Wikipedia talk:Arbitration Committee Elections December 2012/Archive 2
 | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
| Archive 1 | Archive 2 |
Off topic section
| herein lies some mudslinging back and forth on a topic that has very little to do with ArbCom elections NE Ent 20:35, 14 December 2012 (UTC) |
|---|
| The following discussion has been closed. Please do not modify it. |
|
Guides: Beeblebrox and the Civility "Questionnaire" Most of us ignored the civility questionnaire. Did any guide discuss it? I had missed that Beeblebrox was a primary author of the civility questionnaire, and of course regret that he wasn't ranked below YOLO Swag as utterly unsuitable. Kiefer.Wolfowitz 19:43, 14 December 2012 (UTC)
|
Unimportant Question
I am curious, and this is by no means meant to rush or sound impatient, as to whether there's a rough timetable for when we might know something. A couple days, a week, a couple weeks? I'm just curious. Again, I am not suggesting a premature release, I'm just asking out of curiosity. Thanks again to the coordinators for their hard work in orchestrating this process. Go Phightins! 03:58, 17 December 2012 (UTC)
- The results should be revealed before December 20, 2012. — ΛΧΣ21 04:01, 17 December 2012 (UTC)
- It took eight days last year. Sven Manguard Wha? 04:02, 17 December 2012 (UTC)
- (edit conflict)x2 Last year's election also ended on the 10th, and results were posted on the 18th. I wouldn't expect them before then, and it could be a bit longer if the scrutineers are having difficulty with things. As I recall from the Signpost this week, there was a record turnout this year, so that means more work for them. Hersfold (t/a/c) 04:03, 17 December 2012 (UTC)
- Thanks to all. I'm in no particular hurry, but as this was the first election in which I'd participated, I was unfamiliar with how long it generally took. The last thing we need is a Florida-esque controversy! Go Phightins! 04:05, 17 December 2012 (UTC)
- (edit conflict) The scrutineers will be the ones who will post and certify the results. Most of the work in checking voter eligibility and other rules is done or is significantly close to being done. I don't want to say anything concrete, but results will probably be released within a few days, if not sooner. --Lord Roem (talk) 04:06, 17 December 2012 (UTC)
- I agree with Lord Roam's statement. It's also worth remembering that because we didn't have a prescreened voter list this year,the Scrutineers had to look extra-hard at voter eligibility and strike a lot of votes related to that, which takes more time. MBisanz talk 15:56, 17 December 2012 (UTC)
So where are the results?
| We have results now! MBisanz talk 20:37, 18 December 2012 (UTC) |
|---|
| The following discussion has been closed. Please do not modify it. |
|
It's been more than a week now; how long does it take to cook the books? Malleus Fatuorum 01:32, 18 December 2012 (UTC)
|
We have results!
Results are in, but they're unformatted and uncertified, so please help us get them formatted and I'll keep trying to find people to get the certifications in. Thanks. MBisanz talk 18:24, 18 December 2012 (UTC)
- Thanks for all of your hard work, MBisanz.
Poor us :(SandyGeorgia (Talk) 18:28, 18 December 2012 (UTC)- Struck my misunderstanding above. What on earth were these "score" numbers, which seemed to say we had only new arb? SandyGeorgia (Talk) 18:45, 18 December 2012 (UTC)
- It appears to be equal to (Support-Oppose)/(Support+No Vote+Oppose). Not sure why that is relevant though. Monty845 18:55, 18 December 2012 (UTC)
- (e/c) They appear to be "(S-O)/
(S+O)(S+NV+O)", i.e. "Net/Total voters".I believe they would give the same order as "S/(S+O)", but obviously the 50% cutoff wouldn't be the same.--Floquenbeam (talk) 18:55, 18 December 2012 (UTC) corrected to account for the fact that I'm an idiot.- whew. I couldn't have lost to a more deserving group. Congratulations to the new arbs. Beeblebrox (talk) 18:58, 18 December 2012 (UTC)
- I agree : ) - jc37 21:32, 18 December 2012 (UTC)
- Just an observation that there is a lot of overlap between the Guide summary and the actual results. If I read correctly, seven of the eight "predicted" winners(is that the right word?) in the guide summary were elected. Moreover, the order is remarkably close, with David Fuchs a mild exception, my guess because he keeps his head down and is less well known to the community at large compared to the voter guide editors. The observed close concurrence is either a positive, meaning that the guide writers are reasonably representative of the community, or a negative, if it means that lazy voters simply followed the guides. It is not immediately obvious to me how to test which hypothesis is better.--SPhilbrick(Talk) 19:00, 18 December 2012 (UTC)
- I think it's fair to say that the guide writers are diverse enough in profile that, on average, they're pretty much bound to be representative of the community in its broad strokes. — Coren (talk) 19:48, 18 December 2012 (UTC)
- It is near-impossible to tell how much the tail is wagging the dog here, but there were some things that we all knew, guides or not, such as NYB being re confirmed and certain other parties finishing at the bottom. It's too bad there is no way for us to know why there are so many "no vote" results, were people genuinely unsure or were they just supporting and not opposing? There may be some merit to the idea of having random exit polling in the future, if only to give us some idea of why we get the results we do. I get the feeling the candidate's answers to the boilerplate questions had little to no bearing on the results, it is hard to imagine the average voter plowing through such a long series of questions and answers. Beeblebrox (talk) 20:00, 18 December 2012 (UTC)
- I think there are probably a lot of different reasons to select "no vote", including indifference, a laudable disinclination to vote in ignorance, or conflicted feelings about a candidate. —Torchiest talkedits 20:09, 18 December 2012 (UTC)
- Well, I know I got feedback on my answers from people who didn't write the questions, so there are at least some who read them to decide. I'm pretty sure the elections here mirror real life: there are some who do due diligence, there are some who go with gut feelings, and there are those who go with whatever the media seems to say (here, guide writers). Much of this is human nature rather than an effect of our system, I think. — Coren (talk) 20:12, 18 December 2012 (UTC)
- (e/c) I had many NV's, because while I leaned S or O on almost all the candidates, I felt my NV's made my much-more-strongly-felt supports and opposes mean something more, and helped ensure, if only incrementally, that those I really supported rose a little higher compared to the average, and those I really opposed sank a little lower compared to the average. In my mind, the choices were: "strongly support", "weakly support combined with weakly oppose", and "strongly oppose". Someone with a background in game theory is welcome to tell me what an idiot I am for voting that way. And FWIW, I never looked at anyone's guide. --Floquenbeam (talk) 20:16, 18 December 2012 (UTC)
- (e/c) (Also, I can't speak for anyone else, but I voted support for those candidates I wished were on the Committee, I opposed those I felt would harm, and left the others neutral as "not my choice but wouldn't be bad". There were five candidates in that last category.) — Coren (talk) 20:17, 18 December 2012 (UTC)
- I personal voted support for the candidates I supported the most strongly, oppose for candidates whom I thought should never be on the committee, and neutral for the rest (10). @Floquenbeam, as for game theory, you need to decide what your objective is to say whether your an idiot for voting the way you did. There is a correct way to vote to get a particular set of candidates elected, just as there is a correct way to avoid a set getting elected, but there are countless other more nuanced objectives you could have had that would game out differently. Monty845 20:29, 18 December 2012 (UTC)
- I voted the same as you, Coren. David1217 What I've done 01:09, 19 December 2012 (UTC)
- It is near-impossible to tell how much the tail is wagging the dog here, but there were some things that we all knew, guides or not, such as NYB being re confirmed and certain other parties finishing at the bottom. It's too bad there is no way for us to know why there are so many "no vote" results, were people genuinely unsure or were they just supporting and not opposing? There may be some merit to the idea of having random exit polling in the future, if only to give us some idea of why we get the results we do. I get the feeling the candidate's answers to the boilerplate questions had little to no bearing on the results, it is hard to imagine the average voter plowing through such a long series of questions and answers. Beeblebrox (talk) 20:00, 18 December 2012 (UTC)
- I think it's fair to say that the guide writers are diverse enough in profile that, on average, they're pretty much bound to be representative of the community in its broad strokes. — Coren (talk) 19:48, 18 December 2012 (UTC)
- whew. I couldn't have lost to a more deserving group. Congratulations to the new arbs. Beeblebrox (talk) 18:58, 18 December 2012 (UTC)
- Those score numbers are some random system-generated variable, probably from the first iteration of SecurePoll. Given my difficulty in finding Scrutineers over the last two days, when I finally found one, I asked them just to copy and paste whatever was on the tally screen onto the election page and leave it up to us to format the results into something usable. MBisanz talk 19:02, 18 December 2012 (UTC)
- Struck my misunderstanding above. What on earth were these "score" numbers, which seemed to say we had only new arb? SandyGeorgia (Talk) 18:45, 18 December 2012 (UTC)
I wanted to thank all the people who actually brought us the results this go-round, since it sounds like it was a bit more difficult than previous years. I'd also like to thank everyone who asked questions and wrote voter guides. That participation is really the best time to see a broad swath of editors organize their thoughts regarding how arbitration is working and what issues are important to them, and it definitely provided some food for thought. As with real-world politics, an engaged electorate benefits everyone. Der Wohltemperierte Fuchs(talk) 21:24, 18 December 2012 (UTC)
- I'd like to congratulate all the winners in the elections, and I hope they will become good arbitrators in their two-year term. As an additional thing, I have left here the table comparing the guides and the final votes, just for general culture :) — ΛΧΣ21 21:33, 18 December 2012 (UTC)
A graph
I prefer graphical figures (right). NewYorkBrad remains our arbitration golden boy. There are interest-spikes in the data at Elen, Jclemens and Beeblebrox. --SmokeyJoe (talk) 22:01, 18 December 2012 (UTC)
- I have no idea how to read that graph. Can you explain the y-axis to me? Malleus Fatuorum 22:11, 18 December 2012 (UTC)
- It's a bad graph. There's no reason to link adjacent (horizontally) data points with line segments. NE Ent 22:17, 18 December 2012 (UTC)
- Yeah, that graph is very hard to read. May I suggest a bar graph? I think Excel can do three categories (sup, neutral, oppose) so you can have three bars for each candidate. Lord Roem (talk) 22:15, 18 December 2012 (UTC)
- Malleus, I think the y-axis is the number of votes in the respective category (e.g., NYB had about 150 no votes, Worm had roughly 260, etc.), but I agree that it would be much better displayed as a bar graph, or for that matter a series of circle graphs. Go Phightins! 22:26, 18 December 2012 (UTC)
Connected dots imply a variable. The variable need not be time. Here, the variable is candidate ranking. As with any data, deviations in trends imply existence of underlying factors. Inflection points (of which I see none) usually have meaning.
