Calculating helpfulness April 26, 2005 9:58 AM   Subscribe

We know who posts the most, who comments the most, etc. Do we know who the most helpful users of AskMe are (as expressed by the ratio [questions/answers])?
posted by stupidsexyFlanders to Etiquette/Policy at 9:58 AM (52 comments total)

As illustration, jessamyn's ratio is .0115. So without slagging those with middling ratios (me, .1458) who's in the AskMe Hall of Fame?
posted by stupidsexyFlanders at 9:59 AM on April 26, 2005


That doesn't actually tell you how helpful someone's been.
posted by kenko at 10:10 AM on April 26, 2005


It's not a contest.
posted by Witty at 10:12 AM on April 26, 2005


Witty loses!
posted by Floydd at 10:14 AM on April 26, 2005


God doesn't think so.
posted by Witty at 10:16 AM on April 26, 2005


you'd have to break it down by subject i think.
posted by amberglow at 10:24 AM on April 26, 2005


kenko's right, "helpful" usually just means "how many questions do you answer, and/or how useful are those answers", completely independent of how many questions you ask.
posted by cyrusdogstar at 10:39 AM on April 26, 2005


So, if you've never asked a question but you've answered a bunch (like me), is the ratio zero? If you've only asked and never answered you'd have the null set?

I'd agree with cyrus and kenko though; you could determine "helpfulness" without even factoring in the number of questions asked.
posted by LionIndex at 10:42 AM on April 26, 2005


Mine: 11/167 = 6.59%

To calculate yours:
  • Go to your user page, click on the link "N answers to Ask MetaFilter".
  • Exploit a Mefi bug: by clicking on the link for the last page, you'll get a page that shows all your comments to ask Mefi. Comments marked as "Best Answer" will have a checkmark after them.
  • Save that page in your browser.
  • grep the saved page for the checkmark, then use wc to count the matching lines: grep bestcheck.gif search_comments.mefi.htm | wc
  • The first number given is the number of "best answers". Divide this number by the total number of answers you've given. Example: 11 / 167 = 0.065868263
  • Multiply by 100 to get the percentage. Example: 0.065868263 * 100 = 6.59%.
posted by orthogonality at 10:43 AM on April 26, 2005


I'd also tend to argue against the idea that questioners are somehow unhelpful. The people who ask the questions are the ones who get things started.

I could see some kind of ratio of questions answered to answers marked best, I suppose, but I'm not sure what purpose it would really serve.
posted by jacquilynne at 10:45 AM on April 26, 2005


I hope someone automates orthogonality's approach, so that my method of systematically adding "bestcheck.gif" to the end of every answer finally pays off.

bestcheck.gif
posted by Plutor at 10:48 AM on April 26, 2005


Plutor writes "I hope someone automates orthogonality's approach, so that my method of systematically adding 'bestcheck.gif' to the end of every answer finally pays off.
posted by orthogonality at 10:54 AM on April 26, 2005


LionIndex writes "So, if you've never asked a question but you've answered a bunch (like me), is the ratio zero? If you've only asked and never answered you'd have the null set?"

The ratio isn't answers/questions, it's answers marked "best answer"/questions. Where it the first, your ratio would be +Infinity. In truth, your ratio is 2/165 = 1.21%.
posted by orthogonality at 10:59 AM on April 26, 2005


Er, that is, it's answers marked "best answer"/all answers. Sorry.
posted by orthogonality at 11:00 AM on April 26, 2005


I'd also tend to argue against the idea that questioners are somehow unhelpful.

Asking other people to help doesn't count as "helpful." It's not unhelpful, but it's not helpful either.
posted by kindall at 11:06 AM on April 26, 2005


I just want to say that the need to quantify everything and to take complex things and flatten them onto a linear scale, still fascinates me. Maybe because its an urge I dont understand. I've always resisted it because I saw it as a way of losing information not gaining it.
posted by vacapinta at 11:07 AM on April 26, 2005


And I misread the original question; LionIndex is correct, I'm not. And that means my ratio, of questions to answers, is 2.38%.

