« Thoughts on September call-ups | Main | Greatest O's: Introduction »

Greatest O's—for real this time

If you've been following the Orioles at all this season, you've no doubt been bombarded by reminders that this year marks the 50th anniversary of the modern Orioles in Baltimore. The Baltimore Sun has run a year-long series of historical articles on the Orioles of the last 50 years; ad spots have run prominently featuring the Birds and the numbers 50 or 1954-2004; and during every O's broadcast we are exhorted to pick the 50 greatest Orioles by voting on the team's web site.

Considering that I am both a regular Internet user and a close follower of the Birds, it may seem surprising that I did not give much thought to submitting a ballot for the 50 greatest O's until recently, when I came across the voting page during my survey of MLB.com. I guess I don't care as much for "Top N" lists as other people do. So I procrastinated, figuring that I could always refer to the all-time Orioles lists in Rob Neyer's Big Book of Baseball Lineups (one is reproduced here at ESPN.com) to help me fill out my ballot if I ever got around to it.

But what I discovered upon examining the ballot immediately drew my indignation. I was almost scared away by the amount of personal information required to vote. But far more exasperating was the ballot's flawed design. No, it's not what you think: there isn't a confusing, butterfly-style presentation, nor is there potential for hanging chads, and as far as I know, Diebold did not write the code to run this vote. But the inequitable way that the ballot allocates votes for various positions makes it highly likely that many deserving players will be left off the team of the 50 greatest Orioles. I'm not saying that the vote was intentionally rigged to favor certain players or groups of players, but the effect will be essentially the same.

Annoyingly full disclosure

When I finally took a look at the online ballot to elect the Greatest O's of All-Time (sic), I became highly dissatisfied with the way it is constructed. First, even before getting to the actual voting portion, the amount of personal information the voter is required to provide is excessive. Do the Orioles really need to know the voter's full name, mailing address, birth date, and e-mail address?

Certainly, there needs to be some kind of retardant mechanism to prevent overzealous people from stuffing the ballot box—the attempt to hack the 1999 MLB All-Star voting exposed the inherent vulnerability of Internet voting systems—but this is overly intrusive. If the voter were willing to receive commercial e-mails from theorioles.com and MLB.com (the voter must actively opt out of these) or from the voting's official sponsor, a certain heliocentric financial institution, then some personal contact information would be necessary. But otherwise, there is no need to delve too deeply into personal details, and in any case there is no good reason for requiring the voter to supply his or her birth date, which is one step away from handing over one's Social Security number. Most people, I suspect, come to the ballot page for only one reason: to vote for their favorite all-time O's. They don't want to be bothered with marketing hooks or forms asking for invasive personal disclosures. Should I submit a ballot, I will attempt to provide as little actual information about myself as possible, even if it requires forgery.

There are ways to uniquely identify a person on the Internet without asking for a lot of intimate personal data. None of them is completely foolproof, but nothing on the Internet is. One method used in the registration process for some online forums is to prompt the user for his or her e-mail address, then send a message to that address containing a unique password or URL. The user then must use that password or URL to complete the registration process. Each e-mail address is allowed only one registration account—in the case of online voting, each address would correspond to at most one ballot. An industrious voter could use multiple e-mail accounts to vote several times, but most people would find this laborious and time-consuming. Additional security measures can prevent programmed scripts from automatically submitting votes. For example, the site can generate a graphic depicting a slightly obscured string of random characters and require the user to correctly type the characters to continue.

The ballot: doomed by competitive imbalance

The personal information requirement is not even the worst part. The ballot itself is poorly conceived. The ballot page contains spaces for up to 27 possible votes; one may select up to three choices for each of nine "positions." (It is theoretically possible to submit a ballot with zero votes.) The positions are:

  • 1B
  • 2B
  • 3B
  • SS
  • C
  • OF (!)
  • RHP (!)
  • LHP (!)
  • DH

Whoever concocted this ballot did not give enough forethought to the distribution of voting slots among the various positions. The allocation of possible votes almost guarantees that certain positions will be overrepresented (DH) or underrepresented (P, OF) in the final tally. The prefatory text on the ballot indicates that the 50 players with the highest vote totals will be chosen. If this is the case, it is probable that many deserving players will be omitted from the All-Time team because of the inequitable ballot design.

Outfielders get the shortest shrift here. Lumping all outfielders into one category is understandable, as many of them spent time at more than one of the three outfield spots during their careers. But just three possible nominations out of all the outfielders in the Orioles' history? This essentially cuts by two-thirds the chances of outfielders to crack the Orioles' top 50. There are three distinct outfield positions, not one, in the lineup for every game, so it would have made sense to triple the maximum number of outfield votes to nine.

