The Under-Achievers
There seems to be a pattern, in the teams that I have looked at so far, that, while large differences between teams on paper are obviously predictive of performance differences, small year-to-year variations in the expectations for a team are not at all predictive, or may even be inversely predictive. Most teams, once the team is essentially assembled, seem to do best when they are not strongest on paper.
The Yankees were a few points stronger in 1931 than in 1927, ’28 or ’32, but they performed the best in ’27 in ’32.
They were stronger on paper in 1926 than 1927, but performed far better in 1927.
The 1970s Reds were strongest on paper in 1974, when the Dodgers won the division. They were stronger on paper in 1977 than in 1976, and in 1978 than in 1979, although they won in ’76 and ’79 and didn’t in ’77 or ’78.
The Yankees won their division every year from 1976 to 1980 except 1979, but they were strongest on paper in 1979.
The Kansas City Royals were exactly the same. They won their division every year from 1976 to 1980 except 1979—but were strongest on paper in 1979.
The Giants won their only championship of the sixties in 1962, but they were stronger on paper in ’63 and ’64.
The St. Louis Cardinals were stronger on paper in 1947, when they finished second with 89 wins, than they were in 1946, when they won the World Championship.
The Cardinals were stronger on paper in 1965, when they finished under .500, than they were in 1964, when they won the World Championship.
The Dodgers of the 1950s were stronger on paper in 1951 and 1954, when they finished second to the Giants, than they were in ’52, ’53, ’55 or ’56, when they won the pennant.
The Philadelphia Phillies were stronger on paper in 1979, when they finished 84-78, than they were in 1980, when they won the World Series.
The Red Sox were stronger on paper in 1917, when they finished second, than they were in 1915 or 1916, when they won the World Series.
The Reds were stronger on paper in 1960, when they finished sixth, than they were in 1961, won they won the pennant.
The Tigers were stronger on paper in 1963, when they finished 79-83, than they were either in 1961, when they won 101 games, or 1968, when they won 103 games and the World Series.
The Minnesota Twins were stronger on paper in 1971, when they finished 74-86, than they were in 1970, when they won 98 games and the American League West.
Maybe it’s nothing, and I’m not picking on A-Rod; A-Rod is a great player and I usually defend him. But I remember something a Yankee writer wrote after the 2004 playoffs, when A-Rod failed to get a runner home from third with less than two out in one of the games the Yankees could have won. “Say what you want,” said the writer, “Scott Brosius would have found some way to get that runner home.”
I’m not ripping on A-Rod. What I’m saying is that it could be that teams always want more stars and more stars and more stars, but that at some point what they need may be more glue and less sparkle. It’s possible. It’s also possible this is just selective observation, and there’s nothing there.
Let’s get on to the major under-achievers. These are the greatest under-achieving teams that I have found so far:
1. 1925 New York Yankees. A famous team, of course; Ruth missed part of the season with abdominal distress associated with stuffing himself with hot dogs, and the Yankees finished 69-85. The 1925 Yankees are weaker on paper than any other Yankee team between 1922 and 1932, because of the aging of some players, but they weren’t significantly weaker, while their performance was quite significantly worse. The team under-achieved by about 30 games.
2. 1935 Boston Braves. The 1935 Braves were certainly not a good team, but the talent on the team is nowhere near thin enough to explain their 38-115 record. The Braves had finished sort of in contention in 1933 (81-73) and 1934 (78-73), despite a complete lack of power at the corner outfield positions. The Yankees wanted to shed Babe Ruth, and the Braves thought that was the answer to their dreams—a power-hitting corner outfielder and a mega-star for the box office.
It didn’t work, at some historic level. Ruth hit .181 and the team, which on paper should have finished reasonably near .500, lost 115 games.
3. 1980 San Diego Padres. The history of the San Diego Padres from their inception to their World Series appearance in 1984 is absolutely amazing, but I’ll pass it up for now and write about it later.
4. 1962 New York Mets. The ’62 Mets, of course, are famous as one of the worst teams ever.
