Even more thoughts on v4

[Magnate]

Perhaps they should be in chests (or terrain that can contain items securely...it is a vault after all).

I think the idea I like best is that a monster's chance of being able to pick up an artifact is determined by its level, with uniques having an elevated level for this purpose. This should take care of fizzix's fears of abuse, since weak monsters would be unlikely to pick up (m)any artifacts. Ideally I'd like to make it also related to the artifact's power rating, which would make this even more certain, but there might be objections to that that I haven't thought about.

@kaypy: I think you misunderstand the distinction. With EASY_KNOW, each artifact spellbook is visible from afar for exactly what it is. So you don't need to go and inspect it. Without, they're just "Tome of High Magick" or "Book of High Prayer", and you don't know which until you go and pick it up. A small distinction, but relevant to the buggy behaviour Derakon reported, which was that he'd seen the exact same artifact generated twice - this will prevent that. One could go further and make them all generic "Tome of Rare Spells", without giving away which realm they were from. I think I quite like that idea. Anyway, there is no change to the depreservation behaviour - that has only ever happened after an artifact is picked up, and continues to be so.

[fizzix]

I think the idea I like best is that a monster's chance of being able to pick up an artifact is determined by its level, with uniques having an elevated level for this purpose. This should take care of fizzix's fears of abuse, since weak monsters would be unlikely to pick up (m)any artifacts. Ideally I'd like to make it also related to the artifact's power rating, which would make this even more certain, but there might be objections to that that I haven't thought about.

That's fine. Ringil is a shallow item, but whatever, it's a special case. I guess as long as the Witch-King can drop an elven ring and a dread can't pick it up, I'll be (mostly) satisfied.

[Magnate]

That's fine. Ringil is a shallow item, but whatever, it's a special case.

That's why I suggested using power instead of depth. (The "its level" in my previous post was the monster's level.)

[Nomad]

Okay, I've messed about with the affix levels in ego_item.txt and downgraded some affixes from great->good or good->average after a certain depth (mostly after dlevel 50, 70 or 75). It's hard to gauge the impact on top of Magnate's other changes, but some trawling about at those depths with cheat death on didn't seem to produce anything too ridiculously overpowered in the way of floor items or drops from non-uniques. Though unique drops and nest or vault items might well be a different story. Anyway, I've sent a pull request so people who are better late-game players than me can test it out.

[Magnate]

Okay, I've messed about with the affix levels in ego_item.txt and downgraded some affixes from great->good or good->average after a certain depth (mostly after dlevel 50, 70 or 75). It's hard to gauge the impact on top of Magnate's other changes, but some trawling about at those depths with cheat death on didn't seem to produce anything too ridiculously overpowered in the way of floor items or drops from non-uniques. Though unique drops and nest or vault items might well be a different story. Anyway, I've sent a pull request so people who are better late-game players than me can test it out.

Thank you, that's excellent. Unfortunately I accidentally rebooted my machine via ssh while it was in the middle of the 50k stats run, so I've had to start again. It's now about halfway, so I'll leave analysing your changes until the next set of stats. But I'll merge them anyway, because it's undeniably an improvement.

[Derakon]

An update from my hobbit paladin, currently at clvl18, dlvl16.

* I'm using a Bone Maul of Slaying, with which I definitively get 1 blow/round, because it's my best damage source. Before that I was using a Short Sword of Piety, which I got 1.3 blows/round with. Just for all y'all that've been complaining about characters starting with ridiculous melee prowess. I think that's pretty well limited to warriors or the easy-mode races.

* I tried six times to slow a baby MHD with no success (using a wand, device skill 59). I tried twice with Orfax and it worked on the second try. The baby MHD is only 3 levels deeper than Orfax; how steeply does monster level affect chances of slowing? Later I managed to slow Ugluk after three tries. Not that that was enough to make it possible for me to kill him.

* Similarly, it took 6 attempts before Sleep Monster took, on a hill orc.

* Line lengths in comments still appear to be slightly broken -- if you write a comment up to the limit in-game (69 characters), it'll be truncated to 64 characters in the file.

* As I noted earlier, good items are much more available now, perhaps too available. In addition the maul, swortsword, and gauntlets, I also have a Blackrock Shortbow of Accuracy (+18,+8) which means that arrows are pretty much my best source of damage...and I'm a paladin. I also found an Everburning Wooden Torch +2 (i.e. light radius 3). I think at this point my warrior last game was still running around with a (+1,+1) Main Gauche and some basic, acid-damaged armor.

