Changes to make Crawl harder

[grisamentum]

I don't think new runes are the right direction - I think the idea is closer to "more tough choices per minute/turn" than "additional hard sections that make the game even longer". The game is already quite long as it is.

On the contrary, the game can't get any longer, so long as you can win with a mere 3 runes.

[johlstei]

I don't think new runes are the right direction - I think the idea is closer to "more tough choices per minute/turn" than "additional hard sections that make the game even longer". The game is already quite long as it is.

On the contrary, the game can't get any longer, so long as you can win with a mere 3 runes.

The point is, if one wants the game to be more challenge-dense, it's not sufficient for it to be so after the point where one could have finished. If the game should be harder, it should be harder for people who do 3 rune games, not just those who decide to continue playing optionally after that. I'm undecided about whether I feel it should be harder, but if I come to the conclusion that it should be, that should include the 3-rune game.

[dpeg]

Note that additional (i.e. completely extra) content is sometimes added, such as ziggurats or new Hell/Pan monsters. (Also note how ziggurats try to fill a role between "yardstick for veterans" and "can be useful in a 3-rune game".)

On the rest: yes, it's not really about harder. For all the changes discusses, you can consider whether a new/uninformed/unspoiled player would be concerned. Very often he'd not, which is a good sign to me that the change cannot be so bad.

Yes to no more mandatory runes: the rune lock is an attempt to put a little pressure on players, more runes won't. And we have trouble differentiating our postgame content anyway.

[grisamentum]

I don't think new runes are the right direction - I think the idea is closer to "more tough choices per minute/turn" than "additional hard sections that make the game even longer". The game is already quite long as it is.

On the contrary, the game can't get any longer, so long as you can win with a mere 3 runes.

If the game should be harder, it should be harder for people who do 3 rune games, not just those who decide to continue playing optionally after that. I'm undecided about whether I feel it should be harder, but if I come to the conclusion that it should be, that should include the 3-rune game.

Really? Why?

[johlstei]

Really? Why?

The 3-rune game is already a bit long for my liking and generally I don't want to play a single character any more than that, so I would get little out of stuff past it that I can only play after many hours - I simply don't see myself experiencing that content. Maybe it's selfish to view the game this way but it's a really long game.

[grisamentum]

I suppose it's a matter of preference. It seems like such a short game already.

I'm not suggesting more mandatory runes of course. Maybe more intermediate runes, if you feel you'd never see the later runes.

[Brannock]

I generally win Crawl in ~3 hours. For a game with permadeath, that's not really "short" at all to me.

Three hours (likely much more as I'm not always absolutely attentive to Crawl when I play) is a very significant chunk of my free time. Many players take much longer than that to win. I think the length of Crawl right now is pretty acceptable.

[Sprucery]

I think the length of Crawl right now is pretty acceptable.

Me too. My winning games take anything from 18 hours upwards, but I play the game in small patches over several weeks and enjoy every moment.

[grisamentum]

I generally win Crawl in ~3 hours. For a game with permadeath, that's not really "short" at all to me.

Three hours (likely much more as I'm not always absolutely attentive to Crawl when I play) is a very significant chunk of my free time. Many players take much longer than that to win. I think the length of Crawl right now is pretty acceptable.

Interesting. That's exactly how long I thought a good player could win a 3-rune game in, and yes, my wins are much slower (my fastest is only around 5 hours, and my first 15-rune win took almost 20 hours). But hasn't crawl become faster over the years? My first few wins were like 13-14 hours for 3 runes. My first win in .16 trunk was 6 runes and 9 hours, followed by 3 runes in 5 hours.

Regardless, comparing to all games I play, Crawl is relatively short for one "game" or "campaign" of a single-player game. It's much longer than something like FTL, especially for extended. But most other games fall on the "longer" end by comparison.

I'm not saying crawl needs to be "longer" or that it's "too short" - just that it feels like there is room for more optional extended. But I understand if most other people don't share that preference.

[KittenInMyCerealz]

I personally would be interested to see more optional extended (mostly because i find Zigs very unfun, haven't even beaten one fully yet), but i don't think it should be anywhere near top priority in the "to do"-list.

