[E-voting] Odd S'Indo Article - Joe on Newstalk at 14:15 this afternoon

Ciaran Quinn election at polarbears.com
Mon Aug 21 21:54:31 IST 2006

Joe McCarthy wrote:
> Hi all,
> I just now recorded a piece for Newstalk 106FM to be aired around 
> 14:15.  I said Liam Weeks' claim that there was a "major flaw" in the 
> current system was far fetched.
> I pointed out that the article did not acknowledge that the current 
> system is fully accurate for the transfer of the 1st surplus and is 
> statistically accurate for later transfers of a surplus arising from 
> an incoming surplus.  I have not seen any statistical study showing 
> the possible effects of the initial mixing being different.  (Has 
> anyone seen such a study anywhere?)
> I also stated that the Gregory system works fine on a manual basis for 
> the Seanad and in Northern Ireland and I said his claim that such 
> could only be counted by computer is wrong.
The Weighted-Inclusive Gregory System (WIGS) as mentioned in the article 
is not used in Northern Ireland.

The system in NI proportionally transfers the votes of the last packet 
received by the candidate. For example, if a candidate receives 3000 
votes bringing them 1000 over the quota then each of the 3000 votes is 
transferred at a fractional value of 1/3

The WIGS system proportionally transfers a fraction of all the votes 
that a candidate has, regardless of which count is involved. For 
example, if a candidate receives 3000 votes bringing them to 10000 
votes, 1000 over the quota of 9000, then all 10000 votes are transferred 
at a fractional value of 1/10.

The WIGS system would involve much more counting of votes - for example, 
in Galway West -
Eamon O Cuiv was elected on the first count with 9947 votes (quota 8245).
Frank Fahey received 39.71% of his surplus of 1702.Using WIGS this would 
mean 3950 votes to be transferred at a fractional value of 1702/9947 
(approx. 0.171) (assuming all EOC's votes were transferrable)
Fahey was later elected and his surplus of 73 votes was distributed. 
Under WIGS this would mean that the 3950 votes were then transferred to 
other candidates at a fractional value of 73/8318 * 1702/9947 etc. It 
would also mean that about 85% of the transferred votes would come from 
Fahey's initial vote of 7226. In 2002, the 3rd FF candidate received 17 
of the 78 votes. Under WIGS the 3rd FF candidate would probably have got 
about 50 of the 78 votes.

Whether WIGS is better or not is debatable. It should benefit FF because 
FF will more often have a candidate elected by another party's transfers 
with another FF candidate still available to receive transfers, and the 
resulting surplus would under WIGS be dominated by the original vote of 
the elected candidate as in the Galway West example. (In contrast, the 
non-random system used in NI would not benefit FF)

The whole article is rather silly - WIGS would not have had that big an 
effect in 2002, and FF could be in government with the independents 
instead of the PDs if they had wanted



More information about the E-voting mailing list