Suggestion: Review vault / branch generation weights


Although the central place for design discussion is ##crawl-dev on freenode, some may find it helpful to discuss requests and suggestions here first.

User avatar

Lair Larrikin

Posts: 27

Joined: Thursday, 5th May 2011, 18:52

Post Monday, 27th January 2014, 11:37

Suggestion: Review vault / branch generation weights

First off, I must confess I'm not entirely clear on how CHANCE and WEIGHT headers in .des files work so please excuse any misunderstanding on my part.

One of the things I love about Crawl is that I never know what I'm going to get, especially stepping into branch ends, portal vaults and the like. Except in my experience, I feel like I get the same layouts a lot of the time.

Although anecdotal evidence, I'm sure many of you will agree that you see the same (usually older) portal vaults and branch ends again and again. I was just reading up on vault generation (haven't made one in a couple of years) and I was surprised to discover that, as an example, the sewer portal vault contains roughly 16 unique vaults!

https://gitorious.org/crawl/crawl/sourc ... /sewer.des

Now, in the case of sewer vaults, most of them don't have a weight which defaults them to 10, IIRC. But why does the kobold camp vault at the top have a weight of 35? As a relatively casual player this is the one I see again and again. This may indeed be a familiarity bias on my part but if the numbers are correct here, why does it have a bigger weighting (35 vs 10) than the others? It looks like a lot of newer additions forego WEIGHT headers altogether, but the older vaults have been left with specific weightings. In another example in temple.des, why has minmay_temple_forest_path_12 got a weight of 3 and circular_temple_ a weight of 40?

My main concern is that I've played hundreds of games and a large handful of the vaults in these .des files I've never seen. I would like to see more variety, and I think this can be solved by putting all these vaults on an equal chance of generation.

Some questions:

1.What are people's thoughts in tweaking the weights for these .des files?
2. Would removing all the WEIGHT headers put everything on an equal chance of generation? Or is the algorithm based on additional factors?
3. Are there reasons for particular vaults why you would want to increase or decrease the weight?

What I would like to do is see if it's possible to write a script or do something in wizmode that would allow me to generate say, the sewer portal vault again and again so I can get some accurate statistical weightings to support this intuition. Is that easy to do? Even if it's a manual task, if it's relatively quick I could generate a decent sample.

EDIT: Here's a spreadsheet with ongoing analysis:

https://docs.google.com/spreadsheet/ccc ... sp=sharing
Last edited by oiseaux on Wednesday, 29th January 2014, 19:15, edited 2 times in total.

Sar

User avatar

Ziggurat Zagger

Posts: 6418

Joined: Friday, 6th July 2012, 12:48

Post Monday, 27th January 2014, 11:50

Re: Suggestion: Review vault / branch generation weights

Hey, I like rare temples! It's cool that the game has something new to surprise you after hundreds of games, even if it's a small vault.
User avatar

Lair Larrikin

Posts: 27

Joined: Thursday, 5th May 2011, 18:52

Post Monday, 27th January 2014, 11:59

Re: Suggestion: Review vault / branch generation weights

Agreed, it's nice to get something special now and then. I guess the issue we might have at the moment is that you have vaults that are more "common" than others, as opposed to rare. To explain, say we have something like this, where V is a vault variation and their percentage chance of generation:

V1: 40%
V2: 20%
V3: 10%
V4: 10%
V5: 10%
V6: 8%
V7: 1%
V8: 1%

Here, I don't have a problem with the rare vaults so much as I have a problem with V1 and V2 being larger than the majority. Something like this, on paper, looks like a nicer result in terms of variation:

V1: 18%
V2: 18%
V3: 18%
V4: 18%
V5: 18%
V6: 8%
V7: 1%
V8: 1%

So, your rare and uncommon vaults (V6, 7 and 8) are still there, but there isn't a bias towards any one kind of "common" vault. Again, I would love for any suggestions on how to do some trials on vault generation so that I can compare the reality against my conjecture.
User avatar

