Highlighting best answers in profile

January 10, 2007 1:31 PM

Would there be a way to either determine the percentage of a user's responses to AskMe are marked as "Best Answer", and/or be able to list out only those response marked as "best answer" to be displayed as part of a user's profile?
posted by Doohickie to Feature Requests at 1:31 PM (97 comments total)

I think we've been here before. I'm putting money on "no".
posted by boo_radley at 1:39 PM on January 10, 2007

ASKME.
IS NOT.
A COMPETITION.
DOOHICKIE.
posted by Plutor at 1:39 PM on January 10, 2007

Of course there's a way to do it, but the real question is should you.

I say no, because Best Answers aren't useful to anyone except the user who marked them. Everything is just alpha mefiuser bs.
posted by Brandon Blatcher at 1:42 PM on January 10, 2007

I actually think this would have some merit. I imagine it "working" in the same way tags work--lots of noise with a small amount of data but meaningful with a large enough dataset.

However I think instead of "what percentage of answers were marked as best answer", it should be:

What percentage of threads in which a user participated and a best answer was awarded, was a best answer given to the user?

This would count only threads in which a best answer was awarded and not penalize a user for answering multiple times in the same thread where only one of his answers was marked as best answer.
posted by null terminated at 1:46 PM on January 10, 2007

I also think the current system doesn't properly distinguish between answer types for this to work. For example, I would be interested in seeing which user had the highest percentage of "correct" answers (ie, solutions to specific technical problems) but I'm not interested in answers to questions like "What should I name my dog". Both of these types of answers could be marked "best answer" while the former is factually the only correct answer and the latter is the "best answer" only subjectively.
posted by null terminated at 1:49 PM on January 10, 2007

It's a good idea. The reason we've "been here before" is because it keeps being a good idea. It would be an excellent way to gauge the credibility of any given user. Is he helpful? Is she smart? Why should I listen to them? It's like a counsellor providing references. To have earned a "Best Answer" is to have earned someone's trust. Since Metafilter values positive contributions, this would be a healthy way to encourage people to keep contributing positively.

The "not a competition" argument is true, but it doesn't make a lot of sense. We already tally favourites, and I find that great inspiration to strive for better, smarter, more engaging posts and replies.
posted by Milkman Dan at 1:50 PM on January 10, 2007

I'm putting money on "no".

You're on. Something tells me this time Matt will say yes.
posted by Terminal Verbosity at 1:50 PM on January 10, 2007

We could call it "karma", and we could give a "karma bonus" to those who have high scores, so that their answers would be highlighted in some way.

I swear it will work flawlessly.
posted by smackfu at 1:57 PM on January 10, 2007

Or we could call it "Gravitas," and actually consider it as another possibly useful "filter" in the arsenal, rather than assume it was nothing more than an ego thing.
posted by It's Raining Florence Henderson at 2:01 PM on January 10, 2007

Best answers are useful within the thread. They let you know what was genuinely helpful. Best answers aren't really useful per user. It's not a competition and a user can have no best answers but have given lots of great answers, or have some best answers which really weren't that wonderful. This is the way best answers was designed and I think they fulfil their purpose pretty well.

If you're interested in seeing how good a user is go through and actually read some of their answers. Judge for yourself.
posted by shelleycat at 2:05 PM on January 10, 2007 [1 favorite]

We could have something like "Worst Answer," for everyone's amusement.

Seriously folks, Mefi is simple and clean. I for one would like to keep it that way.
posted by muddgirl at 2:07 PM on January 10, 2007

Top ten users by number of "best answers" for 2004-2006.

Apparently jessamyn has the biggest penis. Can we go home now?


93 languagehat
93 occhiblu
97 kindall
100 cillit bang
101 dhartung
104 smackfu
112 iconomy
141 ikkyu2
190 paulsc
191 jessamyn
posted by tkolar at 2:10 PM on January 10, 2007 [1 favorite]

On one hand, I think it would be nifty. I especially like the "karma" bit.

On the other hand, I think the range of questions is too wide.

I think it would be nifty if you could see who had good/best answers by category.

"Oh look. So and so is apparently pretty helpful with 'law & government' related questions. And whatshisface is pretty good with home and garden. And this guy over here is very good at answer health/penis-related questions."

