This message is intended to describe changes that I think need to be made to improve Sangband. Now that SAngband1.0.0 Final is out, and hopefully all the polishing of the interface is done (or mostly), I'd like to spend some time discussing features and changes to the game and game balance. Many of these are from my unofficial version of the game, some are new. Feel free to add your own thoughts and ideas here as well, or to comment on those of others.
Without further ado, here are the changes that I feel are necessary:
1) Skill Proliferation: As is, it is too cheap to keep a skill above the minimum cost limit, which leads to people investing in too many skills (IMO).
2) Shadowstalker: There are currently too many obstacles to playing a shadowstalker type character, many of which can be fixed, but they should be done in a balanced fashion.
3) Martial Arts: Currently aren't very competitive compared to very high end weapons, a little too good in the midgame, and lack flavor
4) Oath Requirements and Restrictions: Aren't completely intuitive yet, and should all start being available at the same time, IMO.
1) Skill Proliferation: This is probably the simplest to fix. If you've looked at the ladder, there are a lot of players and winners who've invested in 10, 20, or even all the skills they're allowed to do (me included). This does a couple of things. First, it means that you get a crappy score. Second, you've now seen everything so there is nothing much new to the game. Third, it means that you're basically always behind in ability, and its hard to get experience without infinite patience. To solve this, we merely need to change one function.
int tmp_pow = MAX(2, (p_ptr->power - 8) * 8 / 10);
This means that the game will pretend that you have a skill equal to (power - 8) * 8/10. In my mind this is too cheap of a cost, and allows for increasing a skill from zero (or other small numbers) too easily, or allows you to maintain too many skills at the minimum level. Thus I propose the following function:
int tmp_pow = MAX(1, -8 + (p_ptr->power * 8) / 10 + (p_ptr->power*p_ptr->power)/400);
This is the same, but adds a quadratic term such that at power 100, you get a final minimum level of 98, which makes investing in new things rather slower (by around a factor of 4), but doesn't penalize early investments of skills overly much. I've played with this function through several winners, and it does help maintain focus with characters.
Alternately, since the experience curve is almost completely exponential (except at very low and very high skills, it is), a simple offset is simple enough to use, i.e.
int tmp_pow = MAX(1, p_ptr->power - 10);
Benefits include being easy to compute on the fly in your head, and fairly straightforward functionality.
2) Shadowstalkers in Angband sound like fun, at least to me. However, they have some serious issues if you want to play them.
a) Inability to use scrolls at all, and devices really poorly, and opening locks/chests
b) Inability to cast spells
c) Nearly impossible to map terrain
In return, they get
a) Less damage from beam/bolt/ball (up to 33%)
b) Some invisibility and stealth and more dodging
c) More (and better) sneak attacks (if oathbound burglar)
Unfortunately, you need to escape to survive in Angband, and penalizing burglars in the dark doesn't make much sense when we're talking about thiefly things, and not being able to map terrain is a real pain. Thus, to make shadowstalkers playable, we need to implement the following:
a) Scrolls that deal with escape, detection, and lightness/darkness call all be read when no_light() is true.
b) Infravision should map the terrain
c) Lower penalties for devices, opening, disarming
Now, some of these penalties or advantages might only apply to oathbound characters. I would suggest that the first two should be available regardless, and the third should be restricted to oathbound characters.
Additionally, it'd be real nice to cast spells while in darkness, heck, it perhaps should even be preferred by necromancers (and disliked by priests?). Unfortunately, I think the advantages of shadowstalking have added up to enough to suggest that allowing this is not a wise course of action, as with a good infravision score, you'll only miss 1/4 monsters (cold blooded), though you'll need to use your light to pick up objects. In that vein, I suggest that if there is only one item you can light or douse, that it is automatically done, to save a keystroke after every battle for the shadowstalker.
This opens up to clear routes for a oathbound burglar, a shadowstalker and a caster, which sounds like reasonable balance to me.