"Net" is reasonably smoothly varying, rapidly decreasing over the first five points, decreasing less rapidly after that.
- I don't see a lot of meaning in the net curve.
"support" decreases smoothly, with a decreasing gradient, with spikes at Elen, Beeblebrox Jclemens. "Oppose" increases fairly linearly (with a significant y-intercept), with spikes at Elen, Beeblebrox Jclemens, until the last two candidates which see a rapid rate of increase.
- The support and oppose spikes of suggest a polarising candidate.
"No vote" shows an intial increase followed by a plateau with negative outliers in the plateau region.
- I'd guess that the "No vote" count for Brad probably marks a baseline value for the remaining candidates. where a "no vote" vote count falls below the trend, it may indicate a special interest in this candidate above the others.
Personally, I feel comfortable with smoothly varying results, and less comfortable with noisy or chaotic results. Therefore, I am quite comfortable with the selection being made from the first nine candidates.
Apologies to anyone upset by the style of presentation. --SmokeyJoe (talk) 06:01, 19 December 2012 (UTC)
Another graph
- I can handle the bar graph. Give me some minutes and it'll be up. — ΛΧΣ21 23:55, 18 December 2012 (UTC)
- How about the comparison between guides and community graph with percentages on the vertical axis -- two color-coded dots for each candidate (guide, actual), and the candidates arranged horizontally. Dots would not be connected. NE Ent 00:08, 19 December 2012 (UTC)
- I can handle the bar graph. Give me some minutes and it'll be up. — ΛΧΣ21 23:55, 18 December 2012 (UTC)
- The percentage trace is in the first line graph is completely useless as is; it should be be either be scaled up (with its own legend on the right) or deleted altogether.—Odysseus1479 (talk) 03:01, 19 December 2012 (UTC)
Congratulating those who made it past the post, as far as I can remember, I only supported 3 candidates. One failed, the other two passed. None passed whom I opposed (8), and the rest were neutral. The large number of neutrals was intended to sway the outcome one way or another. I spent several hours reviewing the candidates and deliberating on my votes. The voter guides were interesting but did not really influence my decisions, but those who wrote them put in some hard and good work. Perhaps I'll write one myself next time round. Thanks to all those who organised and managed the election. Kudpung กุดผึ้ง (talk) 22:26, 18 December 2012 (UTC)
- Funnily enough, having initially agreed that the line graph was inappropriate, I now see that it more easily reveals some very interesting things. Tony (talk) 01:08, 19 December 2012 (UTC)
Thanks to MBisanz
- On behalf of scrutineers team, I would like to thank MBisanz for all his helps and efforts to coordinate this election Mardetanha talk 18:33, 18 December 2012 (UTC)
- Just want to add a "me too" here, although I already did that above. SandyGeorgia (Talk) 18:35, 18 December 2012 (UTC)
- Me also, I'd like to extend thanks to everyone who helped these elections run smoothly WormTT(talk) 18:36, 18 December 2012 (UTC)
- Fourthed. The work of MBisanz and the rest of the volunteers is definitely appreciated. 28bytes (talk) 21:57, 18 December 2012 (UTC)
- Fifthed. Casliber (talk · contribs) 01:40, 19 December 2012 (UTC)
- From behind the scenes, MBisanz was very professional and kind in shepherding us through what needs to be done and when, and provided priceless support and advice throughout the whole process. Thanks! Pundit|utter 15:07, 19 December 2012 (UTC)
Now that the results are in..
My condolences to those elected/re-elected, and my congratulations to those who've successfully avoided a major bit of stress in their life over the next two years.
Ok.. in all seriousness, thank you to everyone who ran, and to those who did their best to make sure everything ran smoothly. Well done, one and all. SirFozzie (talk) 19:08, 18 December 2012 (UTC)
Thank you
Thank you to all of the editors who had the courage to put themselves forward as candidates in this election. In particular, I give credit to those who did not succeed in this election but who took the time to articulate their views and to respond to the many questions posed by members of the community.
Thank you also to those who worked in the background, ensuring that RFCs were completed, that the voting software operated, that the votes were properly scrutinized, and the results posted for the community: the coordinators, electoral commission, scrutineers, and the WMF developers who created the SecurePoll instance. It's sometimes difficult for the community to assess the amount of background work that goes into a successful election, and even more difficult to keep one's cool when having to work with many others. Risker (talk) 20:23, 18 December 2012 (UTC)
- I wholeheartedly agree that everyone involved well deserves our appreciation : ) - jc37 22:29, 18 December 2012 (UTC)
Question about the results
Am I correct that the 8 successful candidates ((presuming JW selects them of course) are all arbs, ex-arbs and arbcom clerks, plus WTT who only missed it by a hair last time?
And if so, could we pretty much call this a status quo election? - jc37 22:27, 18 December 2012 (UTC)
- Almost correct. Worm That Turned actually was a clerk for a short time. However, Timotheus Canens as far as I recall was never a clerk. (He's active as an arbitration enforcement administrator, however.)
- I think it's fair to say that no one was selected who was a complete stranger to the arbitration pages and process. Read into that what you will. On the other hand, several of the new arbitrators have not exactly been supporters of everything the Committee's done in the past couple of years. I also think there will be a healthy balance between continuity and turnover within the Committee from year to year, which is generally for the best. Newyorkbrad (talk) 22:36, 18 December 2012 (UTC)
- Having worked with (and seen) lots of folks who ran, I can say there is a diversity of opinion which probably trumps whether the person has been a clerk or arb previously, so I don't think status quo captures it well, but I can see why someone would think that. Casliber (talk · contribs) 01:39, 19 December 2012 (UTC)
- I wholly agree about there being a diversity of opinion. (Which benefits Wikipedia, as arbcom acts as a committee.) Perhaps "status quo" was not the best phrasing, but I think you (plural) at least understood what I was asking. - jc37 01:57, 19 December 2012 (UTC)
- "Status quo" would not seem to fit an election in which half the incumbents on the ballot were not re-elected. I think it could be said that the results as a whole represent "moderate change" but definitely not "radical change.". Neutron (talk) 03:52, 19 December 2012 (UTC)
- Ah. To clarify, I meant in general, not the specific individuals. Just that we're seeing what I saw someone else call "arbcom insiders". But in thinking about it, aren't such people usually who are chosen in such instances? : ) - jc37 04:04, 19 December 2012 (UTC)
- Those incumbents that were not chosen, it was because of recent events that happened close to the elections. And everyone knows that recent events that happen to be very controversial can severely damage your chances to stay at ArbCom. It has nothing to do with status quo, or moderate change, in my opinion. Here, you make your own luck. — ΛΧΣ21 03:59, 19 December 2012 (UTC)
- "Status quo" would not seem to fit an election in which half the incumbents on the ballot were not re-elected. I think it could be said that the results as a whole represent "moderate change" but definitely not "radical change.". Neutron (talk) 03:52, 19 December 2012 (UTC)
- I wholly agree about there being a diversity of opinion. (Which benefits Wikipedia, as arbcom acts as a committee.) Perhaps "status quo" was not the best phrasing, but I think you (plural) at least understood what I was asking. - jc37 01:57, 19 December 2012 (UTC)
- Having worked with (and seen) lots of folks who ran, I can say there is a diversity of opinion which probably trumps whether the person has been a clerk or arb previously, so I don't think status quo captures it well, but I can see why someone would think that. Casliber (talk · contribs) 01:39, 19 December 2012 (UTC)
Question
Can an editor review thier votes to remind them of who they voted for?--Amadscientist (talk) 02:50, 19 December 2012 (UTC)
- No, sorry, that isn't an option with our current technical setup. MBisanz talk 02:57, 19 December 2012 (UTC)
- So in the middle of next year we can say Don't blame me, I didn't vote for them! and no one will be the wiser? NE Ent 03:20, 19 December 2012 (UTC)
- Even if you could see your votes, none of the rest of us could, so, that would be the case either way... ;) Courcelles 05:45, 19 December 2012 (UTC)
- I even tried to go back in my history but it isn't there. Can someone explain that?--Amadscientist (talk) 04:58, 20 December 2012 (UTC)
- Your vote is logged, just not within your editing history or action log. It's logged within the SecurePoll software itself. Your particular vote is logged at Special:SecurePoll/details/5122. While the system won't let you see the details of it, I can assure you that I can see the details and looks fine. MBisanz talk 11:47, 20 December 2012 (UTC)
- What did you just say about the details of his vote? --SmokeyJoe (talk) 12:41, 20 December 2012 (UTC)
- Wait, how come admins can actually see our votes, at will? One would think this information should be protected as much as our IPs. That's pretty disconcerting. Yazan (talk) 14:05, 20 December 2012 (UTC)
- I find that deeply troubling. How is that justified? How many voters were, or indeed are, aware of that when voting? Where's the banner at the top of the voting page to remind people?? Martinevans123 (talk) 14:14, 20 December 2012 (UTC)
- Admins can't see votes. Only three officials can, MBisanz is one of them.--Scott Mac 14:30, 20 December 2012 (UTC)
- Thanks for the clarification; this makes sense now. Yazan (talk) 14:35, 20 December 2012 (UTC)
- Yes, thank you. At least no-one has to lug those practically-empty heavy steel ballot boxes around to the official counting station... (?) Martinevans123 (talk) 14:39, 20 December 2012 (UTC)
- I should also add that they can see that you voted and that your votes were fine, but they cannot see whom you voter for at all. Nobody can do that except possibly by going to look directly in the database, and I'm certain that no Wikimedia system administrator would even consider doing such a thing (let alone desire to do so). — Coren (talk) 16:20, 20 December 2012 (UTC)
- Ah, I see. Well there go three more conspiracy theories I was planning. But it is rather reassuring. Am stil pondering why illegitimate votes are even possible. Martinevans123 (talk) 16:32, 20 December 2012 (UTC)
- I should also add that they can see that you voted and that your votes were fine, but they cannot see whom you voter for at all. Nobody can do that except possibly by going to look directly in the database, and I'm certain that no Wikimedia system administrator would even consider doing such a thing (let alone desire to do so). — Coren (talk) 16:20, 20 December 2012 (UTC)
