Page 2 of 2 FirstFirst 12
Results 41 to 54 of 54

Thread: Something wrong ???

  1. #41
    OCworkbench Stats Ho
    Join Date
    Jan 2003
    Posts
    519
    Thanks to early suspicions and Buffering like crazy, myself and the others from my group just Uploaded 1.2 M for the day instead of 60,000. I know how happy I am ( excuse the pun )
    I am not a Stats Ho, it is just more satisfying to see that my numbers are better than yours.

  2. #42
    But Stats sites do not (and cannot) Track the generation of EACH of the work units uploaded! this is necessary to accurately recompute the points for each person.
    Not Necessarily. You seem to be assuming that the given points weren't proportional to the sqrt(gen#). From my limited playing with it, uploading a generation or two, we seemed to have been credited 2*sqrt(gen#) instead of the usual 50*sqrt(gen#). Simple mutiplication will do the repair, no generation by generation data needed.

    Of course, that assumes my little tests were valid. It would be nice to hear just what the error was. It only takes two flipped bits to make 50 into 2 - was it a random act of the universe like cosmic rays? Or was there a reason the problem started at 5 pm (quittin' time!)?

  3. #43
    Originally posted by willy1
    I sincerely doubt Howard is the only person running this project. How hard is it to assign rotating "duty" to the team members (including those grad students!) to keep an eye on the system, or even to have a cron job that emails a team member when some metric goes wonky? After all, this is a 24x7 project, not something that goes to sleep on the week-end.
    As a matter of fact, this remains a one-man (and one-woman) show.

    Now I'm glad there's still a few sensible people here, and for the rest of you (Im not naming names ) please stop blowing things out of proportion. I did NOT say I don't care about the stats. If I said this, please find the quote and show me. If you continue to read things that I didnt say into my messages then you will never be happy.

    I said we do not keep a history of the stats, end of story. Should we? Maybe. Could we? Of course. It doesn't matter. It is not important to US to keep this information, and there's plenty of others doing this for us now. If one of those people would kindly like to provide us with a table of UID's and points over the 'outage' period, we can recompute the exact number of missing points and reimburse them.

    And if I see another post whining about this, I'll be sure to exclude that user from the reimbursement

    Howard Feldman

  4. #44
    Administrator Dyyryath's Avatar
    Join Date
    Dec 2001
    Location
    North Carolina
    Posts
    1,850
    I got a friendly email from Elena about the missing points. I will provide them with what they need & it sounds like they'll fix everything up. No worries, all.
    "So utterly at variance is destiny with all the little plans of men." - H.G. Wells

  5. #45
    Administrator Dyyryath's Avatar
    Join Date
    Dec 2001
    Location
    North Carolina
    Posts
    1,850
    StarDragon: You've Got Mail.

    If it's not precisely what you needed, let me know...
    "So utterly at variance is destiny with all the little plans of men." - H.G. Wells

  6. #46
    Senior Member
    Join Date
    Sep 2002
    Location
    Meridian, Id
    Posts
    742
    Originally posted by Dyyryath
    I got a friendly email from Elena about the missing points. I will provide them with what they need & it sounds like they'll fix everything up. No worries, all.
    And they spoke the words of summoning... and "Dyyryath" the God of Statage answered: "You shall have your stats too appease the disgruntled masses"









  7. #47
    OCworkbench Stats Ho
    Join Date
    Jan 2003
    Posts
    519
    All I really want to know is the method in which this actually occured. It is quite funny really, but it is driving me nuts trying to figure how the points suddenly started being calculated differently So if anyone can answer the dummy's question ( and all the others like me wondering ) it would save money being sent to my therapist

    I was glad it got fixed so soon on a weekend
    I am not a Stats Ho, it is just more satisfying to see that my numbers are better than yours.

  8. #48
    Originally posted by Brian the Fist
    And if I see another post whining about this, I'll be sure to exclude that user from the reimbursement

    Yes sir! I mean, no sir! Absolutely no complaints here whatsoever, sir!


  9. #49
    Senior Member
    Join Date
    Oct 2003
    Location
    an Island off the coast of somewhere
    Posts
    540
    <risking a lightning bolt>

    I'd still like to know how the points can be rebuilt without the detailed generation information ...

  10. #50
    Dyyryath,

    I'd like to thank you very much for making this effort. Way to be!

    I kept my machines on line and running, even though I spotted what was happening about 2 hrs. after it began. I did so on the faith that it would be corrected by the admins of this program, in a fair way for ALL. As it's now clear that was a error on my part, I'm very glad to see you step forward and do so.

    Thank you once again.

  11. #51

    Correct points have been added to all those that uploaded results in the "broken stats period". A big thank you goes to Dyyryath for providing me with exactly the information I needed
    The statistics now have the right number of points for all users, those who buffered offline, as well as those who worked online.
    Elena Garderman

  12. #52
    Thank you Elena, good looking out!

  13. #53
    Senior Member
    Join Date
    Jul 2002
    Location
    Kodiak, Alaska
    Posts
    432
    Originally posted by willy1
    I'd still like to know how the points can be rebuilt without the detailed generation information ...
    I'll be corrected if I'm wrong - so here goes my hypothesis:
    1. We know that the points being given for the uploads were not just the number of generations uploaded. The points showing up were less than 1/16th of the normal score - and another estimate given was 1/25th of the normal score. If it was just the number of generations being uploaded that we were being given credit for, then the number would start off at 1/50th of the normal score for Gen 0 and gen 1 and progress to 1/ (50*sqr(250)) or roughly 1/790th of the score.

    2. A possible explanation for the loss in points would be that the uploaded data was being lost. For some reason, only a few of our uploaded generations were making it through the scoring system. If the scoring system was dropping 19 out of every 20 generations offered to it, I'd have expected to see a much larger variance in the points being awarded. (How did it know to keep the right packet for each person so that the estimates for everyone seem around 1/20th?) This option doesn't seem logical.

    3. Given that Howard mentioned that all our uploaded data was safe (and we didn't get massive amounts of complaints about lost generations) - the option I'd go for is some kind of accidental change to the scoring function on Friday night.

    regardless.. It's been taken care of, and a hearty thanks to those involved in correcting the problem.
    www.thegenomecollective.com
    Borging.. it's not just an addiction. It's...

  14. #54
    Junior Member Nanobot's Avatar
    Join Date
    Mar 2002
    Location
    Nottingham, UK
    Posts
    8
    I would also like to add my thanks to Dyyryath, but also to the DF team for taking their usual stance of listening and acting to resolve the problem. How many other projects do this ???

    Although a glitch, and a few moans, the project seems to have come out stronger as a result

    Well done to all who helped

Page 2 of 2 FirstFirst 12

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •