Bugs and complaints on current master

[Gwarl]

Had another quick run, so I have more comments.

First, bugs. I found the same (largely useless) randart ring twice over in the dungeon. Each time I found it it came up in the object list as 'A Ring of & Ring'. So there appears to be a bug there. If you turn selling on in the birth options, shopkeepers will offer you the same gold for an un-ID'd potion as they will after you ID it, allowing you to deduce what it is without selling.

The new targetting system is nice, not having to aim past things in order to hit them. Is that done by adjusting the autoaim or has project() been rewritten?

Also I can't tell if uniques are dropping fewer artifacts, but it seems like it, and I can't tell if level generation has been modified, but it seems like it.

I want to make a case against no-save maledictions again:

It makes optimal fighting the same sequence of moves over and over for every class - slow, confuse, and stun the thing or you're losing HP for no good reason, whether you're a high elf mage or a half troll warrior. Once you've done that, even difficult uniques are trivial - I took down bolg and azog in a corridor on a level they should have murdered me without really breaking a sweat (playing a gnome rogue). Stun in particular is brutally effective. Confuse isn't really all that helpful anymore however, and slow certainly isn't what it used to be - so you could say it's only stun that needs fixing but that isn't the issue. The issue is that unless you're carrying these wands and using them for every difficult encounter now you're doing it wrong - which means optimal play requires more key strokes and inventory management - exactly the problem the trap overhaul was meant to solve!

There are already a good many changes to the fundemental mechanics of the game, certainly a new minor version is already justified without including another far reaching change such as this? As a compromise, since Derakon advocated this idea in the first place, can we get O style potion throwing with no saving throw, and revert wands/status effects to their 4.0.5 form, possibly with PowerWyrm's version of the formula for monster saving throws?

I think the other changes so far are working well and it would be cool to see them all implemented into a stable release before heading out into potentially gamebreaking uncharted territory with the status effect stuff.

[Derakon]

I also just realised that we already have this mechanism in the game for one item - potions of Dragon Breath.

Maybe what we actually need is to implement the ability to throw potions. That would, A, make some use of the under-utilised throw command, B, create a new use for the negative potions like Confusion/Slowness/Poison that currently get squelched after the first 'gotcha' moment when you accidentally drink one, and C, create an additional way for risk-averse characters to ID potions if they're too worried about negative effects to gamble on quaffing them.

I would absolutely be okay with replacing status effect wands with the corresponding potions, at least as an experiment. Add a line to the item description "It can be thrown to affect susceptible opponents", maybe.

Potions of Dragon's Breath are frankly kind of weird though. They're the only potion that prompt you for a direction when you drink them.

[Nomad]

@Nomad-
More or less. Give it fewer charges and it's likelier to blow up per use

Yeah, I suspected there was some sort of internal relationship going on, though it still might be handy to decouple the two to allow more explicit control for different types of wands.

In fact, single-use wands that automatically "burn up" when used might be an idea worth considering.

I also just realised that we already have this mechanism in the game for one item - potions of Dragon Breath.

Maybe what we actually need is to implement the ability to throw potions. That would, A, make some use of the under-utilised throw command, B, create a new use for the negative potions like Confusion/Slowness/Poison that currently get squelched after the first 'gotcha' moment when you accidentally drink one, and C, create an additional way for risk-averse characters to ID potions if they're too worried about negative effects to gamble on quaffing them.

[krazyhades]

The difficulty with rods...is that there's that temptation to sit around doing nothing to let the rod recharge, so you can use it in every encounter.

Yeah this is true. Halls of Mist's intrinsic x-times-per-floor mechanic (for intrinsic ability activation, HoM doesn't do this with devices) is one approach, though of course it only really works with fully non-persistent floors and to some degree HoM's usually-disconnected stairs. Another approach is to have it recharge based on kills or xp gained or some such (and/or a trip to town?).

I'm not necessarily pushing for this sort of change, but the HoM way of recharging a character's intrinsic identify/cure/phase charges only (but completely) through reaching a new floor can be very nice in some ways. Ofc, this doesn't hold as well for the necessary bread and butter utility. It would be really stupid to have to pick and choose where to use your 2 charges of trap detection on a given floor or whatever, at least with the way Angband currently throws traps around randomly.