Pitchers suffer almost as much as outfielders on this ballot. The six slots for pitchers comprise just 22 percent of the possible votes. Although assertions that pitching is 60 or 75 or 90 percent of the game are exaggerated—analyst Bill James's Win Shares typically value pitching at close to 35 percent of the team total, with defense around 15 percent and offense around 50 percent—there's no question that pitchers deserve better treatment than what they got here. It's true that at any given point in a game, the pitcher is just one of nine players on his team, but a starting pitcher has by far the greatest influence over a game's outcome of all his team's players. It would have been more fair to have given pitchers at least one-third of the possible voting slots, and even 40 percent would not be too large a proportion. Fifty percent would probably be too high, though. Keeping to multiples of three, an increase in the number of possible votes for pitchers to twelve or fifteen would have been more equitable. Assuming a maximum of nine votes for outfielders and three for all other positions, twelve votes for pitchers would give them 12/39 = 31 percent of the vote. Fifteen votes would result in 15/42 = 36 percent. That sounds about right.

Not only are pitchers underrepresented on the ballot, but the division between right-handed and left-handed pitchers is also not entirely equitable. On the Orioles' all-time roster, right-handed pitchers outnumber lefties by over two to one (236:97 to be exact), so there is great danger of misrepresentation in giving them equal voting allotments. Instead, the fans should have been asked to vote for the best pitchers based on absolute performance, without consideration for handedness. The rankings of lefties and righties could have been sorted out later if desired.

Another possible snag with the pitchers is that there is no special category for relief pitchers, who run the risk of being totally overshadowed by starters in this contest. There is no question that a great starter is worth more than a great reliever, because the starter pitches so many more innings. But a great reliever who gets critical outs in high-leverage situations is undoubtedly valuable, probably at least as valuable as an average starter. Some people think the very best relievers are worth more than that, while others argue that they are overrated. Regardless of where one stands in that debate, it would be a shame not to see any relievers on the top 50 Orioles team. Requiring that at least one reliever be chosen among the votes for pitchers might help correct this potential imbalance. But I suspect that most fans would not need such a heavy-handed stipulation, and would include a reliever or two in their votes anyway—provided that they had more than six votes to use on pitchers.

Designated hitters are given the most favorable treatment on the ballot. For one thing, designated hitters did not even exist until 1973, nearly twenty years after the Orioles moved to Baltimore, so they were missing in action for about 40 percent of those first fifty years. And the ballot oddly lists just seven DH candidates to choose from; several of the best Oriole DH's, such as Ken Singleton and Lee May, are conspicuously missing from the list, presumably because they spent more time at other positions. With just seven players competing for three possible slots, the danger of overrepresentation is enormous at this position. To compensate, reducing the maximum number of DH selections to two or even one would have been appropriate, even though it would disrupt the multiple-of-three pattern at the other positions.

Implementing my revisions would result in the following ballot:

My ballot vote allocation
Position Maximum Votes Percent of Total
1B 3 7.5%
2B 3 7.5%
3B 3 7.5%
SS 3 7.5%
C 3 7.5%
OF 9 22.5%
P 15 37.5%
DH 1 2.5%
Total 40 100.0%

But there's no way the ballot is going to be changed this late in the season, as the voting period is nearly over as I write this. Can one compensate for the inadequacies of the existing ballot by intelligently modifying one's voting behavior? Not entirely, but here's one solution that comes reasonably close. Instead of voting for the full three players at every position, vote for two players at each infield position and catcher, one at DH, and three at outfield and pitcher. The proportions then would be closer to what they should be, although outfielders and pitchers would still be underrepresented.

Suggested vote allocation using existing ballot
Position Votes Used Percent of Total
1B 2 10%
2B 2 10%
3B 2 10%
SS 2 10%
C 2 10%
OF 3 15%
RHP+LHP 6 30%
DH 1 5%
Total 20 100%

Although I'm still unsure about submitting my ballot online, I've decided to publish my list of the greatest Orioles of the last fifty years, position by position, compiled with the aid of some sophisticated, catch-all statistics like Win Shares and Clay Davenport's Wins Above Replacement Player. More on that will be coming up over the next week.


This page contains a single entry from the blog posted on September 9, 2004 3:20 PM.

The previous post in this blog was Thoughts on September call-ups.

The next post in this blog is Greatest O's: Introduction.

Many more can be found on the main index page or by looking through the archives.

Powered by
Movable Type 3.33