- Admins can't see votes. Only three officials can, MBisanz is one of them.--Scott Mac 14:30, 20 December 2012 (UTC)
- I find that deeply troubling. How is that justified? How many voters were, or indeed are, aware of that when voting? Where's the banner at the top of the voting page to remind people?? Martinevans123 (talk) 14:14, 20 December 2012 (UTC)
- Wait, how come admins can actually see our votes, at will? One would think this information should be protected as much as our IPs. That's pretty disconcerting. Yazan (talk) 14:05, 20 December 2012 (UTC)
- What did you just say about the details of his vote? --SmokeyJoe (talk) 12:41, 20 December 2012 (UTC)
- Your vote is logged, just not within your editing history or action log. It's logged within the SecurePoll software itself. Your particular vote is logged at Special:SecurePoll/details/5122. While the system won't let you see the details of it, I can assure you that I can see the details and looks fine. MBisanz talk 11:47, 20 December 2012 (UTC)
- I even tried to go back in my history but it isn't there. Can someone explain that?--Amadscientist (talk) 04:58, 20 December 2012 (UTC)
- Even if you could see your votes, none of the rest of us could, so, that would be the case either way... ;) Courcelles 05:45, 19 December 2012 (UTC)
- So in the middle of next year we can say Don't blame me, I didn't vote for them! and no one will be the wiser? NE Ent 03:20, 19 December 2012 (UTC)
- To be clear, the interface at Special:SecurePoll/details/5122 looks like the screenshot on the right. It's all just metadata, the actual vote data is not accessible to anyone apart from sysadmins. And having had to extract raw vote data from the database for other wiki elections before (I run a wiki that elects its leadership by Single Transferable Vote, which it a PITA to calculate) I can assure you that even that is far too much work to be done by any but the most committed sysadmin. Happy‑melon 21:17, 20 December 2012 (UTC)
Eligible votes
Is there a way to know whether one's vote was counted or not? Is there a mechanism to inform editors who's votes were disregarded as ineligible of that, so there would be a subsequent appeal process in case it was done in error? From reading this page, there seems to have been a sufficient amount of such ineligible votes; which, one would assume, leads to a sufficient amount of false positives, no? Yazan (talk) 03:45, 19 December 2012 (UTC)
- Any vote with a line through it or in light gray text on Special:SecurePoll/list/259 was not counted. I know people tried to inform those whose vote was struck, but I can't say it was done with 100% compliance. MBisanz talk 03:49, 19 December 2012 (UTC)
- Ha! Many thanks. Yazan (talk) 03:58, 19 December 2012 (UTC)
- AFAICT the lightly struck-out entries are for votes that were superseded later on; the heavier strike-outs are for ineligible votes.—Odysseus1479 (talk) 04:49, 19 December 2012 (UTC)
- Correct. MBisanz talk 05:08, 19 December 2012 (UTC)
- Ah. So we can see who can't a make a decision. Interesting. NE Ent 00:22, 20 December 2012 (UTC)
- Eh, not really. Events during an election can change people's minds and they may then go back and change their votes. Some people might even put in 'temporary' votes with the full intention of returning and changing/confirming them before the end of the election. What I do think should be flagged up is the slight possibility that people change their votes without realising it, by accidentally clicking the wrong button and revisiting the vote page. The way to check that your vote hasn't been redone is to check the log, but not all voters know that. It might be worth considering having a bot automatically leave a message with editors that redo their votes to notify them of this, so that if any votes are redone by accident, those voters will go "hang on, I didn't change my vote", and hopefully get a chance to correct any accidental resetting of the vote (usually to all 'no votes'). I would also (though I suspect it is not possible) suggest that the number of voters that submit all 'no votes' should be either noted in the results tally, or not allowed (you have to support or oppose at least one candidate). I suspect no-one really would need to vote all 'no vote' (all support and all oppose is another matter). I also agree that the voting system really should be set up properly next time, so there is less of the rather embarrassing spectacle of good-faith voters having their votes struck because the software didn't tell them they were ineligible (I know why it happened this time, but it should really be avoided). Carcharoth (talk) 01:57, 20 December 2012 (UTC)
- The UK recently held a set of national elections that, despite the record low turnout, also had, in some cases, a higher than normal level of "spoiled ballots". While I think any intention to "spoil ballots" here would be rather unlikely, I suppose it is still one possible motivation for those who are disatisfied with the process/ whole structure of ArbCom. But why are "ineligible votes" even possible? Would it not be possible and/or preferable to trap these at source and to provide an immediate indication to the voter of the reason(s) for the ineligibilty? Thanks. Martinevans123 (talk) 12:15, 20 December 2012 (UTC)
- It normally isn't; but this year the delay over the selection of the coordinators means the technical ball was started at the very last minute and the step of compiling a voter registry was skipped to prevent the election from starting late(er). — Coren (talk) 12:57, 20 December 2012 (UTC)
- Thanks. I am genuinely surprised. I had imagined that the mechanics would have been, by now, honed down to a sleek reliable process, quite independent of any timetable of personalities or Electoral register. Martinevans123 (talk) 22:24, 20 December 2012 (UTC)
- It normally isn't; but this year the delay over the selection of the coordinators means the technical ball was started at the very last minute and the step of compiling a voter registry was skipped to prevent the election from starting late(er). — Coren (talk) 12:57, 20 December 2012 (UTC)
- The UK recently held a set of national elections that, despite the record low turnout, also had, in some cases, a higher than normal level of "spoiled ballots". While I think any intention to "spoil ballots" here would be rather unlikely, I suppose it is still one possible motivation for those who are disatisfied with the process/ whole structure of ArbCom. But why are "ineligible votes" even possible? Would it not be possible and/or preferable to trap these at source and to provide an immediate indication to the voter of the reason(s) for the ineligibilty? Thanks. Martinevans123 (talk) 12:15, 20 December 2012 (UTC)
- Eh, not really. Events during an election can change people's minds and they may then go back and change their votes. Some people might even put in 'temporary' votes with the full intention of returning and changing/confirming them before the end of the election. What I do think should be flagged up is the slight possibility that people change their votes without realising it, by accidentally clicking the wrong button and revisiting the vote page. The way to check that your vote hasn't been redone is to check the log, but not all voters know that. It might be worth considering having a bot automatically leave a message with editors that redo their votes to notify them of this, so that if any votes are redone by accident, those voters will go "hang on, I didn't change my vote", and hopefully get a chance to correct any accidental resetting of the vote (usually to all 'no votes'). I would also (though I suspect it is not possible) suggest that the number of voters that submit all 'no votes' should be either noted in the results tally, or not allowed (you have to support or oppose at least one candidate). I suspect no-one really would need to vote all 'no vote' (all support and all oppose is another matter). I also agree that the voting system really should be set up properly next time, so there is less of the rather embarrassing spectacle of good-faith voters having their votes struck because the software didn't tell them they were ineligible (I know why it happened this time, but it should really be avoided). Carcharoth (talk) 01:57, 20 December 2012 (UTC)
- There's probably an obvious answer to this, but why was User:Philippe (WMF)'s vote struck (in bold)? Are WMF members not allowed to vote? Black Kite (talk) 12:58, 20 December 2012 (UTC)
- See Wikipedia_talk:Arbitration_Committee_Elections_December_2012/Coordination#Oops. -- KTC (talk) 13:10, 20 December 2012 (UTC)
- I knew there'd be a logical answer, thanks :) Black Kite (talk) 15:40, 20 December 2012 (UTC)
- See Wikipedia_talk:Arbitration_Committee_Elections_December_2012/Coordination#Oops. -- KTC (talk) 13:10, 20 December 2012 (UTC)
Correlation functions?
Is there any data about correlations like the fraction who voted for both X and Y etc.? Count Iblis (talk) 23:33, 19 December 2012 (UTC)
- No, sorry, the system isn't advanced enough to provide such a function. MBisanz talk 23:37, 19 December 2012 (UTC)
- It's not so much that it's not advanced enough, as that those data would be nibbling away at the edges of what counts as a "secret" ballot. Happy‑melon 23:45, 19 December 2012 (UTC)
- Yup, agree totally with Happy-melon. Tony (talk) 14:02, 20 December 2012 (UTC)
- I see no secrecy issues of correlations - if I knew that (for example) X number of users voted in favor of both YOLO Swag and Count Iblis. In fact, what (theoretically) could be done is to give each voter a random number, and then show how each number voted (i.e voter#1 voted in favor of these candidates, etc), with little issues of privacy violation - but I wouldn't go that far, because some users may have a unique combination which would allow third-party users to guess that it's them. עוד מישהו Od Mishehu 21:36, 20 December 2012 (UTC)
- Yup, agree totally with Happy-melon. Tony (talk) 14:02, 20 December 2012 (UTC)
- It's not so much that it's not advanced enough, as that those data would be nibbling away at the edges of what counts as a "secret" ballot. Happy‑melon 23:45, 19 December 2012 (UTC)
How exactly is the result "certified"?
I just hope, for the sake of the stewards that they dont manually check all the votes. In which case, how do they "certify" the votes? I can only think of them just checking to see if all votes are ok, and a computer/bot adding all votes automatically. TheOriginalSoni (talk) 18:17, 20 December 2012 (UTC)
- The technical method is actually fairly simple; once the scrutineers are satisfied that every vote that needed to be struck is (and that there are no erroneous striking left), they press the "tally" button and out comes the sums. The point is that this final act also makes it impossible to strike further votes, and the sums aren't accessible until that button is pressed – this way the actual votes cannot influence which voters are stricken, and you can deduce who voted for whom by elimination.