[ragnarokchu]

If people want the game to be more brutal earlier without more late-game content why not release an "hard" mode or whatever flavor you want to call the mode.

I mean you can already make the game much harder playing certain combos + certain gods, or trying to go for every single rune and thus on. But making the game harder just to "win" when the 3 rune win is suppose to be the bare min (thus the easiest) isn't going to stop Mr. WinManiac who only plays every game with Spriggan Stabber trying to win in 30 mins then claims the game is too easy after his 50th win.

Which like I already discussed before, many people have different ways to play the game and consider things differently. Inherently the biggest issue is people think winning is a sole "metric" on how easy or hard a game is and lose sight of just having fun by making the game intentionally harder.

[johlstei]

If people want the game to be more brutal earlier without more late-game content why not release an "hard" mode or whatever flavor you want to call the mode.

I mean you can already make the game much harder playing certain combos + certain gods, or trying to go for every single rune and thus on. But making the game harder just to "win" when the 3 rune win is suppose to be the bare min (thus the easiest) isn't going to stop Mr. WinManiac who only plays every game with Spriggan Stabber trying to win in 30 mins then claims the game is too easy after his 50th win.

Which like I already discussed before, many people have different ways to play the game and consider things differently. Inherently the biggest issue is people think winning is a sole "metric" on how easy or hard a game is and lose sight of just having fun by making the game intentionally harder.

For me the dicey situations where I have to consider every move carefully are the fun parts, fights where I don't have to think aren't very fun. There's lots of these situations in the early game and I find they get more sparse once the player enters lair.

[and into]

I realize that the idea of this thread itself (which opens with several, not strictly related changes at once) might lend itself to unfocused posts, but I still want to ask of people who are troubled by any of the changes brought up this far, to please try to refer directly to which changes you find troubling, and why. "This would make Crawl too hard and/or unfair" is a potentially legitimate criticism, but in many of the posts it is unclear what proposal(s) the criticism is referring to, and so it is unclear whether or not the criticism is warranted.

[neil]

I realize that the idea of this thread itself (which opens with several, not strictly related changes at once) might lend itself to unfocused posts, but I still want to ask of people who are troubled by any of the changes brought up this far, to please try to refer directly to which changes you find troubling, and why. "This would make Crawl too hard and/or unfair" is a potentially legitimate criticism, but in many of the posts it is unclear what proposal(s) the criticism is referring to, and so it is unclear whether or not the criticism is warranted.

I do think "What should be the overall difficulty/length of the game" is a valid topic, though, and one that should be discussed from time to time, independently of particular proposed changes. To some extent the title of this thread enourages that topic. Perhaps someone would like to start such a thread and move some of the discussion there (not I: I haven't yet worked out my thoughts on the matter well enough to write a post as reasoned as I would like it to be).

[duvessa]

Make unknown traps trigger when they enter LOS, rather than when they are stepped on
In addition to providing a difficulty increase this makes traps much more effective at achieving their purpose (currently you can avoid them by keeping track of which squares have been stepped on).

[Sprucery]

Make unknown traps trigger when they enter LOS, rather than when they are stepped on
In addition to providing a difficulty increase this makes traps much more effective at achieving their purpose (currently you can avoid them by keeping track of which squares have been stepped on).

Combine this with
Bring back Traps skill
This way the player can control how much traps affect their game. The effect on game difficulty depends on how trap detection works with no skill and how it scales with increasing skill. Anyway, training Traps means other skills will be lower correspondingly.

The effect of the skill should be that instead of triggering, the trap will be revealed without triggering. It could also make the effect of the trap less severe, depending on trap type.

[kroki]

(currently you can avoid them by keeping track of which squares have been stepped on).

i dont think anyone ever did this

[Sivar]

(currently you can avoid them by keeping track of which squares have been stepped on).

i dont think anyone ever did this

I have done it before while picking my way through a part of Lair that had a bunch of shafts near some scrolls or something that I wanted

[ragnarokchu]

Make unknown traps trigger when they enter LOS, rather than when they are stepped on
In addition to providing a difficulty increase this makes traps much more effective at achieving their purpose (currently you can avoid them by keeping track of which squares have been stepped on).