Not denying that they’re one of the worst teams ever, but what I had never exactly understood before now is that, on paper, they’re really not that bad. I’d put it in two ways:
a) On paper the ’62 Mets are better than most of the other expansion teams.
b) The ’62 Mets finished 69 games worse than the 1961 Yankees. Obviously the ’61 Yankees were a better team than the ’62 Mets, but my point is that the difference is nowhere near as large as has been generally represented by people who write about the teams.
The ’62 Mets, in my opinion, were a better team on paper than three of the other four 1960s expansion teams that I have checked out, the 1961 Angels, 1961 Senators, and 1969 Royals. The only expansion team I have seen that was stronger on paper than the Mets was the 1962 Astros.
At this point it is easy to interject that the ’62 Mets made a mistake. They gambled on veteran stars like Richie Ashburn and Frank Thomas, rather than gambling on young players with more upside. This made the team better on paper, because of these players’ past credentials, but made the team worse in fact.
It is easy to say that, but I don’t think it’s true. Frank Thomas in 1962 hit .266 with 34 homers and 94 RBI, and he wasn’t really that old. Richie Ashburn had a .424 on base percentage, and honestly, he was all that old, either. I don’t think those are the guys who under-achieved.
Let’s compare the ’62 Mets, on paper, with the 1961 Angels. The Mets went 40-120; the Angels were 70-91:
Catcher Earl Averill 6 Chris Cannizzaro 10 Advantage Mets
First Base Lee Thomas 10 Marv Throneberry 5 Advantage Angels
Second Base K Aspromonte 5 Charlie Neal 10 Advantage Mets
Third Base G Thomas 6 Felix Mantilla 10 Mets
Shortstop Joe Koppe 4 Elio Chacon 2 Angels
Left Field Leon Wagner 14 Frank Thomas 16 Mets
Center Field Ken Hunt 3 Richie Ashburn 25 Mets
Right Field Albie Pearson 10 Jim Hickman 14 Mets
#1 Starter Ken McBride 5 Bob Miller 12 Mets
#2 Starter Ted Bowsfield 5 Roger Craig 10 Mets
#3 Starter Eli Grba 4 Al Jackson 9 Mets
#4 Starter Ron Moeller 1 Jay Hook 5 Mets
Closer Tom Morgan 14 Craig Anderson 2 Angels
The total is 130-87, Mets. On paper the ’62 Mets had one star (Richie Ashburn) and were solid at seven other positions, whereas the Angels had no stars and were solid at only four positions. The performance difference really is pitching. The Angels starters in 1961 basically had no careers, other than 1961—but they pitched pretty well. Remember, the ’61 Angels were pitching in one of the most extreme hitters’ parks of all time. Their young, inexperienced pitchers should have gotten killed—but they didn’t. They held their own. Ken McBride went 12-15, 3.65 ERA. Eli Grba was 11-13, 4.25 ERA. Ted Bowsfield was 11-8, 3.73.
The starters on the 1962 Mets had much better careers than McBride, Bowsfield and Grba—but in 1962, they just didn’t perform. Roger Craig, Bob Miller and Al Jackson all had longer careers with some very good moments, but in 1962 they went a combined 19-56, with ERAs of 4.40, 4.51 and 4.89. Their park, as well, was working against them, but what it comes back to, I think, is the two things we have discussed before; managing, and infield defense. Casey Stengel, by 1962, really didn’t belong in the dugout—and the Mets’ infield defense was extremely poor. The team never gelled, never developed any confidence in their ability to win, and hugely under-performed. Stengel’s defensive theory relied heavily on ground-ball pitchers and infielders who could turn the double play. It just didn’t work with Charlie Neal and Elio Chacon manning the bag.
The ’62 Mets and ’61 Yankees were reported on by the same gang of sportswriters, who almost uniformly portrayed the ’61 Yankees as one of the great teams of all time, and the ’62 Mets as probably the worst team of all time. But I would put it this way: that of the 69 wins that separate the two teams, only about 23 are actually explained by the difference in the quality of the rosters. There are three other causes of the difference:
1. The National League was a far stronger league,
2. The ’61 Yankees dynamically over-performed, and
3. The ’62 Mets dynamically under-performed.
The under-performance of the Mets is the largest of those three factors. We have them under-performing by about 27 games, although some of that is “natural underperformance” due to the quality of the league’s other teams, and is thus a part of (1) above, not (3).
