In 2002, Mike Williams had a career year saving 46 games, almost doubling his previous high (24) and finishing third in the NL behind Cy Young contenders John Smoltz and Eric Gagne (number three and four in the CY vote). He also recorded a 2.95 ERA and 1.22 WHIP. At the All-Star break, he had 25 saves and a 1. 84 ERA, but he was left off the All-Star team.

In 2003 Williams went into the All-Star break with 25 saves again, but he had blown five save opportunities, had a 6.44 ERA, had given up 41 hits and 22 walks in just 36.1 innings pitched for a 1.73 WHIP, and had struck out three fewer men than he'd walked (22 to 19).

And yet he represented the lowly Pirates in the 2003 All-Star game.

ESPN picked him as the worst All-Star of all time. That piqued my interest: Who are the worst All-Stars ever selected? Can it be quantified?

My solution was to use Bill James' Win Shares. I took a look at the All-Stars who had the lowest Win Share totals and cobbled together a team based on that criterion.

The worst of all time will come as a surprise, but it's Lou Gehrig in 1939, his final year. He was named to the team even though he had already played his last game (one of eight for the year) due to illness. And he, of course, was unable to play in the game. He is the only All-Star without a Win Share (his 0.1 total rounded to zero).

There are actually a good number of great players on the team. They just happened to go to an All-Star game long after the had passed their prime. Among them are Mike Schmidt, Ozzie Smith, Johnny Mize, Reggie Jackson, and Willie Mays.

Also, there is a great deal of catchers who were ostensibly named for their defensive skills or just to fill out a roster. Nine of the 21 position players on the list are catchers.

Also, many of the players ended up as starters for less than a full season due to injury or to losing their job.

Anyway, here's the team:

Now the pitchers (by the way, Williams is second):

By the way, if you think that position players who just happened to be injured are being singled out (e.g., Irvin), here's the list of the worst who played at least 100 games:

On the other side of the equation are those players who had great years but missed being selected for the All-Star game. Here are all position players over 30 Win Shares in a year in which they were not selected for the All-Star game:

Currently in the NL, Adam Dunn projects to 34 Win Shares (5th in the NL), J. D. Drew to 32 (9th), Adrian Beltre 15 (13th), Lyle Overbay 32 (21), Phil Nevin 30 (22), and Brian Giles 30 (31). And in the AL, Michael Young projects to 32 (3rd) (All according to baseballgraphs.com.). And yet none of these players was selected for the All-Star Game (Oh, and the rankings reflect BG's system. I know that some have the same WS total and are ranked differently. It may be a rounding thing, I'm not sure).

Here are the pitchers with at least 25 Win Shares who were not selected for the All-Star team:

At least they didn't miss any pitchers this year that would fall in this group.

Now let's have a look at the worst All-Stars based on career Win Shares (Note that young, current players sort of get the shaft in this one). First position players with 35 or fewer Win Shares lifetime:

Now the pitchers, who outnumber the position players:

Finally, here are the best players (by career Win Shares—150 or more) never to make an All-Star Team (i.e., since the All-Star era: 1933 to the present):

Now the pitchers (min. 125 career WS):

You'll notice that there aren't a great deal of Hall of Famers on that list. I'm glad that Abreu finally got the nod after seven years—he's better than anyone on the list.

By the way, if you were wondering why Scotty "Two Time" Cooper and the superannuated hot-corner Cal Ripken didn't make the list, there's good reason for that. Cooper had 12.3 WS in 1993 and 10.1 in 1994. His career total was 40. He was pretty bad: there were just others who were worse. Ripken just missed the list for his 2000 season. Here are his Win Shares totals as third baseman, 1997-2001: 1997 18 WS, 1998 13.4 WS, 1999 11.8 WS, 2000 8.3 WS, and 2001 9.1 WS. Of course, his career totals were more than All-Star-worthy.

For pitchers, some will recall Neal Heaton of all people made the team with the Pirates in 1990. As was pointed out on Baseball Primer, Heaton had one of those first halves that demanded inclusion on the team (10-3). He did finish 13-9 with a 3.45, 4% better than the park-adjusted league average, and 7.8 WS. Yeah, that's more like a decent number two/three starter than an All-Star, but again there were just so many that were worse that he didn't make the list.

My favorite Odd-Star is Atlee Hammaker. He seemed completely overmatched at the 1983 All-Star game giving up seven runs, including a grand slam to Fred Lynn, and then saw his career quickly deteriorate due to injuries and ineffectiveness. I expected him to show up on the career list. He finished with 50 career Win Shares. 1983 was his career year though (10-9, 2.25 ERA to lead the league, and 13.4 WS).

The one person on the shafted list that surprised me was Kirk Gibson, who was a very good, fairly popular player on successful big-city teams (Detroit and LA in the Eighties). He didn't even make it the year that he won the MVP (1988).

It made me wonder how many award winners were left All-Star teams (Cy Young, MVP, ROY). I would expect some rookies to get overlooked but I didn't expect many veteran award-winners to be overlooked. Let's see who else was.

Actually, Gibson is not the only league MVP not to be named to an All-Star team. There were nine in total. There were also 17 Cy Young award winners, 15 Rolaids Relief Pitcher award winners, 74 Rookies of the Year, 28 LCS MVPs, 21 World Series MVPs, and 330 Gold Glove winners who were kept off the All-Star team.

Here's are the award-winning non-All-Stars (excluding the Gold Glove winners who are too voluminous to list here):