So, the certification is on the validity of the votes, and not the counting. :-) — Coren (talk) 18:51, 20 December 2012 (UTC)
- When you say "you can deduce who voted for whom by elimination", what exactly do you mean? Thanks. Martinevans123 (talk) 20:48, 20 December 2012 (UTC)
- 1) Tally, 2) Strike a single vote, 3) Tally again and see that each score is reduced by one point, corresponding to the specific votes of the ballot that was struck; 4) Unstrike the vote if the voter has voted the way you want, otherwise manufacture a reason to leave the strike. To avoid that happening, all the checking and striking is done before the tally is ever generated. Happy‑melon 20:55, 20 December 2012 (UTC)
- So Coren meant "you can't deduce who voted for whom by elimination"? Martinevans123 (talk) 21:04, 20 December 2012 (UTC)
- 1) Tally, 2) Strike a single vote, 3) Tally again and see that each score is reduced by one point, corresponding to the specific votes of the ballot that was struck; 4) Unstrike the vote if the voter has voted the way you want, otherwise manufacture a reason to leave the strike. To avoid that happening, all the checking and striking is done before the tally is ever generated. Happy‑melon 20:55, 20 December 2012 (UTC)
- When you say "you can deduce who voted for whom by elimination", what exactly do you mean? Thanks. Martinevans123 (talk) 20:48, 20 December 2012 (UTC)
- So all scrutineers do a separate striking for every candidate that needs to be struck? Or once some vote is struck, the rest simply verify it? TheOriginalSoni (talk) 19:20, 20 December 2012 (UTC)
- I need to clear up some confusion here. There is only one voter log. The certification of the results is simply the scrutineers agreeing that the numbers reflected in the table are the same ones that show up when you press 'Tally'. Because pressing 'Tally' shows you the results, each scrutineer can independently check to confirm the posted numbers. As per striking, all scrutineers and commissioners worked together to figure out issues of eligibility (and other issues, as described in threads above). In the log, you can see which person struck X vote so we all are on the same page. I hope this answers your question. :) Cheers, Lord Roem (talk) 19:24, 20 December 2012 (UTC)
- So the final certification is no more than checking if a button named "Tally" on Wikipedia works properly or not? O.O TheOriginalSoni (talk) 19:51, 20 December 2012 (UTC)
- The certification represents the agreement by all scrutineers that all the votes were cast by eligible voters, and all the other rules for voting were followed (e.g., no double voting). Then 'Tally' is pressed and we get the results. Lord Roem (talk) 20:25, 20 December 2012 (UTC)
- But by your last reply, doesnt that happen when all the scrutineers are together? Implying that after they are done collectively doing that, their pressing tally button multiple times (one for each scrutineer) is just checking to see if the tally button works? TheOriginalSoni (talk) 20:43, 20 December 2012 (UTC)
- More that the person who copied the results from the tally onto the publicly-viewable page has done so honestly. Since the general public cannot see the tally, in principle there is no way of proving that the results on the page are not entirely manufactured. The certification is a way of having three users independently assert that the results are honest, which is inherently more trustworthy than one. Happy‑melon 20:59, 20 December 2012 (UTC)
- Sounds logical. Thanks TheOriginalSoni (talk) 21:17, 20 December 2012 (UTC)
- Still wonders if 3 scrutineers somehow planned a conspiracy to change the tally marks in the voting sheet TheOriginalSoni (talk) 21:17, 20 December 2012 (UTC)
- I'm curious why you would wonder that. Does something make you suspicious? It's just a strange thing to be saying if you don't have any evidence, since it calls into question the personal integrity of the three scrutineers. First Light (talk) 21:41, 20 December 2012 (UTC)
- It's déjà vu all over again. --Floquenbeam (talk) 21:44, 20 December 2012 (UTC)
- ...by which I mean, TheOriginalSoni, that while most people would recognize that was a (IMHO somewhat unfunny) joke, a sitting Arb has, in a section way up this page, threatened to block anyone who says stuff like that. And a similar comment ended up with someone dragged to ANI and then ArbCom, resulting in retirements and grudges and trillions of slaughtered electrons. I suggest you strike it out; there are more than a few people here with access to the block button, and no sense of humor. --Floquenbeam (talk) 21:49, 20 December 2012 (UTC)
- I guess I didn't get it, because of all the other questions, but I can see how in the right context it could be seen as a joke.... First Light (talk) 22:05, 20 December 2012 (UTC)
- ...by which I mean, TheOriginalSoni, that while most people would recognize that was a (IMHO somewhat unfunny) joke, a sitting Arb has, in a section way up this page, threatened to block anyone who says stuff like that. And a similar comment ended up with someone dragged to ANI and then ArbCom, resulting in retirements and grudges and trillions of slaughtered electrons. I suggest you strike it out; there are more than a few people here with access to the block button, and no sense of humor. --Floquenbeam (talk) 21:49, 20 December 2012 (UTC)
- It's déjà vu all over again. --Floquenbeam (talk) 21:44, 20 December 2012 (UTC)
- I'm curious why you would wonder that. Does something make you suspicious? It's just a strange thing to be saying if you don't have any evidence, since it calls into question the personal integrity of the three scrutineers. First Light (talk) 21:41, 20 December 2012 (UTC)
- More that the person who copied the results from the tally onto the publicly-viewable page has done so honestly. Since the general public cannot see the tally, in principle there is no way of proving that the results on the page are not entirely manufactured. The certification is a way of having three users independently assert that the results are honest, which is inherently more trustworthy than one. Happy‑melon 20:59, 20 December 2012 (UTC)
- But by your last reply, doesnt that happen when all the scrutineers are together? Implying that after they are done collectively doing that, their pressing tally button multiple times (one for each scrutineer) is just checking to see if the tally button works? TheOriginalSoni (talk) 20:43, 20 December 2012 (UTC)
- The certification represents the agreement by all scrutineers that all the votes were cast by eligible voters, and all the other rules for voting were followed (e.g., no double voting). Then 'Tally' is pressed and we get the results. Lord Roem (talk) 20:25, 20 December 2012 (UTC)
- So the final certification is no more than checking if a button named "Tally" on Wikipedia works properly or not? O.O TheOriginalSoni (talk) 19:51, 20 December 2012 (UTC)
- I need to clear up some confusion here. There is only one voter log. The certification of the results is simply the scrutineers agreeing that the numbers reflected in the table are the same ones that show up when you press 'Tally'. Because pressing 'Tally' shows you the results, each scrutineer can independently check to confirm the posted numbers. As per striking, all scrutineers and commissioners worked together to figure out issues of eligibility (and other issues, as described in threads above). In the log, you can see which person struck X vote so we all are on the same page. I hope this answers your question. :) Cheers, Lord Roem (talk) 19:24, 20 December 2012 (UTC)
Scrutineers' checkuser flags removed
Hi. I've just removed (diff) the temporary checkuser flags from scrutineers. Regards.—Teles «Talk to me˱M @ C S˲» 07:40, 21 December 2012 (UTC)
- I had completely forgotten this step. Thanks for remembering. MBisanz talk 11:35, 21 December 2012 (UTC)
Negative (oppose) voting can be used to skew results
Compared with a voter who uses only positive votes, a voter who uses only negative votes has the equivalent of many times the number of votes. Why that is so: Each single negative vote for a candidate has the same effect as a positive vote for all the other candidates, so a negative vote carries more weight than a positive vote. LittleBen (talk) 00:15, 18 December 2012 (UTC)
- I don't get this. What do you mean? MathewTownsend (talk) 00:20, 18 December 2012 (UTC)
- Tactical voting is possible. --SmokeyJoe (talk) 00:25, 18 December 2012 (UTC)
- Each negative vote for a candidate cancels out someone else's positive vote for that candidate, which has the same effect on the results (the ranking or ordering) as a positive vote for all the other candidates. So when you convert negative votes to equivalent positive votes, it's obvious that negative votes carry far more weight. If you know this, you can game the system.
- I hope that negative vote totals for each candidate as well as positive votes totals are published, so that it is easy to work out who would have been elected if it weren't for the unfair advantage that negative voting gives people who are "in the know". LittleBen (talk) 00:29, 18 December 2012 (UTC)
- The totals of support, oppose, and neutral/no votes on each candidate will be published. Lord Roem (talk) 00:37, 18 December 2012 (UTC)
- (ec)A negative vote doesn't really skew in the way you suggest. It has the same effect as a non-vote for all other candidates. Each candidate's support/oppose ratio is individually calculated. Because the ranking of candidates is based strictly on supports and opposes (and non-votes have no effect), and all successful candidates must have more support votes than oppose votes, a single oppose has no more weight than a single support vote and a non-vote carries no weight. Now, it's still possible to vote tactically; some editors oppose all but one candidate in the theory that it will give their most preferred candidate a better shot, and other editors support as many candidates as there are seats open while opposing everyone else. I'm not convinced it is beneficial in the long run, if one feels that there are more "good" candidates than seats open, because if enough editors do that there might not be enough candidates who receive a majority of support votes. On the other hand, I support the concept that editors should be able to actively oppose candidates they believe are unsuited to the Arbitration Committee. This is why I oppose the use of voting processes like the Schulze method, which does not permit "negative" voting, and is also counterintuitive. Risker (talk) 00:45, 18 December 2012 (UTC)
- Tactical voting is possible. --SmokeyJoe (talk) 00:25, 18 December 2012 (UTC)
- (ec) Uh? You might as well say that each single positive vote for a candidate has the same effect as a negative vote for all the other candidates. --Stfg (talk) 00:43, 18 December 2012 (UTC)
Has negative voting been used for past ArbCom elections? Is there a record of who proposed it? LittleBen (talk) 01:10, 18 December 2012 (UTC)
- Support/Oppose voting has been used for years, I'm not up on the history, but I'm not aware of any other voting method that has been used. While support/Oppose voting can be used tactically, its also important to have a way to oppose candidates who you think would be particularly bad arbs, not just support ones you think would be the best of the candidates. While it is possible to game a straight support voting system to the same ends, its not as intuitive. Monty845 01:20, 18 December 2012 (UTC)
- Yes. It has been used since the very beginning of ArbCom elections. Since 2004 and until 2008, it was done a la RFA. Since 2009, the SecurePoll system is used instead. — ΛΧΣ21 01:27, 18 December 2012 (UTC)
- Risker, "a non-vote carries no weight" – in a narrow sense, yes; but it does carry weight in that it reduces the raw proportion of voters who support a candidate. If a thousand people vote, of whom 995 give you a "non-vote", three support you, and two oppose you, you've passed the "more must support than oppose" threshold by a long shot, and with a nice fat support percentage of 60%. You'll be elected with 0.3% of voters' giving you the tick, over a candidate with 599 supports, 401 opposes, and no "non-votes". Usually the S/(S + O) doesn't produce such distortions, but it's quite possible for less extreme "non-vote" effects to make a difference.
