Répondre: Re: Répondre: Re: [Mmwg] Importan t : procedure for runoff election

[Mohamed Tijani BEN JEMAA]

Dear Bertrabd,

I agree that we have to innovate to avoid a winner with 50.01 %. I also
agree that if your proposal has never been used, it doesn’t mean that we
don’t have to try it.

But when you try to go deeper in the system you proposed, you will notice 
the following :

Since the number of subscribers is around 35, I will use figures close to 
this number.

1. First case (All people vote for 2 candidates):
              a.	Voters = 38	result= 13 (A-B)	13 (A-C)	12 (B-C)	          
A= 24		B= 25		C= 25
              b.	Voters = 37	result= 12 (A-B)	12 (A-C)	13 (B-C)	          
A= 24		B= 25		C= 25
              c.	Voters = 36	result= 12 (A-B)	12 (A-C)	12 (B-C)	          
Equality (it’s the case you mentioned of irrelevant result)
              d.	Voters = 35	result= 12 (A-B)	12 (A-C)	11 (B-C)	          
A= 24	                B= 23	                C= 23	  
              e.	Voters = 34	result= 12 (A-B)	11 (A-C)	11 (B-C)	          
A= 23		B= 23		C= 22 
              f.	Voters = 33	result= 11 (A-B)	11 (A-C)	11 (B-C)	          
Equality (the case as c.)
              g.	Etc.    
                             
2. Second case (Some vote only for 1 candidate, while the others vote for 2):
             the result will not reflect the choice of the majority. It
will be as if some vote         
             twice, and the others only once.

3. Third case (1 or more votes twice for the same candidate):
            If we oblige people to vote for 2 candidates, you may find
some of us who don’t 
            have confidence in more then one. So they will vote for the
same person twice.  

As you see, in the first case, and whatever the number of voters is (even 
if it is not a multiple of 3), we can find a one vote difference if the
result is very tight (translation from French) I am sure you understand
Bertrand, but for the others, I mean if the number of votes is not far
from those won by the other candidates. It will be the same case as if we 
vote for only one candidate and get a result like A= 12, B= 12, and C= 13.

You can also notice that, for 37 voters for example, one vote difference
represents in the traditional system 2.7% while in yours, it represents
1.35% only.

I think that it will be an artificial improvement of the number of votes, 
and may give a different result (except the case that every voter has
equal preference for 2 candidates), and in any case, it will not solve
the 50.01% issue. 

What I want to explain here, it that when difference between candidates
is large, there is no problem with both system, but when the result is
tight, both system will lead to 1 vote difference. We have to find the
more representative of the voters’ choice.

Best.

Tijani

----------------------------------
Mohamed Tijani BEN JEMAA
Member of the Tunisian Engineers' Order
Vice Chairman of CIC, World Federation of Engineering Organisations
Phone : + 216 98 330 114            Fax : + 216 71 781 927
--------------------------------------------------------------------------