Lair Larrikin

Posts: 27

Joined: Thursday, 5th May 2011, 18:52

Post Monday, 27th January 2014, 13:59

Re: Suggestion: Review vault / branch generation weights

Ran a quick bunch of tests by macroing wizard commands CTRL+R and CTRL+E for sewer. sewer_kobolds is weight of 35, everything else is 10 apart from sewer_kobold_necro_badplay which is 3. So sewer_kobolds appears 20% of the time, everything else pretty much 6% of the time. I would rather sewer_kobold's extra percentage was distributed across the others. I think something like this would be good for highlighting which vaults and branch endings are being generated throughout the game and pinpointing areas that could do with some balancing (with thousands of interations instead of 150).

  Code:
sewer_kobolds:              |||||||||||||||||||||||||||||||| 32
sewer_co_tree:              ||||||| 7
sewer_co_15_rooms:          |||||||||| 10
sewer_co_purgy_island:      ||||||||||| 11
sewer_fruit_machine:        ||||||||||| 11
sewer_big_baddie:           |||||||| 7
sewer_co_scattered_pipes:   |||||||||| 10
sewer_co_cracked_pipes:     ||||||||| 9
sewer_ribbit_badplayer:     ||||||||||| 11
sewer_frog_island_02:       ||||||||||||| 13
sewer_frog_island_01:       ||| 3
sewer_minmay_w:             ||||||| 7
sewer_the_slug:             ||||||||| 9
sewer_minmay_treatment:     ||||||| 7
sewer_kobold_necro_badplay: || 2
sewer_sea:                  || 2

TOTAL: 151

Dungeon Master

Posts: 553

Joined: Wednesday, 22nd December 2010, 10:12

Post Tuesday, 28th January 2014, 17:58

Re: Suggestion: Review vault / branch generation weights

I used to like how vaults had skewed weights, because in the past there wasn't nearly as much content. So, there was at least some justification for making certain layouts rarer, that way you could have a "standard" version of a branch end, and a "rare variant" for example. Now though, there is so much content that it leads to weird things (like sewer_kobolds showing up too often) and other maps practically not existing. So overall, changing weights around sounds like a good idea to me. I think portal vaults are the main offender here, since the numbers tend to be VERY imbalanced (certain maps are 10-20x more common than others, if I remember correctly).

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Tuesday, 28th January 2014, 18:08

Re: Suggestion: Review vault / branch generation weights

We used to have flat distributions of vaults. Until Lemuel came along and suggested skewed weights, with good reason: for the occasional player it doesn't matter but regular players get to see new content over a longer period.

What you are bringing up, in my opinion, is the fact that chances should not be fixed when vaults get added. Let's take a hypothetical portal vault which starts out with just three maps. We don't want a flat distribution of 33%, 33%, 33% but rather something like 50%, 35%, 15% or similar. However, once we are at nine maps, then we certainly don't want a single map with 50% chance; at that stage a distribution like 20%, 20%, 20%, 10%, 10%, 10%, 3%, 3%, 4% looks better. I would have thought that use of WEIGHT (as opposed to CHANCE) does the trick, but I certainly agree that 20% chance for one sewer map out of 16 is too much. If someone comes up with a system that scales well with adding new maps, we're all ears :)

To sum it up: skewed weights should be kept, but also kept in check.

For this message the author dpeg has received thanks:
oiseaux

Dungeon Master

Posts: 553

Joined: Wednesday, 22nd December 2010, 10:12

Post Tuesday, 28th January 2014, 18:09

Re: Suggestion: Review vault / branch generation weights

Something I should also add, a lot of the lower weight vaults aren't as good as the higher weight ones, or don't fit in as well. Of course, not all content can be 100% equal, but I think there's a lot of stuff that can be removed. There's been a reluctance in the past, but that was when there was less variety in the game. Something I think would be nice is for someone who knows their stuff to do a thorough overview of all the vaults in the game, and purge the worst 20-30% or so. Sounds harsh, but there's so much in the game now that you wouldn't really notice less content, but would notice increased quality. If I ever get the chance in the future to work on crawl again, I might do this myself, although I'd worry that the effort would get reverted...