3) Martial Arts are internally marked as a section for improvement in the Sangband code, and I agree. The fixed number of blows and lack of any accompanying flavor text make them pretty bland. That, and I think wrestling is far better than karate. Both are generally poor in the very late game, and a little too good in the early and middle game. Add in the fact that you are not dependent on equipment, and they critical strangely, and they start to look nothing like the weapons which they replace.
First, we have to decide what we want the skills to do. I like associating karate with dexterity entirely and wresting on strength entirely, and I'd have the skills grant the following:
Karate:
+1 Stealth@50
+1 Dex@30, 60, and 90
+1 Speed@50, 80, and 95
Free Action@75
Resist Confusion@90
Wrestling:
+1 Str@30, 60, and 90
+1 Tunneling@17,50,84 (this is how it is currently...)
+1 HP@ 51-100
Free Action@75
Resist Sound@90
Second, I'd like to:
a) Use the blows table somehow, so that players have the similar/same choices to make
b) Rely entirely on Dex (Karate) or Str (Wrestling)
c) Wrestling should max out at 4 blows, Karate at 6. Wresting shold do a little more damage in general, Karate should do more when branded with scrolls
d) Actually balance this thing at all levels.
A high level character with a big weapon (say, the glaive of pain) can deal up to 5 blows at 6d8 each, with, say, a +150 to 200% increase in damage from deadliness (str, gloves, and . Before criticals, this comes to 405 dpr. If you brand it, you can get another +10 or +14 per blow for up to 475dpr. Other high end weapons (Aule, Deathwreaker, Ringil, Zaracuthra, Crisduin, Eonwe, etc.) do similar amounts of damage, at least when their brands apply. Wrestling does at most 120 twice per round, however, it only takes experience, not drops to get it, so it should be a little worse damagewise than a high end weapon. For wrestling, let's target 320 dpr, or 80 per hit. For Karate, let's target 300 dpr, or 50 per hit.
To make it work the best, the damage should behave mostly like weapons. There should be a base damage based on the skill of the wielder, deadliness should apply (up to 84% from str and +117 with full str and +12 gloves). This means the target number for damage should be half that of previously stated for wrestling and karate. Each should probably be composed of a linear and a quadratic term, and should start at no less than 2.
Thus we might consider these functions for damage, though perhaps there is a better one:
Karate = 2 + skill/7 + skill*skill/700
Wrestling =2 + skill/5 + skill*skill/550
Ideally, these numbers would be tweaked (and lowered) for non-Oath of Iron Characters, but these might be a reasonable baseline, or maybe the difference in to hit will even that out just fine.
4) Oath Requirements and Restrictions
Perhaps I'm the only one who finds it weird that spellcasters get to take their oaths first, followed by burglars, followed by warriors. Personally, I'd like all the oaths to be available at the same time, probably starting around a skill of 20 in any of the appropriate skills (burglary, melee/ranged, mana/spellcasting/spell failure). As it stands, warriors get few bonuses over a non Oathed spellcaster, and that seems funny to me.
Also, I'm not sure I like how the restrictions on oaths are done currently. Back in the day, Oath of Iron characters had a restriction on magic devices at 50%. Now, they simply have a 50% penalty to using that skill. I prefer the latter version. I don't see why spellcasters don't have similar penalties to weapon (and martial arts!) use in that fashion. I'd like to see spellcasters get a flat 20 to 40% penalty onto their melee (and archery?) skill, rather than being capped at an arbitrary number. Of course, I'm not quite sure what those numbers should be, and they'll differ between priests and nonpriests.
Finally, I think the 50% penalty to magic devices is a bit extreme. I think a 1/3 penalty on magic devices should be sufficient to keep warriors from overusing them, and actually allow them some use with them.
--
There are two other pretty reasonable ways to limit a character's versatility:
a) Penalize skills by a fixed amount, as in warriors get a -10 penalty (or whatever) to their device checks
b) Change the costs of a skill post-oath. This is basically similar to #a (at least with exponential experience costs), but doesn't do funny things like make you suddenly worse at something after taking an oath, and means that you can raise everything to the same level with enough patience (which is probably a bad idea in Sangband, but maybe not). It will also induce interestingness in the calculation of power afterward (i.e. the game can think that you suddenly found and got a whole lot of xp)
I'm not sure which of the three options I like the most.