* Any thoughts on replacing Staves of Curing with Staves of Cure Serious Wounds? I've never using the Curing effect, since I always carry strictly-better potions around.

* At 900', 20 unsuccessful attempts to bash down a door before I gave up. clvl19 / STR 17. Are jammed doors meant to be impossible to open now? Then why do we have the bash command?

* Do monsters in labyrinth levels start out more awake or something? Entered a labyrinth, was immediately set upon by several nonliving enemies (okay, no problems so far). Blink away, have to take down some rats and lice that are already awake and breeding (less than 10 turns after entering the level). Start a fight with some forest trolls, get ambushed from behind by a cutpurse (okay, they're naturally edgy) and a baby red dragon (not so normal).

* Use an unidentified Wand of Scare Monster on a Cold Vortex. The wand is not identified by use, but I learn that the vortex cannot be frightened anyway.

* I see that chance-to-hit has been added to monster memory once you learn the monster's AC. Thing is, it takes so long to learn the AC that this is functionally never going to happen for monsters that you care about. Any thoughts on making it easier to discover monster AC? If it were a 1% chance per attempted hit, say, then after 100 swings you'd have a 64% chance of having discovered the monster's AC.

* Should the unique Uruks get arrow attacks? That's Ugluk, Lugdush, and probably Bolg and Azog. To me the most notable thing about black orcs and uruks is that they have bows, making avoiding melee with them rather unprofitable. Then again, you could argue that the uruk commanders have better things to be doing than practicing archery.

* <insert rant about useless prayers clogging up the prayer books here>

EDIT:

* Giant grey ants have +5 speed and 200 AC. Who decided that was a good idea? They're practically untouchable without some form of magical offense, which at the depths they show up at means you have to be playing a mage, priest, or (maybe) ranger, since otherwise you're stuck with Stinking Cloud wands.

[Magnate]

* I tried six times to slow a baby MHD with no success (using a wand, device skill 59). I tried twice with Orfax and it worked on the second try. The baby MHD is only 3 levels deeper than Orfax; how steeply does monster level affect chances of slowing?

It's only 1% per level. But it uses the passed-in timer value (i.e. the random number of turns the monster will be slowed if you're successful), so it's hugely random.

* Line lengths in comments still appear to be slightly broken -- if you write a comment up to the limit in-game (69 characters), it'll be truncated to 64 characters in the file.

Hmmm. Ok. The in-game limit has to go down to 64 then. It's just to do with the number of characters used up by the leading columns in the history (turn, dlev, clev).

* Any thoughts on replacing Staves of Curing with Staves of Cure Serious Wounds? I've never using the Curing effect, since I always carry strictly-better potions around.

Hmm. I rather preferred the other idea of splitting out effect cures from healing potions, thus making Curing staves quite useful ;-)

* At 900', 20 unsuccessful attempts to bash down a door before I gave up. clvl19 / STR 17. Are jammed doors meant to be impossible to open now? Then why do we have the bash command?

I don't think this is new, is it? Artes's changes in 3.3 were about spiking and locking doors, but I think bashing's been pretty rubbish for a while. I think it's somewhere down at around 200 on the to-do list - make bashing a little easier but not pointlessly so ...

* Use an unidentified Wand of Scare Monster on a Cold Vortex. The wand is not identified by use, but I learn that the vortex cannot be frightened anyway.

Dammit. I thought we'd fixed ID-by-use of devices properly. I wonder if this is a regression, or still exists in V.

* I see that chance-to-hit has been added to monster memory once you learn the monster's AC. Thing is, it takes so long to learn the AC that this is functionally never going to happen for monsters that you care about. Any thoughts on making it easier to discover monster AC? If it were a 1% chance per attempted hit, say, then after 100 swings you'd have a 64% chance of having discovered the monster's AC.

Happy to do this - the current numbers seem completely arbitrary.

* Should the unique Uruks get arrow attacks? That's Ugluk, Lugdush, and probably Bolg and Azog. To me the most notable thing about black orcs and uruks is that they have bows, making avoiding melee with them rather unprofitable. Then again, you could argue that the uruk commanders have better things to be doing than practicing archery.

Nice idea ...

[fizzix]

* Giant grey ants have +5 speed and 200 AC. Who decided that was a good idea? They're practically untouchable without some form of magical offense, which at the depths they show up at means you have to be playing a mage, priest, or (maybe) ranger, since otherwise you're stuck with Stinking Cloud wands.

