Post this in the lounge IB...
Team Raiders of the Lost Primes (ROLP) has thrown down the gauntlet glove in our direction.
In fact, they have done this several times and we have not responded.
What does everyone think of running a rally over the Jan. 23rd-25th weekend on that drive? The usual start and end times of 7 PM GMT will be in order.
For a top-5000 range, these will be some quick tests so there should be more results processed than in previous rallies.
I say we do a dog pile on these folks.
If you can toss even 1 core on this project for the rally it would be helpful to the cause.
Server = nplb.ironbits.net
port = 8000
Let me start this off:
I'll toss ~30 cores on this project/port.
Any one else with me???
There are some good STATS to this project now, thanks to AMDave and Bok...
so you don't have any excuses...
what are you waiting for?
You know you want to do a huge push for a BIG dog pile!
It's been awhile since we had a little get together Free-DC picnic, so let's have some fun.
Post this in the lounge IB...
But if I understand correctly, the fact that the candidates are quick to check means nothing right? I thought I'd just done a bucket load on the port 9000 server but if the stats are changing it was less fruitful than I expected.
Count me out of this one, as I said I've already been working on this for a little while recently and have work debts elsewhere. Maybe next time!
Thank you for your help on port 9000. The work that you are doing there was not top-5000 work. For the rally, we're running port 8000, which is. The candidates will be in the range of n=365000-370000. They should be < 180 secs. per test on a fast machine.
Free-DC rules the roost on our stats for # of pairs done and primes found. :-) The stats can be sorted in any manner that you like. The additional column for scoring is just another way to look at them.
Good luck on your other work!
Gary
Essentially, here's how things work with the new stats scoring system: you will gain approximately the same amount in the stats regardless of what work type you run. That is--there's no stats bias between worktypes any more. This is crucial since we don't want there to be an unfair advantage to those who ran our 9th Drive work during its relatively short tenure.
Now, all the drives score equally, so if you run on the port 8000 server for this rally then you should gain just as much as if you ran on port 9000 for the same length of time.
Don't get me wrong, I think it's only fair for stats to be weighted. I just wanted to be clear that it isn't really advantageous for the participant to know that the candidates are quick to test, if it doesn't affect their rate of production.
So will Bok's stats be adjusted to switch over to the new scoring method, or are we keeping them as they are over here?
The "default" is for them to sort by score so if you go to the scoring pages, that is what you'll see. But if you click on the various columns, you can sort them however you want in either ascending or descending order. He's done a very nice job with it.
For me, on the top-5000 site, even though my score is far lower than the "big boys" (41st place I think), I like to compete for total # of primes since that is the only way that I have a chance. The same type of thing applies to our stats. Everyone can choose their method of competing by sorting them however they so choose.
To see the stats, go to http://stats.ironbits.net/statsnew and click on "k/n pairs" or "primes" on the left side.
Edit: I'm pretty unfamiliar with the Free-DC forum. I see that you may have been referring to Boc's Free-DC stats elsewhere on this site instead of the NPLB stats. I wouldn't have an answer for that.
Gary
I take whatever data is thrown at me. What would I need to change ?
Change nothing lad, it's fine as it is... they have it there way we have it ours
I may just have to switch a machine to this for fun...
Semi-retired from Free-DC...
I have some time to help.....
I need a new laptop,but who needs a laptop when you have a phone...
Now to remember my old computer specs..
At least from NPLB's point of view, we'd probably consider the scores to be the primary stats data, since they're more even across the board regardless of what type of work we're doing. The k/n pairs stats gets very, very skewed whenever we load up small work (which is of a very finite work and will be completely gone within a few weeks at least).
Of course, this is just my $0.02 as an NPLB admin--you guys are of course free to go by whichever score you like.
To me, it makes sense to go with the weighted score. It's about total effort, not who can do the most when the work units are small. But that's just my opinion.
ok,
Will change it tonight
It's probably worth wiping all the scores so far (in my DB not the NPLB one) as the history would be worthless..
Bok
no it's not.. I slice and dice the data a little differently to how it's held in the NPLB db. I could potentially recreate it, but it could take a fair bit of time to figure it out and write the statements.
I haven't looked at the tables in there for a while, but is it actually possible to recreate on a daily basis the scores for each user/team retrospectively? Plus the rankings and such at the time...
more than likely
. . . . . ___
. . . . . . .\___/\______
. . . . . . . \__AMD___\\__
-----------------------------------------
I'm putting some on for amd_users gotta have another team or two in there.