|?? cheating at BU

[yoyo]

I think blocking will not work, the cheater then uses a different userid/team/IP.
I told rebirther already a year back that there workunits are much to small. At this time they were running on my laptop only 7 minutes. They should increase the runtime drastically. This much better for the server.
For counterstrike the cheating I would store all results in a database and if a result is returned a second time, twice of the credit which would be granted will be reduced and the account will be blocked.
Don't know how simple it is to implement it.

[Mumps]

Problem with that approach, yoyo, is that, as the fraud went along, the later results were returned with everything a normal cgminer run would produce, simply fabricated by a program that did none of the work. It just created random "Accepted" hashes and calculated appropriate GH/s speeds to make the output look like a real result. Only 4 thousand times faster.

For BU, the biggest challenge is coming up with a way to verify the work is actually done. Aside from the easily faked text only output, there's nothing really obvious that they can use to confirm the work that was done by cgminer through a third party that can be trusted. Other projects will need to work through ways of ensuring the work is done, by the host assigned even. As the fraud over on WCG shows, they even copied the output of their wingman to "complete" WU's for credit, while doing nothing to help ensure the validity of those results.

In a "trust but verify" environment like BOINC, it's pretty likely that people with attitudes like this will come along and make everything the project is working for and towards something that's no more "science" than wikipedia is guaranteed to be "facts."