As most of you know cross project credit parity (for the combined stats) is a bit out of sync lately, with new projects popping up which grant generous amounts of credits or rogue optimized apps that do nothing more than inflate credit. Some people care, some don’t, some are all for it, and all groups fight over it. In my opinion it makes the combined stats kind of meaningless, so a solution is needed.

Another problem is that new users or teams almost have no chance to get top ranked. With the solution I came up with, they can be number one.

I came up with a new way to calculate credit and generate combined stats in addition to the existing combined stats.
The current way of doing combined stats will not disappear!

I call the new credits “points” and BOINCstats will use the points in the BOINCstats World Cup.

How the World Cup works:
1: The total credit generated by a project in the current month is calculated.
2: For every team and user the credit for the current month within that project is calculated
3: The contribution to the total credit is calculated as a percentage
4: This percentage is multiplied by the number of active teams or users of that project
5: And then we have World Cup points

Example:
Project output this month: 1000000 credits
Team XyZ contributed 15000 credits, which is 1.5%
There are 132 active teams, so 1.5*132 = 198 points

Ranking is based on the points.

By using this kind of calculations I hope to increase cross project parity while also taking into account the size of a project.

For combined stats, the points of the projects are summed.

Every month will be a cup on its own. This means, every month another race to the top!
Every year the points of all months are summed resulting in the final ranking for that year.


Now I know this is quite a step away from the original stats. It will need some tinkering with the calculations to make it completely fair. The current World Cup shows the usual suspects in top ranking positions which should be OK.

Note that the World Cup is far from complete. Detailed stats pages have to be updated, compare graphs need to be added, and much much more, but I wanted to release now so you can offer your input and feedback. I’m sure the World Cup can be improved and you can help just doing that.

So here's your chance. Speak now and influence how the World Cup works or remain silent forever.

The World Cup stats can be found by opening a project menu on the left.

____________

Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.

Interesting!! Good way of getting an idea of crunching EFFORT rather than crunching MUSCLE. Nice idea, Willy. Might have to steal this one from yah!!
____________
Seejay *Proud Founder & Member of BOINC Team Allprojectstats.com*

Nice idea, Willy. Might have to steal this one from yah!!

Nice idea, Willy. Might have to steal this one from yah!!

You don't have to steal, I intend to export the stats once they mature.

Interesting approach. Sounds good.

A very fine Idea.
Greats Schmitzbauer

Looking good!

I can't find where my total is though. As far as I can see there's only the 'daily' totals visible, not a combined total.

If it's not there we need a linkable total on the combined pages - one for individual totals and one for teams totals.
____________
Member of The UKBOINC team.

Join us here




Join the UK BOINC Team: http://www.ukboincteam.org.uk/uk-boinc-team.html

Visit the UK BOINC Team forum: [url]http://www.ukboinc

Combined totals is under BOINC combined.

The detailed stats page (for your combined stats) has a pie at the very bottom showing how credit is spread over projects.

I will add more info to the detailed stats pages so you can quickly find your rank.
____________

Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.

so you do separate calcs for team total points vs user total points? or is team (no team) also included in team point calcs? or is team points just a sum of user points?

over simlified example according to the bried description

total daily points per project a 10000
total users 200
total teams 10

so team a has 5 users
team a generates 500 points or 5%
5 * 10 = 50 points

user 1 of team a generates 50 points or .5%
.5 * 200 = 100 points

just looking for some clarification?

................ change is not good when previoulsly unannounced.

@Evilcowgod
No, team credit is not an accumulation of team-member credits (mainly because that will cause errors because users can join/leave a team in the middle of the month) so yes, it's calculated separately.

Your calculations are correct (as in, the same as they are calculated for the World Cup).
____________

Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.

................ change is not good when previoulsly unannounced.

.... OK - I looked up 'previously' (...........and change) ...... I like your Stats site Willy. Let's not quibble.

@Evilcowgod
No, team credit is not an accumulation of team-member credits (mainly because that will cause errors because users can join/leave a team in the middle of the month) so yes, it's calculated separately.

Your calculations are correct (as in, the same as they are calculated for the World Cup).

I think i get it, but my world cup shows that i have over 50% on Mindmodeling @ home, does this mean that i am producing 50% of all the work this project is getting?

I think i get it, but my world cup shows that i have over 50% on Mindmodeling @ home, does this mean that i am producing 50% of all the work this project is getting?

Intersting idea!

I especially like the idea of it being monthly.

