Skills that "top out" early.

[Twilight]

I wanted to open up a discussion on skills being worth raising all the way through level 27. Some skills arguably are almost always worth continuing to increase, whether they're at level 5, or 25, such as Fighting. There are others, which are worthwhile, but see diminishing returns as they get near level 27, such as Stealth (I think). And then there are skills that peak really early, such as weapon skills, especially for lighter weapons.

People tend to get their weapon skill to the minimum delay level, say around 14-16 for a smaller weapon, as I recall, and then ignore it after that. Either there are good reasons to continue to raise that skill level and most people aren't aware of them, in which case I'd like to know what they are, or there aren't really many reasons to continue past 16 or so, in which case I think there should be.

What's the point of skill levels 17-27, if the huge majority of the advantages are already gained before then? Another example might be Poison magic skill, which people often turn off relatively early (Before 10 or so?), and never increase it again. I'm not saying you should *have* to increase these skills, but that there should be potential worthwhile gains to doing so, so that it's a strategic choice of whether to continue down that path and reap the rewards, or to focus on other things. From a gameplay perspective, hitting the minimum weapon delay at weapon skill 14, and turning it off, is boring, choice wise, and arguably somewhat spoilery.

To me, at least, skills seem like things you'd always want more of, and that would always continue to help you out the higher they got. Choosing to stop advancing in a skill because there are other skills you want or need more is a strategic choice to weigh before making it, with reasons to do or not do it. Choosing to stop advancing a skill because it's mostly pointless seems counter-intuitive: "Yeah, you could gain more levels in X skill, but it wouldn't really do much anyway..."

If it does so little after a certain point, then shouldn't the skill's advantages be more evenly spread out over the 27 levels of skill, even if that means giving that skill more benefits at the higher levels?

Basically, every skill should continue to give reasonably useful advantages throughout the 27 level range. This doesn't mean that it's optimal to go down that road, since continuing to gain more Poison magic skill for a fire elementalist might not be the best choice compared to other choices, but it should at least give some kind of advantage.

It just seems to me that some skills give strong benefits throughout the whole range, and others peak pretty early. I would think they all should. Or are there strong, but not necessarily easy to notice benefits for all skills?

Anyone notice this, or have any examples of skills they turn off early, because it's pointless after a certain level? (I mean for what the higher levels give you, not because focusing on higher levels of that skill doesn't suit your character's particular career path...)

Just interesting in hearing other people's opinions on this, and possible solutions (If I'm not just imagining this problem).

Thanks!

[galehar]

Sure, we are aware of the problem. I'm not sure about weapon skills. To reach minimum delay with the biggest 2 handers you need to almost max them. Is this really a problem to stop training axes if you want to stick with a broad axe, and keep training it if you want to use a bigger one? The damage formula could be tweaked to weight the skill a bit more, but stats too need more weight.
It's a bigger problem with short blades. One solution would be to merge short blades and stabbing, something dpeg has talked about for a while.
UC is fine, maybe even too good at high level.
I don't know about ranged combat skills but it could probably use some rebalance anyway.
T&D is never trained very high. There are proposals to make it more useful by allowing the player to plant traps using wands. Even then, you wouldn't train it all the way to L27. It's just a "utility" skill, I don't see a playstyle centered on it being very exciting. Because that's what raising a skill to L27 means.
I don't think pure stealth/stabby playstyle ever disable stealth or stabbing, but I might be wrong.
The defensive skills (armour, evasion and shields) are always useful, no problem here.
For magic schools, poison highest spell is L6. We could add higher level spells, but most things resist poison in the late game. Hexes and Translocation highest spells are L7. Hexes has the same resist problem as poison. Translocation high level spells are already extremely good, higher level ones would be hard to balance. Or we could raise cblink's level.
Evocations has been significantly boosted, but I don't know if it's enough to support a pure evoc playstyle.
Invocations obviously depend on gods. I think skill investment is rewarded accordingly for most, but you can't rely only on your god because of the piety resource, so you can't afford to invest only in it.

[mumra]

The biggest problem with this is the non-obviousness of a lot of the caps. Working out the optimum level for certain skills requires complex spoilers about the various dependent equations, and a new player would probably assume that each level of a skill is an improvement and wouldn't realise there are limits of effectiveness. I think a solution would be to make the skill effects more transparent so you can at least clearly see when a training a skill has no further benefit.