In reponse to the discussion above about "strategic" voting: yes, the power of your support votes is definitely maximised by avoiding the non-vote buttons completely and opposing all candidates for whom you don't vote support. In fact, to assert the most powerful support, vote support for one candidate alone, and oppose for all other candidates; then you're disadvantaging all but your single chosen candidate. Tony (talk) 09:57, 18 December 2012 (UTC)
- Very good. You're a pedagogue, Tony! Notice incidentally this comment by Ten Of All Trades about how a full ballot of oppose votes was (in this particular election) the lazy person's best choice, i.e. an improvement on not voting at all. That hadn't occurred to me, but he's perfectly right. Bishonen | talk 11:14, 18 December 2012 (UTC).
- Tony is largely but not completely right. See Nash equilibrium for further reading. --SmokeyJoe (talk) 12:28, 18 December 2012 (UTC)
- Risker, "a non-vote carries no weight" – in a narrow sense, yes; but it does carry weight in that it reduces the raw proportion of voters who support a candidate. If a thousand people vote, of whom 995 give you a "non-vote", three support you, and two oppose you, you've passed the "more must support than oppose" threshold by a long shot, and with a nice fat support percentage of 60%. You'll be elected with 0.3% of voters' giving you the tick, over a candidate with 599 supports, 401 opposes, and no "non-votes". Usually the S/(S + O) doesn't produce such distortions, but it's quite possible for less extreme "non-vote" effects to make a difference.
- Yes. It has been used since the very beginning of ArbCom elections. Since 2004 and until 2008, it was done a la RFA. Since 2009, the SecurePoll system is used instead. — ΛΧΣ21 01:27, 18 December 2012 (UTC)
All voting systems can be manipulated; see Duggan–Schwartz theorem. NE Ent 12:42, 18 December 2012 (UTC)
- Surely the Duggan–Schwartz theorem states clearly that all preferential voting systems can be easily manipulated (but not the single-positive-vote-(or no vote)-per-candidate system used in democracies). LittleBen (talk) 14:20, 18 December 2012 (UTC)
There is one simple way to reduce the use of "negative bullet voting" - that is to require any ballot to have at least one more "support" than "oppose." In the current system, assuming that otherwise there is a farily normal vote distribution (that is, a significant number of candidates, all other things being equal, will have 50 to 60% support) -- if even 10% of the voters voted "oppose all", the number of passing candidates will be greatly diminished (possibly to nearly zero). It is possible to some that this would be a "good thing" but is decidedly not in the best interest of having a diverse group on the Committee. If we wish diversity of opinion on the committee, the use of "oppose votes" is likely a very "bad thing." Collect (talk) 13:35, 18 December 2012 (UTC)
- Diversity is the last thing I want. I want the best: patient, principled, wise, tough, believer in Wikipedia, able to focus on the big picture ... better to have a smaller committee than one with crusaders and zealots and cynics on it. NE Ent 14:00, 18 December 2012 (UTC)
- (ec) It would allow a group with a loony agenda to provide so many candidates that some would be sure to get in. For me, removing incompetence seems more urgent at present than variety of opinion. --Stfg (talk) 14:07, 18 December 2012 (UTC)
- Not really -- for such a "loony group" to elect bad candidates would require that it find a very large number of "loony voters." Places with "proportional representation" may have some "loony electees" but they generally do not end up running the place. The concept of "oppose" voting is not widely used in the outside world (heck - can you show me any serious place using "oppose voting"?). Meanwhile, the current system appears to encourage "groupthink" which has its own negative attributes. Collect (talk) 14:38, 18 December 2012 (UTC)
- I think you just said that ArbCom is not a serious place
I have no objection to support-only voting, but I'm putting forward that if opposers votes are to be allowed, it should be possible to oppose all the candidates you object to, without restriction on your ratio of supports and opposes. OK, I don't think many of this year's candidates are "loony", but I think more than half of them are at best mediocre. I voted slightly more opposes than supports, not for any tactical reasons, but purely on the basis of what I could see of the candidates. (BTW, I also don't see anything wrong with tactical voting.) --Stfg (talk) 15:25, 18 December 2012 (UTC)
- I think you just said that ArbCom is not a serious place
- Not really -- for such a "loony group" to elect bad candidates would require that it find a very large number of "loony voters." Places with "proportional representation" may have some "loony electees" but they generally do not end up running the place. The concept of "oppose" voting is not widely used in the outside world (heck - can you show me any serious place using "oppose voting"?). Meanwhile, the current system appears to encourage "groupthink" which has its own negative attributes. Collect (talk) 14:38, 18 December 2012 (UTC)
- Now that the results are in it is interesting to see the impact of the oppose votes. I note that myself and Elen both actually had more supporters than the 8th place finisher, but we are both a bit controversial so we also got more opposed and therefore did not qualify. That's the middle, the community seemed a lot more sure about who was at the bottom and the top, with NYB getting downright amazing support and the Arbcom reform party consigned to the dumpster. Beeblebrox (talk) 19:34, 18 December 2012 (UTC)
- It's been said before that the support/oppose system favours those who keep their head down and say nowt. But I suspect every voting system favours one strategy or another. Elen of the Roads (talk) 20:11, 18 December 2012 (UTC)
- This S/(S+O) system has the disadvantage of rewarding candidates who avoid opposes, just as much as those who attract positive reactions in the community. And many or most voters don't even understand it, so we don't even get the advantage of simplicity. The obvious way to go is to make it preferential, which would involve the choice of one of a number of standard, well-oiled systems used both out there and in the WM movement. There's nothing exotic about it, and for voters it just means numbering in cardinal order. Tony (talk) 13:02, 22 December 2012 (UTC)
- Something in that. For an extreme example, look at the pretty impressive support for Giano in 2007. He got the second highest support vote of all (more than FT2, more than Raul, etc, etc: more than everybody except the uncatchable Newyorkbrad). Yet Giano didn't make the cut, because he also got a lot of opposes. To the people falling over themselves to post "How fortunate, then, that we had a support/oppose system in 2007", please don't bother, we'll take it as read. Bishonen | talk 13:51, 22 December 2012 (UTC). (Sorry, Scott Mac, I don't know what happened there; I didn't get an edit conflict.)
- This S/(S+O) system has the disadvantage of rewarding candidates who avoid opposes, just as much as those who attract positive reactions in the community. And many or most voters don't even understand it, so we don't even get the advantage of simplicity. The obvious way to go is to make it preferential, which would involve the choice of one of a number of standard, well-oiled systems used both out there and in the WM movement. There's nothing exotic about it, and for voters it just means numbering in cardinal order. Tony (talk) 13:02, 22 December 2012 (UTC)
- It's been said before that the support/oppose system favours those who keep their head down and say nowt. But I suspect every voting system favours one strategy or another. Elen of the Roads (talk) 20:11, 18 December 2012 (UTC)
The issue is what type of arbcom you want. There is little doubt the current system favours the well-known and least objectionable. So, that's either people who are natural and convinced moderates, or people who'd not really come and said too much, so that everyone thinks the best of them, but no one really knows what they'll do when elected. A different system would almost certainly produce a different arbcom -and probably a more representative one. Strong minority perspectives would be certain to get at least one candidate. You'd have a Giano, a reformist, a BLP crusader, among your 16. But do you really want that? If you were electing a "Gov Com" you certainly would - you'd want all significant minority opinions represented. But arbitration is somewhat different. You need to be able to trust not just that the outcome of any arbitrator vote reflects the community's values, you need to be able to trust each individual with secret information too. Can you do that with a partisan and divided committee? I think not. You can only do that with a committee where most people don't find any member highly repugnant. In short, you need a committee of moderates. The voting system we have is probably the best one for the role arbcom actually plays.--Scott Mac 13:45, 22 December 2012 (UTC)
- That's an interesting perspective, and I don't disagree with it mostly. I think that – from year to year – the results have clearly favoured the known quantities and those candidates that eschew highly polarized positions. There is, however, a cost to pay for this "blandness": more marginal candidates end up being disqualified even when they could have had a positive influence on the committee by bringing in a fresh perspective.
