1977 Ballot

1) Rod Carew - not a surprise as most systems like his performance this year
2) Rick Reuschel - I tend to favor 'rate' over 'bulk'. Dan R adjusts his contribution to account for Chicago's poor defense
3) Frank Tanana - Another strong 'rate' pitcher
4) Joe Morgan - Still the best 2B in baseball
5) Tom Seaver
6) Mike Schmidt - Still the best 3B in baseball
7) George Foster - Originally I had Parker slightly ahead until I was told Foster was playing some CF this year. Foster's bat was better so I gave him the edge
8) Dave Parker - very good defensive numbers in RF
9) Carlton Fisk - usual catcher bonus
10) Reggie Smith
11) John Candelaria - Did Candelaria and Reuschel pitch more effectively because they had Gossage and Sutter to relieve?
12) George Brett
13) Nolan Ryan - bulk can still make the ballot if the rate is high enough

14-20) Gene Tenace, Steve Carlton, Gary Carter, Jim Palmer, Davey Lopes, Bruce Sutter, Ken Singleton

1977 Ballot

1. Rod Carew, 1B, Minnesota Twins: leads AL in OPS+ and RC by large margins, plus defender at 1B
2. George Foster, RF/CF, Cincinnati Reds: Foster had a career year defensively (+17) as well as with the bat (a league leading 167 OPS+)
3. Mike Schmidt, 3B, Philadelphia Phillies
4. Carlton Fisk, C, Boston Red Sox: a great defensive season (+12 fielding runs) and a catcher's bonus bump Fisk into the top four
5. Bruce Sutter, RP, Chicago Cubs: one of the greatest relief seasons of all-time- a 1.34 ERA in over 100 innnings
6. Dave Parker, RF, Pittsburgh Pirates
7. Reggie Smith, RF, Los Angeles Dodgers: 167 OPS+ leads the National League
8. Ted Simmons, C, St. Louis Cardinals
9. Steve Carlton, P, Philadelphia Phillies: the best starting pitcher in the NL
10. Goose Gossage, RP, Pittsburgh Pirates
11. Jim Palmer, SP, Baltimore Orioles: his combination of innings (319) and ERA+ (132) make him the best pitcher in the AL this year
12. Nolan Ryan, SP, California Angels
13. George Brett, 3B, Kansas City Royals: Brett gets the first lucky 13 ballot spot.

------

14. Gary Carter, C, Montreal Expos
15. John Candelaria, P, Pittsburgh Pirates
16. Rick Reuschel, P, Chicago Cubs: they're essentially tied in my system but Candy breaks the tie with better peripherals (whip and k/bb rate)
17. Jim Rice, LF, Boston Red Sox
18. Tom Seaver, P, New York Mets/Cincinnati Reds
19. Joe Morgan, 2B, Cincinnati Reds: It's a steep drop compared to his recent MVP seasons but a 138 OPS+ and 121 runs created at second base are still good enough for the top 20
20. Dennis Leonard, P, Kansas City Royals: I didn't expect this one but Leonard's 50 IP advantage over Tanana made the difference

1. Rod Carew appears to be an easy choice
2. George Foster--OK, now it gets tricky. #5 WS, #3 WAR, #6 and 8 on the above ballots. How come so low?
3. Mike Schmidt
4. Carlton Fisk
5. Dave Parker
6. Reggie Smith
7. Steve Carlton--283 IP vs. 252 for Reuschel, 241 for Tanana....
8. Greg Luzinski--what exactly has the Baby Bull ever done to WAR that WAR hates him so bad?
9. Ken Singleton--while I have him #6 I have to ask how the hell he is #2 on WS?
10. Bruce Sutter

11. Rick Reuschel
12. Jim Palmer--319 IP
13. George Brett

Apologies if this ends up posting twice, but I tried once and it didn't post.
I don't see it now so I presume it has fallen into a black hole.

This is my usual uber-stat blender.
No post season credit.
10% catcher bonus.

