What Eddie Plays

[TJS]

Putting quotes around half the words in your post serves more to disparage your point than your opponent's.



Infinite non-persistent levels have the advantage of being infinite and non-persistent. What do I mean by this blatantly obvious statement?

You get to choose the levels you play, Angband gives you that power, and that is really cool in its own way. Angband primarily fails here because sometimes the level you want to play is 10 levels deeper than the level you are on, and unless you are stashing ?Deep Descent or can cast Stair Creation, getting down there is a pain.

If you want to play a roguelike game where more levels are interesting, Brogue and Infra Arcana do this excellently. Limiting Angband to 127 persistent floors per character is making it into pseudo-IronmanBand and doesn't solve any of the issues you have with it. You can still get a bunch of uninteresting floors, you're just more likely to have to play them.
Making Angband a 50 non-persistent floor game is a much more approachable topic though. I wouldn't be adverse to trying that out. But I could also see how it would heavily disrupt the balance of force_descent (by which I mean probably destroy it).

Is there not a contradiction between wanting to teach the player not to hang around at the same depth too long and also having an infinite number of levels of that depth that you can play?

Force descent is supposed to be very difficult and I'm sure that some people could beat it even with just 50 levels. The game is too easy at the moment.

[Derakon]

Is there not a contradiction between wanting to teach the player not to hang around at the same depth too long and also having an infinite number of levels of that depth that you can play?

Force descent is supposed to be very difficult and I'm sure that some people could beat it even with just 50 levels. The game is too easy at the moment.

Forced descent is and always has been the weaker of the two components of ironman; being cut off from town (and particularly from your home) is the stronger component. I'm not saying that forced descent is meant to be easy necessarily, but I'm sure that when people were winning ironman games back in the 2.9x days they felt the loss of town more acutely than the shortage of dungeon levels.

Angband has always had an issue of being too long: the original devs decided "let's just double the number of levels there are in Moria!" but they failed to come up with enough meaningful content to fill those levels.

[Estie]

Angband has always had an issue of being too long: the original devs decided "let's just double the number of levels there are in Moria!" but they failed to come up with enough meaningful content to fill those levels.

They probably thought 50 levels was too easy.

[Pete Mack]

@Estie--
You need enough levels that 1% risks are enough to kill you in a game. Forced decent and reduced level count between 40 and 75 (speed rings are native) makes the game easier. Of course, since speed boots are so common, this is much easier already than it was in 3.0. That stretch, with bad speed and poor stats, is my favorite part of the game.

[TJS]

@Estie--
You need enough levels that 1% risks are enough to kill you in a game. Forced decent and reduced level count between 40 and 75 (speed rings are native) makes the game easier. Of course, since speed boots are so common, this is much easier already than it was in 3.0. That stretch, with bad speed and poor stats, is my favorite part of the game.

I'm getting a bit lost here. We need hounds to make players play the levels less, but we need more levels so they have enough risk to possibly kill the player.

Why not introduce more risks but also reduce the number of levels?

I think Angband had got itself in a state where although there were a lot of unfun and illogical design decisions there was a sort of game balance that people appreciated and found challenging. Removing the unfun stuff has knocked the game balance out of whack making it too easy, but maybe we could make the game harder in other ways rather than bring back some of that boring stuff.

[Pete Mack]

Of course getting the balance right with fewer levels would be ideal. But it won't be fixed by messing with the margins. Making the levels smaller in the style of Sil is one possibility. Or making all levels cavern/labyrinth.

[Nick]

I think Angband had got itself in a state where although there were a lot of unfun and illogical design decisions there was a sort of game balance that people appreciated and found challenging. Removing the unfun stuff has knocked the game balance out of whack making it too easy, but maybe we could make the game harder in other ways rather than bring back some of that boring stuff.

That's a pretty nice summary, and broadly speaking my plan as maintainer (beyond attempted code improvement) is

1. Re-examine design decisions that were a result of restrictions on computer power and
2. Re-balance the game after changes from 1.

The question of making the game harder is a subtle one. IMHO the changes that have made the game easier were intended to make the game more fun and interesting, and easier was a side-effect. The tricky part here is that everyone's definition of fun and interesting is different, whereas mostly we can agree on hard/easy.

Traditionally it was really hard to get any win in Angband, and it has certainly become easier. Is that a bad thing?

[Estie]

What I would like to play is something like this:

Divide the character hit points by 2 (or 3). Divide the damage of Sauron and Morgoth by 2 (or 3). Leave everything else unchanged.

