Like you all, I have thought a lot about how a hall of fame ought to be structured and what sort of system would ensure a worthy membership. I thought I?d offer some of my suggestions.

The first thing to do is forget any of the rules that govern the current hall. Don?t try and modify a flawed system. In this exercise, let?s start from scratch.

Here are a few of the main issues:

How Many Players

You?re absolutely right in establishing now exactly how many guys we?re going to enshrine each year. I?ve always tried to figure it on a percentage basis.

My Bill James electronic Encyclopedia goes up to 1996. It says that there are 14,555 players up to that time (presumably beginning in 1876). It also says there were 1,006 players active in 1996. So through 1995 there were 13,549 retired players. The HOF website says there are 188 players in the Hall to 2001 (not sure where you got 215; you must be including some blacks and managers). Using these numbers shows 1.388% of retired players as hall of famers.

We aim to eventually elect a similar proportion. The BJE also shows 3,417 players through 1911 with 537 active in 1911, giving us 2,880 retired players through 1910. Applying our percentage gives us 40 players. I?m inclined to agree with the idea that, given the rapid improvements made in the game up to that time, a lower number than that is OK for the early days. How much lower is open to debate.

You might use a structure like this: let?s aim for .8% representation by the 1935 election (44 enshrinees), 1.0% by 1955 (77 players), 1.2% by 1975 (117 players), and 1.4% by 1995 (170 players).

Or just keep it simple. After our original ten (see First Election, below), elect two per year. That would give us 182 hall of famers through 2001. If you wanted a few more, go to three per year some time in the 1980?s or 90?s.

First Election

It?s always seemed to me that when you?re establishing a Hall, the first thing you do is elect an all-time team. I suggest we do that, too.

The first election should be a positional vote. For the 100-player ballot have ten candidates at each position and 20 pitchers. Elect a ten-man team, including two pitchers. Have voters rank #1 to 3 at each position, #1-5 for pitchers.

This helps us in several ways. It gets us immediate representation at every position; it clears away a stack of obvious choices; it ensures a healthy representation for 19th-century players; it gives our hall?s roster a nice membership right away.

Retirement Year

First, an aside: what?s the point of the five-year wait for eligibility? The hall was not founded with this rule. If the point is to enable us to focus the lens of history more accurately, removed from the emotional issues of a player?s career, well, two years would seem to be enough for this.

Anyway, for our purposes this isn?t relevant. We?re starting with players retired in 1910 or earlier. This requires us to define what is a ?retired player?. It?s easy to assume that the definition is ?the year he played his last game?. For 98% of players that?s fine. However, this doesn?t always work so well.

It?s an especially important issue for our first election. Here?s why. Hughie Jennings was long retired in 1910, finishing up in 1903. He became a manager and put himself in one game in 1907, two games in 1909, one in 1912, and one in 1918. Clearly, it?s absurd to keep him off the ballot until 1923. Why restart the clock with every token appearance? It?s illogical and inappropriate. I think a man should appear on the ballot with his contemporaries.

It would be simplest to say we?ll consider a player?s final year to be the last time he plays in at least 10 games or pitches 10 innings. Anything less is just token appearances. However, that may be going too far in the other direction, given that many players make brief appearances in the following year.

I suggest a rule like this: To determine a player?s retirement year, we will ignore seasons with token appearances (less than 10 games or 10 ip) that are more than a year removed from the player?s last season with 10 games or 10 innings. Of course, there are many other ways you can define it, like using PA and BFP.

Take Clark Griffith, for example. His last year with 10 IP or games was 1906. I?m saying to ignore all token appearances in 1908 or after. He played 5 games (8.1 IP) in 1907. Use that as his year of retirement. He made other token appearances in 09-10-12-13-14. We should ignore those appearances. He should be eligible for the first election.

These token appearances were much more common years ago. By adopting this rule we make eligible for our first election Jennings, Griffith, Deacon McGuire, Kid Gleason, Heinie Peitz, and Fielder Jones, among others.

Ten-Year Requirement

The Hall originally adopted this rule to ensure a minimum standard of quality. (At the time, they didn?t have much else.) It still stands as the sole numerical qualifier. We can do better.

