[EM] Entropy in binomial stv

Richard Lung voting at ukscientists.com
Thu Dec 23 13:39:28 PST 2021

Dear Kristofer and group,

The trial example, you gave me for Binomial STV gave minimal allowance for entropy. 
The laws of physics are in principle time-reversible. But in practise, entropy governs times arrow.
So, it is with Binomial STV, the preference election count and the unpreference exclusion count are in principle symmetrical. But in practise, as in Irish elections, the number of later preferences falls off exponentially. There is preference entropy, and preference is more important than unpreference.
There should be some difference with Binomial stv, because its unpreference count gives voters the power to exclude candidates, which they may be expected to use by giving detested candidates last preferences.
Never the less, one might reasonably expect the entropy of unpreference voting to exceed that of preference voting.
Of course, the real test is to hold binomial stv elections and follow the pattern of voting behavior. The hand count is less accurate than FAB STV could be expected to be. But still should afford sufficient indication.

Richard Lung.

More information about the Election-Methods mailing list