What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 134 Stack Exchange communities.

The responses from this question:

What has happened to the sorting of answers on Stack Overflow?

Seem to indicate disagreement with this change. Please return the sorting system to the way it was before based on time instead of randomly sorting answers with the same number of votes.

Edit:

In my typical meta fashion I have applied a ridiculous bounty to this question.


Response to Podcast #66:

Codexon's blog post

Mentioning codexon's blog post about how to game reputation by posting a scaffold answer ("peanut butter and jelly" was the example) and then copying a correct one to fill the spot so that your answer appears on top is a great reason to revert to the old method. Now, you don't even have to post first to game this way. Since the sorting will be random now, just find a decent answer with 0 votes and copy/paste it as your own, and now you can appear on top as if you were copied. With the time sorting method, copying another answer would push the duplicate to the bottom.

Every answer should be judged on its own merits

This was never an issue, answers have always been judged by their merits. A great example is my own answer to a very simple question, which was basically just a collection of other answers with a benchmark added. It came in over an hour after the accepted answer, but it has been voted up higher. It wasn't magically "unread" because it started off lower in the answer order.

Also, some questions just don't deserve an answer with your definition of quality. A lot of jQuery questions, for example, are basically "which selector should I use". How much quality could you possibly want for questions like that, copying the jQuery documentation? If someone is actually willing to make a summary of the docs, they will get upvoted anyway.

I want people to read all the answers

Huh? How does this change by sorting them randomly? If a user would stop reading beyond the first post, what is going to make them start now? Again, this new sorting system only makes it worse! If they just vote for the first post they see (which may or may not be a duplicate), there is now incentive to leave copied answers (instead of delete them and upvote the first one) because your answer might get lucky enough to be on top.

Of course you could edit your post and make it better, but again, it would be upvoted similarly using the old sort order. Again, it also may not be worth a tremendous effort depending on how simple the question is.

There is an incredible amount of angst and hand-wringing about this new sort order

This is blatantly incorrect. Nowhere in this thread is anyone truly upset about the change, we are merely expressing disagreement with it. We're trying to help benefit the site and the community, not scream and whine about the change. Personally, I couldn't care less which sort order you ultimately use, I'm just explaining which way I think is better.

Random is better ... in surveys

StackOverflow is not a survey site, it is a question/answer site.

share|improve this question
2  
it is handy when the dupe is identified in the question. –  EBGreen Aug 26 '09 at 18:44
2  
@Manni: You are incorrect. Everyone here has the ability to go and make their opinion clear there. –  GEOCHET Aug 26 '09 at 18:50
2  
@Manni - did you read those responses? "I find this to be horrible decision", "basically I hate this change", "I'm not sure how I feel about this", "this change is a huge mistake", "This is a terrible decision" - of course there are others that aren't adamantly against it, and only like 1 or 2 who believe it is a positive addition. –  John Rasch Aug 26 '09 at 18:53
1  
@John: Not to mention the very valid arguments against it with logical reasoning presented that no one can address. –  GEOCHET Aug 26 '09 at 18:54
2  
@Rich B, John, he means it doesn't represent every single SO user's opinion which is true until all 67k (or whatever) of them respond. –  hyperslug Aug 26 '09 at 18:58
6  
It's just a matter of which vocal minority you want to listen too. I doubt that 90%+ of users even give a crap. –  EBGreen Aug 26 '09 at 18:58
3  
@hyperslug: Right, and the people complaining against FGITW which caused the change did not represent the 67k of users either. –  GEOCHET Aug 26 '09 at 18:59
2  
@Manni: We are saying so. That is what this thread is about. You are the only one who seems to have an issue with it. –  GEOCHET Aug 26 '09 at 19:07
1  
(-1) I like it just fine the way it is now. I liked it fine the way it was too, I just want to wait a week and see how it actually plays out in terms of behaviour. –  devinb Aug 26 '09 at 19:25
1  
@Manni: I understand you completely. You are being pedantic to be an ass. The claim is correct. There is overall dissent to this change. –  GEOCHET Aug 26 '09 at 19:25
4  
Either all discussion on MSO is pointless because it doesn't include everyone, everywhere and we should just go home... or there's some value in it and we should state our arguments. Pick one. –  ٠٠ Aug 26 '09 at 19:31
2  
@Manni: He did. It just flew far over your head. –  GEOCHET Aug 26 '09 at 19:36
4  
@hyperslug: i've yet to find anywhere on MSO where the "randomize answer order" suggestion has been positively received, and yet it was implemented anyway. So there's an argument in favor of "pointless, go home"... or possibly, "game the new system and then post your writeup on Reddit." But, if we're gonna discuss it anyway, then no sense in arguing about whether it's worthwhile or not, eh? –  Shogging through the snow Aug 26 '09 at 19:45
2  
@hyperslug: do the folks unwilling to fill out a ballot and drop it in the box deserve a vote? –  Shogging through the snow Aug 26 '09 at 19:51
6  
No worries - it's the old conflict between polling and voting: in the former, you track down and question a sample, then theorize that the opinion of the majority... in the latter, you require participants to show up and indicate their opinion. –  Shogging through the snow Aug 26 '09 at 19:54