Rod

i assume team credits are passed as a prebuilt sum from the project rather than you calculating it.

but wouldnt something along these lines be closer to a sum of users.
(continued from my example below)
team a 500 credits/10000 total credits = 5%
5 * 200 users = 1000 points

point allotment will still fluctuate based on total active users but will be standardized rather than 2 independant scales.

I think i get it, but my world cup shows that i have over 50% on Mindmodeling @ home, does this mean that i am producing 50% of all the work this project is getting?

That means that Mindmodeling delivers 50% of your combined cup credits.

i assume team credits are passed as a prebuilt sum from the project rather than you calculating it.

but wouldnt something along these lines be closer to a sum of users.
(continued from my example below)
team a 500 credits/10000 total credits = 5%
5 * 200 users = 1000 points

point allotment will still fluctuate based on total active users but will be standardized rather than 2 independant scales.

If you are suggesting to calculate the team credits based on active team-members (that's how I understand you) then I think it's a bad idea. Big teams would get an enormous boast, even if only one member produces lot's of credits and the others produce 1 credit per day, while a small team which produces the same amount of credit would get much less.

hope that makes more sense.

Interesting approach. So far I like it. I would like to see it run for a couple months to see how people/teams drift up and down the rankings before making a final decision.

Sooooo, got a question

Im looking at my daily stats, it said I did no work yesterday on SIMAP... is that because of the issue that stats page had yesterday or is that because of this cup points.. Cause it says I received "cup points" for it.

I personally dont see the need for this.. I really liked BOINCstats the way it was.. But we will see.

hope that makes more sense.

Yes, I get it.

OK, I'll consider this. But in this case it's just that team points are too low compared to user points. For the team ranking it doesn't matter.

Very interesting idea, let's see whether Willy (and this community!) can get the system going

nice surprise, looks like a nice additional ranking to work on
any chance you're going to create signature images for this?

Sooooo, got a question

Im looking at my daily stats, it said I did no work yesterday on SIMAP... is that because of the issue that stats page had yesterday or is that because of this cup points.. Cause it says I received "cup points" for it.

Sooooo, got a question

Im looking at my daily stats, it said I did no work yesterday on SIMAP... is that because of the issue that stats page had yesterday or is that because of this cup points.. Cause it says I received "cup points" for it.

i also got 0 points on simap today - due to the trouble yesterday. and nothing is telling me that i got any cpp-points on my stats pages..

i also got 0 points on simap today - due to the trouble yesterday. and nothing is telling me that i got any cpp-points on my stats pages..

You should have SIMAP cup credits (you got credit this month), but they're not showing. Will look into that.

I think it's a neat idea although I don't quite see how this is supposed to level the playing field... maybe just give another measure of credits/RAC. Did I miss something or is there a BOINC cup points total for a single user? I looked for it on my page but didn't see it.

One more thing: my LHC@home account is getting 224 cup points but I think it's only gotten like 5 credits in the past week or something like that. Small projects like that can inflate your new-fangled calculations.

Great idea. Let's see how that works out

Something that would make it easier for everyone to judge would be, if next to the "position" field in the ranking there would be a "position cred" field, showing the position from the "old" credit ranking to get an impression of how much this differs.

Thanks!
____________
ThEfT
BOINC@Heidelberg -Admin
http://www.petitiononline.com/phletter/petition.html
http://www.boinc-team.de

Doing stats on the percentage of contribution per project of users/teams rather than simply adding the credits is a good idea as over-crediting has no influence.
Putting in place monthly and yearly rankings is also a good thing as it allows the new comers to have a chance to be ranked.

But, why do you want to multiply these percentages by the number of users/teams of each project ?

There are plenty of users/teams appearing and disappearing every day on some projects, do you really think such numbers reflects the size of a project ?
If two projects are producing approximately the same amount of work, I can't see why one of them should grant much more points than the other one just because the first one has many small crunchers while the other one has few but big crunchers or worse, just because the users are spread into many teams.

Moreover, it opens an easy way of cheating. Just imagine I am a big cruncher with a big farm of computer crunching a significant amount of work on a small project.
To improve my world cup ranking, I can, on one of my host, create a new user per wu crunched. This would just decrease a little bit my overall production but considerably increase the number of active users on the project and so my cup points. I'm pretty sure that this can be done automatically by script. Same for the teams, a lot of fake teams with one single new user for only one single unit can easily be created.