[Nick]

What I would like to play is something like this:

Divide the character hit points by 2 (or 3). Divide the damage of Sauron and Morgoth by 2 (or 3). Leave everything else unchanged.

OK, my first reaction to that is it's a stroke of genius, but maybe I should be a bit more measured.

I had never thought of making the final battles shorter, but now you say it it seems like an obvious possibility. So I would maybe suggest the following modification of your idea:

1. General cutdown in player hitpoints, but also a wider range of character possibilities
2. Significantly reduce S/M hitpoints, but make them more dangerous in other ways, so fighting them is less about endurance.

[Pete Mack]

That makes essentially all higher breaths instakills for all characters. Instead, play RocketBand

[Nick]

That makes essentially all higher breaths instakills for all characters. Instead, play RocketBand

Hmm, yes. If only we could introduce some measure to make breaths less dangerous...

[TJS]

My idea was to get rid of stat gain completely. Game is on easy mode once you're maxed out with reliable detection.

You should have to scrap for every point of each stat.

[fizzix]

My idea was to get rid of stat gain completely. Game is on easy mode once you're maxed out with reliable detection.

You should have to scrap for every point of each stat.

Stat gain is sort of a weird Angband thing. Usually in roguelikes players progress along two (somewhat independent) axes. The first is xp, which provides a relatively constant rate of improvement, the second is through equipment, which has a much more jagged curve. Stat gain essentially makes what we think of as generic domain of player growth, something that usually is flat and tightly controlled, and throws it into the equipment jaggedness. The end result is that the constant climb due to xp is of little importance past level 25 or so (modulo the ability to cast spells) and the more important part is what you find in the dungeon.

In a way this is something of a necessity. If you got everything you needed from xp, then scumming would be even more powerful. The fact that you need to explore, means you will want to get to more dangerous areas. Or at least, that's how I see the intent.

Removing stat gain potions would require some significant rebalance, but it also removes something that really distinguishes angband from other roguelikes. Namely an entirely gear dependent second half the game.

[Derakon]

That makes essentially all higher breaths instakills for all characters. Instead, play RocketBand

It continues to amaze me how useful Rocketband is for didactic purposes.

Angband in general has a lot of weird nonlinearities in it. The fact that going from, say, CON 18/150 to 18/200 is vastly better than going from CON 13 to CON 18 is really weird! The incredible importance of multiple blows for successful melee is also super weird; a character with 2 blows/round is much better than twice as good as a character with 1 blow/round (due to more consistent damage and the fact that you get energy refunds for blows you don't use). Speed has diminishing returns...which is needed because speed is so ridiculously overpowered.

Together, all these nonlinearities and outlandishly powerful abilities combine with outlandishly powerful enemies to make a system that's mostly balanced, but is completely nonintuitive. My inclination is therefore to say "let's replace every table in the game with a formula that's as simple as possible", and then tweak the formulae until the game is balanced. Then we can publish the formulae and the game will be much more transparent and accessible without necessarily being easy.

Speaking of which, Angband being easier is a problem. At least, Angband being too much easier is a problem. Replayability is maximized when the game is hard, but not impossibly hard. Of course ideally it has a learning mode that is less punishing, without teaching players bad habits that prevent them from "graduating" to harder difficulties.

Regarding removal of stat gain, I think fizzix's analysis is good. I like that characters are able to get permanently stronger in a way that isn't wholly reliant on experience and equipment. However, I also agree that characters tend to have broadly similar endgame stats, which isn't very interesting. We could broaden the effects of race and class on max stats, so that e.g. a half-troll warrior's internal INT cap could be 15 instead of 18/30 or whatever it currently is. However, that starts to tie into the nonlinear returns on stat points I mentioned earlier. If, for example, the CON cap of a gnome mage were decreased from 18/50 to 15, then they won't be able to reach "survivable" HP levels for the endgame.

So, nerf the endgame damage values. Probably should be done anyway. But now the game's too easy because the player always has at least one turn before they can die, and a myriad number of ways to avoid being put in serious jeopardy.

So, remove the player's toolkit? Now how do they interact meaningfully with the dungeon?

Everything ties together. If we're really going to fix this, Angband is going to be a lot more broken, balance-wise, before it gets better.

[Estie]

No I wont play Rocketband.

I think my version would be perfectly winnable; but maybe add 1 more change: since most of the uniques cant be killed, reduce their frequency to give the player more breathing room.