RNG is teasing a returning player

[Philip]

Yeah, I agree with Antoine. I may disagree with some of your changes, but you are the one who asks us about what we think about the development direction.

[Magnate]

Yeah, I agree with Antoine. I may disagree with some of your changes, but you are the one who asks us about what we think about the development direction.

Thank you both for the kind words. Takk has suggested that I take a break, so I will. I had a bunch of things I wanted to do for 3.3, but hopefully the important ones will get done better by someone else.

To Eddie: I'm sorry to have annoyed you so much. If I had known about your impending fork/variant I would have reacted positively.

And I never did catch Nick. Ho hum.

[TJS]

I really like the changes that have happened in the past year or so. It is making the game more interesting and I think Magnate has put in loads of effort to make that happen. New randarts are great for example. There seems to be a lot more new players around now then there were a couple of years ago so changes must be doing something right.

I think some people treat the game as a kind of spreadsheet, they crunch the numbers and get their wins and anything that upsets their incredibly optimised play style they will complain about. If you need to watch how many turns you take to make the game interesting then the current game really isn't very balanced and needs to be changed.
I also think there's a lot of rose-tinted nostalgia going on. The game is far more interesting to play now than it used to be.

Constantly detecting every couple of steps is boring, there's no doubt about it. It is fun for there to still be some mystery to what monsters/items are around a corner. Once you've got ESP and reliable detection the game gets much easier than at the start which really ain't right.

However I do think that if detection is nerfed then the player needs some clues as to nearby monsters to prevent too many unavoidable instadeaths. Some messages such as:

"You hear several footsteps to the East" for a pack of novice warriors.
"The ground shakes as if something large is approaching from the North" for an Ancient dragon
"You feel shiver down your spine" a ghost is nearby.

Perhaps perception could help with getting more accurate messages. This will help break up the monotony of the detection minigame and bring in some much needed flavour.

Flasks of oil are now less damaging, brands are removed from elemental rings, and percentile healing has been removed in favour of healing x HP per potion. Also in the past few days: true mimics, destruction removes artifacts, TO is now a bolt. These changes will be in the nightlies sometime in the week.

When I get a chance, I'll look at reducing off-weapon damage too. Enjoy!

All good changes, except I do like elemental rings as they are. They'll be pretty useless otherwise, unless they drop earlier and can be used as a form of double-resist. Perhaps combining fire/cold into one ring and elec/acid into the other would make them really useful again.

In addition to this I'd make banishment only work in line-of-sight. It's still too easy to remove everything from vaults and then walk in a grab the loot unchallenged.

Was there a consensus that monster melee against the player was seriously unbalanced? That's the only reason for the ridiculous AC boost. Don't tell me there is plan to boost monster melee in the future to balance it.

I'm sure in the past you used to argue that AC was completely useless. Therefore how can boosting it be unbalancing? I could have remembered wrong though.

[fizzix]

"You hear several footsteps to the East" for a pack of novice warriors.
"The ground shakes as if something large is approaching from the North" for an Ancient dragon
"You feel shiver down your spine" a ghost is nearby.

I have a lot of ideas for this that extend from having terrain change based on nearby monsters, and allowing for players to search for clues for nearby monsters.

There's some major refactors going on with terrain handling and that's postponed the terrain changes. The second one could probably be implemented, although I think the reason I've held off on it is because I haven't yet figured it out is because I got stuck on the pathfinding problem.

I don't think detection will be nerfed until these, or equivalent, changes come in, with the possible exception of granular ESP.

In addition to this I'd make banishment only work in line-of-sight. It's still too easy to remove everything from vaults and then walk in a grab the loot unchallenged.

Unless you're playing mages you have to use a rare and valuable mass_banishment scroll to clear out a vault. If you are playing a mage, I don't see the problem for them being very powerful in this regard since they get shafted in so many other areas.

[zaimoni]

However I do think that if detection is nerfed then the player needs some clues as to nearby monsters to prevent too many unavoidable instadeaths. Some messages such as:

"You hear several footsteps to the East" for a pack of novice warriors.
"The ground shakes as if something large is approaching from the North" for an Ancient dragon
"You feel shiver down your spine" a ghost is nearby.

I've looked at this for Zaiband. It's very high-effort/low-return, unfortunately, although it would allow eliminate all monsters being telepathic.

Even the simplest sound propagation is painful to implement even with efficient flow calculation (basically instead of just calculuating for the player, every monster that moves needs their own flow calculated, etc.).

