#3004 Zero-scum traps and doors
This is an idea I had while discussing mapping changes with dploog; it seemed tentatively good, so it is here for public debate.
First, the 's'earch command is bad. When one is looking for traps or doors, a turn spent is almost always irrelevant; so search just becomes a ritual action to gain some reward. In other words, grinding. Worse, because search can be spammed, traps and doors skill matters very little.
Therefore, the search command, at least as it is used outside of combat, should go. Here is a proposal for a search-less Traps and Doors implementation.
Each trap and secret door in the dungeon has a difficulty number. This is rolled by a Poisson distribution with a mean related to depth, such that 95% of traps on Zot:5 have a difficulty no higher than 15. Difficulties range from 0 to 27. A trap or door is known if and only if, you have seen it, and your Traps and Doors skill is higher than or equal to its difficulty. Note that your skill at the time the door was seen does not matter.
The travel cache keeps track of the traps and doors you have seen but not yet discovered and that block autoexplore. When your skill increases, it checks if there are any new possible locations to autoexplore, and if so generates a message. "You suddenly recognize an unexplored secret area on D:3.". This ties in to the new ability to see maps of other dungeon levels.
Discovering traps and doors does not disturb them, or reveal them for other actors. Examining a discovered secret shows that it is obvious only to you.
The search command is removed.
Disarming traps are as now. Setting traps will not be affected.
This radical approach has merits but also problems.
First, the terminology: 's'earching is just the same as resting (unlike Nethack). In fact, you are searching on every turn, even if you walk. The only reason to have the 's' command is for compatibility with other games.
Next, this change would demand a complete overhaul in trap creation. If you reach T&D 3 and you're informed about arrow/axe/bolt traps all over the dungeon, that is just message spam.
This approach makes it unclear how T&D can be trained. By running back to the announced traps and disarming them?
Finally, I think the (currently) most cool bit about T&D is how you can, at higher skill levels, detect distant (iei. non-adjacent) traps/doors.
What I think would be fair game and a meaningful improvement: T&D tells you about secret doors you missed but only if they lead somewhere. Crawl generates tons of secret doors and, just like with the traps, announcing them all would be rather pointless. This is still not completely satisfying, but could be a good base for further discussion.
This sounds exactly right to me.
Note that as proposed, there is no message spam. Announcements come only when new, unexplored tiles open up.
As a logical next step, disarming traps should also be deterministic -- if T&D - difficulty > X it succeeds, otherwise it fails, where X depends on the type of trap. (This opens up the possibility of disarming magical traps with sufficient skill.)
Lemuel
1. You currently search more while resting than while moving. This would change.
2. I had not intended for trap realization to generate any messages, since knowing about a trap in an explored area is not useful information. Only doors would be announced, and only if they lead somewhere (IE, autoexplore would open them).
3. I expected that traps and doors would be trained when traps are discovered, either by skill increase or by entering LOS.
4. With this proposal you can discern traps and doors anywhere in LOS, but your probability of noticing them depends on level. Is that too bad? Note that if the distance is limited, people will want autoexplore to keep close to walls. Not good.
5. Isn't that in my proposal?
1. could be changed simply by treating all turns the same way when it comes to searching.
2. Okay, so no announcement of traps. But the vast majority of secret doors is not relevant by the time you'd be informed. This is why there would need to be some code checking that you actually gain something from going to the secret door. Quite often, you see something like
.....
.x=x.
......
where it'd be obvious that the door is irrelevant.
I assume that traps are identified silently? E.g. Ctrl-F "arrow" would list arrow traps you didn't see when you were near them?
The system has a lot going for it. I just feel that something is not quite right. This could also be generation of secret doors, I don't know.
2. No announcement of traps. Doors in islands would be handled as if by autoexplore; autoexplore will not open a door it can walk around, so we wouldn't announce finding them.
6. Trap identification is silent, yes. Perhaps there should be a Lua hook or something, so you can be notified if you have the trap's ammo on autopickup?
7. The main thing I don't like is that it makes very high (20+) traps and doors even more useless than currently. But that will be fixed with trap setting, I think.
Training by discovery will not work with this model.A player raises his T&D to level 2 and spends his last free XP for this.All level 2 traps he hasn't seen are discovered but he can't train because he has no free xp left.NOw he knows all the traps behind him and can only hope that the deeper levels have some low level traps and doors to raise to level 3.But with the trap distibution the chance is small and even smaller when he wants to raise level 3 to level 4.
