I can think of a few ways it could be worked out, but which one is it?

When my computer has "reached a daily quota of 1 task", is that measured midnight to midnight GMT? Or midnight to midnight in my time zone (which for me happens to be the same in winter)? Or a rolling 24 hours since I last received a task?

How is the limit decided? Is it x crashed tasks in a row? Is there forgiveness if I previously did 50 good tasks? Maybe some kind of percentage of the last x tasks?

I'm wondering what effect a bad batch has on the computers. For example if it's worked out by simply putting your computer in the naughty corner after x bad tasks, no matter how good you've been before, a bad batch could knock out every single computer, and you'd for the next day only get 1 task from every host out there.

The BOINC manual and the source code answer that question:
https://boinc.berkeley.edu/trac/wiki/ProjectOptions#Joblimits

<daily_result_quota> N </daily_result_quota>
Each host has a field MRD in the interval [1 .. daily_result_quota]; it's initially daily_result_quota, and is adjusted as the host sends good or bad results. The maximum number of jobs sent to a given host in a 24-hour period is MRD*(NCPUS + GM*NGPUS). You can use this to limit the impact of faulty hosts.


https://github.com/BOINC/boinc/blob/4655fce0b47940e2dacd4269daf70b16904120fe/sched/sched_result.cpp#L50-L53
https://github.com/BOINC/boinc/blob/1e51e98e93eb68cd3f56e84e53e3270ce2cbc96c/sched/sched_send.cpp#L190-L217

I'm not a programmer so I don't understand what's in those links, but looking at the bit you wrote (I assume "interval" should read "range"), that would indicate my 24 core machine would be set to get at least 1*NCPUS = 24 per day. But I got a limit of 1.

Also, where do I see this limit? I found a discussion on the Einstein forum suggesting places other than Einstein with newer Boinc server software have it in the details page of a host. I can't see it here.