12 Answers 12

up vote 18 down vote accepted
+350

This should be a comment to Jeff's response but it doesn't fit in the comment space so I had to make it an answer:

I find the resistance to the new votes sort ordering from a vocal minority very perplexing.

You made an implicit assumption that this is the "minority" and the other way (apparently how reddit guys think) is the "majority." There's no way to prove whether it's true or not in general but as Rich B pointed out in a comment to this question, it doesn't really matter. Everyone is free to express his or her opinion on Meta SO and apparently, the "majority" of the "vocal minority," who represent people that care about SO are against this change. Next, you targeted an implicit ad-hominem attack on people expressing their opinions here. Apparently, you are accusing them of "whining" because "they are at a disadvantage." Most of these guys are high profile users on SO (cletus, Rich B, tvanfosson, TheTXI, ...) who have contributed heavily to the sites. This proves they really care about your Website. I'm sure this whining is purely a result of they caring. I find this implicit accusation on the borderline of being insulting to the high profile users of your site. It's like saying "Shut up or you're one of "them"."

Note that it's not resistance. We didn't boycott StackOverflow. We just expressed our opinions.

To clarify, due to daily reputation cap, it doesn't make any real difference (rep-wise) for most of the users who've "whined." They'll achieve their cap easily regardless of the system in place because they provide plenty of good answers.

First, to clarify: this has nothing to do with tactical downvoting. It is not meant to address that issue at all. Tactical downvoting is the same as it ever was.

I'm not saying there hasn't been any kind of tactical downvoting before. What I'm saying is that it increases the benefit of tactical downvoting. In this system, if someone with a later answer downvotes you, he has much more chance to get upvoted.

Beside that, it now makes downvoting duplicates an ethical and good practice. I, for one, think you should downvote duplicate answers that come later (on all posts, not only your posts.) This is not inherently a bad thing. However, it encourages "revenge votes." You downvote a duplicate answer and he'll get angry and will downvote you.

It's purely a fix for the votes sort order. The answers are sorted by votes. But what sort order do you use when every question has the same score? We had to pick something, so we picked LastActivityDate. I was never happy with this choice (it wasn't even really a conscious choice, honestly), since it had a side effect -- we were implicitly rewarding behavior other than composing answers worthy of upvotes!

It was the natural choice. It was so natural you made that decision unconsciously. You were quite happy with it. Listen to your podcasts. Actually, I think (I'm not sure, though) the new decision is highly influenced by that reddit thread. It was the right thing to encourage quick answers and this was considered "by-design" in SO. If anything, between two identical answers, the older one should be rewarded the most.

Now answers with the same vote score are in random order. Which means answers can correctly be judged based on their merit as answers, not by who happens to end up first in the accidental secondary sort order we didn't even intend to be relevant in the first place.

First, it's not accidental. It's an indirect results of many parameters like when you've seen the question and how much you know about the topic. Obviously, Jon Skeet doesn't need to search about a question about say, a switch statement but some other guy might just Google stuff and copy paste there. Second, the OP wants a quick answer as fast as possible. There is a distinction between a "half-complete" answer and a "wrong" one. I believe posting a "half-complete" answer as soon as possible directly helped achieving one of the hallmarks of StackOverflow, which is "getting answer immediately." Third, this is really the behavior we've had before. If the answers were virtually identical, the fastest one would get upvotes and if the answers were not identical, the better one would eventually float to top.

To the extent that people have "optimized" strategies around that behavior, it's completely a side-effect. A distraction.

