Much obliged. I tip my knitted cap to you.
On Tue, Sep 17, 2024 at 7:16 PM ais523 via agora-business <
[email protected]> wrote:
> On Tue, 2024-09-17 at 00:53 +0100, Katherina Walshe-Grey via agora-
> official wrote:
> > I recuse 4st from CFJ 4094 on grounds of lateness.
> >
> > I assign CFJ 4094 to ais523.
> >
> > CFJ 4094 was called by literallyAmbiguous and reads: "I owe Janet and
> > Mischief one spendie each."
>
> Evidence:
> {{{
> If I have not already made a binding pledge to bribe... I mean *reward*
> everyone who votes for me with one spendie each, I do so now. Spendies will
> be paid to each recipient within one Agoran week of their vote being cast.
> Breaking this pledge is the class 1 infraction of Oathbreaking.
> }}}
>
> Arguments:
>
> As far as I can tell, this is a situation where the pledge (to pay 1
> spendie to each person who votes for literallyAmbiguous, but it's an
> instant-runoff election) is simply ambiguous. Although the rules
> specify that a pledge is violated if the breakage of the pledge is
> indeterminate, the definition of indeterminacy requires a circular
> definition, paradox, lack of information, or indefinite alternation,
> and none of those seem to apply here – it is simply a case of the
> pledge not applying very well to the type of election.
>
> There is no rule directly covering what happens when the wording of a
> pledge is ambiguous – the rules are silent on the matter. As such, the
> relevant rule is 217, which lists tests to be used when the rules are
> silent: game custom, common sense, past judgements, and the best
> interests of the game.
>
> First, the question of what point in time the pledge is triggered at.
> The caller has argued that the relevant time is the time at which the
> votes are evaluated, on the basis that the ballot might be withdrawn,
> [or become invalid during the voting period – but that is not actually
> possible: see CFJ 3749]. However, I think the literal wording of the
> pledge is that requirements on voting check to see whether or not the
> vote was actually made, rather than waiting to see if it was effective
> – and that game custom supports this (both the "if you ask people to
> vote, we just check to see if they voted" requirement, and the "we hold
> people to the literal wording of eir pledge" requirement). After all,
> you could use a similar argument to argue that if a vote didn't
> influence the outcome of an election because the other votes made it
> moot, it didn't count – this isn't conceptually different from a ballot
> being reallocated during an instant-runoff procedure. It is telling
> that at least one player did (after the CFJ was called) try to scam the
> pledge, by voting for literallyAmbiguous, then withdrawing that vote to
> vote differently.
>
> As such, the CFJ needs to be evaluated simply on the wording of the
> pledge, and the content of the vote – is making a ranked-choice vote
> which places literallyAmbiguous third on a list of four voters "voting
> for [literallyAmbiguous]"? In my opinion, that defies common sense. It
> also defies game custom to some extent: IIRC "I vote for X" (in an
> instant-runoff election) is typically interpreted as "I vote for a list
> consisting only of X", and I thought there was even a rule to that
> extent (but if so, I can't find it, so it may have been amended or
> repealed). I could see an argument that maybe putting someone second on
> a list could amount to a vote for em if there was a strong expectation
> that the first-choice vote would be reallocated (and likewise, third if
> the first two votes were expected not to matter), but that doesn't seem
> to have occurred here.
>
> The other relevant tests are past judgements, and the best interests of
> the game. I couldn't find a relevant past judgement other than CFJ
> 3749, which argues against the caller's arguments, but not in favour of
> any particular resolution. As for the best interests of the game, there
> is some balance that needs to be found between letting people scam
> their way out of pledges and letting people scam pledges made by
> others. At one time in Agora's history, there were two sorts of pledge,
> Legalistic pledges in which the literal wording was all that mattered,
> and Equitable pledges which were based primarily on the intent, both of
> which found a balance in different ways. Because so few pledges have
> been made recently, we haven't had much chance to experiment with the
> best balance now that there's only one sort of pledge.
>
> The interests-of-the-game reasoning could matter in CFJs with similar
> but different facts to this one: I think the argument "I pledged to
> bribe players who voted for me, but I'm not a possible vote, only lists
> of entities are, so the pledge does nothing" would be a plausible one.
> However, the caller didn't try to make that argument, and in any case,
> it doesn't matter for the result of this CFJ, because the "third on a
> list of four" votes are not meeting the required standard to trigger
> the pledge anyway, and thus there's no reason to analyse whether there
> might be a scam to avoid them.
>
> I judge CFJ 4094 FALSE.
>
> --
> ais523
>