Also, before putting in place a new ranking system, it is important to think about the impact it may have on the crunchers behaviour.
The impact of the ranking by adding the overall credits is that some people are attracted by the projects that are crediting the more.
This can be a good thing when it is an incentive for the people to crunch preferably on projects optimized for their platform, to choose some specific hardware ( like cards for GPUs ) or for the project to optimize their apps or to go to GPU. But, this is a bad thing when projects artificially increase the credit granted just to be more attractive.

Doing stats by adding percentage of contribution per project ( without the multiplication ) may also be a good thing as it would be an incentive for people to crunch on the small projects ( if the power on a “monster” like Seti or WCG decrease by 1% it does not change anything for them but, with this 1% you can double or more the power on many small projects )

But here, the incentive you put is to try to crunch on the projects that have the more small or dummy users/teams. What a challenge!
Did you make stats on the difference of cup points between the projects ? Do you know what projects will benefit from this new method ? If you did not, I’m sure there are already some persons doing such comparisons and thinking about optimizing their cup ranking.

Willy, I like your idea of BOINCstats World Cup.
Some crucial advantages over Formula Boinc like monthly new chances for new teams/users and not only top 8 is interesting.
As an extra the opportunity to arrange monthly challenges on certain projects by teams/users or users within teams.

Just a suggestion for the statictics displays:

Could you truncate the decimals (displaying only the integers) and align the figure to the right in the columns.
You will get more place for display at least six columns of historical data during that month.
____________

@darealgh:
You have some very good points. yes, unfortunately, some people will try to cheat and I didn't think about that. This will of course change the combined ranking considerably.

One other reasons I did the multiplication is to enlarge the points. Just a percentage would be a very small number. Any ideas about the factor?
____________

Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.

Added the world cup to the search so you can quickly find your rank.
____________

Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.

Added the world cup to the search so you can quickly find your rank.

Added the world cup to the search so you can quickly find your rank.

nice - but where? i can't find a search function here: http://boincstats.com/stats/world_cup_stats.php?pr=bo&object=user&st=0

Added the world cup to the search so you can quickly find your rank.

nice - but where? i can't find a search function here: http://boincstats.com/stats/world_cup_stats.php?pr=bo&object=user&st=0

Uhm I added the World Cup to the search, not a search to the World Cup. Try hitting search at the top of the page.

that was my first try - i ended up there: http://boincstats.com/search/result.php and no sign of cups. but maybe this is my page again..

To quote one of my favorite TV-shows "Excuse me, are you from the past?"

So you always stop at the result page and never go to the page that shows your credits??

Here ya go: your project overview.

Top right, find the link "World Cup position", click it, and behold, Cup Points.

now i got it - as i told you before, i'm much too old for playing paper chase...

Be prepared for some pretty hefty point changes in the World Cup after the next daily update. I changed the calculations based on some of your suggestions.

This will not influence ranking positions in the projects as those are based on work done, but it will change rankings in combined quite a bit I think.
____________

Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.

@[BOINCstats] Willy

It depends on what you want to highlight by your new system of stats and what behaviours you want to encourage/discourage.

You can simply add the percetages mutiplying them by 1000 10^6 or 10^9 if they are too small. The advantage of that is giving all the project whether they are overcrediting or not the same importance. The drawback of that is the strong incentive given to crunchon the small projects or on those the other ones are not crunching. This may even lead to servers overloading or crashes when big teams are racing for that.

A solution to mitigate that other may be to multiply by a factor depending on the project category size ( like multiply by 4 the score of projects in top 5; by 3 the one for those in top 10 ; by 2 the score for those in first half; and all others by 1 ).

If you really want to keep your multiplication by users/teams, I suggest you at least remove all the "small" users/teams keeping for example only those who did at least 10000 credits or 0.1% of the total but this means only granting points to the best crunchers. I anyway don't like the idea of multiplying by the number of teams as, if the number of users may somehow show the importance of a project ( a number of host or better of cores is probably more significant ), a number of team does not mean anything.

@[BOINCstats] Willy

It depends on what you want to highlight by your new system of stats and what behaviours you want to encourage/discourage.

You can simply add the percetages mutiplying them by 1000 10^6 or 10^9 if they are too small. The advantage of that is giving all the project whether they are overcrediting or not the same importance. The drawback of that is the strong incentive given to crunchon the small projects or on those the other ones are not crunching. This may even lead to servers overloading or crashes when big teams are racing for that.

A solution to mitigate that other may be to multiply by a factor depending on the project category size ( like multiply by 4 the score of projects in top 5; by 3 the one for those in top 10 ; by 2 the score for those in first half; and all others by 1 ).