Without further ado, here are the changes that I feel are necessary:
1) Skill Proliferation: As is, it is too cheap to keep a skill above the minimum cost limit, which leads to people investing in too many skills (IMO).
2) Shadowstalker: There are currently too many obstacles to playing a shadowstalker type character, many of which can be fixed, but they should be done in a balanced fashion.
3) Martial Arts: Currently aren't very competitive compared to very high end weapons, a little too good in the midgame, and lack flavor
4) Oath Requirements and Restrictions: Aren't completely intuitive yet, and should all start being available at the same time, IMO.
1) Skill Proliferation: This is probably the simplest to fix. If you've looked at the ladder, there are a lot of players and winners who've invested in 10, 20, or even all the skills they're allowed to do (me included). This does a couple of things. First, it means that you get a crappy score. Second, you've now seen everything so there is nothing much new to the game. Third, it means that you're basically always behind in ability, and its hard to get experience without infinite patience. To solve this, we merely need to change one function.
int tmp_pow = MAX(2, (p_ptr->power - 8) * 8 / 10);
This means that the game will pretend that you have a skill equal to (power - 8) * 8/10. In my mind this is too cheap of a cost, and allows for increasing a skill from zero (or other small numbers) too easily, or allows you to maintain too many skills at the minimum level. Thus I propose the following function:
int tmp_pow = MAX(1, -8 + (p_ptr->power * 8) / 10 + (p_ptr->power*p_ptr->power)/400);
This is the same, but adds a quadratic term such that at power 100, you get a final minimum level of 98, which makes investing in new things rather slower (by around a factor of 4), but doesn't penalize early investments of skills overly much. I've played with this function through several winners, and it does help maintain focus with characters.
Alternately, since the experience curve is almost completely exponential (except at very low and very high skills, it is), a simple offset is simple enough to use, i.e.
int tmp_pow = MAX(1, p_ptr->power - 10);
Benefits include being easy to compute on the fly in your head, and fairly straightforward functionality.
2) Shadowstalkers in Angband sound like fun, at least to me. However, they have some serious issues if you want to play them.
a) Inability to use scrolls at all, and devices really poorly, and opening locks/chests
b) Inability to cast spells
c) Nearly impossible to map terrain
In return, they get
a) Less damage from beam/bolt/ball (up to 33%)
b) Some invisibility and stealth and more dodging
c) More (and better) sneak attacks (if oathbound burglar)
Unfortunately, you need to escape to survive in Angband, and penalizing burglars in the dark doesn't make much sense when we're talking about thiefly things, and not being able to map terrain is a real pain. Thus, to make shadowstalkers playable, we need to implement the following:
a) Scrolls that deal with escape, detection, and lightness/darkness call all be read when no_light() is true.
b) Infravision should map the terrain
c) Lower penalties for devices, opening, disarming
Now, some of these penalties or advantages might only apply to oathbound characters. I would suggest that the first two should be available regardless, and the third should be restricted to oathbound characters.
Additionally, it'd be real nice to cast spells while in darkness, heck, it perhaps should even be preferred by necromancers (and disliked by priests?). Unfortunately, I think the advantages of shadowstalking have added up to enough to suggest that allowing this is not a wise course of action, as with a good infravision score, you'll only miss 1/4 monsters (cold blooded), though you'll need to use your light to pick up objects. In that vein, I suggest that if there is only one item you can light or douse, that it is automatically done, to save a keystroke after every battle for the shadowstalker.
This opens up to clear routes for a oathbound burglar, a shadowstalker and a caster, which sounds like reasonable balance to me.
3) Martial Arts are internally marked as a section for improvement in the Sangband code, and I agree. The fixed number of blows and lack of any accompanying flavor text make them pretty bland. That, and I think wrestling is far better than karate. Both are generally poor in the very late game, and a little too good in the early and middle game. Add in the fact that you are not dependent on equipment, and they critical strangely, and they start to look nothing like the weapons which they replace.