Ironically, the ratio mentioned by the post's author, the ratio I misread, that ratio is a metric I first proposed, here.

Sorry for muddying the waters.
posted by orthogonality at 11:09 AM on April 26, 2005


I would like to suggest that those of you who find this sort of exercise to be meaningful follow baseball. There is a much longer tradition of this sort of questionable use of arithmetic in that context.
posted by anapestic at 11:11 AM on April 26, 2005


Asking other people to help can be very helpful. Ever taught a class where no one wants to ask the question everyone's wondering about because they're all afraid to sound stupid?
posted by transona5 at 11:11 AM on April 26, 2005


God doesn't think so.
posted by Witty at 10:16 AM PST


Oh man that is hilarious!
posted by tr33hggr at 11:22 AM on April 26, 2005


I can't wait until we finally know who the bestest answerer is!

I agree with vacapinta. This idea has been proposed at least three times in MeTa before, and the conversation never seems to end well.
posted by odinsdream at 11:25 AM on April 26, 2005


And I misread the original question; LionIndex is correct, I'm not.

Sure, but that's just me reading the question to the letter, and making sure of what we're looking for. Going by best answers seems like a better way to get at what Flanders wants. Otherwise someone who's done nothing but answer one question can have the same ratio as a (theoretical) person that gets marked for best answer on every question ever posted to AxMe. Not that it matters in the end.

And I love baseball.
posted by LionIndex at 11:35 AM on April 26, 2005


I love baseball, too, though my enjoyment of it is not increased by calculating the ratio of hits with batters in scoring position to walks achieved by being struck by a left-handed pitcher. Also, in baseball, the stats have some chance of being more meaningful, and are based on more reliable information. The statistic being proposed here probably correlates very loosely with how helpful people are and seems more like yet another variation on bragging about penis size.
posted by anapestic at 11:41 AM on April 26, 2005


baseball stats make sense within the game because the purpose of the game is to increase another sort of statistic, the number of runs. Statistics in something like askme are nowhere near as useful because the goal is much less specific and quantifiable.

How many questions you 'answer' does not mean you provided useful answers; some of the questions are just asking for general opinions anyway. Even getting the tick mark at the end of your answer is a little arbitrary - some questioners don't use it, some use it on practically everyone, some don't notice a comment they would have marked, some choose answers that other users would not have chosen...

accept the muddiness! It's okay not to know your precise rank.
posted by mdn at 12:12 PM on April 26, 2005


hay guys look how big me ePenis is

8=======================>
posted by cmonkey at 12:14 PM on April 26, 2005


Long and thin and ballsy and floppy / the ePenis of cmonkey is pointy / and when it points each one it points says, "oh!"
posted by kenko at 12:19 PM on April 26, 2005


A few thoughts -

Of course this ratio has a loose correlation with "helpful." But AskMe is pretty clean -- there's not much posted that's just noise, abuse, or wankery. People who respond to questions are giving of themselves, even if it's not much. So maybe "helpful" is too strong a word for what you are if you attempt answers. But it's not a terribly wrong word, either.

I thought I might get some flak about comparing answers to questions, as oppposed to just tallying people's numbers of answers. I agree that questions are a contribution, too, and the site needs them. They just seem to come from a less altruistic place than answers (or "responses.") Surely we can agree that there's some average contribution ratio between 1.000 and .0000 above which the site would be too question-heavy, and below which AskMe feels rich and interesting. That doesn't make jessamyn a saint for asking few questions, or me the devil for asking a lot.

Finally, I don't mean for this to be a litmus test of people's fitness to participate in AskMe. There's no MetaFloe onto which people should be pushed out to sea if they ask too much relative to what they answer. If there were, I'd definitely be on it. I just thought it would be a fun exercise that would reveal the top contributors to the gold mine of AskMe knowledge, just as a listing of the top commenters or FPP'ers is a fun exercise, even though we all know some of those users would improve the site by posting less, and some others not on the list would make MeFi better by posting more.

So to summarize: It's only fun. Don't be afraid.
posted by stupidsexyFlanders at 12:28 PM on April 26, 2005


If we have to do the math for you, no.
posted by thomcatspike at 12:51 PM on April 26, 2005


Will you parse my math if I parse your english?
posted by stupidsexyFlanders at 1:01 PM on April 26, 2005


The ticked answers as noted are arbitrary yes. And the thing you can't really find out is what the rest of the community thinks. Often there's a cumulative wisdom that shines through whether an answer is ticked or not. And then there are people (other than the original questioner) who walk away with their own sifted 'answer'. So it ends up very abstract in terms of making any judgment calls.
And I agree with the spiked cat.
And on preview...heh
posted by peacay at 1:03 PM on April 26, 2005


dumb idea.
posted by angry modem at 1:20 PM on April 26, 2005


stupid, you’re kidding. Because if your are asking for an index to be made like the one that exits for posts and commenting, your posted statement needs bettering. You won’t find the true answer as you state it. Because every member will have to participate which will not happen. Also, I don't think you are going to do each member yourself since it would be a big workload finding every member’s ranking position yourself.
If you want a pony just ask. Plus the only purpose it would serve, "I'm smarter than you."
posted by thomcatspike at 1:47 PM on April 26, 2005


Asking other people to help doesn't count as "helpful." It's not unhelpful, but it's not helpful either.

What I was trying to convey was that asking questions shouldn't be considered negatively helpful. Your helpfulness rating, should such a thing exist, shouldn't be reduced because you also required help.
posted by jacquilynne at 1:52 PM on April 26, 2005


I apologize for my last comment, that was thoughtless.
posted by stupidsexyFlanders at 2:02 PM on April 26, 2005


I should have just have laughed.
The answer you might be looking for is an idea in an earlier thread.
posted by thomcatspike at 3:52 PM on April 26, 2005


Please take into account when producing your "best of list" that invariably, I'll try and answer a question, hit the post button, and then realise that there is something that I've missed. As a consequence all my answers are over two
posted by seanyboy at 4:19 PM on April 26, 2005


comments.,
posted by seanyboy at 4:20 PM on April 26, 2005


Looking at this it seems maybe there should be a quota on answers and no quota on questions. People would get a limited number of answers unless their answer is marked as the best one. That way we can punish the unhelpful ones with censorship and establish a strong feedback loop against the quality of questions. (Nobody would waste their precious answers on all the sub-par, 'help me find my long lost daughter' posts.) Eventually AskMe would evolve into a highly regimented dialog between the neediest and most precise askers and the most experienced and driven answerers.
posted by nixerman at 4:24 PM on April 26, 2005


nice illustration of your point seanyboy!

and im a zero if you only count best answers, actually either way im a zero. :(

Im also a bad questioner (only 1 question that ended up being deleted.)
posted by schyler523 at 4:32 PM on April 26, 2005


I FUCKING TOLD YOU THIS WOULD HAPPEN.

Goddamn popularity contest.

I'd just like to point out that the Best Answer feature was implemented looong after much of my 1000+ answers.
posted by Civil_Disobedient at 6:09 PM on April 26, 2005


To save you all the bother...it's me.

Now go and do something more interesting instead.

Not that it's a competition or anything...
posted by i_cola at 6:32 PM on April 26, 2005


so the question is: "What's the metric for helpful?"

*thinks hard... gives up*
posted by warbaby at 7:05 PM on April 26, 2005


Like if your question isn't the best answer, but makes it possible for somebody else to come up with it?

*goes back to thinking some more*
posted by warbaby at 7:07 PM on April 26, 2005


Or, many a time I've seen this, along warbaby's logic: The chosen "best" answer is not, IMHO, the best at all. Or, because life can be so complex, there's more than best answer possible, but only one marked. Or someone's finger slipped an is too lazy to e-mail mathowie. That last one is probably the most likely, but who really knows?
posted by SeizeTheDay at 7:51 PM on April 26, 2005


nixerman writes "Eventually AskMe would evolve into a highly regimented dialog between the neediest and most precise askers and the most experienced and driven answerers."

And the trains would run on time.

(We could use them to transport the bad answerers for "Resettlement in the East.")
posted by orthogonality at 8:43 PM on April 26, 2005


Nice Godwin there.
posted by stupidsexyFlanders at 9:19 PM on April 26, 2005


It's only fun. Don't be afraid.

Er, Calvinball aside, sometimes fun's more fun when the rules make sense. Yours don't.

That doesn't make jessamyn a saint for asking few questions, or me the devil for asking a lot.

You still haven't addressed the fact that someone who's asked *zero* questions while supplying endless "best answers" will have a ratio of, er, 0. How is that a helpful measurement of helpfulness again?

orthogonality: it's answers marked "best answer"/all answers

And how would questioners who prefer not to mark a "best answer" for fear it would limit further discussion enter your equation? Sorry, numerologists; you're just being silly in thinking you can quantify this in even a slightly meaningful way.
posted by mediareport at 9:47 PM on April 26, 2005


You still haven't addressed the fact that someone who's asked *zero* questions while supplying endless "best answers" will have a ratio of, er, 0.

I think I have. The idea is that the lower the ratio, the more answers the person had contributed, relative to the answers requested. That person (with a ratio of 0) has asked the least of the community, while giving the most. (assuming some minimum participation level, of say 25 Qs+As) Why is that hard to understand?

No, answers don't tightly correlate with helpfulness, but if person A does nothing but read questions and ignore them, and person B tries to chip in his expertise when applicable, who's more helpful? No, it's not a bad thing to ask questions, but if person A does nothing but ask questions, and person B does nothing but answer them, who's more helpful?

Look, in the past people have found it interesting to see a compilation of the top 50 commenters in the blue. I thought this would be at least as interesting and/or meaningful. I would just do it, but I don't know how to hook into the site to extract the data.

It's puzzling (though entertaining) to see people's fears that this would lead to outcomes ranging from people's feelings being hurt, to genocide. Which is quite a range.
posted by stupidsexyFlanders at 3:54 AM on April 27, 2005


That person (with a ratio of 0) has asked the least of the community

Just strikes me as odd that someone with a zero rating would be your ideal contributor. But given all the valid points raised in the thread - most of which you happily acknowledge - I really can't understand how you think this would be meaningful. And I'm not sure I've seen many "hurt" feelings here, but I have seen lots of folks thinking this over and coming up with the idea that it would be a silly exercise. I mean, you're the one who proposed a "most helpful user" metric, and then a few screens down added, "Of course, this ratio has a loose correlation with 'helpful.'"

Forgive me, I guess, for being confused about your point.
posted by mediareport at 6:14 AM on April 27, 2005


That person (with a ratio of 0) has asked the least of the community, while giving the most. (assuming some minimum participation level, of say 25 Qs+As) Why is that hard to understand?
Because that was the first time you added a minimum participation level to the equation, perhaps? Otherwise, someone who has neither asked nor answered would also have a "perfect" rating of 0.
posted by dg at 6:26 AM on April 27, 2005


Just strikes me as odd that someone with a zero rating would be your ideal contributor.

No odder than that someone with a zero ERA would be the ideal pitcher.
posted by languagehat at 6:47 AM on April 27, 2005


Oh, is that how ERA works?

*shrugs*

This is still a stupid metric, though.
posted by mediareport at 3:00 PM on April 27, 2005


« Older waste of valuable front page AskMe space   |   What a (wedding) day Newer »

You are not logged in, either login or create an account to post comments