Stuff was said
Thanks for the reply and answers.
I try to set up my industry in sectors and set labours accordingly while also using quantum piles, so my tradesdwarves aren't running back and forth very far. I agree that strength and agility shouldn't be weighed for most trades, and instead those dwarves should be reserved for military or haulers. But as you said, some trades benefit from speed and strength; certainly miners, but also mechanics, architects and siege operators, and maybe to a lesser degree growers, masons, stonecrafters, furnace operators, and milkers/shearers/butchers. I think these attributes should add a moderate boost to the ratings for those roles.
As for other attributes, what I understood from the crafting skills and quality thread you linked is that attributes (including creativity) are not significant factors toward quality. They have negligible gains and only at unrealistic levels. And while it mentions creativity by name as an example, it doesn't necessarily give more importance to it than any other. Going by those conclusions, I personally wouldn't include job-affiliated attributes for non-military roles at all if it wasn't for production speed, which is yet unquantified and arguably unnecessary.
Toady's test of creativity with carpentry quoted in that thread is too small scale to be conclusive and is at odds with Blue_Dwarf's take on it. But assuming it was a simple demo of something Toady knows to be factual rather than a quick attempt to remember how it was implemented, while also accepting Blue_Dwarf's conclusions, all affiliated attributes should have an equal weight. Likewise, I agree with Thistleknot that removing any contributing factors unilaterally is not a good idea. Until proven otherwise, I think the assumption that affiliated attributes do affect jobs and do so equally is the logical approach. As they do have some effect, many players will want them included for tradesdwarf roles, but we're on the same page in thinking the overall weight for attributes should be set quite low for those roles. I'd pin it at 0.2, essentially serving as tie-breakers. But this would have to be set in individual roles rather than the master weight control in options as it is now because of military roles. An exception would be str and agi, where there are significant considerations outside the scope of direct bonuses to job performance.
I disagree with weighing preferences that low, however. For players that wish to get the best dwarf on the job right away, high skill and low pref weights would be ideal. But if that's your main concern, you can find those by sorting on the labour view rather than using roles. I put greater importance on preferences and would rather train an unskilled dwarf with better potential from scratch, so I think roles should be used for facilitating that. Again, it's personal play style and you can't please everyone, but I think that makes the most sense if you're carefully tuning roles for best overall matches. That said, I was unaware that skill ratings are capped similar to attributes. I never noticed that, but I haven't trained many dwarves to legendary other than miners. And I thought variable skill learning rates was something implemented by mods, not in vanilla (beyond xp values of different products). If xp rates and potential for skills vary dwarf to dwarf, then that certainly does far outweigh preferences for long term job matching, though I maintain that prefs far outweigh attributes. (BTW, I don't see skill potential or rate represented in the 20.1 options.)
I don't know what percentage of players use mods. I haven't tried any yet. I understand the desire to make the roles universal with broad preference categories, but IMO doing so breaks the utility of roles when they include many inapplicable materials/products. Defining them individually is a bit of work initially, but that's the point of defining roles and is most ideal. I don't know if it's possible to include materials for popular custom mods in DT without crashing, having it just ignore those which don't exist. But that would put maintenance on your shoulders when mods evolve or new ones come out. I think you should stick to roles for vanilla while promoting, documenting and facilitating custom roles so that modders and their fans can create and distribute them for specific mods.
Traits. Yeah, beyond the little that we already know I think it will be hard to unveil these. They are the most abstract. Strength is straightforward as a concept, easy to associate and measure. You might not have damage numbers and hit points, but there are hit ratios, arena outcomes, hauling distance/tick - measurable
outcomes. Things like endurance and focus are easy to qualify, but harder to quantify. Measuring them usually requires observing processes rather than results, and is disrupted by too many factors like thirst, distance, and random "do something else now" AI. Traits, such as modesty and dutifulness, are much harder to qualify beyond speculation and almost impossible to quantify. The interactions of all these things with various aspects of the game are amazingly intricate. I'm not surprised if strength affects muscle mass and therefore reduces critical wounds. I am amazed that someone without a corkboard of post-its outlining design ideas (ie Toady himself) came up with that
and possibly found evidence in game. I think there are many other interactions, even for "understood" variables and principles, which will never be discovered without help from Toady. Still, I believe people will find ways to measure these and discover more when they set down to it, either through ingenuity or painstaking trials.
Thanks for the info on happy thoughts from materials/products. I'd appreciate if anyone can confirm whether or not anvils, buckets, etc used to build a workshop are physically present in the workshop location to be noticed and admired.
I know the inner workings of games generally remain a mystery, and discovery is hard earned by hackers and modders. I guess there's a pride and possessiveness, as well as real intellectual property issues, with all the mathematical complexities that compose good AIs and artificial personalities. Even if within game dev circles it's all straightforward math and code, it will be achieved in different ways to different effect. For players, this instils mysticism and an archaic quality to games, allowing us to just enjoy the unpredictable experience rather than formulating
win. I'm all for that, yet I'm torn because this stuff kinda keeps me up at night, or stuck on pause for hours on end. I'm like that, I over-analyze and try to understand mechanics of any game I respect. I never got deep into programming, and have only been a moderate modder, but I like the puzzle and appreciate the art of how it comes together as much as playing the game itself. So I kinda wish some of these basic things would just be revealed.
Anyway, I'd like to help draft some roles for public scrutiny, but first I'd like to establish some standards and get a sense of what's most appropriate and desired for defaults. So I hope there's additional feedback on this discussion.
Author
Topic: Dwarf Therapist (Maintained Branch) v.37.0 | DF 42.06 (Read 1002121 times)