[Pete Mack]

@Nomad-
More or less. Give it fewer charges and it's likelier to blow up per use

[Nomad]

Is it currently possible to tweak the odds of a specific type of wand blowing up on recharge? It would be good if you could set some sort of 'instability' factor in object.txt so that more overpowered wands have higher odds of exploding any time you try to recharge them. (Then you could effectively create your single-shot wands by giving them a single charge and 100% odds of blowing up on recharge, though that's probably taking it a bit too far.)

Or, to be more radical, you could even remove the ability to recharge staves and wands entirely. The initial number of charges is all you're going to get, and once the last charge is gone, it self-destructs.

[Derakon]

An alternative might be to just have rods rather than wands or staves and give them a suitably long recharge time.

The difficulty with rods, which is already present with Rods of Healing/Restoration and with some artifact activations, is that there's that temptation to sit around doing nothing to let the rod recharge, so you can use it in every encounter. I can't think of any way to mitigate that issue short of having much more heavyhanded "clocks" (e.g. by aggressively spawning in new monsters on the level), so I tend to think that rods work best when their recharge times are long enough to be "once per encounter" but they finish recharging shortly after the encounter ends. I guess that can still work with status effects, but it has implications on how strong the status effects can be, because we have to assume the player is capable of laying them all down on at least one target in each fight.

The secondary difficulty with rods of course is that they represent a permanent addition to the player's arsenal -- at least until they get zapped by an electric attack. That makes their impact on player power correspondingly larger compared to a wand.

[Nomad]

I really like this idea. In fact, single-use wands that automatically "burn up" when used might be an idea worth considering. Basically like a scroll or potion, except it's implicitly aimed at something. Then you can't use them against everything, but they can remain powerful without being unbalanced. Kind of like Potions of Speed are now.

An alternative might be to just have rods rather than wands or staves and give them a suitably long recharge time.

[Derakon]

And yeah, status wands are currently way overpowered, but I must admit there was something gloriously satisfying about zapping a wand of stun monster and having it just work and actually do something useful. So I do think it's definitely worth continuing to experiment with ways of balancing always-successful effects so they're not so overpowering. In addition to fiddling with the effects, tweaking the number of charges might be a factor to think about - we're used to status wands being unreliable and having plenty of charges to spam, but no reason they can't be limited to say just 2-3 at a time before needing recharging.

I really like this idea. In fact, single-use wands that automatically "burn up" when used might be an idea worth considering. Basically like a scroll or potion, except it's implicitly aimed at something. Then you can't use them against everything, but they can remain powerful without being unbalanced. Kind of like Potions of Speed are now.

(Which is not to say that the current effects don't need rebalancing, just that I'd rather see the effects remain useful. )

[Pete Mack]

There's only a low level variety of healer, and they go down to a single ice storm--with useful side effects to all nearby monsters. At native level, their friends die too quickly for them to have significant effect.

[Nomad]

I see healers have snuck into the monster list? Neat, though I haven't really had a chance to see them in action yet, but I don't think they should be the same colour as ninjas - light green the same as druids would be more logical, maybe?

And yeah, status wands are currently way overpowered, but I must admit there was something gloriously satisfying about zapping a wand of stun monster and having it just work and actually do something useful. So I do think it's definitely worth continuing to experiment with ways of balancing always-successful effects so they're not so overpowering. In addition to fiddling with the effects, tweaking the number of charges might be a factor to think about - we're used to status wands being unreliable and having plenty of charges to spam, but no reason they can't be limited to say just 2-3 at a time before needing recharging.

And for a more off-the-wall option, what about combining all effects into a single wand/spell of "incapacitate monster" that gives you a random status effect? Keep them all fairly powerful and effective but make it impossible for the player to pick which specific one they're going to get.

[Gwarl]

I cloned the git and compiled yesterday (10th April) and took a run at it. Here are some comments:

New traps are working excellently. I don't really notice the difference until I feel like 'hey shouldn't I be detecting traps or something?' and the answer's no. Some of the new traps are pretty interesting too.

New ID system is really innovative and looks really good. It's a little frustrating at first, and takes some getting used to, but definitely an improvement.