Another thing mentioned in the other thread is that people are used to the old behavior. They think the answer floating on top is the first one and upvote accordingly. Even if this decision was a good one, it's significant enough that it should have been mentioned explicitly.

Bottom line: if you want to get upvotes, WRITE A GREAT ANSWER.

This has been happening ALL THE TIME. SO is full of great answers at the top of questions. And by the way, if you really believe in this decision, you should really consider stop docking the accepted answer to the question. If the best answer, quality-wise, is supposed to be on top, let it be the most upvoted. (This is just a side note. It's not a part of my argument.)

(comment) "There is no drawback to posting late?" Not so. The earlier you post, the more opportunity for people to VOTE for your answer. If you wait, you're losing potential votes. – Jeff Atwood♦

Exactly, there's more opportunity to VOTE for your answer. Vote is not equal to "upvote." Let's not forget this vote can be both "upvote" and "downvote." Let's not forget that in the "fastest gun" version, if you posted a crap sooner than anyone, you would have been voted down to the oblivion. It's not always been a good thing.

Bottom line: You've not addressed the SCITE problem; the fundamental problem with this issue. So far, your answer was effectively like saying "Hey, I can do it so I did it." IMO, This is a valid response but if you want to use it, you should be open about it. It's not possible to say "StackOverflow is run by You" and this statement at the same time. Definitely, we don't want to deal with "propaganda" on StackOverflow. You can choose to come back to the spirit of StackOverflow (either by providing a good solution for SCITE or reverting to the old system) or jump the shark and be more like "hyphen." It's all up to you. :)

share|improve this answer
11  
I can't take this rant seriously. –  Jeff Atwood Aug 27 '09 at 4:28
4  
nor can I... –  RSolberg Aug 27 '09 at 5:18
9  
@Mehrdad: Very good write up. Shame on @wood and his lap dog for not reading it and considering it. You summed up the dissent very well. –  GEOCHET Aug 27 '09 at 11:43
22  
@Jeff - I for one am incredibly disappointed at your attitude towards this post. The fact that you dismiss Mehrdad's commentary as a "rant" feels very childish. I understand that this is your site and you can do what you wish. I am not calling that into question. I only hope to caution you that your site would be worthless without users like Mehrdad who consistently hold a strong presence on SO. I think he has earned the right to at least have his opinion considered. –  Andrew Hare Aug 27 '09 at 13:29
1  
For the record, I didn't think this was a rant at all. Jeff... I really looked up to you as a great community leader, and an exemplar of how to manage community. Please reconsider how you handle this situation, and other forms of "dissent" as well. –  kanamekun Aug 27 '09 at 23:32
2  
opinions stated thoughtfully and with data and examples backing them up will be considered. People yelling in my face .. well, I gotta be honest with you -- it only steels my resolve. So in that sense, what you're doing here is not only not effective, it's actively working against you. Just so you know. –  Jeff Atwood Aug 28 '09 at 7:25
    
4  
@Jeff: This isn't yelling in your face. He used bold to emphasize points, but in no way is it yelling. This is a reasoned argument, through and through. Moreover, it uses your logic to make the argument, which is ironic in its own right. I implore you to read Mehrdad's arguments. He is not hot-headed nor out of line. He's an intelligent bloke who cares passionately for this site. You're dismissing him because you think he's wrong. Asi es la vida. –  Eric Aug 31 '09 at 2:54
4  
@Jeff: Just curious, did you have data and examples to back up the notion that the previous ordering was an undesirable side-effect that needed to be fixed? If not, why is the standard of proof so much higher for us than it is for you? –  ベレアー アダム Aug 31 '09 at 19:53
2  
@Jeff: You're just being stubborn and not reviewing his points in a sober fashion. He has a valid point. –  Paul Nathan Sep 2 '09 at 16:13

I agree. I don't see any reason why we ever departed with encouraging the FGITW issue, but the solution is far worse than the cure.

share|improve this answer
2  
Rich: I just noticed I can remove my upvote on this answer to make this question an example of the "SCITE" (slowest cheater in the east) problem. :) –  LeakyCode Aug 26 '09 at 19:22
1  
@Mehrdad: Exactly. –  GEOCHET Aug 26 '09 at 19:24

I agree. If we don't revert, you could track Jon Skeet's recent activity and immediately clone his responses (change variable names, at most) and submit :)