If you really want to keep your multiplication by users/teams, I suggest you at least remove all the "small" users/teams keeping for example only those who did at least 10000 credits or 0.1% of the total but this means only granting points to the best crunchers. I anyway don't like the idea of multiplying by the number of teams as, if the number of users may somehow show the importance of a project ( a number of host or better of cores is probably more significant ), a number of team does not mean anything.

Anybody (dis)agrees with this idea?

If I understand you correctly you want to multiply the points by the percentage of credits the project contributed to combined during the last month?
____________

Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.

If I understand you correctly you want to multiply the points by the percentage of credits the project contributed to combined during the last month?

It's a nice idea with the World Cup Points.
Can we expect separate world cup sigs for that also?
____________
I am an Aperture Scientist. The cake is a lie, believe me!!!

Can we expect separate world cup sigs for that also?

Because the world cup stats are Beta I'll place a bug here and not in the bugreports.
When you asked for world cup position on your user page, the cup points displayed are the same as total credit.
E.g.
Project overview for [BOINCstats] Willy total credits - and
Project overview for [BOINCstats] Willy World Cup position
____________

Because the world cup stats are Beta I'll place a bug here and not in the bugreports.
When you asked for world cup position on your user page, the cup points displayed are the same as total credit.

Can we expect separate world cup sigs for that also?

Once we are happy with the stats.

Willy, are you happy with the cup stats as they are now.
I'm not. The outcome of the world cup stats are ridiculous now.
There should be build in a weighting factor oriented on the size of the project as I suggested already earlier in this thread.
The minor, tiny projects have a very heavy weight on the combined cup stats.
Just to show you a random example nr. 3 in BOINC total combined cup stats.
Three high positions in minor projects (in credits and number of users) and you are number 3 in BOINC combined world cup.

Willy, are you happy with the cup stats as they are now.
I'm not. The outcome of the world cup stats are ridiculous now.
There should be build in a weighting factor oriented on the size of the project as I suggested already earlier in this thread.
The minor, tiny projects have a very heavy weight on the combined cup stats.
Just to show you a random example nr. 3 in BOINC total combined cup stats.
Three high positions in minor projects (in credits and number of users) and you are number 3 in BOINC combined world cup.

I noticed it, and no, I'm not too happy with it. This is the result of working with percentages instead of credit. On the other hand, Formula BOINC works on the same principle. There points from smaller projects have the same weight as larger projects. I guess that would be OK when all teams/users participate in all projects.

OK, so we can either use credits instead of percentage or let the size of the project matter. If we go for size of project, then what is the measure? Total credit, participants, teams?

Any thought about this is appreciated.

OK, we'll try that. But it's already late here and I don't want to do a quick fix so I code it tomorrow evening. Tomorrows cup stats will have the same calculation as today.
____________

Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.

I find your approach disappointing in that you weight my credit differently between my projects because these are valued differently. I don't because x units of CPU time on one project is the same as x units of CPU time provided by my systems regardless of the project. Rather the value is how much of the CPU time is dedicated to Boinc.

It is not sufficient just to consider project size. You have to adjust for workload. For example as the March SIMAP draws to a close, there will be no more work until April. So that means over credit for the first week of the week and under credit for the rest of the month. Other projects have limited work that also mean that credit is limited and it is not fair to consider these the same as those with with unlimited work. Further this ignores that people allocate different time to different projects. You also assume that all projects give the same credit for the same work (which includes CPU vs GPU and multiple processors and cores) but there is not much you can do about that.

It is hard to know what you are trying to do here and what you want to achieve in terms of a handicap. At present the information given just appears to allow users and teams with more computer power to do better than others with less. I am surprised that you are not measuring the monthly change in the project or user credit.

I am surprised that you are not measuring the monthly change in the project or user credit.

My proposal is to use the already calculated percentage of the projects during the last month.

My proposal is to use the already calculated percentage of the projects during the last month.

There's a drawback to this method. Again you will let the 'old' credit get into the equation. Calculating this way will result in over-awarding projects looking bigger than they actually are and under-awarding projects looking smaller than they actually are. That's exactly what you were trying to avoid in the first place.

Not that I have an alternative though.

So let's say there are 2 exactly the same projects. They both receive the same amount of work done by their users in terms of cpu-power. But one of the two projects rewards twice the amount of credit to it's users. The total credit of that project will be twice as high as the other one. As a result it will make this project look twice as big as the other one in that pie-chart, and thus you will receive twice the amount of cup-points for it, with the same amount of work done.