For this message the author evilmike has received thanks:
duvessa

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Tuesday, 28th January 2014, 18:38

Re: Suggestion: Review vault / branch generation weights

evilmike: This is true, and actually happens sometime. There's never been a really systematic purge, more that vault makers retracted bad maps of themselves.

A problem with purging bad content is the lack of criteria what this should mean. For example, I really don't like large, early portal vaults. I am pretty sure that many people do love them. I don't feel in a position where I can go ahead and delete those. (I always wanted to make some sewers smaller but never got around to it.)

Vaults Vanquisher

Posts: 508

Joined: Tuesday, 1st November 2011, 00:36

Post Tuesday, 28th January 2014, 21:36

Re: Suggestion: Review vault / branch generation weights

Maybe there could be some kind of "spring cleaning" of vaults? Like, the weightings are flattened for a while to make sure that the obscure ones show up, and players are encouraged to report ones that are bad? I seem to remember a couple of cases of vaults being removed almost immediately just from someone pointing them out on ##crawl.

oiseaux wrote:So, your rare and uncommon vaults (V6, 7 and 8) are still there, but there isn't a bias towards any one kind of "common" vault. Again, I would love for any suggestions on how to do some trials on vault generation so that I can compare the reality against my conjecture.

I agree with this suggestion. Every vault by default could be "common" (ie, would have a standardized weight) and then vaults which are meant to be special and quirky (like, say, the Lugonu temple vault) could be given a lower one. I don't think having "ultra common" vaults is ever interesting.

Ziggurat Zagger

Posts: 8786

Joined: Sunday, 5th May 2013, 08:25

Post Tuesday, 28th January 2014, 21:46

Re: Suggestion: Review vault / branch generation weights

dpeg wrote:There's never been a really systematic purge, more that vault makers retracted bad maps of themselves.
I didn't know vault makers were allowed to do that in the first place. Where should I go to do that?

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Tuesday, 28th January 2014, 21:59

Re: Suggestion: Review vault / branch generation weights

duvessa: You could simply send me a list with vaults you think are subpar, and I look at them.