I'd like to think that this would be useful in some way. Perhaps contacting the user after seeing they have experience or knowledge about a certain thing, instead of using a question only to have it answered by them anyway. Y'know?
posted by CitrusFreak12 at 2:12 PM on January 10, 2007

Must work harder to make top 10, must work harder to make top 10, must work harder to make top 10.....
posted by ThePinkSuperhero at 2:17 PM on January 10, 2007 [1 favorite]

Perhaps contacting the user after seeing they have experience or knowledge about a certain thing, instead of using a question only to have it answered by them anyway. Y'know?

I'm pretty sure this is why some users (like medical expert ikkyu2) keep their e-mail addresses private. Imagine an inbox full of stupid medical questions every. single. day. AskMetafilter is a community for answering questions, not identifying smart users we can bug privately forever after.
posted by ThePinkSuperhero at 2:20 PM on January 10, 2007 [2 favorites]

Milkman Dan: To have earned a "Best Answer" is to have earned someone's trust. Since Metafilter values positive contributions, this would be a healthy way to encourage people to keep contributing positively.

Not everyone uses "Best Answers" that way. We've all seen threads where the funny dismissive answer has been marked, or where it's the type of question for which there can't be a best answer and yet there are eight "best answers" marked. And don't take this personally because I'm just generalizing, but the people who would only answer in order to raise their best answer tally are the kinds we don't want around here. You want karma, head on over to Slashdot, and don't let the door hit you on the ass on the way out.
posted by joshuaconner at 2:20 PM on January 10, 2007

AskMe: Tell me if this hurts.
posted by It's Raining Florence Henderson at 2:21 PM on January 10, 2007

Oh yeah, and of the 9856 people who answered questions in 2004-2006, 5320 never had one of the answers marked as "best". Ha hah! Suck it Losers!
posted by tkolar at 2:21 PM on January 10, 2007

Askme is an exhibition, not a competition. Please, no wagering.
posted by found missing at 2:27 PM on January 10, 2007 [2 favorites]

I'm saving my "Best Answers" for my wedding night.
posted by gigawhat? at 2:28 PM on January 10, 2007

And sometimes the best answer isn't the best answer.
posted by slimepuppy at 2:30 PM on January 10, 2007

Must work harder to make top 10, must work harder to make top 10, must work harder to make top 10.....

Hmmmm, that could solve the too many questions problem neatly. Or not.
posted by IronLizard at 2:31 PM on January 10, 2007

They let you know what was genuinely helpful.

NO THEY DON'T. They let you know what some random poster happened to feel like awarding Best Answer to. "Best answers" can be genuinely good answers, they can be completely wrong, they can be clever jokes, they can be stupid jokes, sometimes they seem to be awarded by throwing darts at the monitor. And many times none are awarded at all, even to thorough, exceedingly helpful answers. As I've said a zillion times before. This idea keeps coming up because it's superficially attractive, but like "hey, let's invade Iraq," it has downsides that are visible in advance if you give it a moment's thought.

*bangs head repeatedly on keyboard*
posted by languagehat at 2:35 PM on January 10, 2007 [2 favorites]

ThePinkSuperHero:
....
Damn.

If I ever come up with a counter to that, I'll let you know. But you make a good point.
posted by CitrusFreak12 at 2:35 PM on January 10, 2007

Slimepuppy presents part of the reason I think it's a bad idea. Very often there are threads where the answer that is the most correct isn't marked, but a lesser answer is.
posted by drezdn at 2:36 PM on January 10, 2007

I wish I could mark languagehat's response as best answer.
posted by drezdn at 2:37 PM on January 10, 2007

In a different sort of way, you just did.
posted by IronLizard at 2:39 PM on January 10, 2007

languagehat is right- sometimes best answer just means Answer I Most Want to Hear. Or funniest crack- in a question full of thoughtful, intelligent responses, I awarded this quip "best" answer just because it made me laugh really hard.
posted by ThePinkSuperhero at 2:39 PM on January 10, 2007 [1 favorite]

By the way, did someone declare this 'Beat the Dead Horse Week' and I missed it?
posted by IronLizard at 2:42 PM on January 10, 2007 [1 favorite]

And sometimes the questioner marks one of their own comments as best answer. Anyone for a game?
posted by Kirth Gerson at 2:42 PM on January 10, 2007

Hrm. Actually, TPSH, I would like to see something similar to what I described.