5. 1971 Cincinnati Reds. Like the 1925 Yankees, the powerhouse Reds just took a year off and didn’t perform.
6. 1971 Minnesota Twins. The Twins had won their division, but bombed in the playoffs, the previous two years. They had two Hall of Famers in the infield, and two if not three starting pitchers who probably should be in the Hall of Fame (Blyleven, Kaat, and Jim Perry. Nobody advocates for Jim Perry, but there are Hall of Fame pitchers who didn’t do any more.) They had solid players in Leo Cardenas, the spider-thin shortstop who hit .264 that year with 18 homers and 75 RBI, and Tony Oliva, who won the batting title at .337, but the offense never really ignited. Rich Reese, who had hit .322 two years earlier, hit just .219, and Brant Alyea, who had pounded the ball in 1970, hit .177.
This is one of the seasons keeping Bert Blyleven out of the Hall of Fame, by the way. Blyleven, only 20 years old, pitched 278 innings with a 2.81 ERA, while the aging Jim Perry pitched 270 innings with a 4.23 ERA—but was credited with more wins. Blyleven finished 16-15, Perry 17-17. The Twins, who scored 4.33 runs per game for Perry and 4.26 for Jim Kaat, scored only 3.42 per game for Blyleven. With 4.3 runs a game to work with, Blyleven would have won 20.
7-8. 1937-1939 St. Louis Browns. The Browns went 46-108 in 1937, 43-111 in 1939. It’s a bad team, a legitimate last-place contender, but the roster is not bad enough to explain 110 losses a year. Their fan base was gone after fifteen years of losing, and I think the team just was going through the motions.
9. 1979 San Diego Padres. See (3) above.
10-11. 1936-1930 Philadelphia Phillies. See 7-8 above. A bad team that played at an unnecessarily bad level. These were two-team cities, St. Louis and Philadelphia, in which the other team had pulled ahead. The Browns competed with the Cardinals and the Phillies with the A’s, and the Cardinals and the A’s were strong organizations that met in the World Series in ’30 and ’31. When the other team in town got good, the fans deserted the Phillies and Browns, and those teams played to near-empty parks.
The list of under-achieving teams is dominated by two types of teams: very bad teams that just got disgusted and quit, like the ’35 Braves and the Browns and Phillies of the 1930s, and strong teams that just had a year when things didn’t click, like the 1925 Yankees and 1971 Reds. Others of those include the 1979 Phillies (underachieved by about 17 games), the 1964 Dodgers (16 games), the 1965 Cardinals, the 1958 Dodgers, and the 1979 Yankees. All of those teams were one year away from a World Championship, one direction or the other, but they all struggled.
Early in the article I discussed the 1963 Detroit Tigers, who went 79-83 despite a roster including Al Kaline, Rocky Colavito, Norm Cash, Bill Freehan, Jim Bunning and Mickey Lolich. That team underperformed, yes, but they underperformed by “only” about 11 games; they probably should have gone about 90-72.
That’s a more significant under-performance than several of these teams, because it’s a team that should have been in contention but wasn’t. Measured in raw distance, 11 games is not a big underachievement. But there are also lots of teams that I just had never focused on that actually under-achieved worse than the ’63 Tigers. The ’63 Tigers rank only 42nd on my list of under-performing rosters.
The San Francisco Giants probably should have won almost 100 games a year from 1963 to 1968. They just never did (except in ’62).
Another team that was as good on paper as the 1963 Tigers, but which performed even worse, was the 1972 St. Louis Cardinals. This is the Cardinals on paper:
Catcher Ted Simmons 22 years old 24 (star)
First Base Matty Alou former batting champ 15 (solid)
Second Ted Sizemore ; 15 (solid)
Third Base Joe Torre MVP in ’71 25 (star)
Shortstop Dal Maxvill Glove Wizard 13 (solid)
Left Field Lou Brock Broke SB record in ‘74 33 (superstar)
Center Jose Cruz Just starting long career 24 (star)
Right Luis Melendez   5 (weak)
Starter Bob Gibson Aging, but still good 23 (star)
Starter Rick Wise 18 (solid plus)
Starter Reggie Cleveland   11 (solid minus)
Starter Al Santorini   3 (weak)
Reliever Diego Segui     18 (solid plus)
On paper, that’s a 90-win team—five stars, six solid positions, two weak spots.