Ziggurat Zagger

Posts: 8786

Joined: Sunday, 5th May 2013, 08:25

Post Tuesday, 28th January 2014, 22:41

Re: Suggestion: Review vault / branch generation weights

  Code:
entirety of ponds.des (water monsters, water)
minmay_shoals_entry_centaurs (bad monsters in lair)
minmay_shoals_entry_sharks (water monsters)
minmay_lair_oklob_tunnel (stationary monsters)
minmay_lair_oklob_items (stationary monsters)
minmay_lair_oklob_chamber (stationary monsters)
minmay_lair_oklob_bushes (stationary monsters)
minmay_lair_end_frog_pond (water, vampire mosquitoes, giant leeches)
minmay_lair_end_enchanted_forest (spriggans)
minmay_temple_entry_oklob (stationary monsters)
minmay_snake_entry_guardpost_small (slow monsters, also bad in lair)
minmay_snake_entry_guardpost_medium
minmay_snake_entry_guardpost_large
minmay_snake_river (water)
minmay_elf_archery (training dummies)*
minmay_elf_blademastery (training dummies)*
minmay_elf_hall_golems (golems)*
minmay_sea_of_fire (too large, lava, fire bats, fire crabs)
minmay_slime_entry_jelly_boxes (jellies)
minmay_slime_entry_pillars (brown oozes)
minmay_slime_entry_oklob_open (stationary monsters)
minmay_slime_entry_oklob_enclosed (stationary monsters, breaks autoexplore)
minmay_swamp_entry_wisps (insubstantial wisps)
minmay_swamp_entry_worms (water monsters)
minmay_swamp_entry_eels (water monsters)
bazaar_minmay_generic_d (lava monsters, monsters in bazaar)* (aren't bazaars with liquids supposed to have the no_monster_gen tag?)
bazaar_minmay_challenge (item test)
bazaar_minmay_oklob (stationary monsters, monsters in bazaar)
bazaar_minmay_mutagenic_fog (harmful fog generator, spoilery, breaks autoexplore)
minmay_ossuary_two_tombs (slow monsters)
ossuary_minmay_tomb_2a (traps, slow monsters)
ossuary_minmay_four_chambers (slow monsters)
trove_dig (item test)
minmay_weird_star (lava, water)
minmay_stone_box (golems)
minmay_skeletal_surprise (slow monsters, autotravel trap)
minmay_separated_statues (stationary monsters)
minmay_statue_hallways (stationary monsters)
minmay_brainsmasher (item test)
minmay_lindwurm_lava (lava, spoilery)
minmay_serpent_palace (slow monsters, lava)
minmay_goblin_outpost (water)
minmay_mini_ossuary (slow monsters, mummies)
minmay_rotating_lava (lava)
minmay_hunger (hungry ghosts, spoilery)
minmay_nine_imps (crimson imps, white imps, shadow imps)
minmay_beelzebubs_servants (orange demons, rotting devils)
minmay_brain_eaters (spoilery, breaks autoexplore)
minmay_statue_bubble (stationary monsters)
minmay_protected_statue (stationary monsters)
minmay_statue_crystal_thing (stationary monsters)
minmay_dual_archer_statues (stationary monsters)
minmay_crystal_box (stationary monsters, golems)
minmay_more_statues (stationary monsters)
minmay_sticks_and_stones (golems)
minmay_metal_thing (lava, golems)
minmay_three_doors (spoilery, breaks autoexplore)
minmay_guarded_unrand_robustness (golems)
minmay_guarded_unrand_zonguldrok (slow monsters)
minmay_guarded_unrand_curses (slow monsters, mummies)
minmay_guarded_unrand_dyrovepreva (stationary monsters, golems)
minmay_guarded_unrand_flaming_death (lava)
minmay_guarded_unrand_mage (stationary monsters, water)
minmay_smoke_and_mirrors (rakshasas, vapours)*
minmay_elf_outpost (low level elves)
minmay_ice_pillars (water, stationary monsters)
minmay_skeleton_keep (slow monsters, curse skulls)
minmay_spriggan_fort (spriggans)
minmay_haunted_quarter (water, Ws)
minmay_crystal_thing (stationary monsters, golems)
minmay_twisted_water (stationary monsters, water)
minmay_hedge_maze (bushes)
minmay_misc_feat_secret_circle (weird no_rtele_into)
minmay_item_behind_shaft (traps, breaks autoexplore, spoilery)
minmay_statues_item (item test)
minmay_bad_closet (spoilery)
minmay_gold_behind_plants (blocking plants)
minmay_opposing_rooms (broken by monster digging)*
minmay_spiral_squares (broken by monster digging)*
minmay_five_octagons (broken by monster digging)*
minmay_imperfect_grid (broken by monster digging)*
minmay_quarter_circles (broken by monster digging)*
minmay_ornate_pinched (broken by monster digging)*
minmay_quantized_circle (broken by monster digging)*
minmay_eight_partial_circles (broken by monster digging)*
minmay_blocking_statues (broken by monster digging)*
minmay_four_inscribed_fans (broken by monster digging)*
minmay_concentric_multi (broken by monster digging)*
minmay_diamond_stairs (broken by monster digging)*
minmay_octo_star (broken by monster digging)*
minmay_central_circle (broken by monster digging)*
minmay_bad_star (broken by monster digging)*
minmay_tri_box (broken by monster digging)*
minmay_textured_paths (broken by monster digging)*
minmay_distorted_concentric_squares (broken by monster digging)*
minmay_containers (broken by monster digging)*
minmay_crescents (broken by monster digging)*
minmay_octagon_city (broken by monster digging)*
minmay_expanding_squares (broken by monster digging)*
minmay_three_patterns (broken by monster digging)*
minmay_acid_items (broken by monster digging)*
minmay_hollow_rounded_squares (broken by monster digging)*
minmay_multi_spiral (broken by monster digging)*
The items in parentheses indicate what is wrong with the vaults. An asterisk indicates that I believe the vault could be salvaged by removing the items in parentheses. (Note: making vault walls undiggable does not count as fixing monster digging: 1. players are intended to dig these vaults, 2. they still dig walls near the vault and cause similar problems.)