Combine this with
Bring back Traps skill
This way the player can control how much traps affect their game. The effect on game difficulty depends on how trap detection works with no skill and how it scales with increasing skill. Anyway, training Traps means other skills will be lower correspondingly.

The effect of the skill should be that instead of triggering, the trap will be revealed without triggering. It could also make the effect of the trap less severe, depending on trap type.

Traps aren't really an interesting enough mechanic, they either do nothing or cause instant gibs. Very few actually create any meaningful interaction.

[Lasty]

The version of this I'd advocate is that each level rolls a number of traps to include (say 0-6), and each time you explore a new tile on that floor, each remaining trap on the floor has a 1 in (remaining unexplored tiles + 1) chance of triggering on you. That way you get the benefit of not having traps arbitrarily block connectivity and the benefit of not giving an advantage to players who tediously keep track of tiles they've stepped on, while also avoiding the situation where autotraveling back through levels random shafts you repeatedly just to add irritation.

[bel]

The version of this I'd advocate is that each level rolls a number of traps to include (say 0-6), and each time you explore a new tile on that floor, each remaining trap on the floor has a 1 in (remaining unexplored tiles + 1) chance of triggering on you. That way you get the benefit of not having traps arbitrarily block connectivity and the benefit of not giving an advantage to players who tediously keep track of tiles they've stepped on, while also avoiding the situation where autotraveling back through levels random shafts you repeatedly just to add irritation.

Does anyone bother to keep track of where you've been? I just autoexplore. Most traps, even if you trigger them, are easily met with some consumable, at worst. It is possible to have bad luck and be surrounded by a shadow trap with no way to escape, but those are very marginal cases. (It did happen to me once, I survived anyway).

[Zeke]

The version of this I'd advocate is that each level rolls a number of traps to include (say 0-6), and each time you explore a new tile on that floor, each remaining trap on the floor has a 1 in (remaining unexplored tiles + 1) chance of triggering on you. That way you get the benefit of not having traps arbitrarily block connectivity and the benefit of not giving an advantage to players who tediously keep track of tiles they've stepped on, while also avoiding the situation where autotraveling back through levels random shafts you repeatedly just to add irritation.

Other options:

-Remove traps altogether. Traps sometimes feel like the cruel hand of fate. They have similar problems to monsters wielding chaos weapons. We already have Zom after all.
-Replace traps with monsters that function effectively as traps. At least we could interact with them.
-Restrict traps to certain vaults sealed with a runed door. That way, a player would have the option of subjecting themselves to getting shafted in some way, shape or form. Traps could be made more severe for these optional vaults.

[johlstei]

shafts are good and even if other traps go away, they should stay forever

[Sandman25]

shafts are good and even if other traps go away, they should stay forever

As far as I know it's possible to get shafted from Zot 2 to Zot 5. Also it's possible to get shafted from D3 to D6, walk a tile, get shafted from D6 to D9, walk a tile, get shafted from D9 to D12, walk a tile, get shafted from D12 to D15.
I don't see how it can be good.

[Lasty]

Does anyone bother to keep track of where you've been?

Sivar just indicated that they do, but aside from individual examples, it's good for the game to make sure that no one feels incentivized to do that, since it's tedious but technically optimal.

[johlstei]

shafts are good and even if other traps go away, they should stay forever

As far as I know it's possible to get shafted from Zot 2 to Zot 5. Also it's possible to get shafted from D3 to D6, walk a tile, get shafted from D6 to D9, walk a tile, get shafted from D9 to D12, walk a tile, get shafted from D12 to D15.
I don't see how it can be good.

Extreme low probability events don't change the fact that the common case of "oh crap, you're out of depth, mad scramble back up before you get murked" is really fun. I play this game because the high tension moments are enjoyable, I don't really care about what the probability of winning or losing are. You've never been satisfied at pulling off a clever escape from D:6 at XL:4 or whatever?

[bel]

Does anyone bother to keep track of where you've been?

