mersenneforum.org  

Go Back   mersenneforum.org > Prime Search Projects > No Prime Left Behind

Reply
 
Thread Tools
Old 2009-06-09, 07:38   #1
MooooMoo
Apprentice Crank
 
MooooMoo's Avatar
 
Mar 2006

1C616 Posts
Default Bug in top5000's score page?

I don't know whether anyone has noticed this yet, but there seems to be a bug on Chris Caldwell's top5000 site. When you click on "top projects sorted by (normalized) score", http://primes.utm.edu/bios/top20.php...by=ScoreNormal

you get this result:

Code:
normalized	project	                primes	    score
69395	       Seventeen or Bust	   11	   51.3943
17 or bust's normalized score has decreased a lot over the past few months - it went from well over 100,000 to under 70,000 now. However, its score has always remained at 51.3943 since at least the beginning of the year. Shouldn't its score have gone down to 50.9 or some other similar value?
MooooMoo is offline   Reply With Quote
Old 2009-06-09, 12:37   #2
AMDave
 
AMDave's Avatar
 
Jan 2006
deep in a while-loop

2×7×47 Posts
Default

The score of the 11 primes they found does not change, but their score might reduce if one of their primes dropped out of the top-5000.
However, since they were quite large that has not happened in a hurry.

The last paragraph below the table on the Top-5000 page sheds some light on this:
Quote:
Note that if a project stops finding primes, its normalized score will steadily drop as the size of the 5000th primes steadily increases. The non-normalized scores drop too, but not as quickly because they only drop when the project's primes are pushed off the list.
Seen here:
http://primes.utm.edu/bios/top20.php...by=ScoreNormal

I hope that helps.

Last fiddled with by AMDave on 2009-06-09 at 12:44
AMDave is offline   Reply With Quote
Old 2009-06-09, 19:40   #3
MooooMoo
Apprentice Crank
 
MooooMoo's Avatar
 
Mar 2006

1110001102 Posts
Default

Does that mean that it would be possible for some other project to have a lower score than 17 or bust even though that project's normalized score is higher?
MooooMoo is offline   Reply With Quote
Old 2009-06-10, 11:51   #4
AMDave
 
AMDave's Avatar
 
Jan 2006
deep in a while-loop

2×7×47 Posts
Default

Have a look at the actual and normalized scores on the table and tell me what you see.

Whilst it does help to ask the question, you do need to read back your own question and attempt to answer it. This way, when you observe the answer you will understand it better than if someone were to tell you it is 'just so'.

If you are very concerned about this and have evidence to support your hypothesis, may I suggest you address the question with Professor Chris Caldwell who runs that website.

In actual fact I did observe 1 tiny 'apparent' inconsistency on a completely different page whilst reviewing your questions. However, that was unrelated and, I suspect, due to the order-by sequence in a database query and not of the score values. Not anything serious enough to bother the Professor about, in my opinion.
AMDave is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
LL results page NBtarheel_33 PrimeNet 7 2016-01-03 23:33
What Murphy score to expect or need [cado-nfs] kosta Factoring 3 2013-04-17 17:30
Connection between rank and digits on top5000 list MooooMoo Lounge 4 2008-03-31 16:31
Requesting SWAG on what's most likely to score me a Mersenne Prime. jasong jasong 1 2007-01-29 21:25
Project where the goal is to get the highest Prime Pages score. jasong Lounge 4 2006-08-22 21:04

All times are UTC. The time now is 01:00.

Thu Apr 2 01:00:14 UTC 2020 up 7 days, 22:33, 3 users, load averages: 1.00, 1.01, 1.15

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.