@Deidara-senpai I wasn't implying that it was a problem if you become top rep user or answer your own questions or whatever else. I don't know anything about Fairy Tail or Naruto, but judging based on the voting you seem to be providing good and useful answers.
I was actually just impressed with how much rep you gained last week. Since December, no one had gained over 500 rep in one week, but you had 865 last week. It's pretty obvious that you'll pass all of the current 3k users in the near future if you keep that up.
@LoganM How did we end up with 272, 284, 272 in consecutive days (as in, how is it mathematically possible)? Moreover, why do the site stats still say 283 while Area 51 says 272?
@Eric I don't know why different places show different numbers for traffic/day. I'm assuming all of them are correct to some degree, but maybe they report different sorts of averages.
As for going 272->284->272, it's certainly possible. What it means is that the day before yesterday our traffic was at least 284 hits. Then yesterday, our traffic was down to 272 hits. Also, our traffic 2 weeks from the day before yesterday was <284, and 2 weeks from yesterday was >272. We can also rule out any instances in the past 2 weeks of traffic in the range between 272 and 284,
I think there may be other things we can say, but I'll have to think about it more. The math associated to reconstructing data from moving 2-week medians isn't terribly difficult, but it's pretty strange so I don't always have a good intuition for it.
@Eric I'm sort of assuming that our actual traffic numbers are tending to fluctuate in the range 200-350. We don't really care about the individual highs and lows, but if it jumps up or down by 30 or 40 hits in one day that would be notable.
@LoganM Fair enough. I'm concerned it's not increasing as steadily as it used to. I suppose we'll see in the next few days; our last Redditing was on the 13th.
@Eric It'd be strange if it increased every day unless we're actively promoting right now. I think right now we're growing at a pretty normal rate, and there's no evidence from the data that we're slowing down.
@LoganM Yeah, but we'll see if, after the reddit questions fade from the two-week average, the numbers decline. I'm wondering if we're still running on steam from our promotions.
@Mysticial On Math, someone got away with downloading 1000 posts on a 10k account by waiting 5 seconds between posts. Granted, 1000 is a far cry from 4.5m, but it's at least a proof of concept.
In a recent discussion Mariano Suárez-Alvarez told us, that if we are aware of users, who delete their questions after receiving an answer, we should bring them to the attention of mods. (IIRC the OP can only delete a question if no of the answers have been upvoted.)
In an older discussion about...
I agree that you wouldn't want to use just one 10k account for it. You'd have to find a way to automate it just to pick out the deleted posts with unregistered accounts, and then use the 10k account(s) to grab all the deleted ones.
Also, depending on what sort of analytics SE is using, you might not even be able to get away with that. If they keep track of how many deleted posts you've viewed, that would make it obvious very quickly that you were doing something fishy.
Do you have an order-of-magnitude estimate of how many deleted posts there are on SO? If it's tens of thousands, that might not be too hard, but if it's millions then it's considerably harder.
@LoganM I don't remember. There were some questions (such as the boat programming question) that were so controversial that Jeff wanted it out of the system.
Controversial as in not appropriate for the site. (super non-constructive)
Either way it's gonna be difficult: 1. Probe all 15 million post ids to see which are deleted. 2. Use one or more 10k accounts to scrape all the deleted posts.
1. You need a well organized botnet... 2. You need 10k accounts.
Strangely, when I go to stackoverflow.com/questions/XXXXXXXX, sometimes it takes me to the correct posts, but a lot of the time it takes me to some random other question.
So then there's about 2m deleted posts (assuming the other types don't count for very much). It's not clear what portion of those are questions vs answers.
2m deleted posts. If you had a 10k account go 1/sec. That'd take 23 days. But sustaining something like that for more than a few hours will definitely get you in trouble for taxing the server.
My IP changes daily anyway since I'm on my school network which assigns IPs pretty much randomly. The most they could tell is that we're both at JHU unless I logged out and then logged back in immediately. But I switch networks before I use that account so they shouldn't be able to track it easily even if they want to.
One of the networks in my office always gives me the same IP, but I've stopped using that one because I don't like being easy to track. I have access to 7 wireless networks in my office (university, university guest, physics, physics guest, particle physics, particle physics guest, and astronomy guest networks), so there's no trouble switching between them.