Sivar just indicated that they do, but aside from individual examples, it's good for the game to make sure that no one feels incentivized to do that, since it's tedious but technically optimal.

That particular example was a very rare case in one particular situation. Sivar can indicate himself if he does that routinely. Most people autoexplore and not care about traps in general or deal with them if they happen.

Autoexplore will always be inferior to manual exploration. I really think that the incentive for manual exploration is very little, even if it is optimal. This is why I said that even if manual exploration is optimal, the reward is too little to bother, since most traps can be escaped from by using consumables. (But often they can be escaped from without using consumables, so it is not a no-brainer to just use one).

[duvessa]

The version of this I'd advocate is that each level rolls a number of traps to include (say 0-6), and each time you explore a new tile on that floor, each remaining trap on the floor has a 1 in (remaining unexplored tiles + 1) chance of triggering on you. That way you get the benefit of not having traps arbitrarily block connectivity and the benefit of not giving an advantage to players who tediously keep track of tiles they've stepped on, while also avoiding the situation where autotraveling back through levels random shafts you repeatedly just to add irritation.

I don't understand the motivation here. Aside from making it worse to explore the more tiles have already been revealed (which seems obviously bad), how does this differ from my proposal?

That particular example was a very rare case in one particular situation. Sivar can indicate himself if he does that routinely. Most people autoexplore and not care about traps in general or deal with them if they happen.

Most people didn't farm D:1 with mummies, carefully equip allies, use evaporate or alter self, etc. but those things were still horrible for the game.

[bel]

That particular example was a very rare case in one particular situation. Sivar can indicate himself if he does that routinely. Most people autoexplore and not care about traps in general or deal with them if they happen.

Most people didn't farm D:1 with mummies, carefully equip allies, use evaporate or alter self, etc. but those things were still horrible for the game.

I started playing in 0.15, so I have no idea of what most of them are. But my guess is that you are talking about abuseable or boring behaviour which gives great rewards. By contrast, avoiding traps via manual exploration is boring behaviour giving trivial rewards.

Sivar's example was one particular case where a large potion of the area is littered with shafts. This is a particular vault, which encourages back-tracking. I have encountered it myself a few times (to be frank, I just autoexplore and don't care, but that is irrelevant here). One can't abuse this or generalize it to other places. If anything, the simpler solution is to get rid of vaults placing traps.

[Sandman25]

Extreme low probability events don't change the fact that the common case of "oh crap, you're out of depth, mad scramble back up before you get murked" is really fun. I play this game because the high tension moments are enjoyable, I don't really care about what the probability of winning or losing are. You've never been satisfied at pulling off a clever escape from D:6 at XL:4 or whatever?

What is "clever" escape? Using consumables which I could save for dealing with my mistakes? Training stealth in plate armour because I suspected I can be shafted? Saving your only scroll of magic mapping for shaft instead of using it to find a timed portal?
I believe having high tension moments should be possible (and plentiful) without traps. Probably "possible" is not right word, I mean what if you are not shafted? Is it a boring game in this case? Fix it by something else than a random chance to make it interesting.

[duvessa]

you can keep shafts without keeping traps you know
in fact crawl already has a shaft that isn't necessarily caused by a trap! it's called banishment

[bel]

Extreme low probability events don't change the fact that the common case of "oh crap, you're out of depth, mad scramble back up before you get murked" is really fun. I play this game because the high tension moments are enjoyable, I don't really care about what the probability of winning or losing are. You've never been satisfied at pulling off a clever escape from D:6 at XL:4 or whatever?

What is "clever" escape? Using consumables which I could save for dealing with my mistakes? Training stealth in plate armour because I suspected I can be shafted? Saving your only scroll of magic mapping for shaft instead of using it to find a timed portal?
I believe having high tension moments should be possible (and plentiful) without traps. Probably "possible" is not right word, I mean what if you are not shafted? Is it a boring game in this case? Fix it by something else than a random chance to make it interesting.

It is not clear to me what is the problem with some random things happenning in a game that is full of random things happenning. It is not instadeath, nor is it permanently crippling, nor is it a no-brainer to use a consumable to escape (and doesn't always work, anyway). Personally I have never saved a scroll of magic mapping in case I get shafted. It is optimal to find a timed portal anyway even if you have only one, since you get lots of net consumables in any timed portal.