Mixed feelings on the new cone breaths. I can see why it's been done but I'm not sure that I actually prefer it. Will have to play more to form a proper opinion.

I think unresistable status effects are the wrong decision. It makes things far too easy. Monsters need the chance to make a save. I can't think of any other game that works this way, and while it's great to see Vanilla getting some degree of character, this is too much at once. I can see that a lot of work has been done on it but I would honestly roll back this change.

Final note, decapitalising many of the monster names is jarring, but this falls under the 'could get used to' category.

[fizzix]

Perhaps hold should work like sleep in that a monster will emerge from it when they are damaged.

Another option is to give high power monsters the ability to "fix" problems. So some monsters should have a "cure_self" which means when they're debuffed, they take a turn to fix themselves. You could also give some monsters "cure_monster" potions, which allow them to cure out of effects some limited number of times.

Basically it would look like this.

Sleep monster: Effective against un-smart monsters (animals etc). No guaranteed effect (can fail). Monster wakes up when it takes damage. A monster awoken by taking damage from the player is temporarily hasted, and cannot be reslept for some number of turns. Sleeping monsters that wake up naturally have no haste, and can be reslept immediately. Sleeping monsters can automatically be "pushed past" by other monsters. Sleep is an early game option, and is primarily meant as an escape.

Hold monster: Effective against all monsters. Monster cannot move, but can still do ranged attacks and melee attacks. Held monsters will always do a ranged attack if available. Can be cured out of by monster abilities. Held monsters can always be "pushed past" by other monsters in back. This is primarily meant as a tactical way to deal with strong melee monsters. It is more powerful than a phase door, but more expensive as well.

Stun monster: Increase spell fail rate and reduces damage from non-spell attacks. Can be stacked, like player stun. Monster has a saving throw. Select monsters can cure out of it. This is meant as a debuff used for tactical battles. Should also be possible to produce from specific (blunt) weapons, wielded with a sufficiently high strength.

Weaken monster: Decrease monster melee damage output by a flat amount (say 3-5 points of damage per hit). Cannot be stacked. Cannot be cured out of. This should be an early level spell for priests. Can also be produced with sharp weapons wielded with a sufficiently high dexterity. (can make a stronger version for later game as well)

Confuse monster: Monster cannot cast spells, and fumbles attacks some percentage of the time. A fumbled attack has a chance to land on a random target. Most effective against dumb monsters (i.e. this is a crowd control option against animal packs.) Select monsters can cure out of it. Brainless monsters are immune. Staff of confuse monsters is eliminated (too powerful). This is a powerful debuff and should be more of a later game item/spell.

Slow monster: Monster speed reduced by some amount (less than one full turn). Cannot be stacked. Select monsters can cure it by spending a turn. No saving throw allowed, but some monsters (those with free action) are immune. This is meant as a weaker alternative to a potion of speed. Most powerful between levels 10-30.

[Pete Mack]

I don't care as much about hold, though it's too much like a directed phase door to be allowed. The others are just crazy strong the way they are implemented now.
1. Power is unaffected by monster power. So a level 5 wand of stunning knocks every monster--even Morgoth--helpless for 5 normal speed turns, corresponding to 15(!) turns at speed 30. Power^2/mlev is a better choice; then make wand of stunning native to level ~15. Morgoth would be held for 1.5 normal turns, or 4.5 effective turns, with one turn lost to zapping the wand.
2. Melee is actually reduced by more than Derakon's estimate, because it's not hit probability that's reduced by 25%; it's the 'hit chance', which is used against the player's AC. For high AC, this is significantly more than a 25% reduction. It makes even dlev 70 monsters helpless in melee. (Level 50 monsters barely leave a scratch.)
3. There needs to be more uncertainty than just device fail rate. Something that uses the ratio of (player + device level)/mlev would be appropriate.

[Derakon]

How about making stunning just reduce the monster's accuracy and maybe spell failure chance? Accuracy and damage means their overall damage output is reduced by ~44% (1 - .75*.75), which is clearly excessive. And my intuition is that a disoriented combatant can still swing plenty hard, they just can't necessarily see what to swing at.