[EM] Re: CR/Approval

[Kevin Venzke]

Chris,

Replying publicly again...

 --- Chris Benham <chrisbenham at bigpond.com> a écrit : > Kevin,
> I  am  not  completely clear  on  how exactly that a candidate  that has 
> been marked  as  "not viable" loses that status.
> By overtaking one of  the "viable"  candidates perhaps?

Right.  Viability is determined anew at the end of each round, and isn't a 
requirement for receiving approval votes, so I don't see a guarantee that a 
candidate, once dead, will stay dead.  He could be "scooped up" by small disappointed 
factions.

I noticed that this method will actually preserve some of Approval's ability
to drop low utility winners, in contrast to Forest's Max Power method.  Take
the "weak centrist" scenario:

48 A 10, B 2
3 B 10
49 C 10, B 2

Initial votes: 48 A, 3 B, 49 C.  B is not viable.
The blocs had expectation 4, 3.3, and 4 respectively.
Second round: 48 A, 3 B (half votes for A and C?), 49 C.  Now A is not viable.
Expectations were 5, 0, 5.

But the method doesn't continue from here, so C wins.  A's supporters didn't
"learn" that A couldn't win.  And just as in Approval, if either the A or the
C faction approved B (rated him high), they would've given the election away:

48 A 10, B 2
3 B 10
49 C 10, B 7

Now someone other than B is marked not viable, and this causes the hopeless
opposing faction to embrace B, which means a B victory (B will be the only
one with majority approval).

You suggested I find out whether this method is original; I think the best bet
would be to ask Forest.  If he hasn't heard of it, I wouldn't know where else
to check.

What do you think of "Gradual Info Approval" or "GIA" as a meaningful name?  Doesn't 
mention the CR aspect, though.


Kevin Venzke
stepjak at yahoo.fr


___________________________________________________________
Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en français !
Yahoo! Mail : http://fr.mail.yahoo.com