The changes were announced in the normal way.

I can only reiterate: WCG have never supported any unofficial clients.

You use an unofficial client entirely at your own risk. Sorry.

During the first trial of the modified point system, we found that one of Didactylos' computers running BOINC as a service was claiming the normal AMD Athlon benchmark scores, but seemed to run the science application only half as fast as it should. So the points claimed per hour were fine, but the points per result were inflated. The new system caught this and awarded him only half-credit. (Half of what others claimed - only a quarter of what his system was claiming.)

So how to explain this?

Any change is going to cause disruption. We only made the change because so many of our members were feeling about the way that things were.

Our goal is to end up with a system that our members feel is fairer.

Of course, there are still a few steps to go.

Lawrence

Carl, in http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=9545 , Kevin point on 11/1 that they were going to implement the new system after the successful completion of the beta of it. On 11/6, he updated that thread to say that the beta was successful and that they were implementing the new system on all projects. Okay, you only had five days there to see that thread in Member News where they normally post announcements. If you don't check in say once a week, then yes, you could have missed this before it started. How much notice is enough notice? If you say a week is enough, what about the folks that only look at the forums once or twice a month? What about those that don't look at all? Okay, send an email to everyone. I suspect there that a decent percentage of the email addresses WCG has aren't accurate. If you rely on email, how far do you have to go to make sure folks keep a current email address on file?

I'm not disagreeing with your point but what's the one answer that avoids the potential for any "carls"?

Spot on Sek. All I have at this point is what I've "noticed" and I without question need to pull hard data on my past WU's to see exactly what if any difference there is. That's why I thinking of it in terms of points per unit of runtime. I want to try to eliminate skewing factors and get it down to the point of "my machine crunches X amount per runtime unit and that's worth Z points for my machine's benchmarks. No, I expect I won't be able to eliminate all of the skewing factors but I'd prefer to pull out as many as I can so I get this right the first time. I don't want to go saying running project A vs. project B on the same machine will get you more points, especially with the current "mood of the masses". I do know that we have a lot of smart folks out there and I'm more that willing to consider their input.

I know I want to look at my points claimed and not awarded. I also want to look at it in terms of CPU time, not clock time. I know I want to avoid considering "per WU" as those clearly vary - especially lately - I've had some HDC WU's run almost as long as is typical for FAAH.