Why would I care about a secret door I missed on D3 if I'm somewhere lower than like D5?
I dont like the utter lack of random factor in that proposal. Otherwise it may be made at side of current T&D or extend it outside of LOS. Something in the lines:
Once in [tweak number of] turns ...
If closest from easiest of undiscovered but seen doors or traps complexity loses [tweak roll algorithm, take distance to it into account] against your T&D skill then the door is discovered and your T&D is practiced. If discovered is a door and is not within your LOS and has area that you have not seen behind it then it is also announced.
On the contrary, the lack of a random factor is precisely what's good about the proposal.
One of the basic principles of Stone Soup is no scumming, which means that there should be no benefit from repeating a risk-free activity. Hitting '5' is risk-free, so doing it 50 times shouldn't be any more useful than doing it once.
With regard to searching,one application of this principle would be to make it genuinely costly, e.g. giving the player a finite number of "search points" refreshed on level-up, or having a much stricter clock. But in the absence of major changes like that, a strictly deterministic searching mechanic is the only way to avoid rewarding mindless repetition.
Lemuel
I fail to see how you could train T&D under these circumstances. Someone with 0 skill will never detect any traps, or only level 0 traps or what? Assuming the hindsight detection (why?) would shove a lot of xp into T&D at once, you either end up with wasted skill points because you didn't have any xp in your pool right then or you lose a large chunk of xp when you're not interested in training this. Either way this leads to victory dancing (try to kill only large monsters when your trap skill is at 95%, or conversely, try to spend all of your xp before killing the small monster required for the level increase).
I really like the current mechanic where you only detect stuff in los. If I didn't notice a trap or door back then on level 1, why should I suddenly in hindsight realize it's there because I've become more observant. Of course, if I go back to level 1 it'll be really obvious but I wouldn't expect my character to memorize the scenery to such a degree they would recognize belatedly there was a trap somewhere in a region they passed through ages age.
What is "risk-free activity"? The tactic that you rest each 5 steps to discover possible traps in LOS? You supposedly:
1) run out of food
2) monsters on level wake up
3) out of depth monsters are spawned
4) even muna or nemelex get mad if you really push it
5) your end score goes down
If that is not the case and it can be still all considered "risk-free" then the risks are simply too low and sitting idle is too profitable. I suspect that it is the other way around and the whole "rest for detection is scummable and spammable" issue is conjured out of thin air.
All what is proposed is to turn all traps and doors into ??? that is 100% immediately discovered when conditions are met and has 0% chance to be discovered when conditions are not met. "conditions" here are (trap complexity <= skill of T&D) , but it sounds like not worth to be in game in any set of conditions i can imagine.
A yet another downside: if you teleport/hatch/etc into any closed area, you have no way out.
To castamir: Hatches are already prevented from leading into sealed areas. Random teleport could be made the same.
To j-p-e-g: At this point I think training by exploration would work better; you have a (say) 1/50 chance of training for each square uncovered. Retroactive discovery is crucial; otherwise players will return to previous levels when their T&D score is higher.
I think castamir was talking about areas closed off via secret doors that you either would or wouldn't discover. The only alternative to prevent players being stranded that way would be to turn all secret doors shutting off an area that is otherwise unreachable into normal doors. Think of the glasswall Temple entrance minivault.
As for training by exploration, that would apply to all characters, so everyone would train T&D because other than for Stealth or Armour there are no further conditions. Is that what you want?
To be fair, the glasswall Temple entry should not be mentioned. The main reason I added this one (and keep it in the game) is to teach new players about searching. This would be pointless under the proposed system, and so would be the vault.
I think that kotk has a good point. While normally the food and piety clocks are too slow to matter, they do here: searching for 5 turns before every move is simply not an option (mummies don't count).
This proposal might lead to something good but I see a considerable risk and the gain is not fully clear to me. Therefore, I'd suggest not to test it, or at least only in a reduced form. Here is a proposal: there could be vaults which obviously have entries, only that players with insufficient T&D skill won't find them. In fact, there could be a message to that effect. This would tell players that there is this thing somewhere in the dungeon which profits from good T&D. Alternatively, you try to bring tools (Shatter) to open it nonetheless. (It is possible to mark an area as no-teleports-into-this.)
Based on reactions here and elswhere, I have an alternate proposal, FR 2892391.