[fizzix]

I've looked at this for Zaiband. It's very high-effort/low-return, unfortunately, although it would allow eliminate all monsters being telepathic.

Even the simplest sound propagation is painful to implement even with efficient flow calculation (basically instead of just calculuating for the player, every monster that moves needs their own flow calculated, etc.).

Why do you need one for each monster? You want to know how likely it is for a player to "hear" a monster and how likely it is for a monster to "hear" a player. For that you need to know the "distance" between the monster and the player and the amount of noise/hearing each makes. This distance is symmetrical. If the monster is 10 squares from the player then the player is 10 squares from the monster. You calculate the flow from the player, and then you have the distances of all monsters relative to the player.

[zaimoni]

Why do you need one for each monster? You want to know how likely it is for a player to "hear" a monster and how likely it is for a monster to "hear" a player.

You also need the incoming direction for the monster.

Recall that the flow-by-sound AI option reads the direction to the player from the player-centered flow. It's been around since V2.7.x (pretty sure about V2.7.9, not so sure about earlier versions). Assuming a flow-based implementation, this has to be reversed to get useful direction information for reporting sound to the player.

Note that calculating flows is very cheap in Zaiband (flow.c). Flowing works great for both hiding and fleeing monsters, and completely solves the problem of monsters that cannot open or bash doors trying to use paths going through doors.

[fizzix]

You also need the incoming direction for the monster.

Recall that the flow-by-sound AI option reads the direction to the player from the player-centered flow. It's been around since V2.7.x (pretty sure about V2.7.9, not so sure about earlier versions). Assuming a flow-based implementation, this has to be reversed to get useful direction information for reporting sound to the player.

Note that calculating flows is very cheap in Zaiband (flow.c). Flowing works great for both hiding and fleeing monsters, and completely solves the problem of monsters that cannot open or bash doors trying to use paths going through doors.

Then you only need to calculate it when the player notices something (and is not in a room) which shouldn't be too burdensome. If you're in a room, well the sound is echoing everywhere so you probably shouldn't be able to tell anyway. Then you only need to call the calculation at specific times.

I guess my approach was not to even worry about directional sound. Your clues for monsters go from:

1) monster race is nearby (optional w/ distance)
2a) monster type is nearby (optional w/ distance)
2b) monster race is located here
3) monster type is located here.

Right now detection gives 3 and 1, 2 and 2a don't exist. 2b has been suggested for nerfed detection. Certainly directional sound is better, but at the very least we should look to add functionality for 1 and 2a.

The biggest problem IMO has nothing to do with actual implementation but has everything to do with allowing the player to ignore messages that are no longer important.

[Tiburon Silverflame]

Yah, one thing that would worry me with this, is the message flood that would seem likely to happen, especially if you're within range of a pit or vault. This isn't a problem with sound-based detection; it's a problem with using message outputs.

And: if the monster's asleep, is there any chance to detect it via sound? What if the monster's aware, but not moving? Say you have several water hounds in the same room with you...but coming up the passage to the side is a drolem. Do you hear the drolem? Say you're in the south part of a moat; the room is a dragon pit, and its door is on the north side. Dragons burst the door open. From what direction do you hear the sound?

Now, some of these issues arise in large measure, because we're calling it 'sound.' Sound has real-world properties, and when we call it sound, we create an expectation that it will behave in a reasonably real-world manner. THAT may be a bad idea.

[PowerDiver]

I'm sure in the past you used to argue that AC was completely useless. Therefore how can boosting it be unbalancing? I could have remembered wrong though.

You remember correctly, but you are missing the forest for the trees [IMO, of course].

If you take something mostly useless, and make it rather valuable, that is ipso facto unbalancing [edit: unbalancing is not quite right, more correctly a change in balance in the direction of easier] unless the previous situation was unbalanced toward being weak. V has turned into easy_band because so many junk items were upgraded.

Balance is about getting hundreds of small details right in how they interact. The change to boost AC makes the player stronger with respect to something he need not be stronger. It wouldn't ruin the game all by itself, but it is 1 of dozens of steps in the wrong direction, and they all add up. You have to be vigilant about every single step if you wish to get balance right.

If you want to make AC interesting from a tradeoff standpoint, the first step is a redesign so that melee is at least a threat comparable to breath weapons. When rFire is optional, it is possible to think about how to balance it against AC. When rFire is required, there can be no question of how much AC you would trade for it. With no possibility of tradeoffs, a boost in AC accomplishes only 2 things. It makes the player stronger, and further diminishes the importance of monster melee in comparison to monster breaths, which IMO is the wrong direction. IMO monster melee is currently too weak compared to breaths, but that has been a base design principle from the beginning, and changing this would constitute variant status.