A case in point this year, I voted for two candidates that did not fit the "usual" ArbCom mold; and I'm fairly confident that they did not really have a shot at getting elected over factors which – in the end – are not directly related to their capacity to hold the role. Our system favours a safe, predictable ArbCom (which is not undesirable) but also risks a certain amount of stagnation (which is not so desirable). It's worth discussing. — Coren (talk) 16:39, 22 December 2012 (UTC)
- Grossly unfair, unconstitutional, and dishonest voting systems effectively give a very small minority of in-the-know people (e.g. people who understand how to manipulate the system by using negative votes) a high multiple of times the number of votes of the majority who do not believe in negative voting. The Giano example, where the second-most-widely-respected candidate could not get in because of the negative votes of a small coordinated group, is a good example of the unfairness of the system. Having honest, fair, and impartial people, who put both users and the diverse community of contributors first, and who don't vote with minority-"elite" group-think, would surely make Wikipedia more fair and representative and reduce bullying, vindictiveness, and cronyism. People who are "too far out" will surely always fail to get in for lack of support votes—without negative voting. LittleBen (talk) 17:08, 22 December 2012 (UTC)
- OK, you are losing me. First, your basic premise is wrong. By your logic the following statement is equally valid, simply switching the "positive" and the "negative":
- Grossly unfair, unconstitutional, and dishonest voting systems effectively give a very small minority of in-the-know people (e.g. people who understand how to manipulate the system by using negative votes) a high multiple of times the number of votes of the majority who do not believe in negative voting. The Giano example, where the second-most-widely-respected candidate could not get in because of the negative votes of a small coordinated group, is a good example of the unfairness of the system. Having honest, fair, and impartial people, who put both users and the diverse community of contributors first, and who don't vote with minority-"elite" group-think, would surely make Wikipedia more fair and representative and reduce bullying, vindictiveness, and cronyism. People who are "too far out" will surely always fail to get in for lack of support votes—without negative voting. LittleBen (talk) 17:08, 22 December 2012 (UTC)
| “ | Each [positive] vote for a candidate cancels out someone else's [negative] vote for that candidate, which has the same effect on the results (the ranking or ordering) as a [negative] vote for all the other candidates. So when you convert [positive] votes to equivalent [negative] votes, it's obvious that [positive] votes carry far more weight. | ” |
- Obviously they can't be both right, but by your logic they are. The reason why is because a negative vote is not the same as a positive vote for every other candidate even in terms of ranking. Consider a simple counterexample with three candidates: A has 18 supports and 6 opposes, B has 5 supports and 2 opposes, C has 10 supports and 15 opposes. So the percentage for A is 75%, for B is 71.43%, and for C is 40%, and the ranking is A>B>C. A negative vote for A will bring the percentage down to 72%, but the ranking is unchanged: A>B>C. Positive votes for B and C will bring B's percentage up to 75%, making B higher than A, while C's percentage becomes 42.3%, still the lowest. Therefore, positive votes for everyone other than A will make the ranking B>A>C, while a negative vote for A will not change the ranking. Your entire premise - that a negative vote will have the same effect on the ranking as a positive vote for all the other candidates - is simply wrong.
- Second, whether or not Giano should have been elected, 200+ opposes are hardly a "small" group. T. Canens (talk) 17:45, 22 December 2012 (UTC)
- I'm curious as to how Ben came to the conclusion that there was something "unconstitutional" about the process. I was not aware that Wikipedia had a constitution. Where is it? Also, the contention that a majority does not believe in negative voting is easily discredited by even a casual glance at this year's results. Beeblebrox (talk) 18:24, 22 December 2012 (UTC)
- Probably as a synonym for "unfair/unrepresentative". My feelings align somewhat withc Coren's. Personally I think Giano would have been good as an arbitrator - balancing some big-picture folks with detail-examiners is good as long as everyone is talking constructively, which for the most part has happened (talking, that is). Having partly served in three "seasons" with different selections of folks on the committee, I'd have to look at the election results and think about how different the makeup would have been if we'd gone on straight "support" votes. Casliber (talk · contribs) 00:39, 23 December 2012 (UTC)
- "Grossly unfair ..." (Ben) – yes, it's distortive on a behavioural level because some folks (like me) use opposes to advantage their chosen few, and others aren't aware of this option or are unwilling to use it. Not transparent, I say. Bring in simple preferential voting. Tony (talk) 00:49, 23 December 2012 (UTC)
- (@Casliber) Interestingly enough, all else being equal, this year would have had but one significant difference: Elen would have knocked David out of the top eight and taken his seat. That said, there's no reason to believe that voters' behaviour would have been the same with straight approval rating either; there are a number of voters who would have supported candidates they had remained neutral on, for instance, to prop them up above those they would have opposed if they could (indeed, I would have done so myself) – so it's not clear that you can take the result of a past election and consider that the support votes represent what would have happened if there was no possible opposition. — Coren (talk) 00:54, 23 December 2012 (UTC)
- Yup. Although I don't think using support only would yield grossly different results. It would still favour candidates who had wide support over those who had deep support. Now, if you ranked candidates you might get a very different result. Again it depends whether you want an arbcom composed of those who are fairly moderate (and thus striking the balance of community opinion), or of you want an arbcom which is balanced because groups with strong views can secure seats, but so can those with the opposite views. In the end, a representative arbcom would be interesting, but I'm not sure it would work.--Scott Mac 01:02, 23 December 2012 (UTC)
- @Tony: Doesn't preferential voting presuppose that the electorate wants all open seats to be filled? NW (Talk) 01:04, 23 December 2012 (UTC)
- @NW: Yes, preferential voting fills the number of vacant seats; let's be practical—in no ArbCom election in recent memory have there not been sufficient candidates above this 50% support-ranking fill all vacant seats. And as people are starting to realise, the 50% threshold (originally Jimmy's idea) is based on a wobbly premise: if enough people start voting oppose for every candidate they haven't chosen (and indeed to minimise the number of their supports to gain even more advantage for their favoured candidates, as I've been doing for years), we could get to an unfortunate situation where voters' strategy is pushing a lot of candidates below the 50% mark; but this would not necessarily be because voters actively disapprove of them. In preferential voting there's no way for voters to vote "strategically" in this negative way.
To those who are teasing out how the results might have been different in this and past elections if another system had been used: I wonder what the point is: doesn't the community decide on the voting system that suits, and the system delivers a result, no questions? By analogy, the massive distortions of the state-based electoral college system in US presidential elections don't seem to get much of an airing in the media—you live with it or you change it. Tony (talk) 03:05, 23 December 2012 (UTC)
- @NW: Yes, preferential voting fills the number of vacant seats; let's be practical—in no ArbCom election in recent memory have there not been sufficient candidates above this 50% support-ranking fill all vacant seats. And as people are starting to realise, the 50% threshold (originally Jimmy's idea) is based on a wobbly premise: if enough people start voting oppose for every candidate they haven't chosen (and indeed to minimise the number of their supports to gain even more advantage for their favoured candidates, as I've been doing for years), we could get to an unfortunate situation where voters' strategy is pushing a lot of candidates below the 50% mark; but this would not necessarily be because voters actively disapprove of them. In preferential voting there's no way for voters to vote "strategically" in this negative way.
- @Tony: Doesn't preferential voting presuppose that the electorate wants all open seats to be filled? NW (Talk) 01:04, 23 December 2012 (UTC)
- Yup. Although I don't think using support only would yield grossly different results. It would still favour candidates who had wide support over those who had deep support. Now, if you ranked candidates you might get a very different result. Again it depends whether you want an arbcom composed of those who are fairly moderate (and thus striking the balance of community opinion), or of you want an arbcom which is balanced because groups with strong views can secure seats, but so can those with the opposite views. In the end, a representative arbcom would be interesting, but I'm not sure it would work.--Scott Mac 01:02, 23 December 2012 (UTC)
- Probably as a synonym for "unfair/unrepresentative". My feelings align somewhat withc Coren's. Personally I think Giano would have been good as an arbitrator - balancing some big-picture folks with detail-examiners is good as long as everyone is talking constructively, which for the most part has happened (talking, that is). Having partly served in three "seasons" with different selections of folks on the committee, I'd have to look at the election results and think about how different the makeup would have been if we'd gone on straight "support" votes. Casliber (talk · contribs) 00:39, 23 December 2012 (UTC)
- I'm curious as to how Ben came to the conclusion that there was something "unconstitutional" about the process. I was not aware that Wikipedia had a constitution. Where is it? Also, the contention that a majority does not believe in negative voting is easily discredited by even a casual glance at this year's results. Beeblebrox (talk) 18:24, 22 December 2012 (UTC)
- (sigh) the other big problem is that delving into previous elections and examining which candidates might have gotten in and what that might have meant would get extremely personal really..err, and all it'd take is one or two diplomacy-fails for this page to descend into acrimony....now where is that ten foot bargepole.....Casliber (talk · contribs) 01:39, 23 December 2012 (UTC)
- The following analysis of the 2007 election may show why negative voting is grossly unfair:
| Name | Support | Oppose | Support+Oppose | Net | Percentage | |
|---|---|---|---|---|---|---|
| 1 | Newyorkbrad | 552 | 15 | 567 | 537 | 97.35% |
| 2 | FT2 | 264 | 33 | 297 | 231 | 88.89% |
| 3 | FayssalF | 249 | 78 | 327 | 171 | 76.15% |
| 4 | Sam Blacketer | 159 | 55 | 214 | 104 | 74.3% |
| 5 | Deskana | 294 | 109 | 403 | 185 | 72.95% |
| 6 | Thebainer | 158 | 77 | 235 | 81 | 67.23% |
| 7 | Rebecca | 233 | 125 | 358 | 108 | 65.08% |
| 8 | Raul654 | 317 | 171 | 488 | 146 | 64.96% |
| 9 | Manning Bartlett | 129 | 84 | 213 | 45 | 60.56% |
| 10 | Giano II | 312 | 229 | 541 | 83 | 57.35% |
| 11 | David Fuchs | 104 | 81 | 185 | 23 | 56.22% |
Six got in: Newyorkbrad, FT2, FayssalF, Sam Blacketer, Deskana, and Thebainer.
If a win were based on the positive vote ranking then Raul654 and Giano II would have been in ; Sam Blacketer and Thebainer, would have missed out.
If it is fair to assume that the number of people voting was not much more than 567 (the support plus oppose figure for Newyorkbrad), then fewer than about half the voters had an opinion about Thebainer, Sam Blacketer, Manning Bartlett, and David Fuchs. It appears that the percentage figures in the above table are Support/(Support+Oppose) for each candidate, a pretty meaningless figure—surely Net/(Total number of voters), or Support/(Total number of voters) would be more meaningful ratios.
Let's consider how many voters would need to target Raul654 and Giano II with negative votes to ensure that they would not get in, despite their very large positive support. If the group of people who are targeting them are well coordinated, they will cast positive votes for all the other candidates and negative votes for these two.
For Raul654 to have a percentage greater than 67.23% he would have needed a Support+Oppose total of 471 or fewer (as few as nine people could have swung the vote). For Giano II, the figure would be 464.