1 carew, rod 9844
2 schmidt, mike 8753
3 foster, george 8404
4 fisk, carlton 8059
5 reuschel, rick 8008
6 parker, dave 7690
7 singleton, ken 7558
8 brett, george 7406
9 palmer, jim 7350
10 smith, reggie 7290
11 seaver, tom 7047
12 page, mitchell 6842
13 carlton, steve 6820

14 candelaria, john 6768
15 sutter, bruce 6758
16 simmons, ted 6706
17 tanana, frank 6666
18 leonard, dennis 6628
19 morgan, joe 6627
20 ryan, nolan 6597
21 gossage, goose 6259
22 bostock, lyman 6223
23 carter, gary 6103
24 niekro, phil 6098
25 tenace, gene 6067
26 hendrick, george 5893

not one nominee from the WS champs--has this ever happened before?

You reminded me that I completely forgot to run Guidry through.

See comment #34 in the Discussion Thread for my comments / explanation:

1. Joe Morgan
2. Tom Seaver
3. George Foster
4. Carlton Fisk
5. Mike Schmidt
6. Rod Carew
7. Rick Reuschel
8. Jim Palmer
9. Steve Carlton
10. J.R. Richard
11. Dennis Leonard
12. Bruce Sutter
13. Mitchell Page

My 1977 MMP ballot:

1. Rod Carew - easy number one
2. George Foster - his small positive defensive contribution lands him here
3. Mike Schmidt - another great season
4. Carlton Fisk - by far best catcher in AL
5. Dave Parker - very good offense and defense in RF (warm up for next year)
6. George Brett - don't forget that young george was pretty good defensively
7. Rick Reuschel - great overlooked season due to crappy cubs offense and defense
8. Reggie Smith - slight bump for 3 HR in world series
9. Jim Palmer - 319 IP, tops in Win Value in AL
10. Joe Morgan - his last great season at age 33
11. Ken Singleton - would be higher if he could play defense
12. Frank Tanana - led AL pitchers in ERA, ERA+, shutouts, and WAR
13. John Candelaria - led NL in ERA, ERA+, and Win Value

1977 ballot

WAR framework, with a higher baseline replacement level than "normal", resulting in a reduction in replacement wins (Wins Above Reduced Replacement). All fielding numbers used are an average of TZ and DRA. NL pitcher's batting adjustment were of positive benefit to all NL pitcher's on my list except Phil Niekro.

1. Rick Reuschel 8.6 WARR value higher than would be expected from raw numbers: they achieved in an extreme hitter's park in front of less than average defense.
2. Mike Schmidt 8.25 WARR superior fielding, sets him above other position players.
3.Tom Seaver 7.6 WARR
4. George Foster 7.5 WARR best offensive numbers in NL, good fielder, but a position adjustment for being a corner outfielder lowers rating.
5. Rod Carew 7.2 WARR best offensive season for position players, but gets adjusted down as a firstbaseman
6. Frank Tanana 7.2 WARR
7. John Candelaria 7.2 WARR
8. Phil Niekro 7.05 WARR rather pedestrian raw numbers look much better when the numbers are adjusted for park (extreme hitter's park) and the fact that he played in front of the poorest defense in the majors
9. Dave Parker 6.85 WARR excellent defender
10. Steve Carlton 6.75 WARR most benefit from NL pitcher's hitting adjustment
11. Nolan Ryan 6.65 WARR
12. Jim Palmer 6.55 WARR
13. George Brett 6.45 WARR

The rest of the top 20
Bruce Sutter
Mitchell Page
Carlton Fisk
Goose Gossage
Reggie Smith
Ken Singleton
Bert Blyleven

Here's Brock Hanke's final ballot. I'm amazed I got this done at all, and even more impressed considering the almost total lack of consensus between WAR and Win Shares. WAR's 5,6 and 7 players, Seaver, Niekro and Tanana, rank 29th, 45th and 42nd in Win Shares. On the other hand, Win Shares' #2 guy, Ken Singleton, ranks 24th by WAR, and WS #8, Greg Luzinski, ranks 45th by WAR. Since my memory of baseball seasons is beginning to return, I was able to get some input from memory and contemporary reputation. Given the lack of consensus of the ubersystems, I decided to give these semi-subjective things more weight. Also, I've finally become a convert to giving catchers a bonus, although a hard 10% is still a bit stiff for me.

