Bugfix release for 3.2?

[Mondkalb]

There is a 3.2 Winner with artifact clothes:

http://angband.oook.cz/ladder-show.php?id=11083

One of them found on level 99.
So they do exist but they seem incredibly rare.
I have found dozens of cloaks of all kinds, normal, mage, stealth, aman, protection, everything but no artifact.

This is my winner -with no artifact daggers and no artifact cloaks at all:

http://angband.oook.cz/ladder-show.php?id=11183

[buzzkill]

IMO artifacts are too common overall. Past a certain point it seems to almost rain artifacts. In my perfect world finding a handful over the course of a game would suffice and so they would hardly ever get discarded. As a matter of fact, I'd like to see what would happen if discarding an artifact decreased the likelihood of future finds.

[sandtrap]

TMA - Great idea

IMO artifacts are too common overall. Past a certain point it seems to almost rain artifacts. In my perfect world finding a handful over the course of a game would suffice and so they would hardly ever get discarded. As a matter of fact, I'd like to see what would happen if discarding an artifact decreased the likelihood of future finds.

Wow, that makes a great deal of sense! And it sounds like it would not be hard to implement.

I have not seen this idea yet. Did I miss it in the TMJ conversation?

[Derakon]

TMJ was about reducing junk items while keeping the frequency of useful items about the same. Buzzkill is talking about making artifacts rarer and useful for the entire game (so you don't want to throw them away), which as far as I'm concerned means either completely rebalancing the entire game or not generating any artifacts until you're near the endgame.

[fizzix]

IMO artifacts are too common overall. Past a certain point it seems to almost rain artifacts. In my perfect world finding a handful over the course of a game would suffice and so they would hardly ever get discarded. As a matter of fact, I'd like to see what would happen if discarding an artifact decreased the likelihood of future finds.

I think the problem is "past a certain point it seems to rain artifacts."

I rarely find an artifact before dlevel 30. But at dlevel 99 it's hard to get to the stairs before finding one. I'd rather that artifacts at dlevel 99 be just as rare as at dlevel 30. Only more are available to choose from.

Will this encourage scumming? maybe. But, it is possible to win without artifacts, so in a sense they're unnecessary.

[Derakon]

For what it's worth, my current nightlies character has over 100k active turns (i.e. a long game, for me anyway), a decent amount of which was spent hanging around in the depths trying to scrounge up some decent ammo, and I've found 38 artifacts, which is significantly more anemic than before. Nowhere near "every artifact is a unique and precious snowflake", but certainly not raining-from-the-sky common. It was a long time before artifacts even made up 50% of my equipment.

[Taha]

My last three or four nightlies games have also found artifacts and quality egos to be far less common than they used to be. Only thing found before level 30 was the phial, once. Last game down into the 60s around 3-400k game turns with 7-8 artifacts seen total; three weapons, one launcher, one armor, one hat, and two light sources. And the artifacts weren't much better than egos in the same spot.

So this may be "fixed" already - at the very least it seems progress has been made in the direction being encouraged on this thread.
________
TiTiCa

[fizzix]

My last three or four nightlies games have also found artifacts and quality egos to be far less common than they used to be. Only thing found before level 30 was the phial, once. Last game down into the 60s around 3-400k game turns with 7-8 artifacts seen total; three weapons, one launcher, one armor, one hat, and two light sources. And the artifacts weren't much better than egos in the same spot.

So this may be "fixed" already - at the very least it seems progress has been made in the direction being encouraged on this thread.

Wait until you start killing lots of monsters wtih drop good flags. These monsters are much more likely to drop artifacts. Personally, I think artifact creation needs to be decoupled from ego creatiion.

If I'm reading the code correctly it looks like the following occurs:

1: Check to make special artifact (lights, amulets, rings)
1a if yes but artifact creation failed, set good and great flags to true
2: Pick an object type
3: apply magic to the object including trying to make an artifact.

I would propose this gets changed to.
1: Roll for *any* artifact creation
1a: caveat. Only great objects get boost to artifact generation chance. And only unique monsters can drop great objects.
2: If it fails, pick an object type.
3: remove artifact creation from apply_magic.

Then artifact creation is decoupled from egos and they can be adjusted independently.

We have a Monte Carlo code so we can adjust the artifact frequency to be whatever we want.

[Derakon]

Have you run said Monte Carlo code on the current nightlies to examine artifact frequency? I feel that there is a noticeable change between old and current drop rates.