We have "Top 10 Tags on Ask MeFi."
Why not a similar feature, except with categories the user has answered in? You click on the number of answers the user has given, and it has it organized by category, perhaps?

I just think it would be, as I have said, 'neat.'

If you question the usefulness of such a feature, it's no different than the top 10 tags feature. If you can see what a user has been posting questions about, I think you should be able to see what the user has been answering questions about.

Then again, a pony wouldn't hurt either. Just my $0.02.
posted by CitrusFreak12 at 2:43 PM on January 10, 2007

What about the OPs who mark one of their own comments as best? Sometimes they really did find the information on their own and are highlighting it as an actual Best Answer, but often it's to draw attention to an update or clarification they've made.* How could a tally discriminate between those types of "best answers"? Obviously, it can't. And maybe the distinction isn't needed. But it would certainly inflate a user's Best Answer Tally, and then the tally wouldn't be an accurate indicator of the quality of a user's responses.

*Forgive me for being too lazy to find examples to cite.
posted by phatkitten at 2:43 PM on January 10, 2007

All of my answers are best. Tragically, they often go unmarked.
posted by klangklangston at 2:50 PM on January 10, 2007

Just a bit of clarifcation wrt the original question- what I'm primarily interested in is my own "personal stats"- how many best answers I make relative to how many total answers. Although someone might think that comparing my percentage of best answers to others in the community (or better yet, the community at large.... another pony to ask for, perhaps?), it's hard to tell whether 0.3% of my posts, for instance, being best answers is simply terrible (and I need to try to improve) our astoudingly good (and I should keep doing what I'm doing).

With respect to the fact that the definition of best answer varies from asker to asker, that doesn't really concern me. What I'm trying to learn from this is who I'm helping, and how to do that on a more consistent basis. Whether I'm helping because it's truly the best response given, or simply because it made the person feel better about the issue they were asking about, doesn't really matter because both goals to aspire to.

Frankly, I wasn't thinking in terms of "competition", Plutor. The fact that it could be used for that kind of sours me on the whole idea a little.... but I still wouldn't mind seeing it implemented.
posted by Doohickie at 2:59 PM on January 10, 2007

If this were implemented, the only answers I would post to AskMe would be in response to my own questions or those of my sockpuppets. There would be a lot of these, and they would ALL be best answers. nowonmai FTW!
posted by nowonmai at 3:00 PM on January 10, 2007

"Would there be a way to either determine the percentage of a user's responses to AskMe are marked as "Best Answer"...?"

Count the number that are best answers, and then divide that by the total number of their answers. Also, make sure to count any deleted answers.
posted by Eideteker at 3:01 PM on January 10, 2007

I'd like to see best answers go. As noted above, quite often it's used in unexpected ways. It's not reliable as data for anything.

Simplify!

And, we could also make favourites go away too - people could just use their browser system to mark favourites and it'd stop being a quasi-popularity system.
posted by b33j at 3:10 PM on January 10, 2007 [1 favorite]

it's hard to tell whether 0.3% of my posts, for instance, being best answers is simply terrible (and I need to try to improve) our astoudingly good (and I should keep doing what I'm doing).

You should answer as many questions as you feel like as well as you can. If that's what you're doing you should keep it up. If not you should do that. It has nothing to do with how many best answers are thrown your way IMNSHO.
posted by sveskemus at 3:12 PM on January 10, 2007

They let you know what some random poster happened to feel like awarding Best Answer to.

