Ranking

[laz0r]

Can someone explain how the Top Hackers ranking works? http://www.hacker.org/util/tophackers.php My ranking's been going up and down like it's in a washing machine lately...

[AMindForeverVoyaging]

I guess adum is the only one who knows for certain. Try writing him a mail?

[laz0r]

I didn't think it was really important enough to bother him with directly just if anyone happened to know a Turing computable way to work out what my ranking would be tomorrow!

[AMindForeverVoyaging]

Well, adum has given some insight as to the scoring system here and there:

http://www.hacker.org/forum/viewtopic.php?t=105
http://www.hacker.org/forum/viewtopic.php?t=1523

I don't know if a complete description of the algorithm used is given on this here forum though. Maybe there is and I just haven't found the topic where it was stated.

[MindFreakz]

For as far as I can tell, seeing my rank go up and down as well, and from what I can read from the previous links and: http://www.hacker.org/challenge/about.php (and that fact that recalculation is necessary)


You get points for each challenge you solve. The points you get for a challenge are probably something like:

Code: Select all

100^(Number of people who solved the challenge)

You then sum the scores of all the challenges to get your personal score; which put you somewhere on in top list.


This implies your rank going down when someone else solved a challenge you also solved.

I'm not quite sure how the games on the homepage go into this, but I expect a similar method is used.

(And I expect that you rank just doesn't make any sense during the recalculation)

[AMindForeverVoyaging]

This implies your rank going down when someone else solved a challenge you also solved.

Honestly, I don't understand this.

If you get points once, in the moment when you have just solved a challenge - then you would not "lose" points when somebody else solves the same challenge.
If your points are constantly recalculated for challenges that you have solved in the past... then I really wonder what's the rationale behind that. It does not seem very intuitive, at least.

[laz0r]

My Challenges rank is pretty much constant (certainly relative to other people) - it would appear not to be to do with people solving the challenges, at least... I haven't solved any bot wars at all, so that shouldn't affect my ranking, I would have thought (I can't get much lower than 0!), and I haven't seen great upheavals in the 'weekly solvers' lists on the games.

[MindFreakz]

@AMindForeverVoyaging:

I didn't think of it myself (from: http://www.hacker.org/forum/viewtopic.php?t=105):

the points are recalculated hourly.

I don't know why this is done (I guess only adum knows). The only reason to do this I can think of is to give new players a fair chance.
If two people solved exactly the same challenges, they would now be ranked equally; otherwise, the player who started playing earlier would probably be ranked higher.

[AMindForeverVoyaging]

Or you simply give the same amount of points to everybody for solving the same challenge.
(adjusted to the challenge's difficulty level, that is)

[contagious]

if you're top in a particular puzzle, you'll be awarded 100 points, and it goes down from there.

This is kinda unfair, older users get more points.

For example check this guy tiwe, no Bitbath or Wormageddon and still has a high ranking.

[AMindForeverVoyaging]

This is kinda unfair, older users get more points.

I have always felt that this does not make much sense in the context that we have here.

You could do it this way if all people, e.g. in a competition run, started at the same point of time. Then, sure why not, he who solves a challenge/puzzle first should get more points than the second guy.

In a field where people start at totally different points of time, giving more points to somebody who started on a challenge way earlier (and thus had the advantage of being able to finish it earlier than the competitors) is simply not a sensible approach.

For example check this guy tiwe, no Bitbath or Wormageddon and still has a high ranking.

And he has only solved 86 challenges! It's got to be a joke to rank somebody with such average achievements that high. It clearly shows that the ranking system used here is pretty nonsensical.

[Zeta]

if you're top in a particular puzzle, you'll be awarded 100 points, and it goes down from there.

This is kinda unfair, older users get more points.

No, they don't. Same Achievements result in the same number of points in the ranking tables regardless of starting time.

For example check this guy tiwe, no Bitbath or Wormageddon and still has a high ranking.

He earned easy 100 points by solving Runaway Robots to the end. The user ferret was the 195th to get this far and gets the same 100 points. Feel free to do the same...

[dangermouse]

laz0r, you need to bring one of the puzzles to the maximum of the levels (Runaway does not count), then you'll jump in ranking. Just happened to me this weekend...

[laz0r]

laz0r, you need to bring one of the puzzles to the maximum of the levels (Runaway does not count), then you'll jump in ranking. Just happened to me this weekend...

I'm concentrating on the Challenges at the moment, but I'll look at the games again thanks for the advice!

[contagious]

He earned easy 100 points by solving Runaway Robots to the end. The user ferret was the 195th to get this far and gets the same 100 points. Feel free to do the same...

Ok then i dont understand how(for example) tiwe is rank 129. I am probably missing something here.

Usually people with higher rank than expected have good bots in Bitbath or Wormageddon, but i think this isnt the case here.