I looked up the postseason. Reggie Jackson was Mr. October in the World Series, but Mr. December - cold as Hell - in the Division championship. If he had hit against the Royals in October, he'd have been on the list with good postseason credit. He ranked 15th in Win Shares, but only 42nd in WAR, without any postseason credit. No one on any other postseason team did anything truly credit-worthy. Right now, for some reason or set of reasons that comparisons of systems will likely resolve over time, the only thing WS and WAR can agree on is that they can't agree on anyone except for Carew, Schmidt and Foster.

Anyway, here's my list:
1. Rod Carew
2. Mike Schmidt
3. George Foster (career year)
4. Carlton Fisk (with a catcher bonus that only moved him ahead of Parker)
5. Dave Parker
6. George Brett
7. Rick Reuschel
8. Jim Palmer
9. Ken Singleton
10. Mitchell Page
11. Reggie Smith
12. Joe Morgan (Singleton through Joe could be in any order, they're so close to each other)
13. Ted Simmons (needed the catcher bonus)
also
14. John Candelaria
15. George Hendrick

I know we're voting 13 this year, not 15, but I thought that people who are looking at my rankings for some sort of pattern should get a little help, and also should know whom I moved Simmons ahead of.

Regarding WAR and WS disagreeing - it seems to fall into two categories: pitching and defense. WS shortcomings on defense have been discussed extensively. I usually cross check defense between WAR calculators and don't bother with WS. The other is probably related to changing pitching usage in this era. Some teams are still running out a 4 man rotation and trying to get every inning out of the starters. Others have migrated to 5 man rotations with fireman bullpens and we're seeing the emergence of the closer. The calculators tend to agree better when everyone is used the same.

DL - Just on the off chance, is there anywhere that I can find a reasonably comprehensive discussion of Win Shares' defense? I missed the extensive discussion, due to a low level of baseball activity for a few years. I've always preferred Win Shares, since the first day I read the book, because I agree with its top-down approach. But right now, I don't use anyone's defensive rankings except as a starting point, because everything I've seen counts both Range Factor and errors. I don't think you can do that. To me, an error is just another "play not made" when computing up the Range Factor. Counting the play not made also as an error is, to me, double-counting it. I'm aware that some errors don't have the same effect as some plays not made, but mostly they do. An error by an infielder usually has the same effect as a single, maybe a double, which is about what would happen if the infielder just didn't get to the ball at all. In the outfield, it's usually a double, maybe a triple, which is about what it would be if the outfielder just never got there. That's not every time, but it's most of the time, and to me, that means you shouldn't count errors, unless you've extensively researched the actual effects of errors compared to plays not made, which would be a big enough pain in the rear that I'm not going to do it. Uberstat stuff like that is done by teams of people now, not by one-guy mavericks like me. My guess is that this idea in my head has larger effects than the differences between the various systems' computations, but I would like to see what the largest holes are that Win Shares has, since I trust it more than any version of defensive WAR that I understand enough to even try to evaluate.

The pitching difference strikes me as much more serious. For example, as I mentioned this year, the main reason that WS and WAR disagree so seriously on where to rank Joe Morgan is that Win Shares has 6 position players ahead of Joe, but no pitchers. WAR has only 8 position players ahead of Joe, but ten pitchers, putting Joe's ranking into freefall. If you forced me to choose whether I thought there were zero or ten pitchers in 1977 better than Joe Morgan, I would say zero. But in actual fact, when composing my ballot, I just averaged the rankings of Joe and then kicked him up a bit because I didn't believe the WAR pitcher evaluations. And I haven't, really, agreed with WAR's pitcher rankings for about the last three years of this project. Something, in my opinion, has gone horribly wrong in WAR in analyzing this time period.

This, to me, is actually one of the BEST aspects of this MMP project. It forces you to focus on one year at a time, thereby picking up problems that might not show in a larger sample, or that you (or at least I) would never see because the systems are so complex that it's really hard to focus on something as small as a single season. I don't know if anyone who works on WAR systems agrees with me at all, but that's what I think, and I think it's also the biggest contribution the MMP project is making to sabermetrics, assuming that anyone who is involved in computing up uberstats pays any attention at all to what we do here. And they should. Just the pitcher problem here is well worth investigating, if you're on a team doing an uberstat. Or, at least, that's what I think, for what that's worth. - Brock Hanke (who has neither the time nor the computer access to attempt an uberstat, so he has to sit here on the sideline and gripe about isolated points)

Look for the WARP versus WS thread in the archives. I think the major issue is the range of defensive contribution is compressed to fit into a certain size box so very good and very bad defenders end up regressed to the mean.