Dungeon Master

Posts: 634

Joined: Sunday, 22nd September 2013, 14:46

Post Tuesday, 28th January 2014, 23:08

Re: Suggestion: Review vault / branch generation weights

what about minmay_newt_pond, it has water
User avatar

Lair Larrikin

Posts: 27

Joined: Thursday, 5th May 2011, 18:52

Post Wednesday, 29th January 2014, 12:40

Re: Suggestion: Review vault / branch generation weights

Thanks all, to summarise, would most agree that:

1. A weight pattern 30, 30, 30, 10, 5, 5 is better than say, 40, 30, 20, 10, 6, 4.
2. A flat pattern like 17, 17, 17, 17, 17, 17 is bad (Not all vaults are created equal, players want surprises)

If this is the case, might I propose that we define a series of tiers, for example: common, uncommon, rare, legendary (for lack of a better term right now) that one can use when rewriting the weightings? Again correct me if I'm wrong but if common was the default (i.e. WEIGHT: 10), and each tier going towards legendary is say, half of that, we'd have:

COMMON -> 10
UNCOMMON -> 5
RARE -> 2
LEGENDARY -> 1

Vaults in a .des file can be assigned one of these WEIGHTs only, which, especially for large .des files, should, if I'm not mistaken, produce a fair distribution but also allow for the possibility of a sufficiently granular system of rarer content? Vault builders can then say, "I think this vault isn't particularly troublesome, but the layout is kinda unique, so I'll make it UNCOMMON" and be able to make a clear decision on what their WEIGHT is set to?

The only problem is if this schema was accepted, it would have to be enforced, and the WEIGHT system doesn't enforce that. But it would be a good guideline for reviewing the current weights and when reviewing vault patches in future?

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Wednesday, 29th January 2014, 12:56

Re: Suggestion: Review vault / branch generation weights

oiseaux: yes to both.
User avatar

Ziggurat Zagger

Posts: 5832

Joined: Thursday, 10th February 2011, 18:30

Post Wednesday, 29th January 2014, 13:25

Re: Suggestion: Review vault / branch generation weights

Did vault generation in Labyrinths get increased in Trunk because last lab I played it was 5 monster encounters in a row.
"Be aware that a lot of people on this forum, such as mageykun and XuaXua, have a habit of making things up." - minmay a.k.a. duvessa
Did I make a lame complaint? Check for Bingo!
Totally gracious CSDC Season 2 Division 4 Champeen!

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Wednesday, 29th January 2014, 13:38

Re: Suggestion: Review vault / branch generation weights

XuaXua: Yes.
User avatar

Lair Larrikin

Posts: 27

Joined: Thursday, 5th May 2011, 18:52

Post Wednesday, 29th January 2014, 13:45

Re: Suggestion: Review vault / branch generation weights

Would any of the following contributions be worthwhile?

1. An analysis of current vault generation percentages in the .des files (similar to my third post) to highlight any vaults that stick out too much

2. Potentially a series of patches that rewrote the weights in the .des files so that they fit into the four tiers suggested above? This would be informed by:
a. The current weight (highly weighted vaults get brought down to the common tier; low weighted vaults get placed in an appropriate rarer tier dependent on what the author gave, relative to the other weights in the file).
b. Any particular notes, comments or clear indicators that the vault should be of a different "rarity"

I would very much like to at least discover if there are any freak occurrences much like sewer_kobold but point 2 would be too much time/effort if there's no particular interest in the tier method.