First, we have to decide what we want the skills to do. I like associating karate with dexterity entirely and wresting on strength entirely, and I'd have the skills grant the following:
Karate:
+1 Stealth@50
+1 Dex@30, 60, and 90
+1 Speed@50, 80, and 95
Free Action@75
Resist Confusion@90
Wrestling:
+1 Str@30, 60, and 90
+1 Tunneling@17,50,84 (this is how it is currently...)
+1 HP@ 51-100
Free Action@75
Resist Sound@90
Second, I'd like to:
a) Use the blows table somehow, so that players have the similar/same choices to make
b) Rely entirely on Dex (Karate) or Str (Wrestling)
c) Wrestling should max out at 4 blows, Karate at 6. Wresting shold do a little more damage in general, Karate should do more when branded with scrolls
d) Actually balance this thing at all levels.
A high level character with a big weapon (say, the glaive of pain) can deal up to 5 blows at 6d8 each, with, say, a +150 to 200% increase in damage from deadliness (str, gloves, and . Before criticals, this comes to 405 dpr. If you brand it, you can get another +10 or +14 per blow for up to 475dpr. Other high end weapons (Aule, Deathwreaker, Ringil, Zaracuthra, Crisduin, Eonwe, etc.) do similar amounts of damage, at least when their brands apply. Wrestling does at most 120 twice per round, however, it only takes experience, not drops to get it, so it should be a little worse damagewise than a high end weapon. For wrestling, let's target 320 dpr, or 80 per hit. For Karate, let's target 300 dpr, or 50 per hit.
To make it work the best, the damage should behave mostly like weapons. There should be a base damage based on the skill of the wielder, deadliness should apply (up to 84% from str and +117 with full str and +12 gloves). This means the target number for damage should be half that of previously stated for wrestling and karate. Each should probably be composed of a linear and a quadratic term, and should start at no less than 2.
Thus we might consider these functions for damage, though perhaps there is a better one:
Karate = 2 + skill/7 + skill*skill/700
Wrestling =2 + skill/5 + skill*skill/550
Ideally, these numbers would be tweaked (and lowered) for non-Oath of Iron Characters, but these might be a reasonable baseline, or maybe the difference in to hit will even that out just fine.
4) Oath Requirements and Restrictions
Perhaps I'm the only one who finds it weird that spellcasters get to take their oaths first, followed by burglars, followed by warriors. Personally, I'd like all the oaths to be available at the same time, probably starting around a skill of 20 in any of the appropriate skills (burglary, melee/ranged, mana/spellcasting/spell failure). As it stands, warriors get few bonuses over a non Oathed spellcaster, and that seems funny to me.
Also, I'm not sure I like how the restrictions on oaths are done currently. Back in the day, Oath of Iron characters had a restriction on magic devices at 50%. Now, they simply have a 50% penalty to using that skill. I prefer the latter version. I don't see why spellcasters don't have similar penalties to weapon (and martial arts!) use in that fashion. I'd like to see spellcasters get a flat 20 to 40% penalty onto their melee (and archery?) skill, rather than being capped at an arbitrary number. Of course, I'm not quite sure what those numbers should be, and they'll differ between priests and nonpriests.
Finally, I think the 50% penalty to magic devices is a bit extreme. I think a 1/3 penalty on magic devices should be sufficient to keep warriors from overusing them, and actually allow them some use with them.
--
There are two other pretty reasonable ways to limit a character's versatility:
a) Penalize skills by a fixed amount, as in warriors get a -10 penalty (or whatever) to their device checks
b) Change the costs of a skill post-oath. This is basically similar to #a (at least with exponential experience costs), but doesn't do funny things like make you suddenly worse at something after taking an oath, and means that you can raise everything to the same level with enough patience (which is probably a bad idea in Sangband, but maybe not). It will also induce interestingness in the calculation of power afterward (i.e. the game can think that you suddenly found and got a whole lot of xp)
I'm not sure which of the three options I like the most.
Comment