[FAD] OCt 02 query

[tmx468]

Why do the listings either have all or since oct 02... and why do some teams lose millions of molecules when from that date... what is so special about it?

[RustyGT]

Gday tmx.

I'm sure someone else will have the complete answer.
But I suspect there was a change in the software about then.

You'll see that folks that started the project early.
Have very high molecule/job ratios.
Rather than the standard 10K mollies per job.

You might say that at about Oct '02.
The project went into full release.

Stand to be corrected.
Cheers Rusty

Rusty is right. All those running FAD prior to oct02 where running beta software which processed mollies differently, hence they had huge molecule scores over a short period of time.

Strictly speaking, these scores should not be counted towards a members' current score, IMO, as not all members had the option of running that software, and it skews the scores somewhat. This issue has been raised in the FAD forums a fair bit.

Somewhere done the track, when points are the main indicator of score, then hopefully we may see more parity for those who weren't in the original beta program. This is my view only, and may not be shared by other members, especially those who were involved in the original beta program.

We have a number of original beta contributors in our team. Any of them like to comment??

Cheers
Mick

[RustyGT]

quote mick: "Somewhere done the track, when points are the main indicator of score"

I hope not that, to be the case.
It has been shown already, that points can be rorted.

UDs 'benchmarking' rewarded high end hardware, rather than hard work.

Jobs returned, is a simple stat.
Why not go with that?

R.

Because the number of jobs processed is not a measure of contibution, as it's random. Some jobs take a long time; some are quick. It's a matter partially of luck, partially of what projects you pick to be involved in. If no. of jobs was the key indicator, then projects which had long jobs would not attract many, whcih may affect the projects potential.

Points are based on time spent crunching and CPU speed, which is a better indicator of actual contribution, IMO. It just needs to be polished a bit.

Hopefully rorting can be reduced, with better randomisation of jobs.

IMO, anyway.

Cheers
Mick

[RustyGT]

Point taken, Mick [no pun]

Prolly find folks killing the hard tasks.
Just to get up in the stats.

quote from above: "especially those who were involved in the original beta program."

Why do these guys deserve special treatment?
It's not like they were breaking rocks!

I wonder whether it's all getting a bit messy at the server end.
Trying to please all the people all the time.

IMO, Team None is a 'copout'
But do you deliberately disenfranchise 1300 members,
because TeamTimo beleive they should be 3rd rather than 4th?

R.

[Wizzard~Of~Ozz]

Rusty, We are in third for organized teams, which is how I look at it.

none is not organized, therefor to call it a "team" is not accurate, team implies organization which it is not (IMHO)

the "beta" issue, I'm pretty sure that while crunching for the beta, they had hugemolecule counts, fast processing speed. it's more accurate to go by # of Jobs.

The points main problem is with Win9x/ME/Dos based. These operating systems do not report CPU time / task, so that complicates things immensly, but other then that, there is no issues with the points, however some of us got a bit of a head start by running beta.

[tmx468]

Do I feel a can of worms was opened... Oops!
Surely to be equal, you should take stats since OCT '02, as the idea of points tables is, in essence, a 'we are better than thou', and so everyone on an equal footiing should be used...?

[RustyGT]

tmx
Dont think that you have caused a revolt in the ranks.
These issues are as old as the project. OK?

Wizz. I've got one box on XP.
All my cows run 98.

IMO: Points are meaningless Atm.
There seems to be little correlation between jobs/mollies/points.

Certainly there is work to be done on this for sure.
I would hate to see FaD adopt the UD benchmark principle.

ChrsR.

There was talk of a benchmark molecule being run on each cow, which then sent a result back to the server, and that's how points would be alloted, based on benchmark molecule.

Only problem is, there is no molecule that adequately covers all that is required by FAD. So while the principle is good, it doesn't work in reality. Doesn't really show someone's contribution to FAD

Trying to use CPU size and time taken is a compromise, as not all OSes report CPU adequately. Still needs a bit of work, but is the best way, IMO.

All the discussions on this are in the FAD forums for all to see (some really old bits of the tale may have been lost when they switched forum software, but most of it's there. )

Cheers
Mick