[fizzix]

sure, let's look (You can play with it too now in nightlies with debug command "f". You may encounter some bugs, I haven't updated it in a while)

for a quick sim of standarts doing a full clearing game. (pick up all items and clear all 100 levels once) I found on average 64 with a stdev of about 6. IIRC previous identical sims had roughly 70 with a stdev of about 6. So you might make an argument that they're slightly less common now.

Randarts are harder to test because randart regen is slow (and not accurate) but in the past randarts were about 60 per game. Now I'm getting something like 70 with a stdev of around 7.

So there might have been some changes, but they're really too small for anyone to notice.

[Derakon]

Oh, neat. Okay, I tried running 50 iterations with randarts on (assuming the sim uses the same options the character you generate stats with uses). I'm not certain what the difference is between diving and clearing, but this gave me 15.16 for diving. When I try running a clearing sim I keep getting the error "find_empty <x, y> failed" (with numbers as appropriate for x and y); looks like something about the Monte Carlo simulator makes the game more crowded than generate.c can handle. I tried running the clearing sim 4 times before getting it to run without generating this error; it returned 68.04.

Of course, I don't play clearing games, and the game I'm playing right now has had far fewer artifacts than usual (38, after hanging out in the depths for quite some time), so in my opinion either my own game is a significant outlier or the simulator doesn't generate representative statistics for my playstyle.

Why didn't the diving sim ask me if I wanted to regenerate randarts?

[fizzix]

Oh, neat. Okay, I tried running 50 iterations with randarts on (assuming the sim uses the same options the character you generate stats with uses). I'm not certain what the difference is between diving and clearing, but this gave me 15.16 for diving. When I try running a clearing sim I keep getting the error "find_empty <x, y> failed" (with numbers as appropriate for x and y); looks like something about the Monte Carlo simulator makes the game more crowded than generate.c can handle. I tried running the clearing sim 4 times before getting it to run without generating this error; it returned 68.04.

Of course, I don't play clearing games, and the game I'm playing right now has had far fewer artifacts than usual (38, after hanging out in the depths for quite some time), so in my opinion either my own game is a significant outlier or the simulator doesn't generate representative statistics for my playstyle.

Why didn't the diving sim ask me if I wanted to regenerate randarts?

So the way the diving sim works is that it simulates each level assuming that you have killed no uniques and found no artifacts yet. It generates the probability of items occurring on any level assuming everything is available to be created. Generally, I would use this to see the probability of stuff occurring on any given level. It won't give you the results for a full game though.

The clearing option goes through a full iteration of the game and after each one remakes the artifacts and then regenerates the uniques and goes through again. While this is not how anyone actually plays, I think it's not too far off from reality. In fact, it's probably a scalar multiple off.

Your error is odd though. What system are you on?

As for why diving doesn't regen randarts. That's probably just a legacy situation. That and the fact that I really didn't want to regen randarts because 1) it takes too long. 2) it's partially broken.

[Derakon]

I'm on OSX, and I last updated my code Feb 13th at 10:55 PST (I'm waiting to update again until my current character's journey ends).

I guess I'd approximate my playstyle algorithmically as follows. Read "clear" as "kill everything whose levels are less than 75% of the current dungeon level".

* Clear 25% of levels 1-35
* Clear levels 35-45
* Clear 20% of levels 46-98
* for i in 1 to 30: clear level 98 (10 + i * 2)%

[fizzix]

I'm on OSX, and I last updated my code Feb 13th at 10:55 PST (I'm waiting to update again until my current character's journey ends).

I guess I'd approximate my playstyle algorithmically as follows. Read "clear" as "kill everything whose levels are less than 75% of the current dungeon level".

* Clear 25% of levels 1-35
* Clear levels 35-45
* Clear 20% of levels 46-98
* for i in 1 to 30: clear level 98 (10 + i * 2)%

That's similar to me. However in the course of your game you'll probably kill the same amount of uniques as the sim player and clear more vaults due to scumming 98-99. You'll also get some summons that the sim won't pick up. I've spent a lot of time thinking about how to make it more approach the "standard" playstyle (read: mine), but never came up with anything good.

I'll have to see if other OSX players get that crash.

[Max Stats]

I want to be able to say that no one who has given no_selling a fair shot has ever said it was a bad idea.

I am playing my very first no selling game, and already I like it a lot. I was always a compulsive seller, even after I had enough gold to never need to worry about it again. I am surprised by how much I like it. The only minor negative I see is that it makes the lesser artifacts completely useless unless they are found early. They used to be worth a good sale, but now there is nothing to do but unceremoniously leave them behind in the dungeon. Maybe we need the artifact museum that occasionally gets discussed?