Print

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

[ancistrus]

Sorry if this came up before. I did not find a mention of this on the development page and I only took a couple of minutes to search this forum.

Let me define some things before stating my suggestion.

A life of a dwarf soldier has 3 phases.

Phase 1:
The soldier is weak and slow. Sending him against goblins is pretty much guaranteed to cripple/kill him.

Phase 2:
The soldier is reasonably strong and skilled. Sending him against goblins will be successful, provided you make tactically sound decisions. There is a small, but acceptable risk of injury or death.

Phase 3:
The soldier is an unstoppable demigod and will obliterate anything you send him against. The risk of injury is practically nonexistent.

My suggestion: Make phase 2 last longer. Change nature of training, or nature of combat in such a way that phase 3 is almost impossible to achieve.

[Manveru Taurënér]

Sorry if this came up before. I did not find a mention of this on the development page and I only took a couple of minutes to search this forum.

Let me define some things before stating my suggestion.

A life of a dwarf soldier has 3 phases.

Phase 1:
The soldier is weak and slow. Sending him against goblins is pretty much guaranteed to cripple/kill him.

Phase 2:
The soldier is reasonably strong and skilled. Sending him against goblins will be successful, provided you make tactically sound decisions. There is a small, but acceptable risk of injury or death.

Phase 3:
The soldier is an unstoppable demigod and will obliterate anything you send him against. The risk of injury is practically nonexistent.

My suggestion: Make phase 2 last longer. Change nature of training, or nature of combat in such a way that phase 3 is almost impossible to achieve.

Personally I'd like to see this changed for most skills, not just military ones. Making legendary skills actually legendary and thus rare is quite a common suggestion.

[ancistrus]

I would prefer that this was treated as a separate issue, since overpowered military affects gameplay more than overpowered industry workers, or planters, or engravers.

[Manveru Taurënér]

I highly doubt Toady would only change the one though, if he decides to do so. More likely he'll let it wait until revamping the whole skill gain system, as that seems to be how he usually does stuff like that :>

[Ramaraunt]

I find the military slow to train. Then again, I keep having problems with naked dwarves so that might be why.

[PrimusRibbus]

Sorry if this came up before. I did not find a mention of this on the development page and I only took a couple of minutes to search this forum.

Let me define some things before stating my suggestion.

A life of a dwarf soldier has 3 phases.

Phase 1:
The soldier is weak and slow. Sending him against goblins is pretty much guaranteed to cripple/kill him.

Phase 2:
The soldier is reasonably strong and skilled. Sending him against goblins will be successful, provided you make tactically sound decisions. There is a small, but acceptable risk of injury or death.

Phase 3:
The soldier is an unstoppable demigod and will obliterate anything you send him against. The risk of injury is practically nonexistent.

My suggestion: Make phase 2 last longer. Change nature of training, or nature of combat in such a way that phase 3 is almost impossible to achieve.

I really really agree with this.

I think the best way to approach this would be a system of smooth diminishing returns on skill gain (for all skills, not just military). This would allow newb dwarves to be trained in a reasonable amount of time, but make demigod dwarves a product of a longstanding fortress.

[GavJ]

Simple fix to make everybody happy all at once--military and non-military skills--with barely any need for playtesting for Toady, and zero need for tweaking balance:

Just add the experience amounts per skill level per skill to the RAWs.

Spoiler (click to show/hide)

Mining
Dabbling: 1
Novice: 1000
Adequate: 2000
etc.

Brewing
Dabbling: 1
Novice: 2000
Adequate: 5000
etc.

...

Done-zo.

[Helari]

Skill gain could be relative to the form of training. While a characters skill increases, the benefits of less intense training would diminish.

Quite high skilled characters would gain less and less skill experience from teaching or being a student as their skills increase. They could still teach or learn as maintenance that prevents skill rust.

Learning from combat or sparring could be dependent on the overall attributes and skills of the opponent. A superhumanly strong grandmaster fighting a weak dabbler shouldn't learn anything new from the fight.

Attaining a legendary level of skill would then require a legendary challenge. This would also fix the common exploit of strangling zombies or throwing rocks forever for lengendary skill. Could also fix danger rooms if the traps have an appropriate threat level.

edit: Example:

a raw recruit gains alot of skill xp from being taught sworddwarfship (the xp gain and also the diminishment curve influenced by the teaching and sword skill of the teacher) but a skilled dwarf gains alot less xp from being taught. Once a dwarf is skilled enough it becomes impossible to increase his skills with teaching and he can only learn by doing. Fighting equally or more skilled foes grants more skill xp while fighting less skilled foes grants less xp. Beating unconsicious, incapacitated or otherwise nonthreatening (comparatively very low skill) enemies grants no xp. After a while strangling that dismembered zombie or dodging those danger room traps is just no longer a challenge.

[GavJ]

^ That would be even better, but even still I'd say throw it in the raws, as just a parameter "exponent by which doing stuff decreases in experience given over time" rather than trying to tweak it all in defaults or hardcoding.

Set to 1, and it's like it is now.
Set to 2, and experience gets quadratically HIGHER the more you do stuff
Set to 0.5, and experience gets quadratically LOWER the more you do stuff

Whatever you want, yeah? (In addition to being able to modify the tiers of experience per skill)

[MDFification]

Right now, skills require more XP per level gained. The only problem with this is that as you level up, most tasks end up being done faster. In the case of weapon skills, where XP is gained per hit (I think) and higher levels increase the chance of getting hits considerably, this leads to growth actually speeding up over time rather than slowing down. So really the system's already in place, it just needs balancing.

It's important to note that most skills do not need a nerf. From what I've seen, shield/armor use does not increase very fast at all. Basically, with the exception of weapon skills and maybe dodging, right now most military skills increase too slow as long as training is considered patrol duty. At current rates if it weren't for the rapidly growing ability to parry and dodge, your military would frankly be crippled as you'd loose too many per siege unless resorting to traps/turtling before they could become competent. Keep in mind that in-game, these dwarves are training for years; soldiers not reaching basic competency in 1.5 years of boot camp is flat-out ridiculous, but it's what's currently in-game now. I can get things taking years to completely master, but right now the lower tiers of military skills take far too long to learn.

@GavJ: The capability to adjust the rate at which a creature learns a specific skill is already in the raws. This doesn't stop exponential growth, sure, but it does slow it down. So skills would get harder to pick up and take longer to master as opposed to just taking longer to master, but it's pretty close to what the thread's suggesting.
Also, I've noticed a lot of sentiment lately that Toady should just put everything in the raws. If he does that, then he's basically just making an engine with its own unique programming language.

 

[Manveru Taurënér]

...
Also, I've noticed a lot of sentiment lately that Toady should just put everything in the raws. If he does that, then he's basically just making an engine with its own unique programming language.

That's kind of the plan afaik. From the dev page:

MORE RAW FILES, (Future): Everything should be moved out of the code and into the raw folders, insofar as the processor will allow it. There needs to be a generalized framework for things like plant and animal related materials and products. Not all of the building types will make it out, but workshop and furnace types are certainly good candidates.

[GavJ]

king longer to master, but it's pretty close to what the thread's suggesting.

Toady thinks that Toady should put everything in the raws:

Core53, MORE RAW FILES, (Future): Everything should be moved out of the code and into the raw folders, insofar as the processor will allow it. There needs to be a generalized framework for things like plant and animal related materials and products. Not all of the building types will make it out, but workshop and furnace types are certainly good candidates.

edit: aw, sniped.

[Putnam]

Total XP required for skills right now is quadratic, 50x²+950x, where "x" is the level you want to reach. That's total experience. The actual increase in experience per level is 100, thus the quadratic boost (linear increase of experience required to gain a level per level means quadratic total experience required per level--calculus!).

Probably a good way to adjust this would be to... make the amount required for levels different. Perhaps something a bit more fast-growing later on. I guess you could always increase the order of the polynomial. Heck, maybe split it up--every level after 10 or 14 requires twice as much experience instead of just 100 more experience, making it exponential rather than quadratic past that point.

Basically, current system total XP:

1. 1,000
...
14. 23,100
15. 25,500
16. 28,000

Half-assed proposed system:

1. 1,000
...
14. 23,100
15. 46,200
16. 92,400

etc.

Though cubic growth might be good, too (using (50/3)x³+(2950/3)x):

1. 1,000
2. 2,100
3. 3,400
4. 5,000
5. 7,000
6. 9,500
...
15. 71,000
16. 84,000
17. 98,600
18. 114,900
19. 133,000
20. 153,000

[Melting Sky]

Skill gain could be relative to the form of training. While a characters skill increases, the benefits of less intense training would diminish.

Quite high skilled characters would gain less and less skill experience from teaching or being a student as their skills increase. They could still teach or learn as maintenance that prevents skill rust.

Learning from combat or sparring could be dependent on the overall attributes and skills of the opponent. A superhumanly strong grandmaster fighting a weak dabbler shouldn't learn anything new from the fight.

Attaining a legendary level of skill would then require a legendary challenge. This would also fix the common exploit of strangling zombies or throwing rocks forever for lengendary skill. Could also fix danger rooms if the traps have an appropriate threat level.

edit: Example:

a raw recruit gains alot of skill xp from being taught sworddwarfship (the xp gain and also the diminishment curve influenced by the teaching and sword skill of the teacher) but a skilled dwarf gains alot less xp from being taught. Once a dwarf is skilled enough it becomes impossible to increase his skills with teaching and he can only learn by doing. Fighting equally or more skilled foes grants more skill xp while fighting less skilled foes grants less xp. Beating unconsicious, incapacitated or otherwise nonthreatening (comparatively very low skill) enemies grants no xp. After a while strangling that dismembered zombie or dodging those danger room traps is just no longer a challenge.

This would be complicated to code but looks like an otherwise very good solution. This solution prevents grinding exploits which tend to suck the fun and realism right out of a game while at the same time it is quite logical and intuitive from a real world perspective.

If I remember correctly danger rooms are on the list of exploits to be fixed at some point and this would cover a lot of potential alternative exploits with a similiar mentality behind them. The only part I don't agree with would be to remove all the experience from defeating inferior foes. Since an enemy can usually only be killed once, there is no need to remove all the experience from killing even an extremely inferior one. The only place I think where exp should be nerfed to zero or really close to it is when the dwarf is doing the exact same thing over and over and over again such as the examples your provided of wrestling the same zombie for a week or dodging wooden spears for eternity.

[Helari]

This would be complicated to code but looks like an otherwise very good solution. This solution prevents grinding exploits which tend to suck the fun and realism right out of a game while at the same time it is quite logical and intuitive from a real world perspective.

I'd imagine you could create a suitable algorithm for skill xp gain and then use the stats of both fighting critters as values (or perhaps the "chance" of the critters action to further simplify).