Do we really need an objective solution, though? So much of the judging process is subjective, and rightfully so, that I don't think much is gained. It's a judgement call. That's why we call them judges.
If you ask me, I'd say we don't need an objective one at all. But several coments seemed to call for it. Maybe somehting like "very apparent/obvious luck manipulation" would be a better name for the tag then. I don't think those particular names are good, but people wouldn't go raise the question of what actually constitutes a heavy one then.
This is just another suggestion. I don't really have a problem with things as they are though.
I suppose the point of this category is to demonstrate a lot of random number generator abuse. "Heavy" or "Signficant" luck manipulation obviously refers to exploiting an RNG numerous times throughout a run, perhaps almost every time the RNG is used, as is the case in many RPG TASes.
I don't know what manipulating AI behaviour is classed under, but it should also go under luck manipulation imo because you are changing the programmed behavior into a more favourable one, it is still using an RNG or pseudo-RNG and therefore still subject to luck manipulation. At least, this is how I see it.
I suppose the point of this category is to demonstrate a lot of random number generator abuse. "Heavy" or "Signficant" luck manipulation obviously refers to exploiting an RNG numerous times throughout a run, perhaps almost every time the RNG is used, as is the case in many RPG TASes.
I wouldn't say that "significant" refers only to the frequency of abuse, but also to the magnitude of the results. For example, if manipulating the RNG saves 0.1 seconds, I wouldn't call it significant, but if it saves 1 minute, then it's very significant. By "magnitude" I also refer to the likelihood of something happening: For example if something happening has a probably of 1 in 1000, and it's triggered 20 times in a row (which would be virtually impossible to achieve during normal play), that's significant.
I don't know what manipulating AI behaviour is classed under, but it should also go under luck manipulation imo because you are changing the programmed behavior into a more favourable one, it is still using an RNG or pseudo-RNG and therefore still subject to luck manipulation. At least, this is how I see it.
It depends on how the AI is implemented. For example, changing your position so that you convince a deterministic AI to move closer to you would count as AI manipulation, but it's not luck manipulation.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Maybe we should have a tag for "heavy AI manipulation" which would be similar to "heavy luck manipulation" but for when the game AI logic is manipulated very significantly?-)
Perhaps a good definition of heavy luck manipulation would be that even a segmented real-time run of the game would not be able to reasonably pull off the level of luck manipulation shown in the TAS. This would mean that the luck manipulation shown really is beyond human limits.