DL - Thanks! I'll look up the thread. What you said about compression of defensive stats is, IMO, the biggest worry about Win Shares and defense, which is encouraging. Starting top down means that you inherently have to define a space for the team's defense before you start assigning defense to players. To define that space in the first place, you have to define a space for preventing runs, including pitching and defense. If you make a mistake in either place, things can go very wrong. The other side of the coin - and the one that sold me on the top-down approach - is that if you start with individuals, you can end up with numbers that don't add up to anything that can even try to resemble the full team effort, which means that your individual estimates are off. If I were to really try to dig into a WAR system, the first thing I'd look at is whether the sum of the individual pitching and fielding wins bears a serious resemblance to the actual team totals. Back when Win Shares the book first came out, this was a very big problem with WAR systems, including the one I was using, which had been devised by Bill James in the first place. More modern WAR systems may have mechanisms to deal with this. I don't know.

There are two really big differences between Win Shares and the WAR systems that I think I understand at least some of. This is one of them: a top-down system can compress things beyond reality, while a bottom-up system can expand them beyond reality. The other one is how you handle what I call "undiscovered value" - things like the difference between a Pythagorean W/L record and the actual team record. WAR systems, at least as I know them, make the assumption that this undiscovered value is all luck; if they anchor to anything, it will be to the Pythagorean. Win Shares works the other way: undiscovered value is assumed to be all skill, so if the individual components don't add up to the team effort, then they are adjusted until they do. Neither approach is completely correct, of course. Undiscovered value is part luck and part skill that we sabermetricians just haven't found yet. How much of the undiscovered value is skill and how much is luck is one of those things that no one will probably ever figure out, but boy, would it improve sabermetrics by boatloads if someone ever could. I keep hoping that one of the really serious mathematicians in sabermetrics - Walt Davis or somebody - will find some theorem in advanced statistics that allows us to estimate how much of undiscovered value is what. But that hasn't happened yet, and I don't know enough about formal statistics to know if it's even possible. - Brock

BJ

One of the defensive systems that agrees totally with your method of handling errors is Micheal Humphreys' DRA system. He states very clearly that he believes that an error is just a play not made.

Mr. C - THANKS! Can you give me a url for the DRA system? THIS one I really want to check out! - Brock

Official ballot (no postseason bonuses):

1) Rod Carew: Best ML player, but not as far ahead of the pack that 12-year-old me thought back in '77. Really fun following his effort to .400.
2) Bruce Sutter: Best ML pitcher and NL MMP - amazing relief performance. While I can understand not placing him as high as I have him, it mystifies me not to see him on some ballots.
3) Ken Singleton: Best ML right fielder.
4) Mike Schmidt: Best ML third baseman and also NL positional player.
5) Mitchell Page: Best ML left fielder.
6) George Foster: Best NL left fielder - not that far away from Page.
7) Rich Gossage: Best Fu Manchu mustache - seriously, a great season out of the bullpen that was overshadowed by an even greater one.
8) Tom Seaver: Best ML starting pitcher. Part of the worst trade ever as far as this Met fan is concerned. Walked around like a zombie for a day. Damn you, M. Donald Grant!
9) Ted Simmons: Best ML catcher.
10) Carlton Fisk: Best AL catcher - could have easily changed places with Simmons as the best behind the plate.
11) Reggie Smith: Best NL right fielder.
12) Joe Morgan: Best ML second baseman - a great season, but not nearly as dominating as he had been.
13) John Candelaria: Best ML pitcher who wore a vintage 19th-century cap.

