Same problem as here: https://www.worldcommunitygrid.org/forums/wcg/viewpostinthread?post=538384

To solve your problem, you could uncheck doing work for FightAIDS@Home (there are more than enough other people to crunch FightAIDS@Home).

First thanks for answers..
I think i will test the advise from Sgt. Joefirst and will report later in some ;-)

Sgt. Joe:
You said it's an issue with BOINC client. May i aask which client you are using?

emphasis mine,

No, the server has trouble in headering new work with proper FPOPS, where Rigid and the rest have 2 entirely different runtime ranges. The client has little issue, WCG made it to have lots of trouble by disabling the DCF function, so it can sense true runtimes for the Rigid and adapt work requesting, so I've resorted to rebuilding making it oblivious to <dont_use_dcf/>, might even spend time on re-developing code to give each app it's own DCF [re-developing, because the code/client was already created years and years ago, but not boarded by Berkeley... was it offered, was it rejected, not privy to the history, so it only exists in privately build distribution in an older client compile, but it is THE answer for umbrella projects like WCG].

Not my problem anymore.

See ya later.

PS WCG has a solution in hand [where's a will, there's a way]... random mix Rigid and shorts in the distributor, and eventually over time a reasonable mean would percolate out, which is somewhere over 2 hours according the WCGART chart.