Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[LordBaal]

How do you prepare for the next "touchy feely" update? I imagine you doing some mental benchpresses or something with the eye of the tiger song in the background.

[splinterz]

attributes are tricky.  as i understand it strength and agility effect the speed of performance of every task. so really they should be everywhere=nowhere.  no doubt you know more about this than i do so perhaps i'm overlooking something? my focus was to get the preferences into somewhat better shape.  for attributes i followed the information on the wiki but, a big but, that itself is based on the assumption, untested as far as i can tell, that skill use which trains certain attributes also benefits from them.

science is needed both for the attributes and for further refining the preferences. not to mention the traits...  i don't feel like i am up for that amount of rigour and counting of tic's at this point in time.  probably means waiting until i am hit by a strange mood and hoping that dwarf therapist is involved.    or maybe someone else will get keen.

don't know how i did that with dodger... 1.01 version now up.

it's true that since they're doing things faster strength and agility can be applied to every job, but i don't agree that it means they should be completely removed. it's not like preferences where either they have it or they don't; you still want to know who will work faster. yes, this often means that military end up with higher ratings in almost every role (especially if you have a high attribute weight), but it's legitimate, and easy to filter out.

the only reason i brought it up was because it appeared you'd removed strength and/or agility from some of the military roles, which in my opinion, are the most important to have high movement speed in.

the preference changes look good, and traits are another unknown. i guess that for sure discipline could be added to military roles, since it's been proven that, in part at least, factors into whether or not a dwarf will flee from combat. there were a few others... bravery and violent maybe?

How do you prepare for the next "touchy feely" update? I imagine you doing some mental benchpresses or something with the eye of the tiger song in the background.

haha, more like, gather up some whisky and pretend i haven't heard anything about it.

[feelotraveller]

it's true that since they're doing things faster strength and agility can be applied to every job, but i don't agree that it means they should be completely removed. it's not like preferences where either they have it or they don't; you still want to know who will work faster. yes, this often means that military end up with higher ratings in almost every role (especially if you have a high attribute weight), but it's legitimate, and easy to filter out.

the only reason i brought it up was because it appeared you'd removed strength and/or agility from some of the military roles, which in my opinion, are the most important to have high movement speed in.

the preference changes look good, and traits are another unknown. i guess that for sure discipline could be added to military roles, since it's been proven that, in part at least, factors into whether or not a dwarf will flee from combat. there were a few others... bravery and violent maybe?

hm, i'm basically in agreement with your comments but not the conclusion.  but let me babble a bit.

first up the reason for the changes i made were that some of the military roles were using the wrong set of attributes.  it looked to me like some time in the past a copy paste had been done from one role to all the others with no regard for differences in attributes, and, at least according to the wiki, there are differences.  so, for instance, endurance was missing from some.  it was on that basis that i made the changes.  it is possible, of course, that i made a dodger type mistake somewhere...

strength (and agility) are used by some roles more than others.  applying a strength weighting to every role (in addition to whatever is already there) is a possibility but in my opinion would constitute noise rather than good information.  the military view(s?) already present the attributes separately.  to be honest i would be more inclined to add in recuperation and disease resistance as relevant military characteristics, probably with higher weighting for melee types, but then i don't know how to start quantifying these.

could a basis other than the wiki could be used for the attributes?  sure.  but personally i would want to do a lot of science before going down that path.  i noted in my last post what i think the fundamental point of departure for that science should be - figuring out if indeed the attributes trained by skills are used by those skills as bonuses. if they are not then the attribute set dwarf therapist currently uses is misguided.  on the other hand if the attributes are used by the skills then it is a matter of figuring out how much, and possibly also when or where.  that is to actually look at proper weightings

someone else might be in a better position to do that.  (note - i also removed the 2x agility weighting for miners, and supplied a sleight of hand creativity weighting of 0.25 for gem cutters to try to cover the cases where they produce crafts.  the engraver i think i left at 2x creativity but it troubles me that creativity has no significance for smoothing... but a big impact with engraving.)  the wiki has links to a couple of bits of research looking at creativity and speed/agility but beyond those i am not aware that anybody has really looked into this.  now who was it who was counting woodcutter chops...? 

Edit: to play devil's advocate... an alternative defence strategy is to dig in and build fortifications, and in this case the speed bonuses are going to be wanted for, say, the miners, masons, and stone haulers.  the point in case being that the speed is an independent vector of attributes which rather than contributing to a role/skill directly is applied situationally.  (i.e. if it makes all faster it then is a question of which should be made faster, and the answer to that is relative.) apologies for having a foot in my mouth (or is that a boot in my face) while trying to explain myself here. 

[splinterz]

let me be clear, i'm not advocating adding strength and agility to every role, nor am i saying that the original roles, especially military didn't require adjusting.

i honestly don't follow your thoughts on the relevance of attributes and their relationship with skills in determining the best fit for a job. i think part of this is due to a different perspective as to what a role is. from your explanations you appear to be looking at the roles as a definition of a particular job, as opposed to what they were intended for, which is finding the best dwarf for a job.

for example, according to the wiki, and my understanding of your points, the furnace operator should not have agility as part of the role, since the job doesn't directly use it. however, a dwarf with higher agility will always perform the job better since they'll gather the required items for the job faster. this is the same point i was making about recuperation/disease resistance in the military roles. they're not going to be trained or directly used for combat, but they will make a better soldier by helping keep them alive longer.

your example about defensive bunkering actually supports the notion of adding str/agi to every role. doing so you could safely ignore the military roles and you'd still be choosing the faster/better dwarves for the jobs based on the role ratings. speed as an example is tricky because it's not only related to attributes, there are a lot of other factors that DF takes into account (inventory weight, curse, body size, terrain...)

although on another look the roles that had strength removed from them appeared to only be ranged roles, which is probably fine since now they focus more on agility.

[feelotraveller]

yes, something of a storm in a teacup.  just the ranged roles and dodger.  (i replied without looking at into the details.)

more generally, i agree that a furnace operator with higher agility will be better at the job.  i don't understand why it would not make a dwarf quicker at getting the items for all roles, or other forms of movement for that matter.  (not sure about bookkeepers though).  my reservation about including this as a relevant attribute is that with higher agility this dwarf will be better at all roles - or at least all roles which require movement and gathering...  (quick thought experiment - leaving skill aside, with a crazy high weighting on agility - say 1000 - the highest agility dwarf will be best at every role.)

i do differ somewhat on the purpose of the roles.  i see them as enabling informed choices/decisions about which dwarf to give which labours.  knowing which dwarf is best for a job is definitely part of it.  however sometimes it means not assigning the best dwarf for a job because they are already overloaded.  urist is already the best armoursmith, weaponsmith, architect and cook, do i really want to make him milker as well?  or deciding that for three dwarfs and three jobs that the best assignment is for each to do what they are close second best at since this best maximises their usefulness to the fortress (maybe avoiding having a dwarf doing something that they are 'really' bad at).

as an exampls from my current game at embark Onul was terrible at just about everything (but by far the best architect) and a decent carpenter.  when she turned out to have a preference for beds she was in.  she is weak and incredibly clumsy.  given a weighting for strength and agility (on top of what is already in the carpenter role) she would have looked like the worst candidate for a carpenter.  another of the starting seven was in the best couple of dwarfs for just about everything i might have spent points on (high strength and agility, excellent spatial sense) but ended up being mechanic/siege engineer because they had terrible kinesthetic sense and only decent creativity while several other dwarfs had really good creativity and passable kinesthetic.  give all roles an extra weighting for strength and agility and this dwarf would end up as outright best for everything and probably assigned as a mason (by me) even though they are better suited doing something else. that is the sort of problem i see arising.

i am more on board with the disease resistance and recuperation for the military suggestion because these attributes are not that relevant for other roles (although weavers, gatherers and fishers come to mind...).  a consequence of adding them, however, would be watering down the influence of strength (or focus), agility, kinesthetic sense, etc on suitability for these roles.

hope that makes sense.  you don't have to agree. 

[splinterz]

yes, something of a storm in a teacup.  just the ranged roles and dodger.  (i replied without looking at into the details.)

more generally, i agree that a furnace operator with higher agility will be better at the job.  i don't understand why it would not make a dwarf quicker at getting the items for all roles, or other forms of movement for that matter.  (not sure about bookkeepers though).  my reservation about including this as a relevant attribute is that with higher agility this dwarf will be better at all roles - or at least all roles which require movement and gathering...  (quick thought experiment - leaving skill aside, with a crazy high weighting on agility - say 1000 - the highest agility dwarf will be best at every role.)

i do differ somewhat on the purpose of the roles.  i see them as enabling informed choices/decisions about which dwarf to give which labours.  knowing which dwarf is best for a job is definitely part of it.  however sometimes it means not assigning the best dwarf for a job because they are already overloaded.  urist is already the best armoursmith, weaponsmith, architect and cook, do i really want to make him milker as well?  or deciding that for three dwarfs and three jobs that the best assignment is for each to do what they are close second best at since this best maximises their usefulness to the fortress (maybe avoiding having a dwarf doing something that they are 'really' bad at).

as an exampls from my current game at embark Onul was terrible at just about everything (but by far the best architect) and a decent carpenter.  when she turned out to have a preference for beds she was in.  she is weak and incredibly clumsy.  given a weighting for strength and agility (on top of what is already in the carpenter role) she would have looked like the worst candidate for a carpenter.  another of the starting seven was in the best couple of dwarfs for just about everything i might have spent points on (high strength and agility, excellent spatial sense) but ended up being mechanic/siege engineer because they had terrible kinesthetic sense and only decent creativity while several other dwarfs had really good creativity and passable kinesthetic.  give all roles an extra weighting for strength and agility and this dwarf would end up as outright best for everything and probably assigned as a mason (by me) even though they are better suited doing something else. that is the sort of problem i see arising.

i am more on board with the disease resistance and recuperation for the military suggestion because these attributes are not that relevant for other roles (although weavers, gatherers and fishers come to mind...).  a consequence of adding them, however, would be watering down the influence of strength (or focus), agility, kinesthetic sense, etc on suitability for these roles.

hope that makes sense.  you don't have to agree. 

to me, ensuring that dwarves aren't overloaded is either a decision made by the player, or the job of the labor optimizer; it's completely separate from a role rating. the rating is there to tell you 'this is the best dwarf for the job based on skills/attributes/traits/preferences'. what the player decides to do with that information has nothing to do with how the rating is derived.

in my opinion the problem with your examples is not with what attributes the roles contain, but the weights that have been chosen. if you value preferences over attributes for a carpenter, then you should adjust your weights to fit what you value more in determining the best fit for a job, not start removing attributes to try and make the preferences contribute more to the rating. now admittedly, if agi/str were part of every role, in your case it would require a lot of fine-tuning of the role weights if, for some roles, you value creativity over movement speed, etc.

it's impossible to have general roles that are going to fit every type of fortress and playstyle, so the goal of the default roles is to provide a reasonable starting/base point for particular jobs. anything beyond that is what the custom roles are for, and specifically why you can override the default roles with them. that said, your changes to the roles so far have been very reasonable and a great improvement over the much too generic original roles.

the concern about watering down the roles can be somewhat mitigated by adjusting the weights; there's no rule that the default roles' individual aspect weights must be 1.0.

adding recuperation/disease resistance to the fisherdwarf is a good example of what i mean with the default roles. is that a reasonable change for a baseline fisherdwarf role, or is it something more customized that should be left to a custom role?

[feelotraveller]

oh the bed preference was just an aside.  the reason Onul came out as a decent carpenter was an abundance of creativity and kinesthetic sense.  in fact the preference contribution to the overall will not change with adding or subtracting attributes.  what will change is the input that each of the attributes has into the overall attribute aspect.  e.g. bumping up agility proportionally devalues the contribution of all other existing attributes to the overall score (since the sum of all the attribute weightings=1 as thistleknot would put it) and yep the aspect weightings could be adjusted for each role to counter this but then the increase of the attribute aspect will need to take away from skill and/or preferences (or traits if they are added back in).  it has got to come from somewhere!

and i see where you are coming from and there is good sense in it.  but there are two distinct problems i see involved in tweaking the attributes. there is the question of proportional contribution to direct skill use and also the question of situational influence on skill.  the first can be solved by science.  but doing so will take time and is not a task i really want to undertake.  (i envisage, at least, a 3x3 fold comparison low, medium, high in both skill and attribute.  multiply that by the number of attributes.  run it a few times to get within acceptable bounds of randomness. then wonder if we might be missing interactive effects... argh.)  really this is just refining and correcting, to proper weights, what is already in the roles attribute-wise, assuming the information on the wiki is broadly correct.  or surplanting them if it is not.  so a massive task, but doable and would definitely help refine the roles.

but there is also what i call situational influence - the agility of the furnace operator (assuming the wiki is basically correct for the above).  yes a furnace operator with higher agility will function better but the scope of this depends largely on where they get their items from.  stockpile right next to the workshop or the other end of the map are the extremes.  and remember the start point of the job will not always be right next to the stockpile.  how to quantify that?  or the fisherdwarf and recuperation.  we need to know how often a fisherdwarf will get injured, or more correctly how much more often they will get injured from being a fisherdwarf.  once again this is going to be highly dependent on playstyle.  (=extrinsic and variable vs. the intrinsic and always involved of direct skill use)  this i think is best left for custom role modifications.

[khearn]

While strength and agility will no doubt make a furnace operator quicker, if you have reasonably set up stockpiles near the furnace the operator really shouldn't have to do much hauling, so the effect should be minor. Ditto for most any other workshop type job. If your shop workers are carrying stuff very far, you're setting things up wrong. If your gemcutter only needs to move 2 tiles from the center of his workshop to a stockpile that surrounds it to grab a rough gem, then 2 tiles back to the shop, being agile or strong won't make much of a difference. On the other hand, jobs that aren't just performed in a shop, such as miner, woodcutter, hunter, fisher, and hauler will benefit much more from agility and possibly strength. These guys have to cover a lot of ground getting from one work site to another.

One other issue which I haven't seen discussed in role ratings is the difference in effect of attributes and preferences. Good attributes will make the dwarf better at doing the job every time he does it. Preferences will make the dwarf better at making/using that particular item/material, and will make him happier, but won't make him better overall. If the dwarf likes an item he'll be making a lot of (beds for a carpenter, for example), it's pretty valuable. But if it's something you don't make very often, or at all (bucklers for an amorsmith), then it's not really worth all that much. A furnace operator who likes raw aluminum isn't as good as a furnace operator who likes tetrahedrite (assuming you have plenty of tetrahedrite and use it). So I think attributes should probably have more weight than preferences, and not all preferences are of equal value.

Some of this is dependant on the player as well. I equip my soldiers with a mail shirt and no breastplate most of the time, so a dwarf that likes breastplates won't make a better armorsmith for me. But if you use breastplates, then it's valuable for you. If I have a fort with no layers of gabbro, then I don't care if my mason like gabbro. But if I've got dozens of layers of it, then it's much more of an advantage. This probably means that I should make a copies of the default roles and grid views and make customized versions for the way I play, and for the fort I'm playing in.

   Keith

[feelotraveller]

Yes many preferences are highly situational.  This is why a while back I suggested having an alternative which highlights dwarf roles which have a relevant preference and turns the preference weighting to zero.  It is still an option I would like to see.    (It is also why i left out furniture preferences for glassmakers and so on, not because they could not be relevant but because they were unlikely to be so.  The assumption is, for example, that an advanced player embarking in the desert and diggng to set up magna glassworks will be savvy enough to set up roles for themselves for use in this situation.)  An armoursmith who likes steel is a great thing, unless you are on a fluxless (or to a lesser degree ironless) embark.  The default role preferences need to cover what can be commonly expected to arise.  Some of them could probably be weighted lower (i.e. how many hatch covers are made in an average game? ... more or less than floodgates, or armour stands?) but others, like steel for example, are more difficult to determine.  Did you have any concrete suggestions in this regard?

I do not agree that the solution is to make attributes contribute more (you can easily do this for yourself through the options by the way) since when you find a relevant preference is can be a powerful thing.  The likes steel armoursmith in my current game, is reliably churning out high quality armour at a low skill level even though on attributes they are far from the best candidate. 

[khearn]

Well, maybe a steel preference should have a large weight, since it is very valuable (most of the time). But a preference for, say, copper probably shouldn't have as much weight as good attributes.

[salithus]

I really need to learn more about how to use DT  effectively...

[feelotraveller]

Well, maybe a steel preference should have a large weight, since it is very valuable (most of the time). But a preference for, say, copper probably shouldn't have as much weight as good attributes.

I'm not so sure.  Copper is much more common (at more embark sites and in greater quantity when present) and easier to supply to your armoursmith  - one furnace job as opposed to five per 4 bars; the steel will also take an extra 4 fuel, or 8 if non-magma forges and a fair few more hauling jobs.  If your goal is to get your armoursmith up to legendary +5, or whatever, as quickly as possible (a valid approach if not one that i favour) then copper might be the better preference.  It might also be better if you want to equip a whole lot of dwarfs pretty quickly.  (If I was going to pick on any of the metal preferences I would be lining up Bismuth Bronze simply because of its general rarity.)  Optimally Dwarf Therapist should be catering for all playstyles.

Speaking of which after a bit of research on my part sparked by splinterz comments i now think that dresses, and probably cloaks and robes, but not trousers and the like, should be included as leatherworker preferences.  Comments?

[splinterz]

Yes many preferences are highly situational.  This is why a while back I suggested having an alternative which highlights dwarf roles which have a relevant preference and turns the preference weighting to zero.

it will be in the next update, along with highlighting in the tooltips:

Spoiler (click to show/hide)

the border highlighting will be optional, and the more matches, the darker the border is so multiple matches on a single preference will really stand out. it doesn't automatically set the preference weight to zero, as some people may prefer to have both a weight and the highlighting.

for now, the whole issue of attribute contribution to skills should be treated as an unknown. i'd focus on giving higher weights to attributes which have known effects on jobs until more is known. giving agi/str to jobs that require a lot of movement in general (hauling, gathering, fishing, etc.) as khearn pointed out, also seems reasonable.

i'm not sure the metal preferences really matter all that much. remember the primary benefit of the material preferences is for positive thoughts when working with or admiring things built from it, not making them work faster or more efficiently. although i suppose a dwarf throwing a tantrum due to unhappiness isn't being very efficient

you'll have to explain your thoughts on why you want to remove trousers, and what other items 'the like' is referring to.

[feelotraveller]

that looks really good.     (highlighting in the tooltips is even better than setting the preference weight to zero, imo.)

my in game observation (quite "unscientific", i know) leads me to believe that metal (material) preferences effect the quality of item production.  i could be wrong about this...  certainly item preferences have this effect.  (Edit; the wiki supports my observation! yay.)

Edit: for the leatherworker the suggestion is to add things back in.  currently the role has only items which are unique to the leatherworker, quivers, backpacks, etc.  i am responsible for this monstrosity.    looking at it yet again perhaps all clothing items should be included in the role?

[thistleknot]

That's cool splinterz. Amazing as usual.