____________

I have a new suggestion. Forgive me if this has been brought up already. It seems to me, Willy, you are looking for a way to level rankings so those of us with few machines to run can see how we'd compare with those with large farms. Let those crunching large have their ranking in the regular stats. How about simply dividing everyone's credits by however many active hosts they have ? Then do the straight percentages for the month to get the rankings.

I have a new suggestion. Forgive me if this has been brought up already. It seems to me, Willy, you are looking for a way to level rankings so those of us with few machines to run can see how we'd compare with those with large farms. Let those crunching large have their ranking in the regular stats. How about simply dividing everyone's credits by however many active hosts they have ? Then do the straight percentages for the month to get the rankings.

Why not just exclude them from this and promote openness (that a word?)

I have a new suggestion. Forgive me if this has been brought up already. It seems to me, Willy, you are looking for a way to level rankings so those of us with few machines to run can see how we'd compare with those with large farms. Let those crunching large have their ranking in the regular stats. How about simply dividing everyone's credits by however many active hosts they have ? Then do the straight percentages for the month to get the rankings.

Lets forget this is about credit normalization/CPP. it isn't. It's not possible either. It's about competitiveness.

Another problem with the host ID is the number of projects your spread over. Not all project get the same amount of CPU time and it would profit people who are spread across a lot of projects.

Therefor project WC stats are based on total credit produced in the current month without any manipulation of the credits.

Back to this is not CPP:
I know there is no way to normalize the credits in a way that's completely fair. But that was not the original idea. For the combined stats this is my new suggestion:
We divide 1 by the rank in every project of a user/team in the WC-stats. No rank = zero point. Then we sum up these numbers and multiply the rsult by 1000000 (this number would give every user a point score above 1).

So, if we would have 10 projects:
Team A ranks: 5, 3, 7, 1, 4, 9, 5, 2, 5, 1:
1/5 = 0.2
1/3 = 0.333333
1/7 = 0.142857
1/1 = 1
1/4 = 0.25
1/9 = 0.111111
1/5 = 0.2
1/2 = 0.5
1/5 = 0.2
1/1 = 1
---------------
3.937301 * 1000000 = 3937301 cup points

Team B ranks: 2, 9, 4, 6, 2, 2, 7, 1, 8, x: (x= no rank)
1/2 = 0.5
1/9 = 0.111111
1/4 = 0.25
1/6 = 0.166667
1/2 = 0.5
1/2 = 0.5
1/7 = 0.142857
1/1 = 1
1/8 = 0.125
1/x = 0
--------------
3.295635 * 1000000 = 3295635 cup points

Team C ranks: 64, 80, 34, 150, x, 70, 230, 91, 341, 111
1/64 = 0.015625
1/80 = 0.0125
1/34 = 0.029411
1/150 = 0.006667
1/x = 0
1/70 = 0.014285
1/230 = 0.004347
1/91 = 0.010989
1/341 = 0.002932
1/111 = 0.009009
----------------
0.105768 * 1000000 = 105768

Ranking: team A , team B, team C.

In a case where two users/teams have the exact same amount of cup points, the one with the highest RAC will be higher ranked.

There will be some of you who will say "So I go to some small projects and crunch the hell out of it and I get the best rank in combined". This is true. But the other users/teams can see you do that and can either crunch the hell out of another project or attack you in the project you picked. We can all run all projects and if someone targets a small project and we all do nothing about it, then he's the clear winner.

So, to be clear:
Combined WC stats is not based on how many credits you are able to crunch in all projects, it is based on your strategy to be high(er) ranked in as many projects as possible.
____________

Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.

Good suggestion, Willy. It's all about science and competition.
I suppose in the next daily update your new kind of calculations will be implemented or did you implemented it already yesterday for the combined stats?
____________

I suppose in the next daily update your new kind of calculations will be implemented.

So, to be clear:
Combined WC stats is not based on how many credits you are able to crunch in all projects, it is based on your strategy to be high(er) ranked in as many projects as possible.

Hello,
the world cup graph is useless.
It shows the cup rank instead of the cup points.

Legolas13
____________

Hello,
the world cup graph is useless.
It shows the cup rank instead of the cup points.

Legolas13

Hello,
the world cup graph is useless.
It shows the cup rank instead of the cup points.

Legolas13

Probably a little late for clarification, but what he meant is that the world cup graph at the bottom of the page shows positions, not cup points. So for me, it looks like SETI@Home is my biggest contributor, but it's actually the project where I'm lowest in the standings.

