OPS+?

2. You're not comparing apples to apples here. Doyle came into the majors almost fully formed--he had a 135 OPS+ in his first full season--and he retired young. So his career consists of his prime, nothing more and nothing less. Concepción accumulated virtually all of his career value in his 1973-82 prime--his seasons before and after that period were near and sometimes even below replacement level. Since you are using an average-based metric, Concepción is being severely penalized for the years outside his prime, while Doyle is not. That rigs the comparison in favor of Doyle quite unfairly in my view. The same is true of Nettles, who was awesome from 1970-78 and had a bunch of crappy years outside of that period which drag down his numbers relative to positional average by this method. I would really appreciate it if you would repeat your study looking only at the top 10 years of each player, to put them on a level playing field.

I generally only care about best 10-12 years for prime candidates.
Don't get revved up in the counting stats, but don't pummel somebody for playing too soon or too late. I figure that's a management error, not a player's.

For a guy like Tommy John, I lop off an entire late section of career, which makes his rate decent but hurts his longevity numbers.

I'm just not having a guy who is a HOM player thru age 36, and then having him play his way out of it.
Same way I don't solely decide how good Concepcion was by waiting to see how many rivals make dumb SS decisions (I rememeber the era, there were many).

Since I happen to have the data in a spreadsheet, I've gone to the trouble of replicating your method and applying this correction. I've made the following adjustments to Win Shares:

1. War credit is given to Rizzuto for 1943-45, and I compare him to the average positional median of 1941-2-6-7.
2. All seasons are straight-line adjusted to 162 games.
3. WS are increased by 5.5% for DH seasons, as per TomH's recommendation.

I removed the players-in-question from the sample, and then took the median (weighted by playing time) batting, fielding, and total WS per 162 games of the other starters at the same league-season-position in question. To smooth out random year-to-year fluctuations, I calculated a 5-year moving average of these medians. I then compared the player-in-question's batting, fielding, and total WS to this moving average multiplied by his playing time in the year in question.

Here are the best 10 seasons for the four players in disucssion:

BWAM: Batting Win Shares above 5-year moving average of positional median
FWAM: Fielding Win Shares above 5-year moving average of positional median
TWAM: Total Win Shares above 5-year moving average of positional median

Larry Doyle



Top 5 average: +11.2 BWAM, -0.4 FWAM, +10.9 TWAM
Top 10 average: +8.2 BWAM, -0.1 FWAM, +8.0 TWAM

Phil Rizzuto



Top 5 average: +1.9 BWAM, +3.2 FWAM, +5.1 TWAM
Top 10 average: -0.4 BWAM, +3.2 FWAM, +2.8 TWAM

Graig Nettles



Top 5 average: +3.9 BWAM, +2.8 FWAM, +6.7 TWAM
Top 10 average: +1.6 BWAM, +2.2 FWAM, +3.8 TWAM

David Concepción



Top 5 average: +8.2 BWAM, +2.8 FWAM, +10.9 TWAM
Top 10 average: +5.2 BWAM, +2.9 FWAM, +8.1 TWAM

I should clarify that I don't think this is a particularly good indicator of Merit at all, but I think this is more or less what sunnyday is looking to measure. Doyle and Concepción come out equal, as both were clearly the best player at their position in their league for many years (although Doyle wasn't close to the best 2B in MLB thanks to Collins and Lajoie while Concepción clearly was the best SS), while Nettles suffers and Rizzuto brings up the rear due to tougher competition. Note that these figures would change substantially if you adjust the standard deviation of Fielding Win Shares to match the empirically demonstrated standard deviation of the impact of defense--top gloves Concepción, Nettles, and Rizzuto would all move up, while Doyle would get clobbered.

I would just say that whatever fluctuations there are among "competition" with this method--which keys off the median player at the position (generally, through history, the 4th + 5th best player/2)--would apply more or less equally to using the 3 worst regulars (FAT). This method uses 2/8 whereas DanR's uses 3/16.

I understand conceptually the "elegance" of the FAT measure--i.e. it improves upon the so-called replacement level by making it more empirical and less theoretical. But as discussed elsewhere it assumes that players as good or bad as those 3 are in fact FA which due to all of the many vagaries of player acquisition/development/transactions, etc. etc., may or may not be practically accurate.

Using the median conceptualizes the idea that in order to win a pennant--i.e. to give his team an edge in a pennant race--a player has to be better than his positional opposite on other pennant contending teams.

Dan has more or less corroborated my little studies, much more elegantly to be sure as I don't have a database of this info except on paper which of course is not what you would call a database. EXCEPT in the case of Concepcion. The only real methodological difference was that Dan used a 5 year average for the median, which I doubt would be the cause of an almost 2:1 differential from my numbers (which showed Concepcion to be about 5 WS better than the median).

As to whether this is a good measure of Merit or not, well, it's discussed above. Conceptually, is it more helpful to compare a "candidate" caliber of player to "replacemtn/FAT" players or to the median? I consider MVP and all-star status very much in my system, which sort of implies that you're not just X wins or runs better than replacement but that you are also in a strictly ranking sense also better than the other all-star "candidate" caliber of players. I like to measure that and I like to know that. Ergo, my total lack of support for the Bob Johnson's of the world.

All the usual caveats re. WS apply of course.

sunnyday, I use 3/8 of regulars, not 3/16.

Remember that in my system the worst-regulars average is not equal to the FAT level. Rather, I compare the worst-regulars average for the 1985-2005 period to Nate Silver's empirically derived FAT levels for the same years at each position, and hold that gap constant over time. So we know it's right for 1985-2005. The unknown is whether the gap between the worst-regulars average and the actual FAT level has changed over the years or not.

Oh, and regarding Concepción, I think it must be because you are penalizing him for all his crappy years outside of his monster prime. Here's the calculation for Concepción. Note that due to endpointing issues I use the average from year 3 for years 1 and 2, and for year 8 for years 9 and 10. 1981 is straight-line adjusted to 162 games.

Glossary

MBWS: 5-year moving average of median Batting Win Shares per 162 games of starters at the position
MFWS: 5-year moving average of median Fielding Win Shares per 162 games of starters at the position
MTTWS: 5-year moving average of median total Win Shares per 162 games of starters at the position
CBWS: Concepción's Batting Win Shares
CFWS: Concepción's Fielding Win Shares
CTWS: Concepción's Total Win Shares
SFrc: Percentage of the season played by Concepción
BWAM: Concepción's Batting Win Shares above median (CBWS - SFrc*MBWS)
FWAM: Concepción's Fielding Win Shares above median (CFWS - SFrc*MFWS)
TWAM: Concepción's Total Win Shares above median (CTWS - SFrc*MTTWS)

the empirically demonstrated
standard deviation of the impact of defense

Do you mean the
impact of the standard deviation of defense?

No, I actually meant the standard deviation of fielding runs/wins above average at each position. I was just trying to phrase it a different way, apparently without much success. My point is that the impact of fielding quality on wins is understated by Win Shares because the standard deviation of FWS is too low, particularly at right-spectrum positions like 1B and LF/RF.