[Lasty]

As far as I know it's possible to get shafted from Zot 2 to Zot 5. Also it's possible to get shafted from D3 to D6, walk a tile, get shafted from D6 to D9, walk a tile, get shafted from D9 to D12, walk a tile, get shafted from D12 to D15.
I don't see how it can be good.

It's also possible to never roll a hit with a weapon or a spellcasting success over an arbitrarily long period. In practice it doesn't happen, and neither does your shafting example. If that happens to some player once, I'm fine with it, because it'll be pretty entertaining. If it happens frequently, then it's a problem we can solve. If, as is the case, it happens only in a purely hypothetical case, I can't imagine why it would be considered to be worth designing around.

[johlstei]

Extreme low probability events don't change the fact that the common case of "oh crap, you're out of depth, mad scramble back up before you get murked" is really fun. I play this game because the high tension moments are enjoyable, I don't really care about what the probability of winning or losing are. You've never been satisfied at pulling off a clever escape from D:6 at XL:4 or whatever?

What is "clever" escape? Using consumables which I could save for dealing with my mistakes? Training stealth in plate armour because I suspected I can be shafted? Saving your only scroll of magic mapping for shaft instead of using it to find a timed portal?
I believe having high tension moments should be possible (and plentiful) without traps. Probably "possible" is not right word, I mean what if you are not shafted? Is it a boring game in this case? Fix it by something else than a random chance to make it interesting.

There's lots of things you can do to manage your chances of escape, like the pattern in which you explore the dangerous level, the way you move and manage line of sight, when you choose to use consumables vs try to run, etc. I do agree that monsters that shaft you are a viable alternative to "random tiles on the ground are shafts".

Having shafts slowly move around the floor would be a viable way to prevent tracking what tiles you stepped on though is pretty bizarre thematically.

It's not a random chance for death, it's a random chance to be put in a situation where you will have to make moves carefully to survive, that cautious play can't prevent from the get go (other than the unfortunate keeping track of what tiles you've stepped on - I agree that is a problem.)

[Sandman25]

It's also possible to never roll a hit with a weapon or a spellcasting success over an arbitrarily long period. In practice it doesn't happen, and neither does your shafting example. If that happens to some player once, I'm fine with it, because it'll be pretty entertaining. If it happens frequently, then it's a problem we can solve. If, as is the case, it happens only in a purely hypothetical case, I can't imagine why it would be considered to be worth designing around.

I am not fine with player being unable to roll a hit over an arbitrarily long period, it is another design problem. I suggest to solve it by giving a warning like "A goblin comes into view. You will be unable to hit it with your weapon during next 54 attacks".
Why frequency is relevant here? How often did Hell effect paralyze player? Is it comparable with shadow traps for Fo or shafts from Zot2 to Zot5?

PS. I am not joking about goblin. Either provide the warning or make it unable to miss 5 (the number is arguable and depends on to_hit vs EV difference) times in a row (monster should not miss player over an arbitrarily long period also)

[Lasty]

The version of this I'd advocate is that each level rolls a number of traps to include (say 0-6), and each time you explore a new tile on that floor, each remaining trap on the floor has a 1 in (remaining unexplored tiles + 1) chance of triggering on you. That way you get the benefit of not having traps arbitrarily block connectivity and the benefit of not giving an advantage to players who tediously keep track of tiles they've stepped on, while also avoiding the situation where autotraveling back through levels random shafts you repeatedly just to add irritation.

I don't understand the motivation here. Aside from making it worse to explore the more tiles have already been revealed (which seems obviously bad), how does this differ from my proposal?

It doesn't make it worse to explore the more tiles have already been revealed, but yeah, this isn't actually different from your proposal, except that the game has to do more work since it isn't pre-placing the traps. So, I suppose I'm actually endorsing your plan. I thought I had a problem with your plan, but now I can't seem to recall what it is.

