Yep. I understand the methodology. But *why* was it declared invalid? And why did I not get awarded the points accordingly?

Also, if it was really invalid, then why award any points at all?

50% of the median is the rule in effect since the beginning of November, same as the Extreme Outlier. Consideration was, that client side the process ran without quality failure indicators, thus 50% granted for work performed.

This 50% is in deviation to most other BOiNC run projects, where an invalid gets granted zero.

Hold on, 50% of the median is the rule, not 50% of what u claim!

There is no guessing.... your result deviated from the other 3, who agree. My machine is 100%, yet it produces 1 in 200 a real invalid.... the larger the model, the more chance there is it hits very high memory regions...... and memory does fail, sometimes just intermittently.... only needs 1 bit to go the wrong way once.

We had 163,265 results yesterday....unless there are 'errors' logged, it becomes next to impossible to determine what went wrong and why it went wrong...... on these volumes impossible to follow up without clear pointers.

PS,

What i can say with probability looking at the times of the 4, the claims average and your deviated claim & time, that u had a likely restart..... have a look in the log that sits in your Result Detail pages.