As a result, IMO, not only downvoting strategically is ethical now, it's even encouraged to fight gaming the system with the Slowest cheater in the east (SCITE) problem. You should downvote any duplicate you see in the new system.

share|improve this answer
    
Agreed. Look at what is happening here: meta.stackexchange.com/questions/17906 –  GEOCHET Aug 26 '09 at 18:51
    
I beat him by a good 7 seconds, but it is difficult for anyone to know that, and his duplicate post is being treated as if it is the same. –  GEOCHET Aug 26 '09 at 18:51
    
Rich: some guy posted identical answer to mine 3 minutes later and posted a comment "beat me to it!" Hey, it's 3 minutes! –  LeakyCode Aug 26 '09 at 18:56
    
Rich: I just saw the link in your first comment. How's that related to this issue? –  LeakyCode Aug 26 '09 at 19:05
    
@Mehrdad: Refresh the page a few times. They appear in random order. Who posted first? –  GEOCHET Aug 26 '09 at 19:07
    
Rich: Got it ;) –  LeakyCode Aug 26 '09 at 19:12
3  
How does this change the stategic downvoting problem? Before, you just downvoted the answers that you wanted to get ahead of. Now you downvote the answers you want to get ahead of. On that front, there really isn't a change. –  devinb Aug 26 '09 at 19:22
1  
@devinb: There is no benefit, only detriment. –  GEOCHET Aug 26 '09 at 19:23
6  
@devinb: In the past you wouldn't downvote others matching your score that you came ahead of since yours would always be at the top. Now that the answers are shuffled, you have a much higher incentive to downvote anybody with the same vote total as you to make sure they don't get shuffled ahead of you on page views. –  TheTXI Aug 26 '09 at 19:24
3  
devinb: On slow hours (and quiet tags) it's very likely that you don't get an initial upvote immediately. In the meanwhile some guy would post a ripoff answer. How do you deal with it? –  LeakyCode Aug 26 '09 at 19:31
6  
@devinb: If there come along better answers, they will bubble up regardless of which came first. All this does is take away the benefit of getting first dibs on a question and allow for people to post the EXACT SAME THING and as long as they have the same number of votes, it will not give the first poster any benefit. That is a detriment. –  TheTXI Aug 26 '09 at 19:32
2  
Why should my answer here: meta.stackexchange.com/questions/17906 by randomly sorted with another user's dupe answer? What prevents me from waiting for the first answer on every thread and just duplicating their answer now? –  GEOCHET Aug 26 '09 at 19:35
2  
@Rich: Do you really think that Jonathan took 7 seconds to copy out your response into a LONGER response of his own? –  devinb Aug 26 '09 at 19:40
5  
devinb: You're previous assumption was wrong. It's a little harder. I've seen many times that a newer answer floated to top. –  LeakyCode Aug 26 '09 at 19:44
2  
@devinb: But that is wrong. If the person posted a good answer you agreed with, you should have been contributing to his answer, or just voting it up, not posting a more fleshed out duplicate. –  GEOCHET Aug 26 '09 at 19:46

When did this happen? I've always used "Oldest" sort rather than "votes". Maybe that's a better solution: registered users see it sorted that way instead, or some variation thereof.

share|improve this answer
2  
Joel: this is not a problem with how you and I see it. The majority of users will use "votes" sort order (since it's the default) and naturally, they are more inclined to vote on answers floating on top. Consequently, it'll be abused by posting dupes. –  LeakyCode Aug 26 '09 at 19:18
3  
We like things sorted by votes. We don't like things that have the same number of votes getting randomly shuffled to try and "level the playing field" for people too slow. –  TheTXI Aug 26 '09 at 19:18
6  
Things should be sorted by votes first and then by oldest. That was the whole idea of the site. This random crap is ridiculous. –  GEOCHET Aug 26 '09 at 19:24
1  
"That is the whole idea of the site" really? you think the site is in some sense based on 'ordering by time'? –  devinb Aug 26 '09 at 19:38
2  
@devinb: Absolutely. It was the whole concept. This was also the official answer of every FGITW thread. –  GEOCHET Aug 26 '09 at 19:47

Completely agree that the old method needs to return. The potential for abuse on both ends is higher now. People will start downvoting much more to move people down the list who tie with them, and you are also giving equal footing to people who post duplicate answers. As long as the vote totals remain the same, any normal user would be unable to determine who actually posted first. Therefore the first answer they see (even if it was the duplicate) has a higher probability of getting the upvote.