[wheals]

shafts are good and even if other traps go away, they should stay forever

As far as I know it's possible to get shafted from Zot 2 to Zot 5. Also it's possible to get shafted from D3 to D6, walk a tile, get shafted from D6 to D9, walk a tile, get shafted from D9 to D12, walk a tile, get shafted from D12 to D15.
I don't see how it can be good.

For the record, Zot (along with Elf, Crypt, and all rune branches) have a special flag that prevents exactly that. You can shaft from Zot:1 to Zot:4 but not from anywhere but Zot:4 to Zot:5.

[Sandman25]

There's lots of things you can do to manage your chances of escape, like the pattern in which you explore the dangerous level, the way you move and manage line of sight, when you choose to use consumables vs try to run, etc. I do agree that monsters that shaft you are a viable alternative to "random tiles on the ground are shafts".
It's not a random chance for death, it's a random chance to be put in a situation where you will have to make moves carefully to survive, that cautious play can't prevent from the get go (other than the unfortunate keeping track of what tiles you've stepped on - I agree that is a problem.)

It means I am forced to use manual exploration. But I can use the same safe tactics without being shafted, right? Nothing is changed in this case. I am already very careful when I don't feel OP for current level so all shafting does for me is provides a chance to die randomly.

[Sandman25]

For the record, Zot (along with Elf, Crypt, and all rune branches) have a special flag that prevents exactly that. You can shaft from Zot:1 to Zot:4 but not from anywhere but Zot:4 to Zot:5.

Great. I am glad the problem has been realized and partially fixed.

[duvessa]

The version of this I'd advocate is that each level rolls a number of traps to include (say 0-6), and each time you explore a new tile on that floor, each remaining trap on the floor has a 1 in (remaining unexplored tiles + 1) chance of triggering on you. That way you get the benefit of not having traps arbitrarily block connectivity and the benefit of not giving an advantage to players who tediously keep track of tiles they've stepped on, while also avoiding the situation where autotraveling back through levels random shafts you repeatedly just to add irritation.

I don't understand the motivation here. Aside from making it worse to explore the more tiles have already been revealed (which seems obviously bad), how does this differ from my proposal?

It doesn't make it worse to explore the more tiles have already been revealed

Sorry, I was imprecise. It makes it worse if you've noticed that noticed that a lower-than-average number of traps has triggered so far. For example if each level has 1d7-1 traps, and you've explored half of a level without finding any traps, you know the remaining squares are more likely to result in traps triggering than the first 50% of squares. The reverse also applies, if you've already triggered 5 traps you know the remaining squares are safer.
However, I just looked at the code and confirmed one of my fears, which is that this is true for the current version of Crawl as well (see traps.cc:1942). So your proposal is actually about the same as the current situation. I was thinking (read: hoping) that each square had an independent chance for a trap. So:

Give each square an independent chance of generating a trap, instead of picking a number of traps for each level
This way players cannot count the number of discovered traps to estimate the number of undiscovered traps remaining on the level.

[Lasty]

I am not fine with player being unable to roll a hit over an arbitrarily long period, it is another design problem. I suggest to solve it by giving a warning like "A goblin comes into view. You will be unable to hit it with your weapon during next 54 attacks".
Why frequency is relevant here? How often did Hell effect paralyze player? Is it comparable with shadow traps for Fo or shafts from Zot2 to Zot5?

This might be a good time to draw a distinction between "design problem" and "something I dislike personally", since there's an awful lot of confusion about that running around the forum recently. If you want to argue that being able to miss an arbitrary number of times in a row is a design flaw, you need to specify how many times in a row you should be allowed to miss and provide substantial and convincing objective evidence that that number is the correct number. If instead you acknowledge that you would simply prefer that it be different, then you can try to convince other people that they would also like it if it were different.

Frequency matters because it determines how often the situation arises. If a system gives excellent results in all but one in one billion cases, and bad results once in one billion times, that's almost certainly preferable to a system that gives acceptable results every time, right? Shafts give what I consider to be good results almost always, and comically bad results like the one you described so infrequently that it might as well be never. Removing shafts to prevent a situation that is so improbable that it almost certainly never comes up is a huge overreaction. If, however, the existence of shafts made results like the comically bad example you gave common, that would obviously be a problem, because it would affect a meaningful number of games instead of a number of games closer to 0 than 1.

