On Fri, 2019-08-02 at 14:56 -0700, Kerim Aydin wrote: > On Fri, Aug 2, 2019 at 2:17 PM ais...@alumni.bham.ac.uk > <ais...@alumni.bham.ac.uk> wrote: > > A bug exploit could /be/ a win as intended, if the bug had been > > placed there intentionally by the proposer. "Convince people to > > adopt a buggy victory condition and win immediately" is one of the > > more common winning techniques at BlogNomic. > > Rant: This right here is the reason we almost never get around to > actually playing by the intent of many subgames. We just crash them > until we're sick of all the CFJs and then repeal. It really > discourages me from bothering to write a long sub-game - debugging in > play-mode is usually necessary, and I don't see much pride/point in > "hey, I won because there was a misplaced comma or because a certain > set of moves is fundamentally completely imbalanced, isn't that > clever." I mean it's fine on occasion but having that be the outcome > of Every. Single. Subgame. just gets tiring. > > Well, I guess the test mechanism is "Tournament" - where you can put > a "judge by the intent" clause in there.
BlogNomic normally (not always) starts its subgames without any victory condition, and only adds one after they've had several iterations of nomicky changes applied to them. It works a bit better than what we normally do in Agora, but has problems of its own (especially in relation to people positioning themselves in an attempt to anticipate what the victory condition would be). Perhaps what we need is some sort of escalating milestone system: run games in multiple iterations, with the first "win" (which may be trivial) being worth one point, the second two points, the third three points, and so on, with the rules for the subgame being amendable only between iterations. A scam could give you an advantage, but consistently good gameplay would be worth more. -- ais523