I'm at fault here. This is admittedly experimental. However the effect is precisely what I was going for (but maybe it was a bad idea). Specifically, a monster that's very difficult to melee, but fairly easy to kill with spells. There are lots of monsters that mages have to escape from because of insta-kills from breaths. I figured there should be one for warriors as well.

I found that warriors, paladins and rogues had a decent chance of hitting 200 AC monsters though anyway. (perhaps with !herosim and ?bless).

[Derakon]

I rather preferred the other idea of splitting out effect cures from healing potions, thus making Curing staves quite useful ;-)

Ehh, that largely just means that everyone is forced to devote another inventory slot to Foo of Cure Ailments. With Foo being a potion for preference since your magic device skill gets utterly clobbered by being confused, thus reducing your chance of being able to use the item that would cure your confusion. That's the other problem with staves/rods of Curing -- when you need them most they're the most unreliable.

I don't think this is new, is it? Artes's changes in 3.3 were about spiking and locking doors, but I think bashing's been pretty rubbish for a while. I think it's somewhere down at around 200 on the to-do list - make bashing a little easier but not pointlessly so ...

Heh. Well, for characters that are strong but not strong enough to outright destroy doors (via tunneling), having bashing not be complete rubbish would be nice. Looks like this is the usual "stats are massively more effective at 18+" issue -- you succeed if d100 < (str modifier - (door strength * 10)), with door strength ranging AFAICT from 0 to 7, though I can't tell how it's actually set.

If we assume that all doors have a strength of at least 1, then this means that you have to get a STR of at least 18 to have even a 10% chance of breaking a door down. Against a strength-7 door you'd need a STR of 18/60 for that same 10% chance.

Similarly, your chance to avoid being paralyzed by smashing into the door is given by d100 < DEX modifier, which means that my 18/10 DEX gives a measly 10% protection against being "knocked off balance".

I note that when monsters try to bash doors down, they reduce the door power until they succeed. In contrast, the player has to bash the door down "all at once" -- the door's power is not modified by failed attempts. It seems reasonable that the player should have good odds of reducing the door's strength even on a failed bash. Here's a patch that should accomplish this:

diff --git a/src/cmd2.c b/src/cmd2.c
index 2cebdc9..e39f713 100644
--- a/src/cmd2.c
+++ b/src/cmd2.c
@@ -1469,7 +1469,7 @@ static bool do_cmd_bash_test(int y, int x)
  */
 static bool do_cmd_bash_aux(int y, int x)
 {
-	int bash, temp;
+	int bash, door_power, chance;
 
 	bool more = FALSE;
 
@@ -1486,16 +1486,16 @@ static bool do_cmd_bash_aux(int y, int x)
 	bash = adj_str_blow[p_ptr->state.stat_ind[A_STR]];
 
 	/* Extract door power */
-	temp = ((cave->feat[y][x] - FEAT_DOOR_HEAD) & 0x07);
+	door_power = ((cave->feat[y][x] - FEAT_DOOR_HEAD) & 0x07);
 
 	/* Compare bash power to door power */
-	temp = (bash - (temp * 10));
+	chance = (bash - (door_power * 10));
 
 	/* Hack -- always have a chance */
-	if (temp < 1) temp = 1;
+	if (chance < 1) chance = 1;
 
 	/* Hack -- attempt to bash down the door */
-	if (randint0(100) < temp)
+	if (randint0(100) < chance)
 	{
 		/* Break down the door */
 		if (randint0(100) < 50)
@@ -1515,23 +1515,32 @@ static bool do_cmd_bash_aux(int y, int x)
 		p_ptr->update |= (PU_UPDATE_VIEW | PU_MONSTERS);
 	}
 
