PG's use of points to influence how the science is done

[rhavern]

Okay, some people have been banging on about "points inflation" and how they don't like it. Let's be clear, there is no points inflation. Points are awarded by PG as a method of influencing how the science gets done. We've been told repeatedly that PG prefers fewer and faster over more and slower. Now they are changing the point scheme to match that request.

Historically, there was no way of guiding users down the fewer/faster path. Along came the GPU client with comparitively short deadlines, days instead of weeks. No point incentive but restricted deadlines. Then we got the QRB that rewards the behavior that PG told us they preferred. Apply that to the bigadv where we now have restricted deadlines to keep the project on hardware that can do the science in the PG desired timeframe and the incentive of the the QRB.

From PG's perspective, that is science nirvana. If you have a research paper that needs to be completed in a short time, put it on bigadv and job done. No rush, then put it on standard FAH. This is not points inflation, it is the use of points to guide hardware to the appropriate work and to further PG's research.

It all comes down to a simple choice, do you believe in PG's project? If yes, get folding and maximize your points, because that maximizes the science. If no, then thanks for stopping by, nice to meet you, don't fold and have a nice day.

Disclaimer: I don't work for PG but I have been folding for over a decade and am a beta tester.

tl;dr Stop whining about points, PG set them according to their requirements. Accept and press on.

[rwh202]

Although you rightly explain the need to have a range in ppd to encourage/direct processing to meet the needs of the science, I think there is also another aspect.

The real inflation issue is that crunching with the best consumer hardware back in 2005 might have got you a few hundred ppd, doing the same today nets you 100k ppd.

Obviously modern hardware does more work, but is that any more valuable to todays scientists than the smaller output was to yesteryear's?

The inflation affects (competitive) donors in that previous year's efforts are effectively rendered worthless when points are that much 'cheaper' today.

I agree that this a tricky aspect, especially as people wouldn't want to see the alternative of identical hardware dropping away in ppd as time goes on, but at least previous efforts would maintain some value.

[7im]

Thank you. Well said. At times, I agree with both positions. However lately, expediancy seems to be winning over other better methods to resolve issues, and that is where some of my concerns have put me solidly in the second camp for now.

[Adak]

Obviously modern hardware does more work, but is that any more valuable to todays scientists than the smaller output was to yesteryear's?

Yes -- more work is being done, and yes DEFINITELY, more work is more valuable to the scientists today, than the smaller output was in yesteryear's.

When FAH started, we folded smaller proteins. That was all we could simulate folding, due to the limitations of the hardware. Today, we simulate much more scientifically valuable proteins, which are much larger, because we now have the computer power to make that possible.

Yes, there has been points inflation. Most of it has been caused simply by hardware improvements, imo. Some of it has been overdue, to get FAH points to more directly equal Pande Group needs.

I know we get wrapped up in concerns about points, and I do it also. But let's keep our eyes equally on the research. It's fun playing with the points, but we are here to serve the research, with the leadership of Pande group. Points should be an interesting and fun sideline to folding, but not our biggest concern.

[bruce]

Some people will argue about anything.

In this case, it has been argued to death and we don't need any more input.

Each one of you claims to know what the Pande Group needs and then states an unsupportable opinion about the points system. How many times more valuable is a BigAdv WU compared to the 110 PPD that I still get on a P4 @ 2.8 GHz when running core_78. Should it be 50x or 200x or 5000x or 200000x or some other number? I don't know and none of you really do either.

The same question applies to a question of how many times the salary of the average factory worker should the CEO earn. That becomes a politically charged question, not one which will help with the primary function of this forum, which is to help folks who are having trouble getting FAH to do what it should be doing.

I'm closing this topic.