Stop punishing those of us who contribute quality answers quickly. As I have said numerous times, bad answers go to the bottom regardless of when they are posted and good answers should bubble to the top regardless of when they are posted.

share|improve this answer
4  
Aye. I posted an answer earlier this morning only to have two people post identical answers after I had answered--20 minutes after the question was posted. And it made me a sad panda. Luckily, I had pancakes for lunch, so I totally forgot about it until this thread came back up. Now I'm a sad panda again. –  Eric Aug 26 '09 at 21:17
2  
I should clarify that I answered 20 minutes after the question, and the other answers came 2 and 4 minutes after I had posted the full answer (which was about 10 minutes after my initial, quick hit answer). –  Eric Aug 26 '09 at 21:18
    
@Eric. Flag them. Make the moderators can demolish duplicate posts. –  devinb Aug 27 '09 at 15:20
4  
@devinb: And waste Marc and Bill's time with, "He copied me!" schoolyard arguments? My hope is that they have better things to do. My point is that this has introduced a host of new problems without solving any problem at all. –  Eric Aug 27 '09 at 16:23
1  
I have yet to see any real evidence that this "he copied me!" problem is in fact a problem. Could you perhaps cite the post you're referring to, oh-so-vaguely, in the above comment? It's helpful to have.. y'know.. examples.. or we could just have these hand-wavy arguments about our feelings and so forth. –  Jeff Atwood Aug 28 '09 at 7:28

I knew that Jon Skeet should never have posted his random shuffle code.

The new algorithm introduces a gaming tactic of downvoting an earlier answer that was upvoted. Previously, this would have done no good since it would still appear higher in the default sort order. Now there is some gain to downvoting this answer (as long as it is only one greater than yours) since you can increase your visibility on the question. As long as the OP can select the best answer and accept it I think we ought to be using the old system. Good, early answers will continue to get votes over later answers, but the best answer will rise to the top due to the ability to accept an answer.

Personally, I don't think that I am often the first answerer -- unless I'm the only one. My tactic is to try and write complete answers with references and examples. More often than not the inclusion of references and examples trumps quickness.

share|improve this answer
2  
Exactly - we've moved from "Be fast or be good" to "be good or be lucky". Exchanging a game of skill for a game of chance... But still open to gaming. –  Shogging through the snow Aug 26 '09 at 19:41
    
Shog9: "Still open to gaming?" I'd argue it's "even more open to gaming." –  LeakyCode Aug 26 '09 at 19:42
3  
@Shog9: More open to gaming! –  GEOCHET Aug 26 '09 at 19:47
    
@Mehrdad: time will tell... It's certainly not any more closed to gaming! –  Shogging through the snow Aug 26 '09 at 19:48
    
@Shog9: You were never able to post a duplicate of an answer and have your answer appear above the answer you just duplicated before without being detectable. –  GEOCHET Aug 26 '09 at 19:50
4  
Someone should post a blog entry titled "Has Jeff Atwood jumped off the shark?" He basically ignored the whole StackOverflow community (in contradiction with his "StackOverflow is You" statement) by making such a sudden decision without discussing on meta :) –  LeakyCode Aug 26 '09 at 19:53
1  
@Mehrdad: I agree. But I have been screaming about his spineless tendencies for a long time. –  GEOCHET Aug 26 '09 at 19:54
1  
@Mehrdad -- just what I was thinking. Is or when does SO get big enough that Jeff and the team have to start letting the community guide the development more explicitly. I'd like to see something like the moderator votes, perhaps, on a regular basis to allow community input. –  tvanfosson Aug 26 '09 at 19:56
    
@Rich B -- I don't know about spineless. It seems more likely that the developer tendency of "oh, that's a cool feature, I think I'll implement that" is more to blame. So any customer that walks in and suggests something the developer likes gets what they want. At some point you have to put the developer's ideas on the back burner and let the customer base drive the feature set, not the random customer who suggests something the developer already has in mind. –  tvanfosson Aug 26 '09 at 20:30
3  
@tv: But that doesn't make sense since Jeff himself was opposed to this whole idea just like the community before. This is a sudden about face on the whole issue with no announcement or community decision. –  GEOCHET Aug 26 '09 at 20:31

I believe that this new change is just fine.

