[EM] "GIB" fails Participation, and not clone-proof

[Chris Benham]

I  found the examples given in "Moulin's proof  that  participation and 
Condorcet are incompatible" (posted by M.Shulze)
http://lists.electorama.com/pipermail/election-methods-electorama.com/2003-October/011030.html
useful for testing  a  method  I suggested  and  once liked  and dubbed 
 "Improved  Generalised Bucklin".
I  have found that it fails Participation and also a  clone example 
(also posted by Marcus S.), and so the method no
longer appeals to me.
But I still  think that my suggested  "first finalist" method , does 
represent a small solid improvement (with little extra
complication) on a method that is not very good.
That rule was: Allow equal preferences with no split votes, which allows 
more than one candidate to have a majority.
When the leader has a majority, eliminate all candidates who do not. All 
ballots which show any preference among
the remaining candidates but do not contribute to the tallies of any of 
them shall  contribute a vote to the tally (or tallies)
of  the remaing candidate(s)  they prefer. After this, the candidate 
with the biggest tally wins.

Chris Benham