Jose Cruz had a terrible year, bombed out of St. Louis, so he is the #1 under-achiever there. Maxvill was getting old for a glove guy, Torre had a nothing year, and, on the “mythic re-interpretation level”, the Cardinals the previous winter had a contract battle with Steve Carlton, a 20-game winner in 1971. Annoyed with Carlton because he wouldn’t sign, the Cardinals traded Carlton to Philadelphia, where he had a historic season. I think, on a certain level, that lingered over their clubhouse for years.
Gussie Busch redeemed himself later on, when he worked hand-in-hand with Whitey Herzog to put the Cardinals back on the map—but he was having a bad decade. The world was changing. The player’s union was growing in strength, and strong-arm tactics by management were on the way out. Busch, 71 years old in 1970, tried to continue to push his players around, the way it had been done for generations. It didn’t work anymore, and it took ten years of hard lessons for Busch to figure that out. The ’72 team was just one of those hard lessons.
Responding to Reader Posts
I spent an hour last night trying to respond to reader posts, but the g—d--- computer wouldn’t post them, so I’ll include them here. These are responses to comments posted by readers about the first three parts of the article:
1) The use of the Hall of Fame doesn’t create circular logic, because I’m not using the presence of Hall of Famers to suggest that players should be inducted into the Hall of Fame. I using the presence of Hall of Famers to suggest that it was a talented team.
You are objecting, I think, that the use of the Hall of Fame introduces bias into the method. The answer to that is, “Yes, certainly it does introduce some bias into the method. It introduces into the system both bias and real information, in a ratio about 15 parts real information, 1 part selection bias. It is not a perfect system, but the question is, is the system better with this information and bias, or better without it?
It is my judgment that we’re far better with it. Tony Lazzeri played 1,740 games in his career. Without the Hall of Fame adjustment he would rank about even with the other second basemen who played 1700 to 1799 games in their careers—Manny Trillo, Steve Sax, Bret Boone, Joe Quinn, Larry Doyle, Dick McAuliffe, Jorge Orta, Juan Samuel and Tony Cuccinello. You may say, not inaccurately, that Lazzeri is in the Hall of Fame in part because he played for this famous team, but the real question is, are we better off listing Lazzeri as even with those players, or ahead of them?
Lazzeri was a better player than any of those other second basemen who played a similar number of games. Thus, in my view, we’re much better off using the information that Lazzeri is in the Hall of Fame than not using it. We get more accurate values that way.
Suppose that you compare Joe Sewell to the other shortstops playing 1900-1999 games in their careers. Sewell was a better player than any of them.
Suppose that you compare Earle Combs to the other center fielders playing 1400-1499 games in their careers. Combs was probably a better player than any of them, certainly a better player than almost all of them.
The inclusion of the Hall of Fame information, even in those cases where there certainly was a selection bias, gives us more useful information than it does bias. In the cases of Babe Ruth, Bill Dickey and Lou Gehrig, it gives us information without bias. It is not a perfect system, but on balance, it is my opinion that we’re far better off with the information than without it.
2) On the issue of the 1979 Yankees, raised by Martin, I confess that I am missing your general point. I don’t understand what you’re on about. Dealing with the specific points. . .
3) “It seems glib to say that the 1979 Yanks were the best of that bunch just because the personnel fit in better on the age timeline (that is most of the reason, right?)” Not; actually the age adjustments are mainly working against the 1979 team, not in favor of them. The 1979 team is stronger on paper essentially because of the additions of Tommy John and Luis Tiant to the starting rotation, replacing Ed Figueroa and Dick Tidrow. The system assumes that Tommy John should be better than Ed Figueroa (which he was) and that an aging Luis Tiant should be marginally better than a prime-age Dick Tidrow (which I think he was.)