For people who want the answers sorted by time, they can simply change their sort to 'oldest'. For those who want to sort by votes, they will receive a list that is (as they requested) sorted by votes.

In general, being the first to post is still advantageous, it means you have more time to gain upvotes while there are no other responses. However, once the other responses appear, why does it matter which is first, the upvotes should be based on whose response is better.

A side effect of the previous system was that if there were many posts with the same number of votes, and mine was the earliest, I was unlikely to edit it for fear that I would lose my top spot. Usually it was grammatical mistakes (I'm aghast I made them in the first place) but occasionally they were links or other such helpful tid-bits. Sometimes I edited and got upvotes, other times I edited and found that the other answers (with less information) got upvoted soon after. In making my response better, I lost reputation.

This new system is meant to combat that.

I don't know if it will work. But I'm willing to try it, rather than explode with indignation after only 8 hours of it.

share|improve this answer
3  
You are incorrect. You can edit your answer and remain on top if you do it inside the grace period or if you have a higher amount of votes. If neither occur, and someone made a better post, you don't deserve the top place. –  GEOCHET Aug 26 '09 at 19:40
4  
What is going to prevent me from duplicating the first answer in each thread and getting upvotes for it? Before this would have been obvious, because you would always be below. Not anymore though. –  GEOCHET Aug 26 '09 at 19:41
4  
@devinb -- the fear of improving your post only means that you care more about the rep than the answer. I'm always improving my posts -- for some reason I don't seem to be able to see spelling errors until after I click the submit button -- and haven't been materially harmed by it. Write better answers and you will be rewarded. –  tvanfosson Aug 26 '09 at 19:41
3  
Especially if I am sneaky and I just down vote your answer I just duplicated. –  GEOCHET Aug 26 '09 at 19:41
    
@Rich B: I'm well aware that there is a grace period, as I'm sure that you are aware that there is a time after the grace period. Since there is a five minute period when my comment applies and an infinite period when it doesn't, I had assumed that the readers would be able to determine which period I meant. Furthermore, since I explicitly stated that I was talking about posts that have an equal number of votes, I appreciate that you mentioned that there is a scenario where they don't. Please rest assured, I was talking about the scenario where they have an equal number of votes. –  devinb Aug 26 '09 at 20:10
    
@tvanfosson: I always edit my responses. I consider StackOverflow to be a place to hone my professional skills, and so I stand behind every response I give. I was just noting that there is a split between being first, and editing your response to be better. It is a point to consider. –  devinb Aug 26 '09 at 20:12
    
@TheTXI, I would hope that most moderators actually look into matters, rather than lazily blaming anyone. –  devinb Aug 26 '09 at 20:13
1  
devinb: If you've an already upvoted answer and you edit it. Them someone upvoted another answer with less upvotes, (which shouldn't happen if the "assumption" you mentioned in a comment to my answer was true) there's a huge possibility that the other answer was better than yours and deserved the upvotes. –  LeakyCode Aug 26 '09 at 20:13
    
@Mehrdad, I agree, which should make it clear that speed has no bearing on quality. If my answer is quicker and gets an upvote, and then their answer is better and gets an upvote, why should mine be up top? The community likes them both equally. –  devinb Aug 26 '09 at 20:17
    
@devinb: Yes, most moderators probably are able to determine what is a dupe and what is not, but I would not necessarily think that a good number of users on SO or any other site pay that much attention. My example stands. Someone could see your post 2nd and then downvote you for posting a dupe even if you were first and you should get credit for it. And we mods can't go and "fix" that and reverse votes or anything, so you are screwed. –  TheTXI Aug 26 '09 at 20:17
    