The random poster in this case being the person who asked the question. Which is to say not a random poster at all.
posted by Blazecock Pileon at 3:17 PM on January 10, 2007

the reason we've "been here before" is because it keeps being a good idea.
posted by Milkman Dan

I'm not sure I've ever witnessed logic that screwed up before. Congrats.
posted by justgary at 3:19 PM on January 10, 2007

I look at best answer tags as a sort of "apples to apples" like way of seeing, in thread, what the OP felt was useful to them. This could be a "get your shit together" retort or an echo chamber-y reply, or a wiseass (but still helpful) remark, or nothing at all. I bet there would be a way to make some sort of user-facing count, but I'd be really leery of making that information public just because of the "blah blah it's not a competition" aspect which I still feel is important to having the place function smoothly.
posted by jessamyn (staff) at 3:19 PM on January 10, 2007 [1 favorite]

100 cillit bang

Number 7 best answerer (or something). I'm going to put that on my business cards.
posted by cillit bang at 3:40 PM on January 10, 2007

ick
posted by cortex at 3:40 PM on January 10, 2007

Nope, not ever.
posted by mathowie (staff) at 3:41 PM on January 10, 2007 [1 favorite]

It would be cool if the person who had the most bestest answers could choose a cool prize, like:

-Getting hit in the face with something
-Having Raquel Welch dropped on top of him or her
-Ability to remove one post per day just cuz
posted by Mister_A at 3:45 PM on January 10, 2007 [1 favorite]

Can we have "mark as worst answer?"
posted by desuetude at 3:50 PM on January 10, 2007

NO THEY DON'T. As I've said a zillion times before.

And a zillion times I've restrained myself, but tonight I'm feeling rambunctious. So, a zillion times, you are wrong.

Best answers are a very useful metric depending on how they are used. Yes, the relationship filter, name my cow, mix my CD, type questions are often awarded best answers via a purely subjective decision. Those awards could be argued as useful too, but I won't. At best, the noise level versus meaningful best answer status is high on those questions.

But not all best answers are subjective. A lot of questions have hard, discrete, quantifiable answers. And best answers reflect a solid and objectively useful response. Yes, sometimes the questioner screws it up and gives it to the wrong person. It's happened to me, it's happened to you, it happens. So what? Over time one-off mistakes reduce to low-level noise against "correct" best answers. Bet me that I couldn't randomly sample objective-type questions from AskMe and derive statistically significant correlations with generally recognized best answers. Or better yet, don't, because you would lose.

That's all terribly nice, but what good is it? For me it is a lot of good. I use my ratio of best answer awards to keep me on the straight and narrow. I internally keep a target best answer percentage I shoot for. Doesn't matter what it is and the actual percentage can be different for everybody. What is important is that by following my own selected percentage, I can tell when I'm slacking off on working my answers, or putting up too many goofball responses (within the confines of AskMe rules, of course), or straying too far out of my comfort zone. The best answer count is therefore very valuable to me. Tracking it makes me a better answer person. Isn't that what we all want for AskMe?

Am I the only one who does this? Obviously not, reading this thread alone. It's not a remarkable idea. And the special snowflake theory on MetaFilter has been beat as dead as Lamarckism. If I do it, lots of people here do it. And if I find it valuable, so do other people. There's where you're wrong about best answer counts not meaning anything useful or being a good metric.

Can you game the favorites? Sure, you can do it consciously or by virtue of your writing style. Funny, inspirational, or emotional posts trigger higher best answer rewards. Always gonna be that way, given the social patterns of the naked ape. But that doesn't mean that they are all useless as a measure. It simply means that one must engage one's brain a little bit more or dig a little deeper than atom-thin surface stats. Or look at it this way: Many people here like to talk about how highly intelligent the average MetaFilter user is; if you agree (and you might not, but give us collectively at least average intelligence) why when tracking best answer counts must you think we all immediately devolve to subhuman reflex bundles?

It's not all about the race to win the gold (plated, presumably) trophy at the end so the winner can wave it over his or her head and proclaim to everyone how smart he or she is. We aren't all avatars for ego aggrandizement just because we may wish to count the awards at the end. Sometimes we just want to know for our own purposes, perhaps just to see what makes a "winner" and ponder why.

Now if you want to talk about what's badly broken and is a totally useless metric, you might try out the screwy favorites == bookmarks == popularity contest == people literally complaining on Meta if a post's count goes down. However, Matt Haughy has indicated here that he's fed up with people griping about how favorites work (or don't), so best left alone.
posted by mdevore at 4:08 PM on January 10, 2007 [3 favorites]

Matt/Jess, what about my question? It isn't about counting the number of best answers or anything like that. Just... wondering.

::ducks::
posted by CitrusFreak12 at 4:09 PM on January 10, 2007

Nope, not ever.
posted by mathowie at 3:41 PM PST on January 10

But don't let that stop you all debating it, eh?
posted by timeistight at 4:10 PM on January 10, 2007

Yeah, just hit the button that looks like this: [!]
posted by joshuaconner at 4:12 PM on January 10, 2007

Previous comment was in response to this.
posted by joshuaconner at 4:13 PM on January 10, 2007

Misspell the #1 admin's lastname, Michael, right when you kicked him in the pants. Couldn't have been a simple typo, oh no. Smooth operator you are.
posted by mdevore at 4:13 PM on January 10, 2007

You guys are the best at what you do.
posted by Mister_A at 4:18 PM on January 10, 2007

But not all best answers are subjective. A lot of questions have hard, discrete, quantifiable answers. And best answers reflect a solid and objectively useful response. Yes, sometimes the questioner screws it up and gives it to the wrong person. It's happened to me, it's happened to you, it happens. So what? Over time one-off mistakes reduce to low-level noise against "correct" best answers. Bet me that I couldn't randomly sample objective-type questions from AskMe and derive statistically significant correlations with generally recognized best answers. Or better yet, don't, because you would lose.

People who state their qualifications explicitly get best answer checks at least twice as often as people who leave it up to the asker to figure out. That makes the best answer count a worthless metric.
posted by Chuckles at 4:21 PM on January 10, 2007

Do it on your end. Click on your name after your comment, save the page locally, then write a perl script or something to count the comments, then count the ones with a checkmark.

It's such a good idea, that you should be able to sell your script to other users and PROFIT!!!.

Good excuse to learn perl, in any case.
posted by ctmf at 4:25 PM on January 10, 2007

People who state their qualifications explicitly get best answer checks at least twice as often as people who leave it up to the asker to figure out. That makes the best answer count a worthless metric

You've made a logical leap after an unproven assertion. I don't accept either as accurate in the absence of hard data, and tend to disbelieve them without it. Even were the initial assertion proved, the conclusion does not necessarily follow. Nothing personal, just addressing your argument.
posted by mdevore at 4:31 PM on January 10, 2007

Sorry, but I'm not going to do the legwork for you. It isn't like you've done anything to prove your assertions either, didn't you notice?
posted by Chuckles at 4:35 PM on January 10, 2007

You've made a logical leap after an unproven assertion.
That's got to add at least 0.8 to the degree of difficulty, but the Russian judge will have seen it all before anyway.
posted by Abiezer at 4:41 PM on January 10, 2007 [1 favorite]

We have "Top 10 Tags on Ask MeFi."
Why not a similar feature, except with categories the user has answered in? You click on the number of answers the user has given, and it has it organized by category, perhaps?

I just think it would be, as I have said, 'neat.'

If you question the usefulness of such a feature, it's no different than the top 10 tags feature. If you can see what a user has been posting questions about, I think you should be able to see what the user has been answering questions about.

This is an interesting thought, but why use the categories at all? Both on the green and the blue, I'd be interested to see the top 10 or 20 tags on posts I comment on. I think that could be a remarkably good map of someone's interests.

But again, this is all just HypotheticalFilter. Since MeFi is not primarily supposed to be a social networking doohickey, I don't see much practical application.

Also, what mdevore said. I like favorites a lot. It's a way to pat someone on the back, even after the argument has passed. And it's easier than saying things like, "Also, what mdevore said."
posted by roll truck roll at 5:00 PM on January 10, 2007

Ooh, and think about this, we're all equal but some are more equal than others. If someone favourites my post, and they have been more favourited than favouriting, surely that makes that favouriting more favourable. So, you need a weighted coefficient in there too.
posted by b33j at 5:07 PM on January 10, 2007

Yeah, just hit the button that looks like this: [!]
posted by joshuaconner at 7:12 PM EST

No, I meant for all the world to see, so that we can point and laugh.
posted by desuetude at 5:21 PM on January 10, 2007

1) click on your name after the post
2) click on the number of answers to see their answers
3) save that page
4) create this file as awkfile:

BEGIN {comments=0; best=0}
/comment\ posted/ {comments++}
/bestcheck.gif/ {best++}
END {printf("%d comments, %d best, %3.2f percent success rate.\n", comments, best, best/comments*100);}

5) awk -f awkfile search_comments.mefi

I get:
24 comments, 3 best, 12.50 percent success rate.

Repeat as necessary for each user you're curious about.
posted by ctmf at 5:42 PM on January 10, 2007 [1 favorite]

Note that it only counts the ones before the jump to the next page. So really, it's how you're doing lately.

For doohickie, I get:
50 comments, 0 best, 0.00 percent success rate.

So there you go.
posted by ctmf at 5:50 PM on January 10, 2007

I am very fond of the word 'rambunctious'.
posted by stavrosthewonderchicken at 6:17 PM on January 10, 2007

I think part of the difficulty, especially in the mdevore/languagehat controversy, is the discrepancy between questions that have obvious best answers and questions that don't. The fact is that both kinds of questions exist on Metafilter; and, though questions which don't have obvious answers, especially questions that have difficult or complex answers (which are often called 'subjective'), tend to get deleted as 'open-ended' or 'chatfilter,' there are still a slew of questions for which the answer isn't instantly recognizable. "Where can I buy this widget?" has an instantly recognizable answer. "How can I tell my daughter I'm divorcing her mother?" does not.

Personally, it seems to me that neither languagehat nor mdevore has it exactly right. I mean, I agree with languagehat that, because of the existence of questions with many and difficult answers, 'best answers' really aren't a good metric. (And I have to say, mdevore, that your mechanical approach, while useful if you only answer, say, linux questions, doesn't really work for a huge chunk of the questions asked here.) But I don't think it's utterly unuseful.

I say all this because I support the continued existence of the "mark as best answer" option. If I may make a suggestion: keep the "mark as best option" for categories like 'travel & transportation,' 'law & government,' and 'technology,' but get rid of them completely for categories like 'society & culture,' 'religion & philosophy,' and 'human relations.'
posted by koeselitz at 6:19 PM on January 10, 2007

The problem is confounded by the asker being the only judge of the "best" answer. Sometimes the right answer isn't marked as the best (where such can be objectively determined), but the one the querant wants to see (as others have remarked). A quick browse through medical questions often shows this to be true. Harmful advice even, has been marked as "best", while the best advice (usually "See your doctor!") is set aside.

When you don't know enough about a subject to know the answer to your question, how can you know enough about the same subject to judge which answer is best? That's the dilemma of AskMe. Sometimes it works, but often times it doesn't.
posted by bonehead at 6:32 PM on January 10, 2007

plz vote for me
posted by loiseau at 7:47 PM on January 10, 2007

What I'm trying to learn from this is who I'm helping, and how to do that on a more consistent basis.

If only all of one's answers were collected in one place together, and if only there was some sort of indication in that one place of whether any given answer were marked 'best' or not, and if one could keep track of the answers marked 'best'—maybe count is the word best represents what I'm imagining here—and then if all that could happen and at the same time you could click on the indicator for an answer and find out who posted the question, then maybe—just maybe if you use your browser's back button right and if you could muster some critical thinking time from the great gray mush above and slightly to the back and put it all together in perfect synchronicity—maybe maybe you could learn about who you're helping, and how to do that on a more consistent basis.
posted by carsonb at 8:10 PM on January 10, 2007

280 people marked their own posts as best 280 times during 2004-2006.

37 people have done it more than once.

delmoi and geoff. are tied for top self-bester at 6 times a peice.
posted by tkolar at 8:29 PM on January 10, 2007

Apparently jessamyn has the biggest penis.

Yeah, we knew that. It's magnificent. Utterly infalliable. Not to mention hot and outright turgid.

That's not the problem. The problem is that she won't share it with us and let us adore it and worship it as it so rightly deserves.
posted by loquacious at 8:39 PM on January 10, 2007

Do me! Do me!

*Not addressed to jessamyn's huge penis*
posted by Alvy Ampersand at 8:53 PM on January 10, 2007

I wrote...
280 people marked their own posts as best 280 times during 2004-2006.

Whups, that should have been "137 people marked their own posts as best for a total of 280 times during 2004-2006"
posted by tkolar at 8:55 PM on January 10, 2007

"The problem is that she won't share it with us ..."

No, the problem is that it's stuck in a chicken.
posted by mr_crash_davis at 8:55 PM on January 10, 2007

Don't look at me.
posted by stavrosthewonderchicken at 9:06 PM on January 10, 2007 [1 favorite]

Doohickie wrote...
it's hard to tell whether 0.3% of my posts, for instance, being best answers is simply terrible (and I need to try to improve) our astoudingly good (and I should keep doing what I'm doing).


53 people have 100% best answer rates. You've never heard of any of them, because 51 of them have only ever answered one question, and the other 2 have answered 2.

You, personally, Doohickie have 17 bests for 1023 total answers in 2004-2006, for an overall rate of 1.6%. Clearly you will never catch up with the 53 allstars. Perhaps if you opened a new account each time you answered a question you could aspire to a higher percentage.

Still, if you're seeking inspiration, here's the top five users with more than 200 comments by best post percentages:

(percentage, user, total comments)
12.00 eriko 364
12.00 paulsc 1513
12.00 chrismear 538
15.00 RichardP 348
25.00 mdevore223
posted by tkolar at 9:19 PM on January 10, 2007 [1 favorite]

Nobody likes my pony idea.
Eh whatever.

53 people have 100% best answer rates. You've never heard of any of them, because 51 of them have only ever answered one question, and the other 2 have answered 2.

That's the problem with some of this data. For instance, the Contribution Index Thingamabob. I'm #2. Wtf? What does that mean? Nothin'.
posted by CitrusFreak12 at 9:25 PM on January 10, 2007

Oh, and because I'm feeling mischievious I'll also point out that languagehat's percentage is 3.7 percent for the same period, which might explain why he doesn't find it as useful a feedback mechanism as mdevore.
posted by tkolar at 9:26 PM on January 10, 2007

53 people have 100% best answer rates. You've never heard of any of them, because 51 of them have only ever answered one question, and the other 2 have answered 2.

Maybe a cutoff is useful, for reducing outliers?
posted by Blazecock Pileon at 9:37 PM on January 10, 2007

I've got ~80 BA's, but keep in mind that several hundred posts were before BA's were implemented. Go me.
posted by Civil_Disobedient at 9:38 PM on January 10, 2007

Somebody said I had a good answer once.
posted by loiseau at 11:17 PM on January 10, 2007

My penis is double your foot size.

(I'm on 5%, woo!)
posted by liquorice at 11:28 PM on January 10, 2007

I'm at about 10% with 185 answers. I'll really have to make sure those next 15 answers are really good, and only answer confirmed "best answer" givers, and stop commenting in my own questions, and and not participate in "the purification" type threads, or relationship questions for that matter, if I'm going to make that completely arbitrary top 5!
posted by team lowkey at 12:38 AM on January 11, 2007

delmoi and geoff. are tied for top self-bester at 6 times a peice.

Every time you self-best, God kills a kitten.
posted by amyms at 1:13 AM on January 11, 2007

"It would be cool if the person who had the most bestest answers could choose a cool prize, like:

-Getting hit in the face with something
-Having Raquel Welch dropped on top of him or her
-Ability to remove one post per day just cuz"
posted by Mister_A at 6:45 PM EST on January 10

If employees of the site are ineligible to win prizes, I'll take prize #2 (twice, if it's an annual award, as the generated stats are for a 2 years span), please. See my profile for drop coordinates. I'll be standin' on my patio.

Y'all take AskMe way too seriously.
posted by paulsc at 1:46 AM on January 11, 2007

But I don't think it's utterly unuseful. I say all this because I support the continued existence of the "mark as best answer" option.

I don't think it's utterly unuseful either, and I certainly support its continued existence. I don't support its use as some kind of metric of "who's best at AskMe," still less its use to determine who should be allowed to ask more questions (as some goofballs have suggested).

languagehat's percentage is 3.7 percent for the same period, which might explain why he doesn't find it as useful a feedback mechanism as mdevore.

Nice try, but I have never tried to find out what my "percentage" is, nor do I have the least interest in the subject. I was mildly surprised to see that I was in the top ten of Best Answerers, but I don't draw any conclusions from it except that 93 people liked my answers to their questions well enough to bother clicking the doohickie. Best Answers is what it is, it's nice to have your answers checked and it's nice to check others, but it is not a measure of anything useful, and I wish people would stop pretending it is.
posted by languagehat at 5:52 AM on January 11, 2007

Apparently jessamyn has the biggest penis.

Yeah, but it is circumcised, so she can't experience the full pleasure of sex and needs to resort to the kinky stuff.
posted by Meatbomb at 5:55 AM on January 11, 2007

b33j writes "And, we could also make favourites go away too - people could just use their browser system to mark favourites and it'd stop being a quasi-popularity system."

Favourites on comments I wouldn't mind seeing go away (I'd prefer they stay but I wouldn't be sad if they left) but favourites on posts should stick around. It's a valuable feedback mechanism for some of the best posting here (IE: links to good stuff that doesn't garner a lot of discussion).

CitrusFreak12 writes "That's the problem with some of this data. For instance, the Contribution Index Thingamabob. I'm #2. Wtf? What does that mean? Nothin'."

It means for a noob your fairly prolific and you haven't went on holiday since signing up.
posted by Mitheral at 7:02 AM on January 11, 2007

I have submitted questions in which I appreciate all the responses given (I tend to appreciate that one makes an effort, even if sometimes only marginally helpful or just good for a laugh), and as a result did not want to select any particular responses as 'best answer.' And marking them all as best seems a bit silly.

But I think all the people who answer my questions are the best.
posted by troybob at 7:39 AM on January 11, 2007

paulsc writes...
(twice, if it's an annual award, as the generated stats are for a 2 years span),

If you're referring to my recent statistical effluvium, it's actually 3 years. 2004, 2005, and 2006.
posted by tkolar at 9:15 AM on January 11, 2007

Mitheral: It means for a noob your fairly prolific and you haven't went on holiday since signing up.

I think the main problem is that I've been on holiday since I've signed up. I've had very little else to do!

I expect more of my time to be taken up by school when I go back. I'm just not looking forward to detoxing...

PS
troybob: Are you by any chance the same troybob from the Desert Combat forums?
posted by CitrusFreak12 at 9:47 AM on January 11, 2007

Just for the record, not that matters, my personal target is a hell of lot less than 25%. It's worked out that way because I haven't had much free time lately to actually go in there and push it down. No doubt I'll fix that in the near future with a post barrage of glaring stupidity and drooling idiocy. Nor do I think even a 1% figure is low, especially if one takes on relationship questions.

That assumes one concerns oneself with the metric at all. Simply because I say it isn't useless doesn't mean it universally applies either. If one thinks they are answering well, throttling their output to reach an artificial ratio of best answer count is a simply terrible, and damaging, idea.

Plus, I think one or two (or three or something) questions I got a best answer when maybe I was coasting. Not that I'm suggesting a retraction. No way, man! I might lose my top (meaningless statistic) position.
posted by mdevore at 11:32 AM on January 11, 2007

Metafilter: a post barrage of glaring stupidity and drooling idiocy

shut up, I only get one a year
posted by cortex at 11:43 AM on January 11, 2007

Oh damn! I posted three times in this guidance on san francisco thread and the guy didnt even award best answers!!

What a waste. I should just wait until someone asks for ID on a book/movie I know AND I can swoop in before all the other best-answer-grubbers.
posted by vacapinta at 9:16 PM on January 11, 2007

I don't draw any conclusions from it except that 93 people liked my answers to their questions well enough to bother clicking the doohickie.

I've noticed that these days more and more of life comes down to clicking the doohickie. The great responsibilities of life are increasingly about choosing the right doohickies to click. I'm not sure how I feel about this.
posted by stavrosthewonderchicken at 12:15 AM on January 12, 2007

resist the urge towards a button marked life.
posted by carsonb at 6:20 AM on January 12, 2007

Oh yeah, and of the 9856 people who answered questions in 2004-2006, 5320 never had one of the answers marked as "best". Ha hah! Suck it Losers!

*cries*
posted by Penks at 10:34 AM on January 12, 2007

My really best answers - the ones where I've climbed onto a pinnacle of insight previously unscaled by mortal man - invariably don't get marked.

The ones with the check are the ones that made someone on the other end of the internets tubes feel like a better person. That's all well and good - God love you, you 141 people - but I prefer a purer metric.
posted by ikkyu2 at 2:21 AM on January 13, 2007