I actually discovered DRA through Humphreys' book "Wizardry". But he wrote a series of three articles in 2003 and a followup in 2005 at the Hardball Times. The URL below should access the 2005 article which in turn references the articles from 2003. I'm sure that google will find other references to DRA.

http://www.hardballtimes.com/main/article/defensive-regression-analysis/

Grandma -

I thought about Sutter, too, but there was a catch-22. Because of how he was used, the leverage of his innings was not as high as for a modern closer. He did have more IP than they did, but the leverage wasn't high enough to get him on the ballot. If his leverage had been anywhere near a modern closer's leverage, he'd have been on there. - Brock

No ballots yet from OCF, Yardape, fra paulo. Will accept a final ballot from JJ1986

I'm not going to be able to get round to this without an extension. I had an unexpected work-related problem this week.

Added a little credit (or demerit) to pitchers for the defense behind them, otherwise the same.

1. Rod Carew - Clear #1.
2. Jim Palmer - At least 40 more innings than any of the other top pitchers and 70-80 more than many of them.
3. Mike Schmidt - Similar offensive value to a few other guys, but the best defender.
4. Joe Morgan - Not nearly as good as his last two years, but much more playing time than '76 closes the value gap a bit.
5. Goose Gossage - Sutter was better, but Gossage had 25% more innings.
6. George Foster - I hadn't realized he was ever this good.
7. Tom Seaver - Best pitcher compared to average.
8. Carlton Fisk - 149 starts at catcher, plus very good hitting.
9. John Candeleira - Led the majors in ERA and ERA+
10. Frank Tanana - Led the AL in ERA and ERA+
11. Rick Reuschel - I thought he'd rank higher. Maybe I should bump him up because of the Cubs defense.
12. Bruce Sutter
13. Nolan Ryan

How long do you need?

Not sure. I'm waiting on an e-mail. Could be one day; could be three.

I can extend until Monday if you're available to post the results then.

Yes, OK. I can post results Monday.

Election is extended until Monday 9/12 at 4pm EST. OCF, Yardape and fra paulo are eligible to vote during that time period.

I use wOBA converted to runs as my standard for batting, and I've gone over to Humphrys' DRA for this election, as opposed to my own system. For 'offensive Wins', I calculate the average of runs above average for position and runs above average for league. These runs and DRA are converted to wins and added together. For pitchers, I have calculated a wOBA, converted it to runs and then made a very basic adjustment for fielding.

1) Rod Carew
2) Dave Parker
3) Nolan Ryan - Well, really Ryan should be #1, but I'm not going to do something so controversial at this point. Carew is without doubt the best hitter this season, but Parker is probably the better all-round player.
4) George Brett
5) George Foster - Foster is the better hitter, but Brett plays a more important position. Otherwise I have these both rated equally at 4.9 wins above average.
6) Dennis Leonard
7) Tom Seaver - Another pair rated equally, at 4.4 wins. Leonard was in the less-pitcher-friendly league, so gets the spot ahead of Tom Terrific.
8) Mike Schmidt - Like Dave Parker, he gets a lot of help from fielding. There are better hitters below him on my ballot.
9) Jim Palmer - His third 20-win season in a row. Next to Mickey Lolich, he was my favourite pitcher of the 1970s.
10) Mitchell Page - A superb rookie season, but a height he never reached again.
11) Goose Gossage - My pitching metric says he is better than Bruce Sutter for this season.
12) Garry Templeton - Not so good with the glove, but 200 hits at a weak-hitting position give him the most 'offensive wins' of any NLer.
13) Steve Carlton - Better wOBA than Rick Reuschel.

Just missed: Bert Campaneris and Frank Tanana.

Election is closed

You mentioned Reuschel, what didn't you like about Fisk?

Fisk's omission is entirely down to my adding-in hitting above average for the league where in the past I used solely hitting relative to position.

Fisk had a .357 wOBA, behind Schmidt, Foster, Parker, Templeton, Reggie Smith, Hendrick, Brett, Reggie!, Carew and Page. This is good for a catcher (.282 wOBA in the AL) and under my original system he would be at 2.3 wins. Page would have been at 0.5 wins, as a left-fielder, a position with a .318 wOBA. However, by adding value against league average, Fisk remains relatively static at 2.6 wins, while Page goes up to the same 2.6 wins. Under DRA, Fisk's 0.1 is only #12 of all catchers with at least 1,000 innings;Page is rated 1.1 wins in the field, and #4 among left-fielders.