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Wednesday, 29th January 2014, 14:00

Re: Suggestion: Review vault / branch generation weights

oiseaux: Of course! I'd be delighted to see any of these :)

I wonder if we should set weights in such cases (Temple maps, Sewer maps etc.) not in each map individually, but globally, just as we can do with depth. Then we would have one block of "common bazaars", followed by the block of "rare bazaars" etc. Numbers would perhaps be clearer that way.

For this message the author dpeg has received thanks:
oiseaux
User avatar

Ziggurat Zagger

Posts: 5832

Joined: Thursday, 10th February 2011, 18:30

Post Wednesday, 29th January 2014, 18:21

Re: Suggestion: Review vault / branch generation weights

dpeg wrote:XuaXua: Yes.


It needs to be scaled back a bit. 6 gargoyles plus 2-3 invisible stalkers + a hungry ghost all at once was a crock of shit.
"Be aware that a lot of people on this forum, such as mageykun and XuaXua, have a habit of making things up." - minmay a.k.a. duvessa
Did I make a lame complaint? Check for Bingo!
Totally gracious CSDC Season 2 Division 4 Champeen!

Dungeon Master

Posts: 634

Joined: Sunday, 22nd September 2013, 14:46

Post Wednesday, 29th January 2014, 18:49

Re: Suggestion: Review vault / branch generation weights

I spent a while in a new lab and encountered maybe 1 thing, I think it was either bad luck in having a lot generated or landing far away from the exit.
User avatar

Lair Larrikin

Posts: 27

Joined: Thursday, 5th May 2011, 18:52

Post Wednesday, 29th January 2014, 19:15

Re: Suggestion: Review vault / branch generation weights

Just started adding the vault analysis to a spreadsheet, at temple vaults at time of posting. Things I've noticed so far:

* sewer_kobold is as we know too high
* The ice caves have an almost totally equal distribution - that could be changed as per our discussion
* There are two temple entry vaults with the same name!
* Temple vaults seem to favour some of the classical vaults. circular_temple_12 appears 17% of the time

https://docs.google.com/spreadsheet/ccc ... sp=sharing

For this message the author oiseaux has received thanks:
dpeg

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Wednesday, 29th January 2014, 19:38

Re: Suggestion: Review vault / branch generation weights

oiseaux: For the temples, this is on purpose. That dominating layout is the ultra-classical one.
User avatar

Lair Larrikin

Posts: 27

Joined: Thursday, 5th May 2011, 18:52

Post Wednesday, 29th January 2014, 19:57

Re: Suggestion: Review vault / branch generation weights

On second thoughts having talked about this, the only real benefit this brings is spotting clear outliers and flat distributions - so most other things are fine, although I still think the distributions should be tiered, rather than gradated so finely as they are with big files like Temple. So it's probably easier just to search for WEIGHT in the .des files and see if anything looks out of place, which doesn't seem to be the case in everything else I've gone through. So yeah, not much point going forwards, since it sounds like the older vault files need to be looked at and revised as a whole rather than just weights, is the impression I get. What a voyage of discovery I'm on today...

Re: Temple, what's the rationale behind that? There are some nice water ones in there I've only seen in wizmode :(

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Wednesday, 29th January 2014, 20:01

Re: Suggestion: Review vault / branch generation weights

oiseaux: You don't have to agree with this, of course, but my thinking was like this: back then, there was exactly one map, the round one. So I took that as the staple. The variations are more rare, and some very much so, but that's because the Temple gets visited very often, and thus you can get a new (to you) map even after a long time of playing.

Feel free to change the numbers in any way you like, though.