Although I like the idea of putting together a 10k account just for the purpose of viewing deleted stuff when I'm away on insecure connections and such.
But it will take months to repwhore up to 10k. Not to mention that I need to be careful not to post any answers that could potentially go viral - since those belong on my main account.
I think it's tricky, but should be doable in under a year.
For reference, here's the second account of the guy I was talking about. On chat, he admitted that another 10k account was his, and that this wasn't his first 10k account either:
Most of what he does is answer undergrad homework questions, which is the easy way to get lots of rep on MSE. He has some highly upvoted posts, but most of his posts are +3 or less.
Obviously, it's the best way to get rep. But restricting only to easy FGITW will make sure that none of the surprise massively upvoted answers end up on the wrong account.
What is the fastest, paper-pencil method of finding $$\sum \limits_{i=1}^{69} \sqrt{\left( 1+\frac{1}{i^2}+\frac{1}{(i+1)^2}\right)}?$$
This is actually a quantitative aptitude problem, and hence the solutions should be fast enough and probably under a minute.
Using wolframalpha, the sum seems ...
@Mysticial It's a good answer to a trivial question. If something is tagged as [algebra-precalculus] then the majority of the high-rep people can answer it almost immediately.
The question I'm looking for a list of humorous anime similar to oruchuban ebichu led to a debate on the correct interpretion of our policy on list questions. To avoid taking that debate all over the place, I opened this meta thread.
I copy the list questions policy below, along with my inte...
@Mysticial Now that we're thinking about this, I'm wondering if there's an easy way to find out what that guy's other accounts are. I know they can't be posting on the same questions, but that probably will only eliminate half of the 10k users. For the other ones I'd have to try to compare writing styles.
His answers are all in tags which anyone could answer in. He doesn't appear to have any specialties, just a good grasp on math up through undergraduate-level courses.
Does his main account has any specialty topics? If yes, then he's really paying attention...
Although I'm completely theorizing. On smaller sites, if you create a second account that gets a lot of rep, having no cross-voting might give it away to the mods.
Another thing to look for in an account are the user's first posts and comments.
Do they show any indication of already knowing the site?
It's hard to tell. His first question is very basic, to the point that he was pretty clearly not understanding a fairly basic concept. However, his first answer is fairly impressive technically. His future questions show improvement, and he's now asking about more technical things.
Either he's deliberately trying to fool people, or I can pinpoint almost exactly where in his education he is. That should rule out most of the 10k users, though not all of them.
If the user's first post has very aggressive editing on a FGITW with no comments from other people, then that's an indication that the user is experienced.
And pay attention to how fast the first few answers were.
@LoganM I'm tempted to think that it's not that easy to completely cover your tracks.
@Mysticial His first answer was on a question that was open for a while and was very hard (I read it and tried and couldn't answer it). He didn't answer it fully either, but he did provide a nice observation which could be used to make progress. The question was later answered by someone else on MO.
I haven't seen many of his answers being edited. He is going after easy questions, but they mostly don't get lots of answers and he's writing long complete answers right away. I don't think he's trying to do FGITW.
Then again, he could be doing what I used to do on Math where I'd delete my answers which weren't the fastest. I prefer deleting it and moving on to a different question rather than keeping lots of 0- and 1-upvote answers.
He posted something in chat that implied to a mod who was also in chat that he had another 10k account. I suspect it was something regarding flagging. I couldn't tell from that conversation what that account was. The mod then asked if he really had a second account, and he replied that the "anon" account wasn't even his first one.
Reading my first few answers, I now feel like deleting the first 5 or so. Those were the ones where I was basically trying to get to 200 rep before Theoretical Physics beta started so that I could get the association bonus when it did.
@Eric Your visit/day chart is somewhat inaccurate (compared to the actual analytics data), even if we factor in the moving average. I can say momentum from the Pokemon post spiked on the 13th and continued waned until the 15th (where it returned to typical levels). Visits are up slightly compared to last month.
@Krazer Very likely it's a statistical anomaly that has to do with how the 2 week moving averages work. Though you can probably tell from your mod tools if I'm right or not about that.
If 272 is no longer the median, then the median jumps to the next lowest traffic day after 272. If there weren't any days with traffic 273-283 then it'll give 284.
