3.a advocacy

[PowerDiver]

Are there any volunteers to create such a savefile? Otherwise we'll probably delay an EddieBand comp until such time that I can sit down in someplace civilized and do the tinkering.

I could create a savefile easily enough, but it would make more sense for someone else to do it. Monster Memory is not an issue, since everybody gets perfect MM. I would recommend killing Maggot before making the savefile, so that people have no incentive to restart repeatedly until they get a good Maggot drop.

There are lots of choices. Hard as possible or easy as possible? Easy as possible might still be harder than 4.0 for all I know. Emphasize the figure-out-restores-by-using-them mechanic or deemphasize that? Which spell realm? I separated the spell realms to remove most overlap. E.g. priestly casters do not get blink, which makes shoot-and-scoot more difficult for them.

You might want to wait until someone posts a winner and ask their opinions.

[debo]

Can we do something more melee-oriented that emphasizes whatever changes to the ID game were made? Don't worry about the maggot thing, people rarely take the competitions that seriously.

[Derakon]

I'd say aim for something reasonably "middle of the road", which probably means a human/half-elf/half-orc rogue or paladin.

[t4nk]

How about human warrior. Ironman, all other options default.

[PowerDiver]

Can we do something more melee-oriented that emphasizes whatever changes to the ID game were made? Don't worry about the maggot thing, people rarely take the competitions that seriously.

Maybe a dwarf paladin would be good with default point-based stats. Healing items don't compare to what most people are used to, so paladin would mitigate that. Classes are better differentiated if you play with birth_fewer_slots or whatever I called it. You'd need a policy on easiness options, the most obvious being don't allow any.

There are no gain-1-lose-1 potions, so if you want more blows before the end game use birth_higher_stats.

I'm utterly hardcore, typically play ironman with fewer slots, but that might be extreme even for a comp.

[edit] most importantly, do not play a kobold, there are too many cure poison items it cannot learn

[debo]

Ok, would you mind making an ironman dwarf paladin, but let's leave fewer slots off for now? (I feel like I'm ordering at a very strange sort of restaurant.)

[Pete Mack]

How about not ironman for the first attempt? Keep it *relatively* easy to ID things--it's much easier with a house.

[PowerDiver]

I generated 4 savefiles, dwarf paladin, human warrior, ironman and not, default everything. Hopefully you can get the files at

https://1drv.ms/f/s!AgmkWQjFit4MgzfTAiaFr-ntwXRJ

I'll leave you to try to convince debo which is best. Be sure to check the squelch settings when you start. 3.a.0 starts with restore potions set to squelch among other things, counting on the unsquelch useful item setting to show them when you need them.

[debo]

Thank you!!

[t4nk]

I compiled a version with a fix to targeting line (you'll know it when you see it). So if it annoys you, use this one:

https://mega.nz/#!BAImFAhT!3zsh0GvOfpFNQxFgaaUWCu6f537AkrZZLA8T7gEwlQ4

[t4nk]

There are so many great ideas, from how arrows stick to monsters to removal of most identification (unfortunately not complete removal...). But, ultimately, in my opinion, 3.a.0 doesn't give average players (like me) good things with enough frequency. I think that's the symptom of the underlying problem that is also present in 4.0.x - the game is just too long for the amount of content it has.

I (roughly) measure "frequency" in the number of keypresses that is needed to get something done. Notice how Angband tries to minimize that number - auto running in corridors, macros, autoinscriptions, attempts to simplify identification of items, war on hounds; in your version, arrows sticking to monsters, detection detecting the whole levels, books getting autoinscribed with numeric tags, etc, etc. Other games go even futher: autoexplore, autofight, fancier/easier to use macros, smaller levels... Most of the improvements in 3.a.0 that I noticed are of that nature (and there are many! monsters being preidentified is a big one). Clearly, not all keypresses are equal, but roguelikes in particular require rather careful play - at least, if the player tries to actually win.

Cleary minimizing the number of keypresses is good. Some might say "so the game which require doing nothing at all is the best?" Yes, if a game could provide entertaining gameplay while requiring no input, that would be the best. This ideal seems impossible (considering that "gameplay" implies some kind of interaction). So I view keypresses and other input (like mouseclicks - but Angband is pretty unplayable with a mouse, so that doesn't matter) as necessary evil, or, perhaps, as a limited resource. Judging by changes in 3.a.0, you're probably not completely averse to this kind of thinking?

Let's talk about identification in 3.a.0. It's a pain. Say, I want to identify a potion of restoring strength and I'm wearing ring of sustain strength. I have to: unequip the ring, eat a strength draining mushroom, play with drained strength for some unspecified period of time (supposedly taking longer to kill monsters), find an unidentified potion, quaff it - whoops, that's not a potion of restoring strength; find some other potion, quaff it and finally get str back, and 1 xp as a reward. That's a lot of player's input for rather questionable reward.

Or consider hounds. Clearly the problem with them is that they're exceptionally keypress-intensive monsters. It usually takes a lot of time to kill them. It also takes quite a while to avoid them (as in run away). And they don't drop anything. They might have some positive qualities (such as making the game more difficult), but that have to be measured with their bad qualities (press a billion of keys to get rid of them).

