On Tue, 4 Nov 2014, Eritivus wrote: > On Tue, 2014-11-04 at 13:52 +0000, Kerim Aydin wrote: > > Pretty much the same thing that happens if a proposal took effect that > > says "If the last digit of pi is 3, Eritivus gets one point." This > > sort of thing has happened many, many times. > > > > In the current ruleset, your score would become Indeterminate. > > But there is no rule that says "pi has a last digit" (and that would > be ineffective anyway). There _is_ a rule that says "rule changes > always occur sequentially".
I knew I picked a bad example but I was in a hurry. A simple (better) example would be "if Goldbach's conjecture is true..." Anyway, my point was not to defend Murphy's argument, but to point out that "something turning out indeterminate as a result" is not a particularly compelling counterargument. An example that actually happened was with currencies. Currencies were, by explicit definition, fungible - legally, no one tracked individual unit histories. However, at one point, a particular unit of currency picked up a rules-defined trait, that would result in a "score" if you held it at the right time. But the rules let you transfer chunks of this currency, and described them as fungible. If I said "I give Eritivus 5 Units", and one of the units had that trait, and then you transfer 3 Units to a third person and hold onto 2, all the transfers would succeed, but the owner of the "special" unit would be INDETERMINATE (UNKNOWN or UNDECIDABLE at the time). No problem. > Without such an amendment, though, can we really just agree to let the > order be Indeterminate because it doesn't seem to matter (yet)? This is one way that Agora differs from both real Courts and standard games (which emphasize finding an answer by using burden of proof, or common sense). Here, we allow judges to shrug and say "nope! can't tell the answer." > If so, what prevents us from doing so when the order does matter? Nothing prevents us from doing so. For example, if the Rules say: "Eritivus gets 5 points divided by the number of players who have X". If no players have X, the result is mathematically undefined, so the resulting score is INDETERMINATE. Now, that can cascade! If another rule says "any player with a score > 100 wins" you either have to have (1) a rule describing how to handle comparisons to indeterminate values (like we do with Unanimous votes), or it can cascade (comparing an INDETERMINATE value to a number results in an INDETERMINATE result in standard mathematical and computational methods). This has happened, as I said, many times. At one point I remember a card uncertainty led to another card uncertainty that led to a Vote Count uncertainty, and I (the Officer in charge) was trying desperately to get a proposal through to fix it before one of the uncertain votes "mattered" for something and cascaded further. The only real defense for locking up the whole game up is that R1698 (Agora is a Nomic) protects the proposal system, and the proposal system can fix the problem as far as it spreads. We all hope... -G.
