21

Screenshot

About

StackRating.com tracks user ratings which is a measure of users “skill”. It does so by comparing the number of upvotes on the answers of each question and uses the formula of the Elo rating system to update the users ratings.

I created it because I was curious to see if there was any correlation between skill and reputation. (Here is a separate post on describing my findings in this topic.)

For a detailed explanation of how it works and how to interpret the rating numbers, see the about page.

URL

http://stackrating.com

Contact

The webapp was created by me, aioobe, with feedback from dacwe.

You can reach me at [email protected].

Code

https://github.com/aioobe/stackrating

Updates

2015-05-08: Fixed bug due to SE API outage.
2015-05-13: Score no longer depends on age of answer.
2016-01-04: Big rewrite. Automatic monitoring of new questions working properly now.
2017-06-25: Added HTTPS support. Update your badge links!

15
  • 7
    Good‌‌‌ ​‍​‍​‍​‍​‍​‍​‍​‍Job! May 3, 2015 at 0:17
  • Can you explain my ratings on these two questions? SQL First Match or just First and How to stop Excel storing the absolute path to an XLA?. The former is my question, but the higher answer (if accepting my own answer is not given a +1) is by a 90k+ user compared to my 5k. The latter has a leader of 17k.
    – Mark Hurd
    May 8, 2015 at 4:36
  • The first one I can't explain. I'm going to investigate that (and run a rating-recalc which requires me to take down the site for a couple of hours). As for the second one, (Excel) you get a negative rating because you "lost" against Dick Kusleika (and the other two answers that "lost" are posted more than 90 days after the question).
    – aioobe
    May 8, 2015 at 5:20
  • That they are negative makes sense (probably for the first one). But why the magnitude?
    – Mark Hurd
    May 10, 2015 at 17:48
  • 1
    Ok, I see. The magnitude is determined by the K value and depends on how many answers you and your opponents had given at the time the question was posted. (See the about page for the formula used.) For example, if you had posted less than 100 answers when the question was posted, your rating update will be somewhere between -8 and +8 depending on the difference between the expected outcome and the actual outcome.
    – aioobe
    May 10, 2015 at 20:59
  • I love the ratings. I am a little puzzled over how some of the scores are calculated, however. What happened here for an extreme example; why did that post cost me 1.6 points? May 12, 2015 at 14:11
  • Kind words, thanks. Means a lot coming from you Martijn :-) For me it shows +0.83 on you right now. The scores depend on votes, and votes changes over time, so it may very well be the case that you had -1.6 at some point in time (such as right after you posted for instance). The site scrapes the questions frequently in the beginning, and then less frequently and stops scraping when the question is >90 days old. That being said, if you find some odd values it might be due to a bug. It's on my todo to give account for the computation on the page. (Great answer by the way ;-)
    – aioobe
    May 12, 2015 at 14:26
  • @aioobe: thanks for looking! The -1.6 was there 3 hours ago with a score of +378 or there-abouts; checking my rep log I received just one vote in the past 3 hours, with another 2 in the hour or so before posting my comment, 15 votes in total so far today (it's a very busy post as it was posted to Reddit). May 12, 2015 at 17:54
  • ha, your app is cruel to me: rep / rating I rank 6k / 3M because I have only necromancers. Some of them are in fact due to new information which became available since, but not all, I'd say about half. May 12, 2015 at 18:37
  • @aioobe the scores on the post are definitely unstable. I saw -0.12 before switching to 0.87, and my overall score was above 1900 (and me at 5th place) but is now lower again in spite of my only having had positive or 0 scores for all new posts. Something appears to be oscillating here! May 12, 2015 at 22:51
  • @MartijnPieters, the rating does indeed oscillate quite a bit. If someone votes on a question that was answered 89 days ago, all questions are rejudged from that day onwards. But it shouldn't vary as much as you have observed, and if you have the best answer and get a negative rating update, that's a bug. I have a couple of plausible explanations (but none are very satisfactory). I keep the entire database in memory (-Xmx13gb) to solve pagination efficiently and unfortunately I can't afford storing all intermediate values for the rating computations...
    – aioobe
    May 13, 2015 at 6:27
  • 2
    ...and, for everyone: If you have the most upvoted answer, but a negative rating update, please let me know, preferably over email. (Please also include screen shots of stackrating page and stackoverflow page since votes / ratings change over time.)
    – aioobe
    May 13, 2015 at 6:38
  • 1
    I saw on the about page that you're not interested in developing this any further. Fair enough, what you've done is great! I was wondering if you'd make an exception for what is, I presume, a relatively easy modification: expand this to more (all?) sites on the network. If it's just a 10-minute hack, could you do it? And if not, could you point me in the right direction to do it myself?
    – terdon
    Jun 15, 2015 at 22:38
  • @terdon, I did in fact spend some time with this, and the app has now been rewritten to be more stable and less of a memory hog. The code is pretty uncomplicated (plain Java + SQL) and available on github. If you'd like to launch it for another site it should be pretty easy.
    – aioobe
    Jan 8, 2016 at 11:56
  • Sadly, I know no Java so this won't be quite as easy as that for me. Still, might give me the incentive to dive into the SE API. Thank you :)
    – terdon
    Jan 23, 2016 at 10:43

0

You must log in to answer this question.

Browse other questions tagged .