Hello,
the world cup graph is useless.
It shows the cup rank instead of the cup points.

Legolas13

Probably a little late for clarification, but what he meant is that the world cup graph at the bottom of the page shows positions, not cup points. So for me, it looks like SETI@Home is my biggest contributor, but it's actually the project where I'm lowest in the standings.

Jedirock (& Legolas13): On the 22nd of March this issue has fixed by Willy as you can see on your personal graph page.

Well, I'm happy with the WC rankings.

As of 9:00 PM UTC today (the time of posting this message), I'm in position:

977 for WC
2,044 for RAC
22,619 for credit

out of 1,638,046

I think I'll be inclined to quote my WC rank at the moment (although it sounds even better if I say850th for International, but that's out of less people).

I noticed one tiny error and probably one that is low on the list of priorities: A world cup pie chart with two wedges of the same color: http://boincstats.com/charts/chart_uk_bo_object_pie3dcuppoints_teams_2188929.gif

These World Cup Points make absolutely no sense at all....

It seems to put high percentages to projects I've actually abandoned long ago and ranks those I continue to work with very low. As an example it shows my World Cup pie chart % for Mind Modeling at over 40% of the chart. I haven't done anything with that project for some time and only accumulated 4 credits when I did.

So what is it these points are supposed to represent? I can't see where they are of use as they don't reflect anything but the lowest points accrued that I can see.

How can one lose world cup points?
From 20090323 to 20090329 my total has dropped by roughly 5,000 points and my position commensurately by 4. Am I correct in thinking that this is just problematic with respect to the random online/offline activity of the projects I am actively engaged in on any given day? During this period Enigma, Einstein and FreeHAL all had prolonged server disturbances and all active completed wu's were unreportable.
If this is a proper explanation, I would think that the total points would tend to realign and adjust the rank back to where it was prior to the offline activity of the projects involved. I guess 20090331 will prove or disprove the validity of my thinking, I look forward to it.


bill brandt-gasuen
Team ACC - Arthur C Clarke Fans

How can one lose world cup points?
From 20090323 to 20090329 my total has dropped by roughly 5,000 points and my position commensurately by 4. Am I correct in thinking that this is just problematic with respect to the random online/offline activity of the projects I am actively engaged in on any given day? During this period Enigma, Einstein and FreeHAL all had prolonged server disturbances and all active completed wu's were unreportable.
If this is a proper explanation, I would think that the total points would tend to realign and adjust the rank back to where it was prior to the offline activity of the projects involved. I guess 20090331 will prove or disprove the validity of my thinking, I look forward to it.


bill brandt-gasuen
Team ACC - Arthur C Clarke Fans

How can one lose world cup points?
From 20090323 to 20090329 my total has dropped by roughly 5,000 points and my position commensurately by 4. Am I correct in thinking that this is just problematic with respect to the random online/offline activity of the projects I am actively engaged in on any given day? During this period Enigma, Einstein and FreeHAL all had prolonged server disturbances and all active completed wu's were unreportable.
If this is a proper explanation, I would think that the total points would tend to realign and adjust the rank back to where it was prior to the offline activity of the projects involved. I guess 20090331 will prove or disprove the validity of my thinking, I look forward to it.


bill brandt-gasuen
Team ACC - Arthur C Clarke Fans

Did you lose credit in combined or in a project?

Combined is based on your ranking in the projects thus if you rank lower, you get lesser points in combined.

Just a reminder as already explained in this thread:
Points is 1 divided by your position times 1,000,000. Your FreeHAL position 55 means 1/55*1,000,000= 18182 points.
So movements in the higher ranks have an enormous influence on the points: position 1 get 1000000, position 2 500000, position 3 333,333 and so on.

Made a small change:

A project needs to have at least 50.000 credits granted in the current month to be included in the combined stats.
____________

Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.

Well, I dropped to 922 today in the world cup, which makes far more sense. Given that a large chunk of my rating was from hydrogen@home, which is in early alpha, closed membership and not yet able to supply huge amounts of work for anyone It was nice being recognised for my contribution there, but it is the one project I work on for the science, opposed to all the others where credits and RAC are just as important

WorldCup stats are now available for download.

Download directory is http://boincstats.com/export/, the last update time and an overview of files is in http://boincstats.com/export/tables.xml.

Stats are exported once a day after the daily update. Please check the last update time before downloading.
____________

Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.