Before I was finally upvoted past the other answer, who came first here? meta.stackexchange.com/questions/17906 –  GEOCHET Aug 26 '09 at 20:18
3  
@devinb: I am not going to switch to an oldest-to-newest view because then I completely lose the vote sorting ability. And your example of two wildly divergent answers...one has to ask: "Is the answer wrong?" Yes? Then downvote it! If it's not wrong, then it either deserves an upvote or no vote at all. If your answer is better, it will climb higher without the need to artificially shuffle it about in hopes that someone will see it first in the pecking order. This change is going to allow for more problems than fixing any potential or theoretical ones. –  TheTXI Aug 26 '09 at 20:35
1  
@devinb: The community issue here is hypersensitive newbies cry that their late answer that duplicates someone's faster answer is upvoted while theirs is not. That is not going to be fixed by anything except for telling them to "grow up" and "try harder next time". –  GEOCHET Aug 26 '09 at 21:06
1  
@Henk: It is being proven all over the site, have you even read the reasons why people are opposing this? –  GEOCHET Aug 26 '09 at 21:21
4  
"If your answer is better, it will climb higher without the need to artificially answer first in the pecking order." –  Jeff Atwood Aug 27 '09 at 2:57

I find the resistance to the new votes sort ordering from a vocal minority very perplexing.

First, to clarify: this has nothing to do with tactical downvoting. It is not meant to address that issue at all. Tactical downvoting is the same as it ever was.

(edit: this other change is meant to address the tactical downvoting.)

It's purely a fix for the votes sort order. The answers are sorted by votes. But what sort order do you use when every question has the same score? We had to pick something, so we picked LastActivityDate. I was never happy with this choice (it wasn't even really a conscious choice, honestly), since it had a side effect -- we were implicitly rewarding behavior other than composing answers worthy of upvotes!

Now answers with the same vote score are in random order. Which means answers can correctly be judged based on their merit as answers, not by who happens to end up first in the accidental secondary sort order we didn't even intend to be relevant in the first place.

To the extent that people have "optimized" strategies around that behavior, it's completely a side-effect. A distraction.

Bottom line: if you want to get upvotes, WRITE A GREAT ANSWER.

share|improve this answer
6  
The entire crux of the argument lies with the (purported) SCITE phenomenon. Since there is now no drawback to posting late, it's tempting for someone to copy a good existing answer and re-post it while all answers are at zero. It is no longer readily apparent who is copying from who, hence there is actually an incentive to wait before answering (to read other people's answers) rather than to answer quickly. –  ベレアー アダム Aug 27 '09 at 2:54
9  
"There is no drawback to posting late?" Not so. The earlier you post, the more opportunity for people to VOTE for your answer. If you wait, you're losing potential votes. –  Jeff Atwood Aug 27 '09 at 2:59
3  
Right, as I said, the whole problem only exists within the time period before anyone has earned any votes. (The prhase "while all answers are at zero" should apply to my entire statement.) Once votes have happened, we're into the realm of strategic downvoting, which is not dissimilar to the current situation. –  ベレアー アダム Aug 27 '09 at 3:02
1  
"if you want to get upvotes, write a great answer" - that's always been a good strategy. That this showed up positioned as a fix for a specific combination of FGITW and questions with tens or scores of answers just seems... bizarre (how many questions collect 30 answers with the exact same score?). Surely a better solution to that "problem" would be to randomly sort all CW answers, regardless of score, or periodically pop a new answer to the top... Or simply the ability to search answers: meta.stackexchange.com/questions/1274/… –  Shogging through the snow Aug 27 '09 at 3:31
4  
This isn't a fix for FGITW -- it's removing a side-effect of a secondary sort order that users have been "optimizing" for. It is completely accidental, and was never by design. –  Jeff Atwood Aug 27 '09 at 4:29
5  
I think the change is a mistake, but what bugs me most is that despite the claims that "you run the site", it doesn't appear to be true when it comes to fundamental changes. –  tvanfosson Aug 27 '09 at 5:00
18  
...more to the point, what is really "unnatural" is to hit refresh and have the order of the answers change when there hasn't been any material change to them. Reminds me of the definition of insanity -- doing the same thing over and over and expecting a different result. Apparently, now, insanity is the expected outcome on SO. –  tvanfosson Aug 27 '09 at 5:05
4  
@tvanfosson if you feel so strongly about chronology, why are you on the votes tab at all? nothing is stopping you from browsing answers by the "oldest" or "newest" tab. And that's sticky, so it'll stay that way until you change it. –  Jeff Atwood Aug 27 '09 at 7:27
9  
@Jeff: Citing TheTXI:"I am not going to switch to an oldest-to-newest view because then I completely lose the vote sorting ability." –  Ladybug Killer Aug 27 '09 at 7:42
3  
@Jeff: just so it's not lost -- my primary concern is over how the decision was made, not the implementation. Given the amount of controversy inside SO, I think this points very clearly to a break down in the feature selection process. –  tvanfosson Aug 27 '09 at 11:31
2  
@Jeff -- now that's out of the way -- my opinion is that the fix introduces more problems than it solves. These have been documented pretty clearly elsewhere. I suggested on a different thread that a better way to treat this would be to truncate the time when sorting rather than sort randomly within votes. Votes within a few minutes of each other could be treated the same, in this respect, but it would still retain the basic time based character (and historical record) of the previous system. I'll elaborate in a question. –  tvanfosson Aug 27 '09 at 11:35
2  
@kev: I don't need analytics and statistics, when I see what has happened in this thread: stackoverflow.com/questions/1341847 –  GEOCHET Aug 27 '09 at 16:36
1  
"now, insanity is the expected outcome on SO" well, voting is the expected outcome. I guess it depends on your definition of insanity. –  Jeff Atwood Aug 28 '09 at 8:57
3  
@Jeff Atwood: I, for one, think the new randomized secondary sort is a good idea. –  Kip Aug 30 '09 at 19:34
3  
@Kev: As I observed in my comment on Mehrdad's post, was there extensive methodical analysis of the previous sorting order which backed up the decision to change it? If not, why is the burden of proof so much higher now, after the fact? –  ベレアー アダム Aug 31 '09 at 19:56

If you write great answers, what do you have to lose by this change? Doesn't random imply that if you have the same number of votes, you are just as likely to be on the top as on the bottom with those who have the same vote/score?

I tend to think that this will help people who answer 2nd or 3rd with a more complete answer compared to number one who people give votes to for being first. Being first doesn't make you the most correct. It is the content.

Since content drives this site, I ask, what are all of you so concerned about? You have high rep due to content, not the fastest gun in the west, right?

share|improve this answer
6  
This seemed so much more intrinsically fair to me. You're sorting on votes, so the sub-order is irrelevant and any effects of that sub-order should be minimized. –  Jeff Atwood Aug 27 '09 at 6:21

I rarely look at meta and I am perplexed by the few people who are constantly obsessed with strategic downvoting. For what it is worth, I have posted several answers late on various threads and they got upvoted pretty fast.

If I post my answer and then see that an equivalent answer was posted while composing my answer, I just delete mine.

I get hit by the occasional drive-by (or revenge) downvoter. Usually, I post a comment asking why my answer was downvoted and that elicits a couple of upvotes. All in all, things balance out.

If someone is fast enough to copy & paste my answer in a few seconds, get a few upvotes and cost me a few potential upvotes, so be it.

I like my points, I like my perl badge, but, at the end of the day, I answer questions because that motivates me to learn new things, read docs I have neglected for a while.

Thank you Jeff. Starting meta was a great idea so these kinds of discussions do not end up dominating SO.

share|improve this answer

There can only be one answer.

Revolution!

share|improve this answer
    
Wow... That is one ugly flag-bearer! –  Shogging through the snow Aug 30 '09 at 19:39
    
+1 - Nice... –  John Rasch Aug 30 '09 at 20:29

At the risk of being flamed some more, I am offering my opinions on this matter, and a possible solution. Scroll down if you just want to see the solution.

  1. I have noticed that users with a large amount of reputation, including some of the people in this very thread, have a vested interest in the old style. A random sampling of answers for these high rep users show that they are often the first to answer, usually within 1-5 minutes from when the original question was asked.

    The "they would have reached the 200 rep cap anyway" argument is wrong because its all about how fast you can get 200 rep, and then how much extra time you have to get accepted answers afterwards.

  2. SCITE: From what I've seen recently, this hasn't been a problem. Voters who really care about who answered first will be refreshing often, and will know who answered first. Voters who don't care will equally distribute the points to whatever answer is on top. But in case this becomes a bigger problem in a future, see my suggestion below.

    The timestamp is also accurate unlike the old FGITW system where people can steal answers in the first 5 minutes.

Solution

Remove the 5 minute free edit time.

You can then reinstate the last activity ordering which will still encourage fast but good answers.

How will this solve the problems?

  1. FGITW: If someone posts a fast but bad answer, they will get ignored or downvoted. If they try to copy someone's answer, they will get bumped down.

  2. SCITE: If someone posts a slow and copied answer, the sort by last activity ordering will relegate them to a lower position.

share|improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .