What are proficiencies trying to model?

Game Talk The Garage
1

I’m having a bit of trouble wrapping my head around what proficiencies are supposed to be modeling.

I mean, I get the idea that blacksmithing might be a specialization of fabrication obviously - but its the implementation that confuses me. It doesn’t exist as an advanceable sub-skill, it’s just kind of a binary one-and-done thing? At some point I have banged upon enough metal and the epiphany occurs? One minute ago I wasn’t a blacksmith, and now I are one?

Granted, we run into that same problem with all the martial styles, but I thought that was considered to be a problem with those to eventually be corrected - not a model to emulate new skill systems on.

Wouldn’t it be more straightforward to create sub-skills that advance on their own tracks, and feed back into the parent skill’s advancement?

So, I can’t start on blacksmith until I find a blacksmithing book, or maybe not until I reach fabrication 3 (or both), but then when I do blacksmith specific tasks or read blacksmith books, I advance the blacksmithing skill, just like a regular one - AND I advance fabrication because that’s the parent skill.

It comes down to the fact that as a player, I don’t get what the proficiencies are trying to accomplish, as they don’t seem to gate anything, they just make the first task or two in a given proficiency category take a lot longer - and then they never come up again?

1 Like
2
3

All the proficiencies are fairly easy to learn individually. The idea is that they represent having done related things enough to know how to organise your tools, pattern your work, select the right components, and not make beginner mistakes.

In almost all cases I’m adding a little more granularity with some earlier level proficiencies, but there’s not enough here to merit representing with a full leveled skill. Each proficiency is a relatively small knowledge set. In the case of blacksmithing though, I’m adding a general metalwork proficiency you get first, that spreads the penalties and bonuses out over a little more time.

1 Like
4

Cool. That first step is a doozy though.

Looks like you’d be jacking up skill learning times by a factor of 10 or even 100? More realistic certainly, but it’ll have dramatic effects on many, many areas of game-play to make adjustments for it.

I’ll go read through the discussion in the actual topic. Thanks!

5

Yeah once proficiencies are in a fairly good place, skills are going to become way way harder to learn. That will require having enough proficiencies, and also making it so that proficiencies are a bit gated by skill level, so you still get advancement when stuck at level 5 for a week.

6

Ok, that wasn’t as long a read as I feared.

Couple quick thoughts not so much to do with proficiencies, as with the MUCH longer skill learn times that are proposed. That’s fundamentally far more important than the proficiency system itself, with much more halo on the game.

  1. Bump the skill learning tracks, probably bump skill/recipe granularity from 10->100.

Doesn’t REALLY change anything, but it lets you break up the math a bit more nicely, and more importantly it gives the player a sense of concrete advancement over a much, much longer grind, and that kind of perception is quite important. Yeah, the ding from 26->27 doesn’t really do anything but improve my fail rates slightly, but at least it DID something.

  1. I think you want to get rid of ‘shared pool’ character gen options. Stats, Skills, and Traits should have permanently separate pools. Why? Because right now most characters are created as amnesiac imbeciles with great stats and no life experience whatsoever - because doing otherwise is foolish and crippling.

You can learn most of the skills you’ll need for basic survival within 24 hours, and develop advanced skills within 1-2 weeks - obviously that will be impossible with dramatically slower skill development, so the option to create the amnesiac imbecile should be a bizarre outlier challenge profession, not normal. Splitting up the pools permanently ensures that players have points to spend on starting skills and professions appropriately, without sabotaging their permanent stats and traits, and will make the earlier phases of a game with a much longer skill learn cycle far more realistic and playable. A standard character should probably also start with a fairly sizable skill pool, rather than a tiny handful of points.

  1. Probably will have to rework focus considerably. It’s a frustrating mechanic as is, and dragging out the focus maintenance game-play over MONTHS rather than days is a recipe for tooth grinding tedium of the worst kind. Not saying trash focus, but seriously rethink how the pool works, how fast it drains/recovers, what factors affect it and so on. It isn’t scaled for the kind of time factors you’re suggesting.

One possibility is to attach focus to specific skills. It burns down quickly when you are practicing a given skill, allowing you to advance the skill very quickly for a relatively limited amount of time each day, then your pools recover overnight, encouraging non-monotonous gameplay cycles for players. Do project A for a bit, go kill some stuff, come back, work or project’s B & C, go loot, etc. If your character is able to maintain a generally high mood, their focus pools recover to higher values so that the next day’s training is more productive.

As long as it’s not… read, eat, read, read, read, read, eat, read, read, read, sleep, eat, read read read…
(repeat for next 60 game days to master fabrication, begin again for tailoring)

That’s only something we put up with because we get past that phase of the game relatively quickly and get on into the endgame - but now that phase could easily extend into several months making it untenable.

  1. Consider changing books into ADJUNCTS to skill use and practice, rather than an either/or replacement. You always advance skills by use, but if you have a book on you that covers the relevant skill levels, it significantly speeds xp gain and reduces fail rates for challenging projects - again, some effort needs to be made to tone down the emphasis on sitting still and reading for skill gain if the process is going to be exponentially longer.
7

Honestly its going to be tricky. The skill system in many ways sets the pace of the entire game, and so any major change to it fundamentally changes the game.

For the most part it will make advancing skills (other than combat skills) far less important, and it will make found gear and weapons far more important than crafted ones, because unless a great many other things change, you can expect to get power armor long before you can craft survivor armor, and you’ll have trunks full of medieval weapons scrounged from mansions and pawn shops months before you could consider crafting your own, unless you happen to start a character with the correct proficiency at birth. Players will likely want to spend a lot more time focused on developing combat skills using found gear until quite a long time into the game, rather than learning to craft - which, to be fair, IS relatively realistic.

Likewise found and barely maintained vehicles will of course become the norm for a much longer span of the character’s life, unless they start as a mechanic. Again, makes sense - but a big change going forwards.

On the plus side, seasons have already changed to a far longer cycle - but that had almost no game effect. For most games it just means eternal spring weather because a character is likely to die or be abandoned in the endgame before summer arrives - so those won’t have to change.

8

To me, this just makes it more grindy, because it reduces the accomplishment of gaining a level. It turns ‘I got from level 6 to level 7, now I have a bunch more I can do’, into ‘I got from level 71 to 72, nothing’s really different, and I’ve got a few more levels until there’s anything meaningful’. It basically kills the reward in advancing a level.

Having more distinct steps makes the march towards where you want to be just seem longer.

I think we should ditch both systems.

Yep, focus should turn into a long term monitor of skill gain. It should turn into a ‘I learn X minutes worth of this Y, so I lose Z focus’ sorta thing - representing your capacity to learn things over some sort of period.

This one’s easy - books no longer teach skills, they contain recipes and practice methods. You learn skills from performing this practice, not just reading the book.

Awesome, that’s one of the primarily goals.

The quantity and quality of the various medieval gear you find is probably too high and far too good.

And yeah, if you want to be an apocalypse savant who wants to do everything, that’s a significant time expenditure. But you aren’t alone. You don’t have to be good at everything - you have friends. Maybe you can’t keep your couple of cars all up and running, but hey, Natalie, who was a mechanic before this all happened can. Or you can’t make and repair a bunch of the clothing you need, but Dave, who was a student before the apocalypse who enjoyed tailoring in his spare time, learned how to.

9

Sim already basically covered it but yeah. We want it to start feeling more like an apocalypse game, less like an apocalypse skinned Minecraft. Part of that means using a lot more salvaged junk!

The other side of it is that we’re trying to make timeframes a lot longer. Making items from scratch takes days, but those days should pass quickly. When you need to spend two weeks doing a big project, and it passes pretty much automatically over a short span of real time, then having a skill level take a month in game is much less of a big deal. However this will make food security get more important again. Basically we’re moving towards the last few stages of time decompression, something that’s been going on since seasons were changed away from being 14 days.

2 Likes
10

If were going to be spinning huge amounts of time in abstract mode, the concept of zed respawns and wandering hordes is going to have to become a good bit more concrete and understandable.

As it stands now I quite regularly wake up bright and early each game day and need to go out and literally MOW THE LAWN OF ZEDS that have magically sprung up around me each night, regardless of how many I killed the night before, or so it seems.

Not that I have any issues with new spawns or mobile hordes in principle, but the seem extremely vague atm, and if we’re required to camp for months at a time to train rather than just tooling around in yon mobile fortress, that magical vagueness will be a lot less amusing.

I’m reminded of a game I played a couple years ago when I got hunted down and shot in my sleep by an npc, savescummed it several times setting up more elaborate traps and barricades to my cellar hideout to no avail. He’d always magically telespawn in while I slept and kill me. :joy:

1 Like
11

I think you can make this whole extended timescale work - as long as it doesn’t turn into a real player timescale, or become a micromanagement nightmare, or a muddled hidden mess.

Time abstraction has to be simple and natural in appropriate environments (fully cleared neighborhoods, safe LMOE shelters, etc), and abstract threat detection has to improve greatly, such that it functions much more naturally than paging through a few thousand interrupt messages per hour. Long range threat detection events like ‘There is a horde approaching from the SE.’ Would probably become important when dealing with these timescales.

The player’s activity cycle also has to naturally support a lot of ‘down time’ for low strain activities, such that each day tends to require a fair amount of waking rest, rather than 16 hrs of nonstop combat/running, followed by 8hrs of sleep - but still allow for at least a few high intensity days in a row without becoming crippling. I know for example that I can do two very high intensity days in a row without too much detriment, but if I go for a third I’m feeling it pretty hard. When I’m in good shape the recovery from a stint like that takes 2-3 days, though they don’t have to be fully inactive days, just not more crazy days.

Most importantly, if the game is going to insist on more complex behavior cycles enforced by arrays of overlapping conditions like that in order to make space to spin time, then FOR GODS SAKES STOP HIDING CONDITION DATA.

So much of the player frustration over the new mechanics is because every new health/condition mechanic feels like its intentionally hidden or so vague as to be misleadingly useless. Plz, for the love of god stop doing that if we’re going to have 6 of these condition systems overlapping each other. Get rid of Self Aware and just let the player have a decent picture of how they are doing - otherwise testing and balancing so many overlapping systems will be functionally impossible. The fact that these systems overlap and interact will introduce sufficient uncertainty without having to actively hide information.

Doesn’t have to be hard numbers, but the display bars at least are far more viable than ‘you feel meh’. That doesn’t mean anything, because it offers zero sense of scale - it’s just some arbitrary designer breakpoint that could be 5 turns away from crippling disability, or 5 days. Unless you have a serious medical condition, most people can reasonably judge how tired they are on a scale of 1-10, and whether they are likely to suffer serious impairment - doesn’t stop them from doing stupid things anyway, but that’s the player’s decision too when they see their exhaustion bar dropping.

1 Like
12

One of the main oversights I see in this is that intelligence doesn’t play a role. If the time scale for learning skills goes form days to weeks and months than intelligence should play a role in how quickly you gain XP. As it stands now the asthmatic imbecile would be able to master advanced electronics and car mechanics about as quickly as a 200+ IQ cephalopod mutant.

This would be a big nerf to intelligence since it effectively gets rit of the benefit of faster reading. The only real reason to take it would be to prevent skill rust (might be a minor problem one these time scales). This could be fixed by making the XP bonus you get form books dependent on how much more or less intelligence you have than is needed to understand the book well. This way the asthmatic imbecile might have very little use for an advanced chemistry book beyond the recipes as he can’t understand it at all while the cephalopod is already brewing up meth and dynamite after just a few weeks of practice. It would also make the slow reader essentially free points and quick reader completely irrelevent since you would only read for leasure.

I also wonder how quick and slow learner interact with the new skill and proficiency gain system. A change in the point cost of these traits might also be in order because they would allow you to craft certain items significantly easier or latter with these traits.

13

Pretty excited for once proficiencies are fully implemented, though I have a question - Will old characters who have already mastered their perspective skill trees (8+ survival, 12+ fabrication, etc.) in older versions automatically get these proficiencies or will that not be the case?

A nice buff to INT would be that proficiencies will have less of a nerf to time (in the case of crafting) and a slight increase in success in other tasks like hacking or speech, skill learning getting a slight buff would be nice too. That way, if the reading buff that INT provides gets tossed out you get something to replace it.

14

Judging by current state of the game, no they won’t.

1 Like
15

Whew, I went and read through that character generation discussion over in github and there’s a lot to unpack there.

Point based build systems have always created tension between role-playing and power-gaming - which is frustrating - but they ALSO very importantly provide the essential floor for the game’s overall balance. In a completely free-form system, your feedback on difficulty from players could turn into completely useless garbage, as you start having to field feedback from players who are creating characters who aren’t just more or less effectively min/maxed, but are in fact playing fundamentally different games.

Balancing point systems is HARD, and doing it when your granularity is so coarse is even harder. If the game were designed to report back to a central db where the designers could see day by day what the most popular stats/skills/traits were, then they might be able to tweak those various attributes and costs in such a way as to bring some comparative value to them. Right now there’s no real feedback, and not much in the way of tools to make general changes to stats/traits/skills (or their costs) in order to effectively balance them.

Bear in mind that no matter how you do it, there will always be a ‘meta’ regarding what builds are considered to be the best, but that’s a social construct even more so than a game balance one, and the only way to combat that is to change the game frequently in order to keep players exploring the balance.

As for creating a system that is based around the idea that you can ‘describe’ a character effectively based on a hidden point system - that sounds far more complex and difficult to balance than a point build system.

This is describing an attempt to render multiple programmatic analytical judgements about combinatory systems that are hard for humans to readily analyze, and then jam that through some kind of value-to-text system that will have almost no granularity at all, likely reducing the player feedback into incomprehensible mush (think of the feedback difficulties of the current hunger/fatigue systems, which are examples of value-to-text UIs).

K.I.S.S.

Point build systems are a pain in the ass, but ultimately simple, devs just have to dedicate themselves to actually balancing them.

16

Now, if you ask me what I would do personally with CDDA, I’d do this:

Create a batch of ~30 random characters AT WORLD GEN. They’re permanently attached to the world they are generated in. You can’t create or reroll more without rerolling the world.

The character randomization adheres to a close variant of the existing stat pool system, and is templated based on the various professions.

So you create a world, and then you are presented with a reasonably long list of characters of different professions you could take into that world. If one dies, you can go and play with one of the other ones that are also stored with that world.

You can’t reroll them individually or create new ones in an existing world, you have to reroll the whole world and all the characters generated with it to see a new batch. Why? Because the worldgen process takes time, and that time cost encourages the player to actually examine the character list with some care before mashing the reroll button endlessly looking for a ‘perfect’ character.

This examination of a group of pre-gen’d characters has the potential to titillate the imagination and present concepts that the player may not have considered themselves, especially for the powergamers.

This exposure to randomized character concepts could lead to a greater variety of playstyles, which is what you’re ultimately going for with a character gen system, to help prevent the game from becoming stale.

You should still leave a completely free-form character build option available just to save time and frustration for anyone who just has an exact scenario they want to play and damn the system - no point making them jump through endless character randomization loops if they’re bound and determined to play a certain character, but the UI flow should make it clear that the presented characters are the ‘main’ way to play the game.

None of this really obviates the need to balance the system a bit better. If no-one ever takes a trait except for RP reasons, then it is over-costed or over-penalized and should be balanced better. Likewise for stats. And skills should NEVER be balanced against stats/traits, because the latter are permanent, and the former are learned, making them apples/oranges in game balance terms.

1 Like
17

Traits, stats, and point balance aren’t really related to the proficiency issue.

My original recommendation for proficiencies at start was to have them linked to the level of the associated skill, but since they’re not linked to skills anymore, we should probably have proficiencies another thing you can purchase with skill points.

Having intelligence affect proficiency penalties might be interesting, it’s worth suggesting in an issue.

18

Yeah, tying proficiency into whatever character building becomes shouldn’t be too difficult.

My musing in this particular thread isn’t really about proficiency despite starting there - it’s more the game-wide shift to this much longer character development scale and the general halo effects that will need to be sorted out as a result of it, which are numerous.

In terms of chargen, the halo is mainly that if skills take a very long time to develop, then chargen will almost certainly have to be more generous about starting skills in some manner or other, because we can’t simply spam skill development to meaningful levels on the first day of play, as is strongly encouraged by the current system.

19

It would make more sense to have intelligence affect how quickly you learn a proficiency. The bonus shouldn’t be something to insane but something like a 30 % decrease in proficiency gain if you have intelligence 4 and a 25 % increase in proficiency gain if you have intelligence 12 with the bonus leveling off for any extra point beyond that to simulate deminishing returns. This would really have a meaning full impact on the game given the timescales we are talking about here. It would also make intelligence as a state more meaningful since there currently isn’t any meaningfull impact from having more or less intelligence.

A more nuanced system would be to emulate the current system we have with books only do the same with proficiencies. So you give every proficiency a intelligence difficulty, every point of intelligence that you are under it will incur a decrease the speed at which you learn that proficiency and every point of intelligence above that score will increase the speed at which you learn a proficiency. To keep this sane and prevent insanely smart mutants form mastering nearly every advanced skill before the end of winter we should give diminishing returns to every intelligence point you are above the difficulty with the cap being somewhere between 40-50%.

1 Like
20

The main issue with trying to balance Intelligence is that in real life, Intelligence is the most OP stat you can possibly have. Smart people learn things faster, they make more money, they have better planning skills, so on. It will be hard to realistically implement Intelligence without making it the only stat you really need. Because the smarter you are, well, the better you are. At basically everything.

So, let’s take inspiration from another aspect of in real life intelligence: derangements!
Very smart people in real life are usually odd to varying degrees of severity. Nikola Tesla was in love with a pigeon. Terry Davis was… well, Terry Davis.
It’s worth noting that these are extreme cases - both in terms of their intellect and their issues. Most smart people are just a little off. A guy I knew hated the color blue to the point he refused to eat or drink out of any containers that were colored blue and refused to get in cars with blue paint.

But this is the cataclysm, with zombies and magic floating eyeballs and who knows what else, and other things about humanity have already been made extreme, like the healing rate. It stands to reason that if someone who’s naturally smart is going to be a little off, someone who became smart(er) by injecting themselves with weird foreign genes or shoving metal in their brain is probably gonna be very off.

Maybe for every point you gain in Intelligence above 10, you have a chance to gain a permanent, irremovable derangement, ranging from slightly annoying (like being unable to wear any form of underwear or gaining mood bonuses from drinking milk but mood maluses from drinking water) to game changing (like being unable to use firearms because your character just hates them). Lower levels of the above-average Intelligence spectrum (10-13) would only result in minor derangements, like the aforementioned hatred of underwear, and you’re unlikely to get more than one and highly unlikely to get all three, while the higher you go (and the more extreme your methods of obtaining that Intelligence boost) the higher your chance to gain something a bit more serious.

Maybe these extreme derangements could even be lessened through some kind of late-game surgery/bionic/craftable medicine/RNG nether thing/whatever, but only lessened.

I have a lot of time on my hands now that real world events have halved my working hours so I’ve been taking C++ lessons, I’ll get around to trying my hand at implementing this eventually if no one else does.