Swamp Slogger

Posts: 146

Joined: Saturday, 24th March 2012, 02:07

Post Wednesday, 29th January 2014, 20:09

Re: Suggestion: Review vault / branch generation weights

I don't know if this is on topic, but in the review of the temple vaults, could some of them be adjusted for size? Some of them take 30 turns to explore, some of them take a couple hundred and since auto-explore does not stop for temple altars it feels like a waste of turns. Granted, this is only situationally relevant (i.e. when one cares about turncount) but it still feels wasteful.
User avatar

Lair Larrikin

Posts: 27

Joined: Thursday, 5th May 2011, 18:52

Post Wednesday, 29th January 2014, 20:11

Re: Suggestion: Review vault / branch generation weights

oiseaux: You don't have to agree with this, of course, but my thinking was like this: back then, there was exactly one map, the round one. So I took that as the staple. The variations are more rare, and some very much so, but that's because the Temple gets visited very often, and thus you can get a new (to you) map even after a long time of playing.

Feel free to change the numbers in any way you like, though.


I totally see it from a low vault count perspective, but do you think there's a case now, for maybe putting it on a level with some of the other vaults? I mean, looking at the sheet, there are 16 vaults that you would only see once in every 200 games, and 10 you'd only see once in every 100 games (I hope my maths worked there)! With this in mind, it's clear you're still going to get your rare vaults - but the super common vault should at least be on a par with the top 10 vaults - which altogether account for 50% of the vaults you see!

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Wednesday, 29th January 2014, 20:26

Re: Suggestion: Review vault / branch generation weights

If I remember correctly, some of those really are maps come in a bunch of similar types, so I always considered their sum.
User avatar

Pandemonium Purger

Posts: 1341

Joined: Monday, 24th October 2011, 06:13

Post Friday, 31st January 2014, 03:40

Re: Suggestion: Review vault / branch generation weights

Tenaya wrote:I don't know if this is on topic, but in the review of the temple vaults, could some of them be adjusted for size? Some of them take 30 turns to explore, some of them take a couple hundred and since auto-explore does not stop for temple altars it feels like a waste of turns. Granted, this is only situationally relevant (i.e. when one cares about turncount) but it still feels wasteful.


Someone with a concern over turncount wouldn't use autoexplore at the temple in the first place
seattle washington. friends for life. mods hate on me and devs ignore my posts. creater of exoelfs and dc:pt

Swamp Slogger

Posts: 146

Joined: Saturday, 24th March 2012, 02:07

Post Friday, 31st January 2014, 06:10

Re: Suggestion: Review vault / branch generation weights

twelwe wrote:
Tenaya wrote: Some of them take 30 turns to explore, some of them take a couple hundred and ....


Someone with a concern over turncount wouldn't use autoexplore at the temple in the first place


Try again at reading.
User avatar

Dungeon Master

Posts: 4031

Joined: Thursday, 16th December 2010, 20:37

Location: France

Post Friday, 31st January 2014, 08:55

Re: Suggestion: Review vault / branch generation weights

Why would you explore (auto or not) the temple in a speed run? If your concerned about wasting turns, don't explore the temple. There is absolutely no issue with varying temple size.
<+Grunt> You dereference an invalid pointer! Ouch! That really hurt! The game dies...

Swamp Slogger

Posts: 146

Joined: Saturday, 24th March 2012, 02:07

Post Friday, 31st January 2014, 17:31

Re: Suggestion: Review vault / branch generation weights

galehar wrote:Why would you explore (auto or not) the temple in a speed run? If your concerned about wasting turns, don't explore the temple. There is absolutely no issue with varying temple size.


Sure, I would. Even in a speedy game I will choose a diety. I need to explore the temple enough to find the altar. Some of the temples have the altars spread out over a much longer linear path. With the open floor plan temples it only takes a few (maybe 30) turns to find the altar that you want. With the linear maps (I'm thinking in particular of the one where it is a forest path and one altar is found in each forest clearing) it takes much much longer to find the altars.