Or another example, now from Nick's Angband (while I'm at it...): traps. So detecting traps was "tedious" and "press A to not die" kind of thing. Now there are traps everywhere, in every corridor, and they're worse than they ever were. For example, a confusion trap: move into trap, press shift+r, press enter, move into trap, press shift+r, press enter, make a keymap, keep bumping into the trap and resting... Go somewhere else, you say? That's also pressing more keys, besides, I've already explored that "somewhere else", there is nothing interesting there (maybe some hounds spawned...) Nick, how about removing all that crap?

So what was I talking about... oh, yeah. The length of the game (3.a.0). It's long, and upgrades (equipment, stat potions) are few and far in between. I'm sure that with more skilled play I would be getting them more often, but I'm not as good as you are, so I didn't... the most fun part was looting big vaults on level 90+, but that, of course, also required a ton of input, so I didn't want to do it more than a couple of times.

Giving the player good things more often, naturally, makes the game a lot easier (like what actually happend with Angband). Maintaining some desired level of challenge (and Angband is indeed too easy and dull for me; I found 3.a.0 a big improvment in this respect) then requires buffing monsters, presumably by improving their offense (since improving their defense will lead to increase in time to kill them, and, therefore, "tedium"). However, monsters in Angband are already very strong, but detection/tele-other/banishment/run-away is stronger, and changing THAT would change the nature of the game to such extent that it shouldn't be called Angband anymore... Perhaps another option is to reduce the number of levels - e.g., back to Moria's 50? (and rename to Moria ).

Another very reasonable option is to do nothing and leave things as they are. People are, of course, different, and, while I'm sure the general principles that I outlined here apply to pretty much everyone, there are significant variations in the level of "tedium" that individuals are willing to tolerate. In simpler words: I only have 100000 keypresses in me to spend on a game, and only 1500 between "some interesting things" happening, but that other guy perhaps has 500000 and 10000, respectively. I think Angband's player are on the higher end of that spectrum, and the game is basically good enough for them.

tl;dr version: 3.a.0 needs more items, also, Angband is a grind, makes me type the equivalent of "War and Peace" while playing and I'm outta here.

[Philip]

Well, while we're doing comp-based commentary, thought I'd add my observations.

I agree with what t4nk said about ID. It's an interesting gimmick, once, setting up experiments and trying to remove variables. Where I enjoyed it the most was probably with staffs and wands, sort of with rods. With Scrolls and Potions, it would have required me to either inscribe items a whole lot, which I'm not fond of, or have a base system for the order in which to try to ID things. Either way, not super fun. I don't know if there's a solution - the purpose of the ID game is to put a layer of tedium between the player and the stuff they find, so maybe it's entirely irredeemable.

I found that the frequency of good stuff wasn't too bad, though it was interesting that I only ever seemed to get good equipment by killing uniques. In standard V and in O I used to find stuff on the ground, here it's basically only consumables.

I liked the AI changes, much less annoying, but actually probably more dangerous at the same time.

[PowerDiver]

I did specifically ask if the comp wanted use-id. I could easily have added an option to id unaware items on pickup too. I am agnostic about use-id, but I do believe it is idiotic to have a system where you use a spell to identify things or mimic that using shops. Either learn on pickup or force use-id.

Use-id when diving forces many hard choices. That's why I play it, but I understand it's not for most. I had learning code that kept track of everything you knew about consumables, but I foolishly wrote it for the 3.1 or 3.2 codebase, and I won't bother writing that again.

Various things described as changes here and in dumps, are basically just 3.0 [before the game was munchkinized]. The only distribution changes I made that come to mind were related to boots, DSM, and !remembrance. All of those were increases in generation, not decreases.

[edit] I have a vague memory that I made some changes to vault generation, maybe to make it easier to generate greater vaults, or maybe just frequencies, but that was years ago and I don't even remember if I made vaults more or less common.

[debo]

I'm sad no one commented on the experience scheme, it seems like the most compelling change to me (even after having played Sil a bunch.)

[Philip]

The experience scheme was interesting. I did find myself engaging in more interesting fights more often, but it was also pretty weird hunting down early monsters I hadn't murdered on dlvl 35. It does a nice job of ending XP bag focused play (e.g. Chaos hounds with resist), and it reinforces the already correct idea that the only good reason to kill things is because you want their stuff and can get it reasonably easily, it just adds one step before that, which is to kill it exactly once if it doesn't cost you too much. This is not much of a change - what it means is that instead of killing 0 of every monster that I hate fighting and comes after Teleport Other I would kill exactly one.

It is an interesting modification, and one I would like to see applied in main fork V. I liked it as just an idea flying around, I liked it in Sil, and I like it in V. The reason I didn't comment on it was that I didn't actually perceive the difference that much. I expect it made the game a bit faster - Dwarf Paladin normally can't take most of the downstairs they find and expect to have a clvl/dlvl ratio appropriate to stealth.

I might try an auto-id character that has better stealth and see if I can get out of my league. I have a suspicion that monster generation rules will keep me getting xp at a reasonable rate, but I'd like to test it.

Finally, as for ID, I do agree that any levels of ID outside of sustained player activity and automatic ID is useless. If you just carry around a _Perception to handle all your ID stuff, or a large stack of ?Identify, then all you're doing is eating a bit of pack space. Use-ID or no ID, and I suspect I just prefer the latter. Maybe because although PowerDiver makes hard choices when diving with use-ID, I just dive a bit slower and end up not use-IDing a whole bunch of stuff. Maybe even recall up to donate the stuff to a town.