The age adjustment system creates essentially a flat-line peak for each player. When you mix a group of flat-line peaks together, you get an even longer and flatter peak. The system would never cause a team to “peak” to a meaningful extent in a particular year. It DOES cause a team to fade when a meaningful majority of the players were aging.
4) ”They added Bobby Murcer to that team; does that actually mean anything?” No, and our system doesn’t think it does. Mickey Rivers, a regular in ’78, scored at 15; Murcer replaced him in ’79, and scored at 18. That really has nothing to do with the team scoring better in ’79 than ’78.
5) “Since no team has ever won six titles in a row, it seems silly to say that the system proves that even world-beating teams sometimes don’t win.” The system doesn’t prove anything, and I’m pretty sure that I never said that it did. The system suggests conclusions, suggests ways of understanding and comparing teams; it is up to you to buy into those or not.
Perhaps this conclusion was something you already understood, but it looks new to me. I wouldn’t have known, before doing this study, whether the Reds didn’t win in ’71 and ’74 because they under-performed, or because the team was weaker in those seasons. Now I do.
6) Some reflection there about holistic or group consideration of teams. . .”you get away from the problem of stating that the 1973 Reds objectively should have beaten the crap out of the 1998 Yankees.” Well, A, I never said that, or anything remotely close to that. What I said about the 1998 Yankees was that our system was of little use in evaluating that team at this time. B, I don’t understand how you’re getting away from it, or why you want to get away from it.
7) My apologies if I’m misunderstanding you, but you seem a little wobbly about the history of the Reds. The Reds won 102 games in 1970, 95 in ’72, 99 in ’73. 1973 was the geographic center of their period of dominance; it was NOT two or three years before it started—thus, it would seem to me that the conclusion we should draw from this would be that even the strongest teams don’t always win, rather than that dynasties take time to cook. The World Championships won by the Reds were at the end of their strong period, but is that meaningfully different from the 1929-31 A’s, whose World Championships were at the start of their strong period? I don’t see that it is.
8) On the issue of careers getting longer after free agency, you want to assume that this creates a BIAS in the data. It doesn’t create a bias in the data; it creates additional valid information for the system.
Statistical analysis tends to perpetuate the illusion that all eras are the same, that teams from 1956 MUST be as good as teams from 1976. That’s not necessarily true. A system that recognizes the difference is not necessarily biased. It could be insightful. The “pounding flat of history” is one of the things that I am trying to get away from—not something that I need to import into the method.
9) On the issue of Aparicio vs. Mantle, obviously our system’s relative comparison of those two players is not accurate. I have directly acknowledged numerous times that such things will happen in the method. This system—like the Young Talent Inventory—makes entry-level approximations in order to position us to address higher-level questions. There will always be people who trip over the doorstep and can’t get into the room.
10) But we scored Mantle in ’61 at 36 points. The maximum is 40. The problem with the ’61 Yankees isn’t Mantle; it’s the fact that every pitcher on the team except Whitey Ford had a losing record in the rest of his career, mostly in short careers. OK, Jim Coates.
11) On the other Mantle/Aparicio comparison, Win Shares/Loss Shares. ..that’s an after-the-fact analysis, assuming the outcomes as known. We’re talking about how good the players might reasonably have been expected to be, not about the actual outcomes in a particular season. Many players are dramatically better or dramatically worse than they might reasonably have been expected to be—thus, comparisons such as the one you make are not tremendously unusual.
12) Rowen--Issue of the Negro Leaguers was addressed in Part I.
13) John Rickert—Dennis Eckersley is treated as a closer in years when he was a closer, and as a starter in years when he is a starter. His “base value” as a starter would be 36, although his maximum value as a starter would be 36; his “base value” as a closer is 49. The maximum for any player is 40, but Eckersley’s base is so high that, even off-peak, he still maxes out. Which is OK, since he was a dominant closer at 37.
14) Responding to SRoney, Ernie Banks ranks as far better than Mark Belanger because Ernie Banks was a vastly better player than Mark Belanger, and it would be a failure of the system not to recognize this. But Belanger still ranks as a very, very good player. Belanger ranks ahead of Maury Wills or Jim Fregosi. I don’t see that he has a basis for complaint.
I appreciate all of your comments, and I very much appreciate your support.
Bill