Autopickup/always_pickup issues

[BlueFish]

I hadn't been aware of MarvinPA's "variant". (Which is really just a collection of personal tweaks.)

Add a prompt for stepping next to molds

This is clearly good. Presumably it also prompts for melee attacks against molds. As it stands, you're reasonably likely not to have found a two-radius light by the time purple molds come into depth. I guess this is intentional, which makes it intentional to be unable to reliably avoid them in all cases. But at least you should be able to reliably avoid moving into them after stepping next to them.

Round weapon and armour weights to a multiple of 0.5

Clearly good.

Don't increase hunger rate with regen while at full HP

Debatable. Clearly makes Regen substantially more powerful. I guess the logic is that careful and boring play can simulate this, which makes optimal play diverge from fun play. But there are tactical considerations and compromises having to do with inventory space and turns taken swapping it on and off.

Recolour broken doors to darkgrey

I "fixed" this in my own version by changing the ascii character for broken doors to a period, just like blank floor, which they basically are, other than not being able to hold an item.

Map the gates and throne room upon entry

Very strange service to expert players, at the expense of a sense of discovery for those all-important first wins.

Don't hide traps or webs placed by vaults

Very strange service to expert players, or maybe just edit-file diving players, at the expense of a sense of discovery/danger for everybody else.

Identify equipped weapons and bows when using self-knowledge

I must admit, I'm all for making the ID game easier. As it stands, I won't play a build that doesn't include loremaster.

Make Lore-Keeper identify herbs, potions, staves and trumpets

Interesting. What's left, amulets, rings, weapons, armor. I suspect this would diversify builds people might consider. But I'm not in love with it. (Presumably lore-master still also identifies those things - if not, then I'm really not in love with it.)

[debo]

.5 weights are not clearly good. Consider a 0 str assassin who wants to forge shortswords, for example. Alternately, consider any character who wants to use rapid attack with less than 4 str.

I personally like the range of weapon weights. Mpa-sil's fixed weights are fun for a different reason, I don't mind it, but I like the vanilla sil way better.

[debo]

Also, loremaster still ids everything if e.g. you forge it on a helm or something.

[BlueFish]

.5 weights are not clearly good. Consider a 0 str assassin who wants to forge shortswords, for example. Alternately, consider any character who wants to use rapid attack with less than 4 str.

I'm not sure what those ramifications are, but if there are gameplay effects of weights beyond the .5 breaks, then I agree it's not clearly better to round like this variant does.

[taptap]

I'm not sure what those ramifications are, but if there are gameplay effects of weights beyond the .5 breaks, then I agree it's not clearly better to round like this variant does.

See http://angband.oook.cz/ladder-show.php?id=14927 for an example, basically you need less than 1 lb weight to avoid strength penalties with negative strength (i.e. with rapid attack for every char under strength 3), in MarvinPA-sil this is impossible for shortswords. There are also some more subtle effects on smithing difficulties.

[half]

I don't quite see how this differs substantially from how it currently works, as long as "pick up by default" is off. (I can't imagine playing with it on!)

I think it'd be fine to remove "pick up by default" as an option entirely and keep the ammo auto-pickup on. Then we wouldn't be faced with the decision whether to feel like we're "cheating" by grabbing things from dragon hoards using auto-pickup.

The difference from now is that the autopickup would be costing a turn. I also can't really imagine playing with autopickup and no prompt, but I wanted to put it out there to see how many people couldn't stand anything but that.

[half]

I hadn't been aware of MarvinPA's "variant". (Which is really just a collection of personal tweaks.)

It is an interesting collection of tweaks though! I think about half of them are good for the game, and have been porting those over into the main branch.

A few quick observations. These are biased towards the changes I don't like as there is more to say on those. Don't take this to mean I'm generally critical of Marvin's work though!

I understand the idea of the prompt for molds, but I don't actually think it is good for the game. The explanation for this is quite detailed, but the gist of it is:

(1) I don't think it is good for new players.
(2) I think it is babying the player (a rather complex thing that I think is bad overall)
(3) If you want to baby the player, then this is an inconsistent way to do it. e.g. it would be much more useful to ask the player when she is adjacent to a Giant and takes another adjacent step to it whether that was a mistake as this is (a) more likely to be a mistake, and (b) more important. Should I add that? Where does this end?

The rounding of weights to multiples of 0.5 has almost no positive effects and has some negative effects. It is also a bit odd given that the internal game unit is clearly 0.1 lb.

The regen one might be good. I'd need to think on it.

I also find the colour of broken doors too close to open ones. However there are some potential plans for iron doors or iron gates, so I wouldn't choose dark grey. Note to taptap, there are further differences from broken doors and floors, most notable is that staves of imprisonment can fix them. There are plans for this to become more relevant (i.e. a Morgoth song effect like the imprisonment effect).

I definitely don't like the automapping effects. I think they are bad for new players, and for most experienced players too. Also note that traps in vaults aren't always in the same places.

Weapons you could easily pick from self-knowledge will ID in the new version (I've overhauled self-knowledge).

We do plan to improve Lore-keeper in a new version, though we had a different effect in mind.

What I really like about Marvin's version is that he's thought about a lot of these changes, has come up with some that I think are clearly good, and I can include those ones.

[MarvinPA]

Essentially my approach for weight rounding is just to not care about the few edge cases, the interface improvement of not having to sort through a ton of items that are just a tiny bit different is easily worth it in my experience so far. Certainly those downsides do exist, but I'm not convinced that they're a big factor.

The Lore-Keeper change is just an experimental thing for the moment - on the one hand it might mean some builds end up getting Lore-Keeper and not Lore-Master, which would be nice, but I think it probably ends up being too cheap for its effect. In general I like the idea of splitting identification across the two abilities so that Lore-Keeper actually does something on its own other than curse identification, but I don't know if this is the best way of doing it.

I'd prefer to have heavy randomisation of such vaults than to pre-reveal fixed vaults and traps but unfortunately Angband's vault definition system makes that difficult. Currently the only trap randomisation (if I've understood correctly) is that the "place trap" glyph just has a 50% chance of doing so, which doesn't seem anywhere near sufficient to me.

Incidentally another difference between broken doors and floor is that _Imprisonment can repair and lock broken doors with a good enough will roll. (edit: half beat me to it!).

I'd be interested to know more about the new Self-Knowledge effect, since I've been trying to come up with a reworked effect myself but haven't really been happy with any of my ideas. The simplest was to have it just fully identify one (chosen) worn item, or some small random selection of worn items (but that has the problem of allowing you to just equip the specific items you want identified, ignoring curses which admittedly can be reasonably important factor in Self-Knowledge usage).

And in a vague attempt to be on-topic: I agree that it is unfortunate that pointing out the existence of this issue lets people know about a potential method of abuse but I would have thought that makes it all the more important to be known about, so that an appropriate fix can be found - my intention in starting the thread was just to try and find the best solution to the issue, since I wasn't happy with the ones I'd come up with. Having only thrown/fired items be picked up for free sounds like a good fix. I just played a stealth character and was frequently annoyed by knowing that I could be picking up items for free, so chances are I will end up implementing some kind of temporary fix along those lines while waiting for the next release anyway.

[Scatha]

I just played a stealth character and was frequently annoyed by knowing that I could be picking up items for free, so chances are I will end up implementing some kind of temporary fix along those lines while waiting for the next release anyway.

Does it work for you to simply regard using autopickup as cheating? Then you get a fix with no coding effort. (This is what I do.)

[MarvinPA]

Does it work for you to simply regard using autopickup as cheating? Then you get a fix with no coding effort. (This is what I do.)

Fortunately here it's pretty easy to just implement a quick fix. Obviously I can get away with doing somewhat lazy fixes since mine isn't an official version, though - as half mentioned, a proper fix would also adjust the "always pickup" and "prompt before pickup" options. In general I just don't really like forcing things like this onto the player, when it's clearly the game's responsibility. Certainly it would be easier for me if I could just consider it cheating and forget about it! And realistically if the game were closed source I probably would try to do that, but I'm sure the knowledge of it would still bug me.

[taptap]

Essentially my approach for weight rounding is just to not care about the few edge cases, ...

I tended to agree with your reasoning for weight rounding and got the idea for no strength, light weapon, rapid attack builds from the thread where the rounding was proposed. As I believe one strength of Sil is the sheer variety of characters that can win, removing a cool, viable build from the game is always sad to me.

Not sure whether good or bad, but if you find perfectly weighted weapons in 50% of the cases you find a lot more good weapons during the game. 1.9 lb deathblade instead 2 lb effectively removed a damage side, 2.3 lb deathblade removes a notable amount of critical hits compared to a 2 lb. I wouldn't claim axe users gain an advantage that way, with heavy weapons it probably is just more convenient, but for light weapons builds finding the convenient balanced weapons more often can make a difference.

I wonder whether it would be possible to make modularized vaults in some cases, they are still a group of rooms etc. but the variety comes not from more whole vault variations but from vault "blocks" that combine in many more ways, maybe at some point the game could go without the rigid distinction between vault and room/corridor areas at all with simply lots of interesting areas procedurally generated.

P.S. Sorry for being completely off topic. Has been a long day. But I have been arrow collecting for some time today (in real life) arrow and throwing item collection really should remain faster than other kind of pickups. It isn't like a robe that you have to fold and put in your back pack...

[BlueFish]

(1) I don't think it is good for new players.
(2) I think it is babying the player (a rather complex thing that I think is bad overall)
(3) If you want to baby the player, then this is an inconsistent way to do it. e.g. it would be much more useful to ask the player when she is adjacent to a Giant and takes another adjacent step to it whether that was a mistake as this is (a) more likely to be a mistake, and (b) more important. Should I add that? Where does this end?

It's clearly all about purple molds, which show up at a depth where you can't reliably see them in time, and which have attacks which can easily hobble your character effectively permanently. Their attacks are completely out of whack with every other monster at that depth, which IMO is the real problem. Stat drain in this game sucks, especially at that depth, where you're likely to be several hundred feet away from you first means of restoring it.

there's an interesting thread in the V forum about making stat drain temporary. It heals over time all by itself. I think it's a good idea and I think it would work even better in Sil. The time limit in Sil makes temporary effects very important regardless. Regaining one stat point per 1000 turns in Sil would be a step forward, IMO. Purple molds just aren't fun.

[half]

I've just written a fix for the autopickup issue, but in testing it I've noticed that autopickup + the confirm pickup prompt seems to be a strictly inferior interface to non-autopickup. i.e. normally you walk on something, see a message saying what it is, and can press (g) (or control-5 or control-z...) to pick it up, or just ignore it. With autopickup and the prompt, you see a prompt saying what it is that asks a y/n question and can press (y) to pick it up, or other keys to dismiss it. This just seems worse than having no autopickup. Any objections to just ditching autopickup altogether?

[half]

Essentially my approach for weight rounding is just to not care about the few edge cases, the interface improvement of not having to sort through a ton of items that are just a tiny bit different is easily worth it in my experience so far. Certainly those downsides do exist, but I'm not convinced that they're a big factor.

It sounds like I've been too quick to dismiss this. I was thinking that the advantages were:

(1) making the smithing weight interface take fewer key presses.
(2) not having your weapon be outclassed by one that is identical except for the weight quite so often.

I see both of these as really small improvements relative to other issues with the game (e.g. making the early game more varied), and I don't like the fact that it feels a bit artificial for weights to be so discrete when it is clear the system can represent smaller weights.

However, maybe you are also thinking of:

(3) having fewer of each type of sword feels a bit simpler and more boardgamey
(4) having fewer weights makes the function from weapon weight to overall usefulness less spiky (i.e. the jumps from 2.9 lb to 3.0 lb are big quality changes over small weight increases, while 2.5 lb to 3.0 lb is a smaller quality change and over a larger weight increase).

I see (3) and particularly (4) as more important than (1) and (2) and maybe enough reason to do this. I could maybe avoid my other concerns by changing the very light items like rings to weigh either 0.0 lb or 0.5 lb. Other people's thoughts?

Note that this won't solve taptap's problem, but he could try a dagger smith instead (if they are moved to 0.5 lb). This would make daggers a bit more useful by giving them this niche. We've also considered making daggers to be like light shortswords but without the [+1]. i.e. Dagger (+0, 1d7).

taptap is also right that this changes game balance a bit, making things a bit easier, but this is less of an issue if it is in the main branch rather than a fork that only some people use.

[sim]

I understand the idea of the prompt for molds, but I don't actually think it is good for the game. The explanation for this is quite detailed, but the gist of it is:

(1) I don't think it is good for new players.
(2) I think it is babying the player (a rather complex thing that I think is bad overall)
(3) If you want to baby the player, then this is an inconsistent way to do it. e.g. it would be much more useful to ask the player when she is adjacent to a Giant and takes another adjacent step to it whether that was a mistake as this is (a) more likely to be a mistake, and (b) more important. Should I add that? Where does this end?

I think mold prompt works because molds are stationary. They're not going anywhere, unlike most every other monster which are mobile. You can't really gauge other situations in the same way, since it can involve groups of monsters, ranged attackers, etc. all in the same fight. Once you see a mold though, you know it's going to stay there.

Traps have the same behavior. It prompts you to step into traps once you know they're there. What's the difference between this and mold prompt?