I have no idea why you think hell effect paralysis or shadow traps on formicids are relevant to what either of us said.

[Lasty]

For the record, Zot (along with Elf, Crypt, and all rune branches) have a special flag that prevents exactly that. You can shaft from Zot:1 to Zot:4 but not from anywhere but Zot:4 to Zot:5.

Great. I am glad the problem has been realized and partially fixed.

I think that got changed simply to prevent shafting people into LOS of the orb, since you're supposed to have to work a bit harder to get there.

[Sandman25]

I think that got changed simply to prevent shafting people into LOS of the orb, since you're supposed to have to work a bit harder to get there.

In this case it would be impossible to shaft to Zot 5 even from Zot 4. Now you still can shaft from Zot1 to Zot 4 and then from Zot 4 to Zot 5 so I hope devs simply don't want players die on next turn after shafting (because of -cTele and rune/orb vault full of monsters)

[Lasty]

You can also use hatches on Z:4 (unless those were removed) or tele immediately on getting to Z:4; you just don't get to skip Z:3 and Z:4 when you do that.

Edit: neil says that the intention was to prevent shafting players into branch ends that are significantly more dangerous than the floor shafted from, so I'm probably wrong about the intention.

[neil]

You can also use hatches on Z:4 (unless those were removed) or tele immediately on getting to Z:4; you just don't get to skip Z:3 and Z:4 when you do that.

Edit: neil says that the intention was to prevent shafting players into branch ends that are significantly more dangerous than the floor shafted from, so I'm probably wrong about the intention.

I should point out that generated-known shafts are still allowed on branch:$-1 ... it's only unknown ones (that might shaft an unaware player) that are prevented, and only at level-gen time.

[crate]

Some clarification on the "design vs preference" thing:
I don't see how random things happening during exploration (effects like traps or like hell effects) is objectively bad design. Nowhere in crawl's design philosophy is it laid out that the player should be in control of all the situations she finds herself in, and it's in fact pretty clear that to some degree the opposite is a goal. You may disagree with it as a design decision, but I think you would have problems trying to argue it is objectively bad.

Now, the current implementation of traps has problems that are undeniably design problems. Crawl explicitly tries to avoid "grindy" behaviours; tracking which spaces your character has stepped on is precisely behaviour like this. Since traps favor tracking which spaces your character has traversed, and this contradicts a design axiom, it is objectively bad. In addition, there are ways to implement the "player steps on unknown trap" aspect of traps while completely eliminating the problem.

The difference is that in one case we find a design axiom--something that is the actual starting point for the design ideas behind crawl--and find it contradicted (and in a way that could be fixed without violating other design axioms or causing other problems, at least in theory). This is the game design problem. In the other case ("I don't think players should have permanent Xom wrath" for example) it's something that, yes, probably most players disagree with, but it's not definitively wrong. You can still argue against the latter, but you can't say that it must be changed.

(In theory, traps should be changed on the bad-design basis, but in reality the devs are people doing this pretty much as a hobby so even things that definitely contradict design principles might be overlooked.)

Not being a dev myself, and additionally since I by choice avoid ##crawl-dev (so I don't have a lot of interaction with most of the devs), there are surely times where I think something is an axiom of crawl design but then it turns out, no, it's not. But it's still a good standpoint to keep in mind, and if the basis for your argument turns out to be incorrect (I thought "never let the player miss 5 times in a row" was an axiom but it wasn't) then sometimes you have to just accept it and move on because the devs don't agree.

---

Personally I'd suggest just making all actual traps pre-revealed, and then if you want to save the random-effects-while-exploring bit implement it separately. But this is a lot of work to solve a problem that is, admittedly, not that big, so I think it is unlikely to happen.

[Rast]

You're not alone in having these kinds of ideas. Lots of the dorks you see on youtube with channels about game design have the same focus group tested ideas and there are even authors out there who express roughly this idea in relation to literature.

Fiction. Not literature.