[Galefury]

There is an entirely different view about useless high level skills held by some people, which I think is also valid: not every skill needs to be useful at high levels. Investing into something that is strong early and weak later is an interesting character development decision. This mostly applies to poison and to some extent to throwing, short blades and slings (if you don't find slaying), not to most weapon skills and T&D which just have diminishing returns at some point but stay very useful. Of course this part of character development requires knowledge of the late game, which can be seen as a problem.


About poison, assuming making it useful in late game is desired:
Add high level acid spells to the school? Interesting effects would be possible, like a high damage acid cloud that spreads and eats items, corpses and walls. Also the equipment destruction potential from acid spells could be interesting.

I think acid has a natural synergy with poison, both thematic and in gameplay. Equipped enemies rarely resist poison, while demons and most (all?) other enemies that resist poison come without equipment, meaning acid spells have no penalty against the enemies poison does not work against. Some enemies would need to get acid resistance of course, or damage could be lower than for frequently resisted elemental spells. So putting both poison and acid into the same school could be seen as bad (one school works against everything). But it would fix the poison school's uselessness at high levels. Adding acid as another element easily accessible by the player would require some balancing of course, and adding interesting new high level spells is a lot of work too.

Another option would be resist piercing like poison arrow has (still poisons resistant enemies, I'm not talking about the physical "arrow" component). Resist piercing Olgreb's Deadly Halo might be a nice level 7 or 8 spell. Poison's everything in LOS every turn, even resistant enemies, does not affect self, duration of around 20 turns. Kind of like a poisonous version of Tornado with lower damage and longer duration. Of course this has some of the same problems Tornado has.


About T&D:
I don't think Crawl is made for setting traps. It would work somewhat well with Ash, but having to scout first, then retreat when something dangerous turns up, then spend finite resources to create a trap and then lure the dangerous enemy into the trap just seems like an extremely tedious way to kill something before you normally could. Just blasting things with a wand avoids the tedium and is just as interesting, at least to me. Stuff like placing turrets based on the information announced to you to fend off a superior foe works well (the entire tower defense genre is built on that premise, and there are often scripted sequences in action games built around this). In situations that are not explicitly built around traps they always seem to facilitate AI abuse (BG2 is a prime example of a game with bad traps).

Making T&D more relevant is easy, just bump up frequency of dangerous traps in endgame and extended endgame areas or add new ones. Train it or get mutated, banished and killed by zot traps. Adding interesting new traps would be much better than just turning up frequency IMO.

Moving hell effects to a trap-like mechanic might be cool. It could work like this: hell effects do not immediately happen, instead a trigger area is placed randomly in LOS (connected area, somewhat irregular form, about 10 tiles total, possibly placed directly on the player, leading to an immediate effect). When you enter any of the affected tiles the effect goes off and the trigger area disappears. Next hell effect means a new trigger area is placed and the old one disappears. Effect frequency would have to be raised to compensate for the effect not always happening. High T&D could allow you to detect those areas (you sense a malignant aura), allowing you to walk around them.

Something like this could also be added to other areas with malignant powers, with lower frequency. Zot, the orb run and pan are all good candidates for this.


Tloc:
Raising CBlink to level 8 sounds okay. Maybe add an intermediate (level 6 or 7) semicontrolled blink spell that can be upgraded to controlled with ctele.


Weapons:
Doubled weapon speed means doubled damage for all weapons, regardless of base speed. So why should it be easier to reach min delay with lighter weapons than with heavier ones? Having some progression from lighter weapons to bigger ones with raising skill level is good, but I think it currently is a bit too steep. I think all weapons should reach min delay at skill level 20+, heavy ones at a higher level (20 plus min delay could work nicely). The speed progression should be fast at low levels, then slow down for higher ones.

Alternately keep it simple: all weapons reach min delay at a fixed level, and attack speed increases linearly with level. Increase the to hit gain with skill level, tie damage more closely to stats. The progression from light to heavy will be caused by lack of accuracy with large weapons at low skill levels.

Another problem in my opinion is that twohanded weapons are heavier than onehanded ones. The progression from light to heavy is also a progression from onehanded to twohanded, so if you use a shield (which most people do) there is no reason to train skills further. Making some onehanded weapons heavier and slower and some twohanded ones lighter and faster (without changes to damage and min delay) might work. This would give shield using chars something to train for (cant use that demon whip without high skill anymore) and make twohanded weapons a bit more attractive (you can start using them with lower skill). Decoupling min delay from base delay would open up some balancing possibilities by allowing direct tuning of the amount of skill a weapon needs without affecting endgame performance, but would make calculating delay more complicated. An overview of combat performance with various weapons in inventory (delay, damage, accuracy, this has come up in talk about making combat stats more transparent) would be necessary to make this work. Of course delay is not the only thing that could change, there are many weapon stats. I just think delay is a pretty natural fit because it is currently the stat that most affects how effective a weapon is at low skill.


Charms:
Charms is another skill that is frequently trained only to medium levels. Haste --> 7? High level charms other than Death's Door?

[galehar]

About poison, assuming making it useful in late game is desired:
Add high level acid spells to the school? Interesting effects would be possible, like a high damage acid cloud that spreads and eats items, corpses and walls. Also the equipment destruction potential from acid spells could be interesting.

I think acid has a natural synergy with poison, both thematic and in gameplay. Equipped enemies rarely resist poison, while demons and most (all?) other enemies that resist poison come without equipment, meaning acid spells have no penalty against the enemies poison does not work against. Some enemies would need to get acid resistance of course, or damage could be lower than for frequently resisted elemental spells. So putting both poison and acid into the same school could be seen as bad (one school works against everything). But it would fix the poison school's uselessness at high levels. Adding acid as another element easily accessible by the player would require some balancing of course, and adding interesting new high level spells is a lot of work too.

There's a proposal on the wiki (by Eronarn) about changing poison to alchemy and adding acid to it.

Charms is another skill that is frequently trained only to medium levels. Haste --> 7? High level charms other than Death's Door?

They are not as commonly used as haste, but death's door and ring of flame are powerful spells, so I wouldn't say that charms stop being useful once you can cast haste at excellent.

[Bim]

+1 to acid and poison, it seems to be the only spell school that is easily resistible (other than hexes, although thats slightly different) by nearly everything after Hive.

I've mentioned it before, and I can't seem to find the wiki (galehar, if you can find it that'd be really useful) of having a level cap 'bonus' when you reach that high. For instance having 27 stealth making monsters occasionally forget they saw you when in plain sight, or having 27 M&F giving you an additional weak area of effect attack. This would get round the problem in a more interesting way, but I firmly agree that making all spells more useful all the way up would be better.

I think traps is a difficult one, at the moment it seems abit of a waste as other than the odd Zot trap they don't really do much, and increasing frequency I think would just be annoying. Perhaps if traps could be moved? It'd make interesting kill corridors, which wouldn't necessarily be overly powerful or tedious, and would add some nice tactical gameplay.

[galehar]

I've mentioned it before, and I can't seem to find the wiki (galehar, if you can find it that'd be really useful) of having a level cap 'bonus' when you reach that high.

Here you go. I haven't commented it and haven't had the time to think about it much, but I support this idea.

[nrook]

I think it's okay that certain skills top out early, as long as they do so in an intuitive, predictable way. High Poison Magic skill does exactly what a player would guess: it makes poison magic more effective and powerful. It has only niche relevance because of the lack of high-level poison skills, but that is fine. The player's intuitive understanding of what high Poison Magic skill should do (make poison spells easy to cast and powerful) fits with what it actually does. Similarly, if Traps and Doors still scales to 27, it's okay that a T&D of 15 detects 99% of traps, as long as T&D 27 detects 99.9% of traps. (I don't actually know if this is accurate. If T&D 15 detects 99% and T&D 27 detects 99.1%, that is bad.)

If a skill has opaque effects and tops out early, though, that's a problem. Weapon skills are the main culprit here: an unspoiled player would likely never figure out the mechanics by which weapon skills increase melee damage, and thus might train Axes to 27 just to use a broad axe.

[mumra]

Similarly, if Traps and Doors still scales to 27, it's okay that a T&D of 15 detects 99% of traps, as long as T&D 27 detects 99.9% of traps. (I don't actually know if this is accurate. If T&D 15 detects 99% and T&D 27 detects 99.1%, that is bad.)

The thing here is that if a situation like this exists, it's a HUGE waste of exp that, if you knew about it, you'd much rather put into another skill than get a .9% increase in T&D reliability. I have no idea if the real numbers are like that for T&D but I get the impression there are some skills in Crawl that work exactly like that; i.e. massively diminishing returns for an exponentially increasing investment! (which of course is completely counterintuitive...)

What I'm thinking would be great, interface-wise, is if the skills screen had an "info mode" where you can press the skill letter and see a list of all the stat changes you would get from the next skill level with your current gear, spells, etc. Now this would obviously not be at all easy and might even require revealing certain stats that are not currently available in the game. But a lot of the skills currently require spoiler information to figure out your optimum build and that's obviously not a good thing.

[Galefury]

Micromanaging skill levels will not save you, it will only let you handle dangerous situations earlier. Realizing when you cant handle a situation and getting the hell out of there is what actually keeps you alive and lets you win. Absolute strength is much less important. Micromanaging skills is unnecessary and extremely tedious (and actually takes away from time that could be spent learning when to run away), so I think it should not be encouraged. I don't want to have to think about whether the next level of T&D or the next level of ice magic will help me more, and if I can easily look up the stats I have to decide and think about this. The status quo is fine, just switch the skill off when you feel satisfied with its performance. If you're not sure how many levels you need you can either just switch it off and see what happens, or look up spoilers.

Edit: I agree that some information should be made accessible in game. For example a rough measure of weapon dps would be extremely useful. But the increment in skill performance from gaining a level is certainly not among those useful stats.

[mumra]

This isn't a discussion of how to win the game; escape tactics have nothing to do with optimum ways to build your character. Any character, no matter how tough, needs to know when to run away. I've lost three awesome XL:27 characters making that mistake! But even so, the better a character's skills are, the less situations they'll need to run away from; and knowing optimum skill levels will help that. This naturally encourages spoilers, although in this case you need a certain amount of mathematical ability to understand some of those spoilers! There are spreadsheets like the Spell and Weapons Labs which can help figuring out some of this, but even those are pretty complicated.

[Galefury]

I know about that. I was just stating the reasons why listing the bonuses from the next skill level would be bad. Most people, including me, and apparently you, have the urge to min/max, which is not necessary to win and is boring. So exact numbers should be kept out of the game as much as possible.

[lordfrikk]

Weapons could have a line stating how fast is the respective weapon skill allowing you to attack, e.g. "Your skill with axes allows you attack at top speed with this weapon." - "Your lack of skill in attacking with axes hinders your attack speed with this weapon." Literally the same as is now done with converting magic resistance/stealth numbers to adjectives.

P.S.: Perhaps even add a line when you reach a milestone ("You Axes skill is now 14! You feel more confident in using your current weapon, allowing you to attack at top speed.")

[MonorailCat]

Weapons could have a line stating how fast is the respective weapon skill allowing you to attack, e.g. "Your skill with axes allows you attack at top speed with this weapon." - "Your lack of skill in attacking with axes hinders your attack speed with this weapon." Literally the same as is now done with converting magic resistance/stealth numbers to adjectives.

P.S.: Perhaps even add a line when you reach a milestone ("You Axes skill is now 14! You feel more confident in using your current weapon, allowing you to attack at top speed.")

You should even be able to compare weapons in some easy way.
Compare two weapons and then "You feel as if your +2 quick blade would do more damage over time than your +3 mace"

[Shade]

You should even be able to compare weapons in some easy way.
Compare two weapons and then "You feel as if your +2 quick blade would do more damage over time than your +3 mace"

My biggest wish would be for an in game 'Comparinator' that let's a player compare any two things of a given type (insofar as you have them identified or they are know). My fantasy Comparinator would list the relative positives of any two given things of a given type in easy to read list form. This would be based on your current skills and attributes, mention and hidden maximums, and mention any potential convergence points.

[Bim]

I think a 'comparinator' is a great idea, although I know some people would call it spoilery, I don't think it is really though, as most RPG's have a 'damage rating' or something, and different things are better in different situations.

I think the problem again is micromanagement, the trend seems to be to go away as far as possible from micro, and a comparinator is just asking for compulsive testing of everything. However I firmly agree that transparency should be a top issue.

(Thanks Galehar for the link!!)

[galehar]

This is a frequent request. Danr considered implementing but instead did the Weapon Comparison Spreadsheet.
What could work is a screen showing all the weapons you are carrying with some stats similar to the I screen. It would show an adjective for accuracy, and dash bars for speed and damage. The dash bars would be like the spell power bar so you could tell if you can keep improving by training your skills more.

[Bim]

What could work is a screen showing all the weapons you are carrying with some stats similar to the I screen. It would show an adjective for accuracy, and dash bars for speed and damage. The dash bars would be like the spell power bar so you could tell if you can keep improving by training your skills more.

+100 to this, as long as it didn't encourage hours or trawling through trying every different weapon up against each other, I think it would be great.