interesting statistics.

Code:

---

mysql> select nick,rankpoints1,rankrank0 from temp_boinc_combined_user order by rankpoints1 desc limit 0,20;
+-----------------------+------------------+-----------+
| nick                  | rankpoints1      | rankrank0 |
+-----------------------+------------------+-----------+
| zombie67              | 779110.264264381 |        1 |
| zeitgeistmovie.com    | 739505.166102081 |        2 |
| PoorBoy              | 736562.711482215 |        3 |
| Nasicus              | 734297.886879788 |        4 |
| [B^S] Ralfy          | 709570.241425383 |        5 |
| KAMCOBILL            | 698890.539487113 |        6 |
| vaughan              | 693506.415913394 |        7 |
| Shann                | 681965.846018752 |        8 |
| Ram Raider            |  669054.48357307 |        9 |
| user312              | 666866.927957689 |        10 |
| [B^S] Morgan the Gold | 661070.186967151 |        11 |
| Bok                  | 660826.998077637 |        12 |
| Viking69              |  659650.27956312 |        13 |
| [SETI.USA]Tank_Master | 654519.709749795 |        14 |
| [B^S] Spydermb        | 650183.261296516 |        15 |
| Grizz                | 637378.306663701 |        16 |
| Eliot Meadow          | 634981.054607436 |        17 |
| manu                  | 634102.367273942 |        18 |
| dduggan47            | 623448.526196029 |        19 |
| Sven Giese            | 620794.744410701 |        20 |
+-----------------------+------------------+-----------+
20 rows in set (0.12 sec)

---

bonus points* for anyone who can see what this is :)

Oh! Oh! Oh!
Pick me! Pick me!
I know!
I know!


"THIS" is an English pronoun often used to direct attention to a singular item, object or circumstance

:clap::rotfl::guntotin:

I really do have a proper answer...but I have an infinitesimally small wee tiny unfair advantage so I'd better defer. :rolleyes:

lol, yes you do. It's posted on Ars anyway :)

The question I'm facing though is that is this algorithm th best one to use.

It's not how I used to do it and IMO it's wrong to give the same weighting to a project that has 10000 users as to one that has 100 which is what this does...

Project score = 10000 - ((Position - 1) * (10000 / Total number of users in the project))

Should it not be based on something like

points = (score / total score) * total users

in order to make it ranked on a users percentage of the total score in each project i.e. the effort put into the project?

I think you would have to take the total score and total users into account to get the right math you are after.
Same with Teams, total score and total teams and how much of the pie each team gets.

So, do you think that formula I laid out above would work ? It's the same one I think I used for the DCR way back when or at least very similar...

Taking at random, MindModelling@Home that would give me

(100,027 / 13,377,206) * 2375 = 17.76 points

Lauren in the same project would get

(658,518 / 13,377,206) * 2375 = 116.91 points

whereas for say World Community Grid I would get

(509,922 / 14,888,784,044) * 244,840 = 8.38 points..

Try it and see. Sounds right. :)