It seems that it doesn't really matter how small a proportion of the total number of voters support a candidate, provided that the Support/(Support+Oppose) ratio for that candidate beats that of a more widely-supported candidate. Is that considered fair or representative? LittleBen (talk) 14:59, 23 December 2012 (UTC)
- LittleBen, that's a whole lot of text to say "If things had been different, they wouldn't have been the same." You're making the very much unwarranted presumption that people who voted support and oppose on candidates would have kept their voting patterns the same in the absence of opposition voting. That's a very dubious supposition in a secret ballot; it's completely nonsensical with raised hand voting like what occurred before 2009. The "positive vote ranking" of the candidates in 2007 cannot be guessed from the support votes of a support-oppose vote; without the option to oppose, people would have voted differently (including people voting who did not and the opposite). — Coren (talk) 17:21, 23 December 2012 (UTC)
- The question is, did the 432 users (more than half, again) who had no opinion about Pgallert, or the 394 (enough to tip the results in favor) who had no opinion about RegentsPark, say this due to lazyness or simply due to the fact that they thought the positive and the negative more or less balanced each other. And if th voting options were different, we may end up with users who truely don't have an opinion about some user being forced to say something - and it woulsd be something they didn't mean to say. עוד מישהו Od Mishehu 20:53, 23 December 2012 (UTC)
- The key issues, as I see them, are that, (1) in practice, the current negative voting gives people effectively two votes against people they dislike (one negative vote against the candidate, and one vote for all the others) but only one positive vote for people they like (the great majority of honest people are not going to use negative votes against people they have nothing against, just to manipulate the results), and (2) the current percentage criteria is meaningless, as it is not a percentage of the total number of voters (those voting for any candidate) but just a percentage of those voting for or against the current candidate. LittleBen (talk) 04:02, 24 December 2012 (UTC)
- On the other hand, I think we're better off allowing a user do some deepresearch into a few candidates, and form opinions about them, than force them to do research on all 21 candidates and vote in a single sitting. עוד מישהו Od Mishehu 11:07, 24 December 2012 (UTC)
- Little Ben, your assumptions are simply flawed. It is possible for someone to do as you say - but, in doing so, they lose the ability to positively express any preference for those they actually support. In practice, I suspect it is the exact opposite that often happens. People vote "support" for those the like and then oppose the rest to maximise their support vote. Indeed, many people have expressly said that's what they are doing. There is simply no real problem here.--Scott Mac 11:14, 24 December 2012 (UTC)
- On the other hand, I think we're better off allowing a user do some deepresearch into a few candidates, and form opinions about them, than force them to do research on all 21 candidates and vote in a single sitting. עוד מישהו Od Mishehu 11:07, 24 December 2012 (UTC)
- The key issues, as I see them, are that, (1) in practice, the current negative voting gives people effectively two votes against people they dislike (one negative vote against the candidate, and one vote for all the others) but only one positive vote for people they like (the great majority of honest people are not going to use negative votes against people they have nothing against, just to manipulate the results), and (2) the current percentage criteria is meaningless, as it is not a percentage of the total number of voters (those voting for any candidate) but just a percentage of those voting for or against the current candidate. LittleBen (talk) 04:02, 24 December 2012 (UTC)
- The question is, did the 432 users (more than half, again) who had no opinion about Pgallert, or the 394 (enough to tip the results in favor) who had no opinion about RegentsPark, say this due to lazyness or simply due to the fact that they thought the positive and the negative more or less balanced each other. And if th voting options were different, we may end up with users who truely don't have an opinion about some user being forced to say something - and it woulsd be something they didn't mean to say. עוד מישהו Od Mishehu 20:53, 23 December 2012 (UTC)
- If people's concern about the S/N/O voting system is principally that the uninformed cast weaker votes than the informed because they do not realise that their vote can be (minutely, by previous calculations) increased in strength by tactical voting, then why not remove the neutral option? Happy‑melon 11:48, 24 December 2012 (UTC)
- Now that is a very good idea, H-M. Brilliant lateral thinking. But better still would be to abolish the negative buttons as well, so you just click support for the candidates you support. Even better would be to number them in order of your preference. Simple for voters, high level of matching vote result with community feeling. Tony (talk) 11:54, 24 December 2012 (UTC)
- I disagree that making it a straight support vote wouldbe a desired thing, When it comes to something like ArbCom, the last step of DR, you do NOT want someone who is very controversial. Worries about stagnation aside, ArbCom works best when the focus is on the issues before it, not the people who make up the Committee. Neutral votes are useful in that they provide an option for the folks that you don't know that they won't have problems with it, but you don't know that they WILL have problems with it either. It's unfair to folks who aren't familiar faces at the usual locations to remove the Neutral vote, otherwise you'll just have folks voting for the most vocal or well-known at the usual places. SirFozzie (talk) 12:16, 24 December 2012 (UTC)
- I agree with all your sentiments, but your conclusions (that we should keep oppose, and also neutral, options) do not necessarily follow from them. It is a mathematical fact that no voter should ever vote neutral for a candidate, since doing so reduces their happiness with the final outcome by somewhere between zero and a negligible amount. An argument to keep neutral must address the question of why it is acceptable to allow uninformed voters to be disenfranchised by 0.00013% of a vote. Happy‑melon 13:17, 24 December 2012 (UTC)
- The neutral option makes it easier to proceed with voting while still undecided about some of the candidates. Requiring a decisive vote for every candidate may reduce the number of votes cast. --SmokeyJoe (talk) 14:04, 24 December 2012 (UTC)
- Yup. And you will have some voters who've no wide knowledge of the community and no interest in reading any statements, but may have interacted with one or two candidates, and want to say simply "I think [User:Thoughtful] would be be good", or equally, "whoever you pick, please don't choose [User:Utterbastard]". That type of contribution is worthwhile, and requires S/O/N options all to be available.--Scott Mac 14:13, 24 December 2012 (UTC)
- The neutral option makes it easier to proceed with voting while still undecided about some of the candidates. Requiring a decisive vote for every candidate may reduce the number of votes cast. --SmokeyJoe (talk) 14:04, 24 December 2012 (UTC)
- I agree with all your sentiments, but your conclusions (that we should keep oppose, and also neutral, options) do not necessarily follow from them. It is a mathematical fact that no voter should ever vote neutral for a candidate, since doing so reduces their happiness with the final outcome by somewhere between zero and a negligible amount. An argument to keep neutral must address the question of why it is acceptable to allow uninformed voters to be disenfranchised by 0.00013% of a vote. Happy‑melon 13:17, 24 December 2012 (UTC)
Convenience break
Every year a few people remark on how little sense it makes for anyone to cast a "No Vote" vote, and yet at every election there are hundreds of "No Votes" cast, so in the eyes of the voters they do serve a purpose. (A year or two ago someone made up a nice pie chart showing the breakdown of Support/Oppose/No Votes, and I don't remember the breakdown, but "No Vote" got a substantial proportion. I suppose I could just add up the columns for this year and take the percentage, but I don't feel like doing it right now.) I myself use the "No Vote" for candidates to whom I am indifferent. I don't really know the mathematics of it, but it makes some sense to me, since I don't necessarily know whether all the candidates I "Support" are going to win, to differentiate among the other candidates and "Oppose" only those candidates who I truly oppose, which gives my "No-Vote" candidates an advantage (by one vote) over those I "Oppose." As I said, I don't know whether the math works out, but there does seem to be some logic to it, and obviously I am not alone in making use of the "No Vote" option. As for the use of the "Oppose" button, it is true that those who use the "Oppose" button more probably have more of an impact on the election, but that is what it is there for. Everybody has an equal opportunity to use it.
As for "ranked voting", Tony, you have advocated that for the past few years, but there are a number of different "ranked" or "preference" voting systems. Which one are you suggesting? There are some that are "proportional" (like "single transferable vote") and some that are not.
Overall, the fact is that while there are suggestions for change every year, the annual RfC's keep producing majorities for the current system. It seems that the community as a whole finds some comfort in knowing that the elected candidates were not only able to attract raw numbers of "positive votes" but also had relatively little opposition as well. Neutron (talk) 16:28, 24 December 2012 (UTC)
- Neutron, if I may reply to your points ... Just because we use the three options in the voting system doesn't mean we support its continuance. You really have no choice to use it unless you're going to refuse to vote as a protest against it (I know someone who does that, actually).
The three-option system—cast in the first-past-the-post system developed for the British parliament, which the US adopted—is weird, just weird. It seems to have arisen in the en.WP system from the pre-existing RfC, RfA, and other less formal polling process we have, where quite naturally people fell into explicit support, oppose and neutral sections on the page. But those are contexts in which we typically write comments against our explicit !vote, in which it is a !vote, not a numerical vote (well, not formal one). Here in ACEs, we have a proper democratic vote, yet a system that has morphed from page !votes/comments into a formal vote-counting system it was never designed for.
Yes, I'd opt for one of the single transferable voting systems, which are not uncommon in the outside world and in the movement. In one way it's more work for the scrutineers, but it's not rocket science: you just take the candidate who received the lowest number of "1s" and redistribute their votes on the basis who got "2" on their ballots. Then you take the next-lowest 1-scorer and do the same. You repeat this in successive rounds until you get down to the number of candidates needed to fill the vacancies. No one's vote is "wasted", and everyone's highest preferences are counted. In ACE2012, then, YOLO Swag's 45 first prefs would habe been re-allocated to the candidates his voters placed second. Then Count Ibis's 87 first prefs (plus any s/he got from YOLO as second pref) would no doubt have been next to fall. There would have been 13 rounds to get down to 8 candidates, taking into account much more deeply what the community prefers.
