During the past couple of weeks I've made a number of changes to the BOINC scheduler, specifically designed to help prevent machines from downloading more work than they can complete by the 'Report Deadline'. However as has been pointed out in this thread, there is no substitute for intelligent and responsible management by you and other users. Keep an eye on your work queue and deadlines, and try and set your project preferences, work caches, and resource shares in a responsible way. This helps to maximize your overall contribution to the project.
Cheers,
Bruce

0 credits granted
)
> So how about this one:
>
> http://einstein.phys.uwm.edu//workunit/347835
Looks like two windows machines agreed on their results and outnumbered you.
This can happen from time to time (the next victim will be someone else) due
to numerical issues: the results may be of same size (so they basically
contain the same information) but their checksums will be different.
Remember that different OS platforms use different approaches to trigonometric
functions and so on...
Your result was not "close enough" to the canonical one.
Of course it takes a lot of thinking to come up with an acceptable definition
of "close". Perhaps ours is a bit on the narrow-banded side...
S