In any case, IMO traffic isn't our #1 problem right now in terms of staying alive. Plenty of sites have been kept around for a while despite lower traffic than us. The one number that we're very poor on is the total # of users.
Nevermind, I can't upload it. Anyway Feb 11th we were 10th from the bottom in traffic. We're still at the same place; no one's jumped us, but we haven't jumped anyone either.
@Krazer We're beating politics except in total users and answer ratio (roughly tied).
I think we should set our goals higher. Let's try to beat Robotics
@Mysticial Our current level is fine for % answered, but for answer ratio we're one of the lowest in the network. I'd like to try to get to at least 1.8, and honestly we won't ever graduate unless we're at close to 2.5
We've got lots of old questions which aren't fully answered, but have an upvoted answer. We need to go back and figure out which of those need more complete answers.
I asked this before, but do either of you know how the sites are ranked at stackexchange.com/leagues? It doesn't seem to be any of the stats I have access to, and it's worrying that we're just 3rd from the bottom.
@Mysticial Those are not criticisms, they are the current policy. And you're not being targeted. Targeted = personal attack and no-one is attacking you personally.
@Alenanno if it is relevant to the question and the OP wasn't aware of it initially, I suppose it wouldn't do any harm (which is the case in these identification questions)
I recently asked for help with a bug, where I had little idea what the problem was or how to fix it.
I tagged my question with some basic things, like what language I was using, etc. And the person who answered my question (correctly. thanks!) also deleted a couple of my tags, on the theory that...
@Alenanno The problem is that strict adherence to that policy closes off an avenue for people to find the question. The best example of this would my top answers to performance questions on SO. The asker has a performance question X. The answer is due to Y. People learning about Y would be interested in examples of where it could impact performance.
Some times an answer is very detailed and expands on the question, it would be good at this point to have separate tags for that answer.
Maybe just 2 (giving a total of 7 when combined with the question tags).
This would allow for some funky searching and help link answers to different question...
In any case, and getting back the point, being counter-intuitive for new users is indeed a possible way for it to be harmful. But as of right now, I'm not convinced this is actually the case.
Community FAQ
For sites in the Stack Exchange 2.0 network
To see a list of commonly used words and phrases, see the glossary.
Asking questions
How do I write a good title?
How can I get answers fast?
Is there a limit on how many questions I can ask?
How should we deal with Google questi...
One last word before I leave. I'd like to actually try this retagging for a while. If it actually does turn out to be a problem without benefits then it proves your point.
On the bus now. Beta also means that we can experiment with new things that could make the site even better. I don't see any harm in trying it. So if no other site has tried it we might be in for a surprise. Of course if SE explicitly bans experimentation then that's a separate issue.
So the consensus is to have a trial period so we can see if tagging sereis tags to ID questions has any impact on SEO traffic? I'll be keeping an eye on search traffic, but at a glance it doesn't seem to be doing much
I understand the points on both sides. Identification questions don't usually drive much traffic, because they are sometimes overly ambitious. We don't get much search traffic from our ID questions at the moment, based on the analytics.
@krazer both ways actually. Not just to see if it helps as much to see if it hurts. If we start getting users complaining, then we know that it's hurting.
@Mysticial I don't see this being a problem, what do we do about ID questions that we later find out is of the same series? E.g. one giving reference to the main plot another giving reference to some side story or subplot within the main series
I don't think it's gonna draw much traffic in. Bur I like the benefit of being able to search a show/tag with the id-request tag to see if there were any such hits.
It's a small benefit that would make sense to have. I went with that in the first place because there's no clear drawbacks.
@Krazer The losses are on more of an internal component than on a search component. I don't know that it'll subtract from search results due to putting prominence on a term that users who have the problem will not know to search for, but that is something slightly implied by Gilles's answer.
@Krazer I'm not convinced that we really have nothing to lose. Specifically, if we tag new series with these, it opens those tags up for tag wikis that we otherwise couldn't have. Most of the answers to these questions probably reference wikipedia, which would contain most of the same information, so this isn't crucial (especially for series which only have an identification question tagged). But it isn't nothing.
As of right now, there's only one tag wiki in this category, which is pretty good IMO: