Angband 3.5-dev

[Djabanete]

Re: Sil abilities: when I first played Sil one of things I really enjoyed was that whenever I picked an ability, I could immediately see how it impacted the game. Transparency and balance are key. I never felt cheated when I picked an ability. Skill trees can be noob-friendly (also, see Diablo 2).

That said, I don't think Vanilla needs skill trees.

Edit: ... for the reasons Mikko just mentioned. If anything, it needs to figure out what the heck is the deal with Charisma.

[Derakon]

One interesting take on crafting is what Dungeon of Dredmor does. There, crafting essentially is a hedge against getting screwed by the RNG equipment-wise. Crafting ingredients are plentiful enough that you should always be able to make the items you need to fill in the gaps in your equipment. A strong investment in crafting skills will also let you access high-level gear earlier than you "should" be able to get it -- but you've attained that skill in crafting at the cost of skill in other areas, and eventually that high-level gear will drop for non-crafting players anyway. In short, crafting smooths out the difficulty curve.

Diablo 2 is a bit problematic in my view mostly because there are many skills that don't scale well to the endgame. Most characters will tend to invest their skillpoints as soon as they level up, which results in a Sorceress with 3 points sunk into Firebolt, for example -- functionally useless as that spell is almost entirely ineffective after the first chapter. They did eventually mitigate this issue by making every skill give passive bonuses to related skills, so your points in Firebolt weren't completely wasted so long as you stuck with fire skills, but it's still disappointing to get to level 30 (or whatever), look at your skill tree, and wish you had done things completely differently.

[Magnate]

Originally Posted by Mikko Lehtinen

I like the Sangband skill system, but I find it a bit inelegant that the old, non-functioning stat & skill system still exists, too. The two systems don't really connect IMO.

You need to do a bit of source-diving in Sang - this criticism isn't really fair. The "old" skills (melee, shooting, perception, disarming etc. etc.) are very carefully governed by the player's chosen skill expenditure - they're basically just ways of communicating the results of the new skill system. Stats are also carefully meshed with the skill system. It's really very well done indeed. The main reason I don't think V needs any kind of player-selected skill development is because it's already been done - Sangband *is* V-with-skills (or rather, O-with-skills).

I love how Sil reboots everything and starts with a brand new system without any legacy problems.

I really must play Sil. It sounds as if it has borrowed the best bits of angband and mixed them with the best bits of Crawl.

JFTR, I am actually quite interested in v4 (not V) one day having some sort of forging or crafting mechanism. But again, it's done very well in S (where it is inextricably tied to the skill system), and I haven't yet thought of any improvement on that. I was hoping that Craftband would mature and then we could just nick the best ideas from that ...

[Magnate]

Originally Posted by Derakon

Diablo 2 is a bit problematic in my view mostly because there are many skills that don't scale well to the endgame. Most characters will tend to invest their skillpoints as soon as they level up, which results in a Sorceress with 3 points sunk into Firebolt, for example -- functionally useless as that spell is almost entirely ineffective after the first chapter. They did eventually mitigate this issue by making every skill give passive bonuses to related skills, so your points in Firebolt weren't completely wasted so long as you stuck with fire skills, but it's still disappointing to get to level 30 (or whatever), look at your skill tree, and wish you had done things completely differently.

I've never really understood the idea that being "new player friendly" somehow means "new players can optimise immediately". To me, part of the fun and replay value in a game is that I can play a few characters, know that I'm making suboptimal choices, and then play my next characters knowing that I'm making much more effective choices. This is precisely why D2 lived for so long on my hard drive (and ditto Titan Quest, though that allowed re-speccing). It seems to me completely self-defeating that a game would steer a new player towards optimum choices, as that diminishes replay value.

Now that's not to say that the choices should be in any way obscured or misleading - they weren't in D2, it's just that you needed the experience of playing the higher levels to understand what your optimum choices would have been (if you could have survived without spending all your points early - another feature which rewarded experience and better play).

To me, "new player friendly" is something different. It's about the choices being clear and easy to understand what they're about, not necessarily easy to see what the best choice is.

[ekolis]

Originally Posted by Derakon

Diablo 2 is a bit problematic in my view mostly because there are many skills that don't scale well to the endgame. Most characters will tend to invest their skillpoints as soon as they level up, which results in a Sorceress with 3 points sunk into Firebolt, for example -- functionally useless as that spell is almost entirely ineffective after the first chapter. They did eventually mitigate this issue by making every skill give passive bonuses to related skills, so your points in Firebolt weren't completely wasted so long as you stuck with fire skills, but it's still disappointing to get to level 30 (or whatever), look at your skill tree, and wish you had done things completely differently.

Perhaps this could also have been overcome with a "respec" feature - pay some price (say, gold or XP) to delete some of your skills and regain the skill points for use elsewhere? Dungeon Defenders and Space Pirates & Zombies both feature this...

[Derakon]

Originally Posted by Magnate

I've never really understood the idea that being "new player friendly" somehow means "new players can optimise immediately". To me, part of the fun and replay value in a game is that I can play a few characters, know that I'm making suboptimal choices, and then play my next characters knowing that I'm making much more effective choices.

Mm, there's a difference between suboptimal choices and unplayable choices, and D2 will certainly lead you to pick the latter if you aren't careful. I'm all for the tightly-optimized everything-planned-out character having a (much) easier time then the "screw it, let's just throw darts at the levelup screen" character, but I do think the game should be playable, and more importantly, fun either way.

Part of this, in D2's case, is that many skills are just out-and-out terrible (Magic Arrow, anyone?), and/or get completely outclassed a few levels after you get access to them (c.f. Firebolt). Better-designed skill systems ensure that no skill is useless, even if some skills are better than others.

Respecs are a common answer to this problem, and they do solve it in a way, but they're kind of an obvious rules patch. "Whoops! We did a crap job of designing the skill system, so let's let players fix their characters after we let them break them!" The better solution is to keep the problem from happening in the first place.

[Mikko Lehtinen]

Originally Posted by Magnate

You need to do a bit of source-diving in Sang - this criticism isn't really fair. The "old" skills (melee, shooting, perception, disarming etc. etc.) are very carefully governed by the player's chosen skill expenditure - they're basically just ways of communicating the results of the new skill system. Stats are also carefully meshed with the skill system. It's really very well done indeed.

I know! S is one of my all-time favorites really.

You can call the Sangband system "elegant". It really is a masterwork. But it is elegant in an extremely baroque way that reminds me of RPG systems of yesteryear. The way that it works is non-transparent and requires lots of effort to understand. Thankfully the manual does a really, really good work at explaining it all.

I don't think anyone else except Leon could have pulled it off.

I have a really hard time communicating what I think of Sangband.

***

Angband's main strengths are resource management and interesting equipment choices. IMO Angband's stat/skill system should play to these strengths as much as possible.

I see the way that your equipment interact with your primal statistics as the key to the system. There is lots of untapped potential here. If every stat was important to every class, and the system was transparent to the player, how much more interesting would equipment choices become!

I'm afraid that any kind of skill tree system just hinders this interaction or makes it more opaque. In Angband, stats should be king.

[d_m]

Originally Posted by Derakon

Respecs are a common answer to this problem, and they do solve it in a way, but they're kind of an obvious rules patch. "Whoops! We did a crap job of designing the skill system, so let's let players fix their characters after we let them break them!" The better solution is to keep the problem from happening in the first place.

I don't totally agree.

Respecs do have this property but have another important one--they reassure the player that skill choices, etc. can be reconsidered.

Apart from whether e.g. summon golem is "good enough" to justify being a skill, is whether I enjoy casting it. If it turns out I don't enjoy summoning (which I couldn't have known until I tried it) it's a drag if I'm committed and can't reconsider.

[Magnate]

Very well said on both counts - "baroque" is an excellent description of Sangband - especially impressive if English isn't your first language. As with music, some like baroque and some don't, but it's undeniably elegant.

I also agree with you about V - the 'skills' are really just ways of summarising the interaction between your race/class/equipment and your stats. We certainly need to do more with all three of races, classes and stats (including re-working or removing Charisma) - hopefully someone will leap in and try some stuff in v4 - it's an as-yet unclaimed area.

[Magnate]

Originally Posted by d_m

I don't totally agree.

Respecs do have this property but have another important one--they reassure the player that skill choices, etc. can be reconsidered.

Apart from whether e.g. summon golem is "good enough" to justify being a skill, is whether I enjoy casting it. If it turns out I don't enjoy summoning (which I couldn't have known until I tried it) it's a drag if I'm committed and can't reconsider.

But it isn't though! If you had fun trying summoning, and stopped when you decided you didn't like it, you start a warrior or a rogue or a non-summoning mage and have more fun - what have you lost? I don't think we can plan for every player to enjoy every build, so there will definitely be those which don't work for some. The key is making this experience of discovery fun, and allowing the player to choose easily to start over with something more suited to their playstyle.

I really hate the re-spec concept. I agree with Derakon that it's a horrible kludge which breaks the fourth wall. To the limited extent that these games are RPGs (and I know they're not really) it really ruins it for me that you can just 'unlearn' stuff.

That said, I don't think you can make every skill useful to every build either. I agree that every skill ought to be useful to *some* builds, but there's no point aspiring to proof every possible choice against redundancy. Sometimes players will make bad choices (e.g. a ranger keeping Cubragol and chucking Bard) - as long as the game allows them to learn and understand which choices were wrong and why, and that that process is fun, I don't see a problem.

In case anyone is wondering, I think this is relevant to angband development, e.g. for spell choices if you can't learn all spells. It doesn't only apply to skills.