[Derakon]

You clearly missed the part where Magnate rebalanced the chance-to-hit and AC damage reduction algorithms in favor of requiring more armor. The difference between prancing around the dungeon in a robe vs. full plate is now much more pronounced than it used to be. In short, you should be at roughly the same place as you used to be if you're wearing heavy armor, but with light armor you're now worse off.

This doesn't change the fact that breath weapons are miles more dangerous than melee attacks, of course.

[PowerDiver]

You clearly missed the part where Magnate rebalanced the chance-to-hit and AC damage reduction algorithms in favor of requiring more armor. The difference between prancing around the dungeon in a robe vs. full plate is now much more pronounced than it used to be. In short, you should be at roughly the same place as you used to be if you're wearing heavy armor, but with light armor you're now worse off.

I did miss that. I am surprised. [I thought] I noticed that monsters were not hitting me as much, but what you say means that I am mistaken.

[edit] How is this possible? If everything doubles, then proportionally there is no difference before vs after. If things actually change in ways that matter, doesn't there have to be a difference in what monsters hit you how often?

Another problem is selling prices. The primary need of an early warrior is a stack of staves of teleportation. If an avg heavy armor sells for 50, I'll keep using it until I find an armor with a resist or I get slowed. If it sells for 1500, it should be sold as that is a significant step toward buying a staff of teleportation. I usually avoid this issue by playing with no_selling, but I was reminded of it in a recent comp.

The recent change to make rFire lessen the chance of things burning in inventory also makes a robe of resist fire preferable to good plate mail for a warrior even if you are at a point where the hp aspect of fire breaths does not seriously concern you.

[TJS]

You remember correctly, but you are missing the forest for the trees [IMO, of course].

If you take something mostly useless, and make it rather valuable, that is ipso facto unbalancing [edit: unbalancing is not quite right, more correctly a change in balance in the direction of easier] unless the previous situation was unbalanced toward being weak. V has turned into easy_band because so many junk items were upgraded.

That is true if the game is already perfectly balanced. I would argue that having something like AC as relatively unimportant compared to other abilities is unbalanced. There are a lot of other changes that are making the game harder to compensate any change in AC which I think should be made more useful.

If you want to make AC interesting from a tradeoff standpoint, the first step is a redesign so that melee is at least a threat comparable to breath weapons. When rFire is optional, it is possible to think about how to balance it against AC. When rFire is required, there can be no question of how much AC you would trade for it. With no possibility of tradeoffs, a boost in AC accomplishes only 2 things. It makes the player stronger, and further diminishes the importance of monster melee in comparison to monster breaths, which IMO is the wrong direction. IMO monster melee is currently too weak compared to breaths, but that has been a base design principle from the beginning, and changing this would constitute variant status.

Being at melee distance is a threat in comparison to dealing with monsters at a distance. Think about how many monsters you'd much rather take on with ranged weapons/magic from a distance rather than melee because:

a) They have no long range attacks or if they have they don't do as much damage as their melee attacks
b) Their melee has some side effects you'd like to avoid such as paralysis, drain exp or stats, disenchantment, stunning, inventory damage, confusion etc.

I'd like to see AC acting as a saving throw for some of these side effects.

I find that you spend most of the game avoiding coming into melee distance of monsters, especially for mage and priest classes.

Also remember monsters can breath even when at melee distance so the potential danger is only increased when you get close by monsters extra choice of attack (and they also don't waste moves walking towards you).

[Derakon]

Given that monsters use the same AI when in melee as when at a distance, with the only change being that "move closer to player" is replaced by "attack player", it would seem to me that, ignoring terrain, melee is strictly more dangerous than ranged. At range, after all, monsters can waste turns by moving closer to the player; there's no way to take damage from that.

The main difficulty with ranged combat is that summoners can render themselves impervious to non-bolt/beam attacks by summoning; in melee range this isn't possible.

[TJS]

Unless you're playing mages you have to use a rare and valuable mass_banishment scroll to clear out a vault. If you are playing a mage, I don't see the problem for them being very powerful in this regard since they get shafted in so many other areas.

I just don't like the idea that you can sidestep 90% of the danger of a vault using one spell or scroll. I try not to do it since it feels like cheating (although I do get tempted sometimes).

By the time mages get the mass banishment spell they're one of easiest classes with the best escapes already. Priests get Banish Evil which is considered a very powerful spell and that only works in line of sight.