| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Suggestions / Feedback
Thread: New team ranking formula??? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 12
|
|
| Author |
|
|
nerd01
Cruncher Joined: Mar 19, 2005 Post Count: 41 Status: Offline |
Some people mentioned recently, that it would be interesting to create a new formula for team rankings, which would be based on number of Members, Run Time, Points Generated and Results Returned...
----------------------------------------Here's mine proposal. First of all we use Global Statistics to calculate: 1. Avg. Run Time Per Calendar Day Per WCG Member 2. Avg. Points Per Calendar Day Per WCG Member 3. Avg. Results Per Calendar Day Per WCG Member That's our base, now we can do the same for every team: 1. Avg. Run Time Per Calendar Day Per Team Member 2. Avg. Points Per Calendar Day Per Team Member 3. Avg. Results Per Calendar Day Per Team Member After that, we should divide: 1. Avg. Run Time Per Calendar Day Per Team Member by Avg. Run Time Per Calendar Day Per WCG Member 2. Avg. Points Per Calendar Day Per Team Member by Avg. Points Per Calendar Day Per WCG Member 3. Avg. Results Per Calendar Day Per Team Member by Avg. Results Per Calendar Day Per WCG Member and take avarage of these three values. The result is our searched factor, different for each team. It allows us to compare results achieved by an avarage team member to those achieved by avarage WCG member. Value less than 1 indicates less involvement/commitment than avarage on WCG, value 1 exact the same and value more than 1 higher. This factor can be also used to compare different teams. However, don't forget that this calculation involves current number of members (on WCG and in teams), so it doesn't tell anything about all-time points, results etc. of teams. It's just kind of indication of how committed members of different teams are... These are the results calculated using statistics from 15.04 for first 5 teams: Slashdot Users......1,41 IBM Corporation....2,38 IBM Rochester.......7,28 Canada.................1,91 MyOnlineTeam.......4,21 TribalWar.com.......1,44 What do You think about it? Waiting for some feedback...  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Good work nerd01
 ,This sure would show team member commitment and it would be nice to see these stats, at least alongside the current ones. With the existing project and stats having run for some time, unfortunately, I am not sure whether WCG will adopt ANY suggested changes to scoring/ranking that will replace the current format .. (I can just see the flood of protest from those teams that suddenly find the numbers aren't so good  ). It sure would give added motivation and increase competition though!!Can any Community Admins reading this please put it to WCG Tech / Admin (particularly 'nelsoc') for consideration  , as something to run alongside current stats. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
SAKU,
As somewhat of a statmonger myself, I can say that we will definitely take this into consideration. As somewhat of a WCG tech myself, I can say that it will be pretty low on the To Do list...  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hey nelsoc,
Thanks for the reply .. kinda expected that  If you guys do adopt some variation of this in the future, all the credit should go to my MOT team mate 'nerd01'. |
||
|
|
nerd01
Cruncher Joined: Mar 19, 2005 Post Count: 41 Status: Offline |
Thanks for interest guys!
---------------------------------------- I'll try to post here updated results maybe every week or two... Regards, |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Some people mentioned recently, that it would be interesting to create a new formula for team rankings, which would be based on number of Members, Run Time, Points Generated and Results Returned... Here's mine proposal. First of all we use Global Statistics to calculate: 1. Avg. Run Time Per Calendar Day Per WCG Member 2. Avg. Points Per Calendar Day Per WCG Member 3. Avg. Results Per Calendar Day Per WCG Member That's our base, now we can do the same for every team: 1. Avg. Run Time Per Calendar Day Per Team Member 2. Avg. Points Per Calendar Day Per Team Member 3. Avg. Results Per Calendar Day Per Team Member After that, we should divide: 1. Avg. Run Time Per Calendar Day Per Team Member by Avg. Run Time Per Calendar Day Per WCG Member 2. Avg. Points Per Calendar Day Per Team Member by Avg. Points Per Calendar Day Per WCG Member 3. Avg. Results Per Calendar Day Per Team Member by Avg. Results Per Calendar Day Per WCG Member and take average of these three values. The result is our searched factor, different for each team. It allows us to compare results achieved by an average team member to those achieved by average WCG member. Value less than 1 indicates less involvement/commitment than average on WCG, value 1 exact the same and value more than 1 higher. This factor can be also used to compare different teams. However, don't forget that this calculation involves current number of members (on WCG and in teams), so it doesn't tell anything about all-time points, results etc. of teams. It's just kind of indication of how committed members of different teams are... These are the results calculated using statistics from 15.04 for first 5 teams: Slashdot Users......1,41 IBM Corporation....2,38 IBM Rochester.......7,28 Canada.................1,91 MyOnlineTeam.......4,21 TribalWar.com.......1,44 What do You think about it? Waiting for some feedback... While i can see the motivation and it isn't a bad idea as such... it is quite negative to alot of members of the WCG (all the little men). It highly favours those people with multiple boxes. A person who runs it 24 7 and has no access to boxes other then their own can be just as dedicated as someone who does have access to several boxes. Hell they would be more dedicated to WCG then a user who has access to multiple boxes, but only runs them sometimes. That person would rank well in this stat for their team, but really are not as dedicated to WCG. So while it is an interesting metric... it does reflect badly of the vast bulk of WCG members and teams. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Maybe nerd01 would like to come up with a way to calculate points per member per machine. This would take care of Beren's issue of some members having multiple machines but only running them part of the time.
Of course, then we run into the problem of having a new device created every time the agent is uninstalled/reinstalled. Since that is a common work around and/or method of solving problems with the Agent, it often turns out that members have many more "devices" as listed in their member profile than they actual have running the Agent (eg: I have 9 devices in my profile, but only 3 computers). (Also: we are working on a way to merge devices, so the device number reflects the actual number of devices. It'll get done... eventually) |
||
|
|
nerd01
Cruncher Joined: Mar 19, 2005 Post Count: 41 Status: Offline |
Well, you're right Beren. I created this formula only for teams, not for members... And it's true that it works best for top teams.
----------------------------------------I'll try to create a formula that would reflect more dedication/commitement than the last one. However, we should remember that people with multiple devices can process more data, so their input in the project is actually bigger... I think we have to find a compromise between power and dedication, or create just two different formulas... Regards, |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Well, you're right Beren. I created this formula only for teams, not for members... And it's true that it works best for top teams. I'll try to create a formula that would reflect more dedication/commitement than the last one. However, we should remember that people with multiple devices can process more data, so their input in the project is actually bigger... I think we have to find a compromise between power and dedication, or create just two different formulas... Regards, There are so many variables, if you attempt to account for all of them you will simply end up with gobble-de-gook. We need to focus on increasing the competitive and enjoyment elements rather than attempting to "hatch the perfect formula." Besides, there are so many equally important intangibles that can not be quantified such as Graham's banner fetish; certain WCG Techs' committment to excellence in technical support; the contribution of uplifting and stimulating posts by certain members ... ad nauseum. There are no awards. Just a lot of rewards. And that's just fine.   |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Besides, there are so many equally important intangibles that can not be quantified such as Graham's banner fetish;  |
||
|
|
|