[Cs5500] Reputation computing

Daniel Marcucci marcucci.da at husky.neu.edu
Sun Mar 27 23:43:38 EDT 2011


Hi Sapna,

To answer your first question, I believe the qualities here are the quality
from the claim (i.e., C) and the quality from the solution (i.e., C').
Therefore, when computing the result, you take the absolute value of the
difference between those two qualities then add or subtract reputation
accordingly (i.e., the |quality(C)-quality(C')| portion of the formula will
be the same for both players but how it is factored into their new
reputation will depend on who won and who lost).

To answer your second question, I think you would confirm the strengthening
based on the protocol that was described in the rules document (
http://www.ccs.neu.edu/home/lieber/courses/se-courses/cs5500/sp11/projects/rules/rules
).

I could be wrong on both of these, so I would appreciate it if others could
correct any mistakes I may have made. Thanks.

 - Dan

On Sun, Mar 27, 2011 at 2:43 PM, Sapna Krishnan <krishnan.s at husky.neu.edu>wrote:

> hello,
>
> I had a doubt regarding the computation of reputation.
> The formula for calculating reputation in strengthening claim,
> When bob wins ,
>
> Bob  + ClaimConfidence + |quality(C)-quality(C')|
>
> What quality does it refer to in the above formula?
>
> Whether it refers to the quality in the claim(i.e the price in case of csp) or the actual quality of the solutions provided for those claims.
>
> Secondly,
>
> How do we confirm before sending the strengthened claim to the opponent that the strengthening is valid or not?
>
> For example :
>
> Alice claims (22,0.7)
>
> Bob strengthens it (22,0.6)
>
> How do we confirm whether Bob has fairly strengthened it or not? That is Bob could always strengthen any claim he gets.
>
>
> Thanks
>
> Sapna Krishnan
>
>
>
>
>
> _______________________________________________
> Cs5500 mailing list
> Cs5500 at lists.ccs.neu.edu
> https://lists.ccs.neu.edu/bin/listinfo/cs5500
>
>
-------------- next part --------------
HTML attachment scrubbed and removed


More information about the Cs5500 mailing list