Yes, I will continue to advocate changing this inept triple-headed beast, because I don't like the invitation to negativity, the fact that it's already complex without reason, and the disregard of voters' deeper preferences for candidates. At the very least, please consider a minor fix that would remove the unfairness among voters of strategic voting: get rid of the neutral button, as Happy-Melon suggests. Tony (talk) 01:29, 25 December 2012 (UTC)
- Neutral is one thing we cannot remove. Lots of Wikipedians are not widely knowledgeable about most candidates, and don't have the time, or the knowledge to a made decisions on more than on a few candidates they've interacted with and have an opinion on. If you force such into a position where they must express an opinion on all candidates, then you will discourage them from registering the opinions they do have. That's not acceptable.--Scott Mac 02:46, 25 December 2012 (UTC)
- Scott has a point - in teh Real World, most of our candidates are aligned to parties, making voting a matter of voting for a party rather than a candidate per se. We don't have that here (thankfully) Casliber (talk · contribs) 06:39, 25 December 2012 (UTC)
- Agreed, as I said on a brief soliloquy on three years as an arb (available on my user page if you want to read), the only parties Wikiepdia needs involve Wiki-meetups ;) SirFozzie (talk) 13:21, 25 December 2012 (UTC)
- Neutral is one thing we cannot remove. Lots of Wikipedians are not widely knowledgeable about most candidates, and don't have the time, or the knowledge to a made decisions on more than on a few candidates they've interacted with and have an opinion on. If you force such into a position where they must express an opinion on all candidates, then you will discourage them from registering the opinions they do have. That's not acceptable.--Scott Mac 02:46, 25 December 2012 (UTC)
- Please remind me why the presence or absence of parties has anything to do with the utility of preferential voting. Tony (talk) 04:44, 26 December 2012 (UTC)
- Tony, if I could put it simply, in every other kind of "election" we have on this project (and indeed in almost every Wikimedia project election, save the Board of Trustees vote), voters always have the option of "voting" neutrally by not voting at all. Keep in mind that, except for a very tiny number of elections that involve SecurePoll, and the Board of Trustees election, all voting is done publicly with support/oppose voting only. The absence of a vote is equivalent to the no-vote that we currently see on Arbcom elections. I've yet to see a good case made for preferential voting. It is used for WMF Board of Trustees elections for three of the 7 seats, and it has resulted in perverse results in every election I've observed because it uses the Schulze method, which is counterintuitive and places any ranked candidates above unranked candidates, thus favouring tactical voting and making non-votes essentially the equivalent of an oppose. From what I understand from your comments here, the only kind of votes you want people to be able to make are support votes. That would, you realise, eliminate the ability to use SecurePoll. Now, we could go back to the process of public voting, which was used in every election before 2009. In the one Arbcom election where only support voting was permitted, it was a rather significant fiasco. The ability to say "no thank you" to candidates is essential to a huge percentage of the members of the community. Risker (talk) 06:02, 26 December 2012 (UTC)
- I guess that, atfer reading this, the issue is with neutral votes, which, as Risker says, "The absence of a vote is equivalent to the no-vote that we currently see on Arbcom elections." So, why shouldn't we just stop reporting neutral votes on the final table and start pretending they don't exist? This way, we'll only see two columns (support and oppose) and we'll be all happy. — ΛΧΣ21 06:24, 26 December 2012 (UTC)
- No comment on your other points, Risker, but every system we've heard proposed on this page or any other can be implemented in SecurePoll, so to say that switching system would "eliminate the ability to use [it]" is not correct. There are half a dozen different talliers built into the system itself, and it's very extensible to write new ones; I wrote an Alternative Vote tallier for it fairly recently that was not particularly difficult. Switching voting systems does not imply a return to public voting. Happy‑melon 10:06, 26 December 2012 (UTC)
- This discussion has gotten all jumbled up among various proposals and tweaks to come of the proposals, and currently involves just a few editors, so I don't think we're really going anywhere here as far as resolving anything. Evidently all these ideas will need to be hashed out again in an RfC. I was sort of hoping that we had the basic voting system resolved and that we could just go with MBisanz's codification for next year (even though there are other parts of that that I would still like to see tweaked) without a full-fledged RfC. Apparently not. It seems to me that we spend a lot of words and energy re-inventing the wheel almost every year, although I thought that the lack of a serious counterproposal on the election method in this year's RfC meant that we had it settled. Evidently a couple of people were just taking a year off. :) I will just say a couple things, one is that I do not agree with Tony that getting rid of the "neutral" button is a "minor fix." Among other things, I think it could "break" the system by producing so many additional "oppose" votes that not enough candidates would be elected. (I haven't checked the numbers for this year, but I know that for some past elections, if you turned all the "neutrals" into "oppose" votes, not many candidates would reach 50 percent.) I also don't see any reason to take away the voters' option to "abstain" on a candidate, even if such an action is not "mathematically correct." (And I am not convinced about the math anyway, because if at least one of the candidates that a voter "Supports" does not get elected, voting "neutral" on some and "oppose" on others does allow that voter to make a difference as to who wins.) As for "single transferable vote" (a proportional representation system), I am sure it could be worked out technically, but my question is, is it really the system that the community wants, in light of the continuing support for a threshold of at least 50% support? STV does not require that a majority of voters support a candidate; just the opposite, it is likely to result in the election of one or more "minority" candidates over one or more "majority" candidates. This is especially true given the "demographics" of ArbCom elections, in which we are electing seven or eight people at a time and there are some candidates in the current system who achieve almost enough support, but not quite. In an STV system some of the past losers would be winners, and vice versa. Is that really what the community wants? In the 2011 RfC there was very nearly a majority for increasing the threshold to 60%, in fact I believe that was only avoided when the "votes" of the zero-threshold people were added to the 50-percenters on the theory that they wanted "no more than 50". So now suddenly you're going to try to get a consensus for a proportional representation system? I'm not necessarily opposed to it, but I just don't see it happening. Neutron (talk) 19:26, 26 December 2012 (UTC)
- I see absolutely no need for another RFC. You've got a few people who want to change the system (you always have and always will) but there is and will be no consensus to do that. I'd also strong oppose MBisanz's "codification", which rather than be a simply codification is the worst example of instruction creep I've seen a while.--Scott Mac 20:10, 26 December 2012 (UTC)
- I'm personally perfectly happy with the status quo, except being very bemused to watch people continue to flog their preferred member of a herd of horses that are, while not entirely dead, at least badly bruised already. As far as I'm concerned "the math" is correct, and indicates that the tactical disadvantage to using the "neutral" option in our SNO system is less than one part in 750,000, making that particular voting system probably no more susceptible to tactical voting than any other. Whether SNO is the right broadly-fair voting system for ArbCom is not something I have a particularly strong opinion on, but I would like, as you say, to see those who do have a strong opinion to set out their thoughts clearly and in a way which actually recognises the points on each side rather than just talking past each other. In the dozens of screens of discussion on this issue over the years there has been a lot of presentation of hard, empirical evidence on certain points which is often simply ignored in future discussion, which is a shame. Happy‑melon 20:42, 26 December 2012 (UTC)
- I see absolutely no need for another RFC. You've got a few people who want to change the system (you always have and always will) but there is and will be no consensus to do that. I'd also strong oppose MBisanz's "codification", which rather than be a simply codification is the worst example of instruction creep I've seen a while.--Scott Mac 20:10, 26 December 2012 (UTC)
- This discussion has gotten all jumbled up among various proposals and tweaks to come of the proposals, and currently involves just a few editors, so I don't think we're really going anywhere here as far as resolving anything. Evidently all these ideas will need to be hashed out again in an RfC. I was sort of hoping that we had the basic voting system resolved and that we could just go with MBisanz's codification for next year (even though there are other parts of that that I would still like to see tweaked) without a full-fledged RfC. Apparently not. It seems to me that we spend a lot of words and energy re-inventing the wheel almost every year, although I thought that the lack of a serious counterproposal on the election method in this year's RfC meant that we had it settled. Evidently a couple of people were just taking a year off. :) I will just say a couple things, one is that I do not agree with Tony that getting rid of the "neutral" button is a "minor fix." Among other things, I think it could "break" the system by producing so many additional "oppose" votes that not enough candidates would be elected. (I haven't checked the numbers for this year, but I know that for some past elections, if you turned all the "neutrals" into "oppose" votes, not many candidates would reach 50 percent.) I also don't see any reason to take away the voters' option to "abstain" on a candidate, even if such an action is not "mathematically correct." (And I am not convinced about the math anyway, because if at least one of the candidates that a voter "Supports" does not get elected, voting "neutral" on some and "oppose" on others does allow that voter to make a difference as to who wins.) As for "single transferable vote" (a proportional representation system), I am sure it could be worked out technically, but my question is, is it really the system that the community wants, in light of the continuing support for a threshold of at least 50% support? STV does not require that a majority of voters support a candidate; just the opposite, it is likely to result in the election of one or more "minority" candidates over one or more "majority" candidates. This is especially true given the "demographics" of ArbCom elections, in which we are electing seven or eight people at a time and there are some candidates in the current system who achieve almost enough support, but not quite. In an STV system some of the past losers would be winners, and vice versa. Is that really what the community wants? In the 2011 RfC there was very nearly a majority for increasing the threshold to 60%, in fact I believe that was only avoided when the "votes" of the zero-threshold people were added to the 50-percenters on the theory that they wanted "no more than 50". So now suddenly you're going to try to get a consensus for a proportional representation system? I'm not necessarily opposed to it, but I just don't see it happening. Neutron (talk) 19:26, 26 December 2012 (UTC)
- Tony, if I could put it simply, in every other kind of "election" we have on this project (and indeed in almost every Wikimedia project election, save the Board of Trustees vote), voters always have the option of "voting" neutrally by not voting at all. Keep in mind that, except for a very tiny number of elections that involve SecurePoll, and the Board of Trustees election, all voting is done publicly with support/oppose voting only. The absence of a vote is equivalent to the no-vote that we currently see on Arbcom elections. I've yet to see a good case made for preferential voting. It is used for WMF Board of Trustees elections for three of the 7 seats, and it has resulted in perverse results in every election I've observed because it uses the Schulze method, which is counterintuitive and places any ranked candidates above unranked candidates, thus favouring tactical voting and making non-votes essentially the equivalent of an oppose. From what I understand from your comments here, the only kind of votes you want people to be able to make are support votes. That would, you realise, eliminate the ability to use SecurePoll. Now, we could go back to the process of public voting, which was used in every election before 2009. In the one Arbcom election where only support voting was permitted, it was a rather significant fiasco. The ability to say "no thank you" to candidates is essential to a huge percentage of the members of the community. Risker (talk) 06:02, 26 December 2012 (UTC)