-	/* Saving throw against stun */
-	else if (randint0(100) < adj_dex_safe[p_ptr->state.stat_ind[A_DEX]] +
-	         p_ptr->lev) {
-		msg("The door holds firm.");
-
-		/* Allow repeated bashing */
-		more = TRUE;
-	}
-
-	/* Low dexterity has bad consequences */
 	else {
-		msg("You are off-balance.");
-
-		/* Lose balance ala stun */
-		(void)player_inc_timed(p_ptr, TMD_STUN, 2 + randint0(2), TRUE, FALSE);
-	}
-
+        /* Hack: Have a chance of reducing the door strength to make 
+         * future attempts more likely to succeed. Only if the door is 
+         * not already at minimum strength. */
+        if (door_power > 1 && randint0(20) < bash) {
+            /* Reduce the power of the door */
+            cave_set_feat(cave, y, x, cave->feat[y][x] - 1);
+            msg("The door creaks.");
+        }
+        
+        /* Saving throw against stun */
+        if (randint0(100) < adj_dex_safe[p_ptr->state.stat_ind[A_DEX]] +
+                 p_ptr->lev) {
+            /* Allow repeated bashing */
+            more = TRUE;
+        }
+
+        /* Low dexterity has bad consequences */
+        else {
+            msg("You are off-balance.");
+
+            /* Lose balance ala stun */
+            (void)player_inc_timed(p_ptr, TMD_STUN, 2 + randint0(2), TRUE, FALSE);
+        }
+    }
+    
 	/* Result */
 	return more;
 }

I note that it took me a good 30 minutes of play to find a single jammed door after I wrote this. It went down in 5 bashes with my 18/12 STR.

EDIT: tweaked the patch very slightly -- added a message when the door weakens, and removed the message for when the bash fails but you aren't knocked off-balance. Otherwise you'd get "The door creaks. The door holds firm." which seemed a bit odd.

[fizzix]

+1 on Derakon's door bashing patch

I would like rods/staves of curing to have 100% success rates regardless of confused. Of course...this would be a special case and problematic in its own right.

[buzzkill]

I would like rods/staves of curing to have 100% success rates regardless of confused. Of course...this would be a special case and problematic in its own right.

What if they were made very low level items (DL1), but extremely rare. Wouldn't then even warriors would be able to use them with a good/great chance of success even while confused, especially by the time such things mattered. Maybe not 100%, but much better, and no special casing involved.

[fizzix]

Magnate, I think you definitely overshot on the last update where you made affixes more common.

two items that I found on dlevel 40:

golden ball and chain of slay evil: 2d4 +11, +7, +2 wisdom, tunneling, +1 light, slays evil, protection from fear. (this is artifact quality) found in a cavern

hard cap of telepathy: +1 infravision, resist blind, regeneration, telepathy. (also artifact quality) dropped by a master rogue.

edit: In the BM power dragon kite shield of protection (+9, +9) [11 +22]. unfortunately it costs 20x more money than I have.

[Zyphyr]

For some reason Rings of Escaping have shot up massively in value. Previously the sold to the magic shop for the somewhat absurd price of a bit over 4000. Now they are around 12500.

[Derakon]

That'd be because Magnate made speed much more expensive after people complained that it was too easy to buy speed items. I'm not sure what the solution is here; perhaps Escaping and Teleport items should be considered "bad" items (due to the panic / random-teleport aspects) and thus unsalable? Is that even possible?

Buzzkill: AIUI item level is now disconnected from allocation, so you could indeed make Rods of Curing a level-1 item that nonetheless isn't generated until level 20 (or whatever). However, your failure rate is currently capped between 1% and 75% (that is, your chance of success is always between 25% at worst and 99% at best). Also, your device skill is temporarily modified by heavy stun, stun, bless, heroism, berserker, fear, confusion, amnesia, poison, and hallucination; all of these have multiplicative values (being confused has the worst impact, reducing device skill to 75% of what it otherwise would be). Given that, I suspect that even if you removed the 1% minimum failure rate you'd still end up with situations in which you couldn't reliably zap a Rod of Curing / use a Staff of Curing.

[fizzix]

Buzzkill: AIUI item level is now disconnected from allocation, so you could indeed make Rods of Curing a level-1 item that nonetheless isn't generated until level 20 (or whatever). However, your failure rate is currently capped between 1% and 75% (that is, your chance of success is always between 25% at worst and 99% at best). Also, your device skill is temporarily modified by heavy stun, stun, bless, heroism, berserker, fear, confusion, amnesia, poison, and hallucination; all of these have multiplicative values (being confused has the worst impact, reducing device skill to 75% of what it otherwise would be). Given that, I suspect that even if you removed the 1% minimum failure rate you'd still end up with situations in which you couldn't reliably zap a Rod of Curing / use a Staff of Curing.

There was a suggestion a while back on the dev-channel to make potions fallible. You would have a fumble chance that would increase if afraid or confused. I didn't like this idea, and no one has implemented it yet. In this way, the rods could be made to be more reliable than a potion.

Personally, I like the idea of one-use items being failsafe. The drawback for the items is the availability.