We have no use for the ten-year rule. It mainly serves to disqualify players who were afforded only a brief opportunity to show their stuff, especially in the 19th century. Guys like Bill Joyce, Dave Orr and Bill Hutchison deserve a hearing.

A Ballot(?!)

As a service to its electorate, the Hall of Fame provides a ballot of eligible candidates. I suggest we do the same. Sure, it?s noble and fair to proclaim ?everyone is eligible?. But if the voters are going to do the research we want them to do, it behooves us to streamline the task to some degree. I suggest 100 players. There is zero chance we will omit a viable candidate with this size ballot.

If you really want to be sure, since we have until April, let?s get the blog up and running and solicit for nominees for the first election.

Negro Leaguers

People can?t talk about religion or politics. Or race.

To me, the Negro Leagues exist in a different continuum from the majors. We are in a classic apples-and-oranges scenario if we deliberate their worthiness along with players in the white majors. Anecdotal evidence rules the day.

When some statistical alchemist succeeds in transmogrifying oranges to apples, we will have a basis for comparison. Until then, I find it hard to grant much validity to BJ?s rankings of them in the new HBA.

IMO, we should not include Negro league stars without a whole lot more data. We need someone to translate Negro stats into Major League stats. Can?t this be done to some degree? Look at the stats of players who played in both Negro and Majors and adjust? Of course, we need park effects, too.

How Often Should We Vote

I don?t know what schedule you have in mind, but I think two votes per month might work well. The project would be finished in about four years. If we vote every three weeks, we?ll need about five and-a-half years. And so on. Admittedly, I don?t know everything involved with running this sort of thing.

Delaying the start until the spring may prove to be a blessing. We have time to hash out the ?best of all possible? structures. Oh, on the issue of positional quotas: we don?t need it. Given an informed electorate, we should be confident that the best players will be elected. If we (like the HOF) don?t have as many third basemen and catchers, maybe that?s right. In the new HBA, James has fewer players at these positions in his top 100.

Good work Dan. Your interest and thoughts on this are greatly appreciated.

Dan sent this to me in an email just before Christmas. Below is pretty much my response that I sent at that time.

I did include Negro Leaguers in the 215 number. I also added two for this year as a guess.

I very much agree w/not counting token appearances.

I think the 5 year requirement does a few things:

-Allows us to as you say, look back with perspective.

-Makes the election a little more nostalgic. For example, we'll have 5 years to forget Ripken and Gwynn and Raines and McGwire, which will make their election a little bit more special.

-Most importantly, it will align our elections (once we are caught up) w/the BBWAA, so we can always have a relevant comparison, which will be good if this ever picks up in popularity.

I think 4 years might be better, so we'll be a year ahead of the BBWAA. Maybe some of those guys will see our vote and we could "influence" a vote or two. I realize I'm dreaming here . . .

Actually a 5-year requirement with our election being held a few weeks before theirs would serve this purpose even better. Let me know what you think. I think your 10-man team is an interesting option.

Unfortunately, your email response never made it to me.

As many folks don't know, there will be no VC elections this year. (No new negro leaguers will be enshrined.) The new VC votes in January 2003 for the first time.

Do you then agree with my definition for ignoring token appearances?

You said re the five year wait: "Most importantly, it will align our elections (once we are caught up) w/the BBWAA, so we can always have a relevant comparison...." This will only be relevant for the first-time candidates after the mid 50's.

As for the dream of our efforts achieving credibility, this can happen. We must maintain the highest standards of reasonableness and fairness, as well as promoting it creatively and ubiquitously.

To this end, including Negro leaguers would help us gain acceptance, being fair and PC and all. I'm just not sure they can be accurately compaired in the same balloting with the white leaguers. With the regular Majors, we have the data to make our own studies of players; for the Negroes we would all be in thrall to the opinions of the few experts who have attempted a comprehensive study of the issue.

There was a mention of the lack of good new candidates for election #2. Obviously, there is Cy Young. If we use my definition of ignoring token appearances, Fred Clarke also comes on. I also noticed that Bill Dahlen is eligible for the first election under my proposed definition. Otherwise, he's eligible for the second election.

Dan