[EM] Markus & Manipulability

[Forest Simmons]

On Mon, 4 Feb 2002, Markus Schulze wrote:

> Dear Forest,
> 
> in so far as IRV meets majority for solid coalitions and independence
> from clones, IRV can hardly be called "erratic" compared to primary
> with runoff.
> 
> Markus Schulze

I suppose there are as many meanings of erratic as there are of
manipulable.

Evidently the "erratic" that I have described is different from the
"erratic"  that is measured in terms of solid coalitions and independence
from clones.

The "erratic" that I described has more to do with the complexity of
strategy and the sensitivity or strategy to variations in information.

If the output (water temperature) from one shower valve with hysteresis
were to be the controlling input to another shower valve with hysteresis,
and you were to try to adjust the water temperature in the second shower
by adjusting the first valve, you would increase the erratic nature of the
result.  Chain together several such shower/valve combinations and the
result is hopelessly erratic.  The longer the chain the worse it is. 

Top Two is like the one valve/shower case compared to IRV, whose
chain length increases with the number of candidates.

Forest