[EM] Re: How to break this tie?

[Forest Simmons]

> From: Brian Olson <bql at bolson.org>
> Subject: Re: [EM] How to break this tie?
>
> On Feb 10, 2005, at 8:06 AM, Forest Simmons wrote:
>
>> A1>A2>A3>B>C
>> B>C>A2>A3>A1
>> C>A3>A1>A2>B
>
> I don't suppose it would help to know that just about every system I've
> implemented answers "C", eh?
>

The A's form a clone set, which collapsed gives

ABC
BCA
CAB

So any method that gives C more probability than B is clone dependent.


To be clone independent the method would have to give both B and C 
probability 1/3 each.

FWS