Quote Originally Posted by Bok View Post
That's an interesting notion, not sure how we'd categorize an overall MM, perhaps lowest average projrank?

Code:
mysql> select cpid,avg(projrank) as avpr from boinc_milestone_makers where mtype like 't%' group by cpid order by avpr asc limit 0,20;
+-------------------------+---------+
| cpid                    | avpr    |
+-------------------------+---------+
| L'Alliance Francophone  |  1.8182 |
| SETI.Germany            |  1.9091 |
| SETI.USA                |  2.0909 |
| Czech National Team     |  5.8000 |
| BOINC Synergy           |  7.2222 |
| Ars Technica            |  8.3000 |
| BOINC@Poland            |  8.3000 |
| BOINC@AUSTRALIA         |  8.3333 |
| Planet 3DNow!           |  8.3636 |
| BOINCstats              |  8.4000 |
| Team 2ch                | 11.3636 |
| The Knights Who Say Ni! | 12.6667 |
| Team Norway             | 13.5556 |
| Canada                  | 13.8750 |
| AMD Users               | 15.5556 |
| UK BOINC Team           | 16.8889 |
| Russia                  | 17.6667 |
| TeAm AnandTech          | 18.5000 |
| USA                     | 19.0000 |
| BOINC.Italy             | 19.1250 |
+-------------------------+---------+
20 rows in set (0.01 sec)

I think you need to do an outer join against the milestones where a null equates to the max value so that it gives a truer picture. (e.g. Team USA is ranked #19 overall above when they only have 7 MM's. I'm not sure that holds up against other teams which have more MM's (11 or more?) but lower numbers.

For example, if team X has less than a billion credits, then the ranking for the 1B MM would be the count of teams that have 1B plus one (e.g. tied for last) rather than not counted in the stats at all.