There is an issue with varying temple size; the disagreement is the impact and importance of the issue. I believe this issue was brought up before by someone else; maybe it was determined to be too trivial. That's fine; I'm just trying to help and thought it would be a good opportunity to look into some conformance (or designed lack thereof) of the temple sizes.
User avatar

Barkeep

Posts: 1788

Joined: Saturday, 29th June 2013, 16:52

Post Friday, 31st January 2014, 18:57

Re: Suggestion: Review vault / branch generation weights

I'm not sure how this is problematic, Tenaya. You're describing the kind of variability that ends up playing out on nearly every map in Crawl. Removing large Temple maps would constitute the tiniest buff to speedrunning (maybe 100 turns or so for supremely unlucky characters) and would otherwise accomplish nothing except removing content.
User avatar

Lair Larrikin

Posts: 27

Joined: Thursday, 5th May 2011, 18:52

Post Friday, 31st January 2014, 22:43

Re: Suggestion: Review vault / branch generation weights

Even if you changed the vaults themselves, there's still the classic vault entries where you have to search for the secret door in order to get into it. My first runs back in 0.5 wasted god knows how many turns trying to get in. Hehe

Tartarus Sorceror

Posts: 1888

Joined: Saturday, 9th July 2011, 20:57

Post Saturday, 1st February 2014, 19:34

Re: Suggestion: Review vault / branch generation weights

oiseaux wrote:Even if you changed the vaults themselves, there's still the classic vault entries where you have to search for the secret door in order to get into it.


Secret doors have been gone for several versions.

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Saturday, 1st February 2014, 19:50

Re: Suggestion: Review vault / branch generation weights

I don't think that we should take extra measures for speed runs. Some vaults are tiny, others are huge... a speedrunner might get especially (un)lucky, and that's part of the challenge. My beef with the large temples is not the turncount, but how autoexplore doesn't stop -- it should for finding altars of new gods. Anyone knows how to do that from the rc file?
User avatar

Dungeon Master

Posts: 4031

Joined: Thursday, 16th December 2010, 20:37

Location: France

Post Saturday, 1st February 2014, 21:57

Re: Suggestion: Review vault / branch generation weights

explore_stop += altars does just that, and it's default. But it's explicitly disabled in the temple. Not sure if it should.
<+Grunt> You dereference an invalid pointer! Ouch! That really hurt! The game dies...

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Saturday, 1st February 2014, 21:59

Re: Suggestion: Review vault / branch generation weights

galehar: Thank you, that explains it. So the problem is that for the large maps I'd really want autoexplore to stop, but the for the smaller ones it is annoying?
Personally, I'd rather take stopping (and gods being announced one by one) for all them.

For this message the author dpeg has received thanks:
XuaXua
User avatar

Dungeon Master

Posts: 4031

Joined: Thursday, 16th December 2010, 20:37

Location: France

Post Saturday, 1st February 2014, 22:54

Re: Suggestion: Review vault / branch generation weights

I think this special case can be removed. It seems to predate all the new temple vaults, so whatever rational was behind it is probably obsolete by now.
<+Grunt> You dereference an invalid pointer! Ouch! That really hurt! The game dies...

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Saturday, 1st February 2014, 23:09

Re: Suggestion: Review vault / branch generation weights

galehar: Could you do that? Perhaps this was back when there were only the standard twelve temple gods, so the information was clear: as soon as you entered the Temple, the full pantheon was available to you. It's not like this anymore, and that is why I'd like altar stopping in the Temple, too.
User avatar

Dungeon Master

Posts: 4031

Joined: Thursday, 16th December 2010, 20:37

Location: France

Post Sunday, 2nd February 2014, 20:48

Re: Suggestion: Review vault / branch generation weights

ok, done.
<+Grunt> You dereference an invalid pointer! Ouch! That really hurt! The game dies...

For this message the author galehar has received thanks: 3
dpeg, WalkerBoh, XuaXua

Return to Game Design Discussion

Who is online

Users browsing this forum: No registered users and 29 guests

cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group.
Designed by ST Software for PTF.