cTele

There are several seperate problems with the control teleport effect, which were recently discussed on IRC (http://s-z.org/crawl-dev/%23%23crawl-dev-20121119.log from 17:31 to the end). This is a summary of the problems, and further down the proposed solutions are listed and can be discussed.

1. Controlled Teleports
   • For escaping dangerous situations these are a lot less interesting than random teleports.
   • They are also simply too good for escaping, because they guarantee safe escape if you survive until the teleport kicks in, at low or no turn cost, taking much of the danger out of dangerous fights.
   • Having to target it away from walls to avoid downgrading to a random teleport is spoilery and not intuitive.
2. The ring is better than the spell (swap time = casting time, no mp cost, no skill cost) and only costs an inventory slot.
3. Semicontrolled Blink
   • Semicontrolled blink is very strong in many situations, arguably far too strong for a level 2 spell.
   • The targeting interface is spoilery and bad.

One solution to all of these problems is simply removing controlled teleports (spell, ring, mutation). Several people are strongly in favor of it and can write here why. I think it is an interesting mechanic and worth keeping, and don't see the advantages of removing it, so I'm not exactly qualified to write this section.

This would require changes to a few vaults (not many), and the entrance to the Elven Halls would have to be guaranteed to be reachable on foot (which should be done anyway).

It would be the simplest solution, and outside of Orc issues the only big downside here would be scumming repeated teleports if you want to get into a particular area. This can reward persistence in a bad way, especially if you have no means to avoid miasma on Swamp:5 or there's an enclosed vault on the level. This might be mostly alleviated by a large-scale vault review. — KiloByte 2012-11-21 19:17

There is another source of controlled teleports: cards.
Even if we agree to remove ring, spell and mutation (I agree that it is worthwhile to look for appropriate nerfs before), I think cTele will make an interesting consumable, so we might come up with a rare potion/scroll for that effect. — dpeg 2012-11-26 18:25

The problems can be solved in a less drastic manner, but this would require quite a few changes to address all of the issues. There are a lot of suggestions, feel free to add yours. Some combination of several of these would be needed.

An example of a set that might work: 19:37:49 <evilmike> anyway, what I'd do to cTele: add a 5-15 extra delay to teleports, make the ring an evokable item, and make the cTele status end after teleporting. There are other ideas I like here too, of course (such as a proper targeting prompt)

Slightly increased delay (for example 5-15 aut) for controlled teleports. The delay would be increased if the cTele effect is active at any point during the teleport delay. With a controlled teleport you are much safer after the teleport, but you have to spend more time in the situation you want to escape from.

Bigger fuzz. If controlled teleports were less controlled, they would also be less boring and less guaranteed to take you somewhere safe.

Unspoilery targeting: instead of downgrading to a random teleport if the target location is inside a wall after fuzzing, one of the x closest valid target locations could be chosen randomly and the player placed there. This could be weighted by distance from the target. Also there could be a flat chance to downgrade to a random teleport (~10-20%). The current targeting has the advantage that certain locations (for example runes) can be made hard to cTele into by placing a lot of occupied space near them. So this change might require reviewing some rune vaults, because it might make it too easy to reliably teleport very close to some runes.

(Controlled) teleports could sometimes take adjacent monsters along with you (something like 10-20% chance).

The ring could be made evokable, or be removed. The required evocations level for reliable evocation is to be determined. Evilmike suggested “very small”. The mutation could get a similar treatment, granting an ability instead of permanent cTele.

Instead the ring could work similar to Gourmand amulets, only giving cTele after being equipped for some time. Or a smooth version: give an increasing chance to control teleports depending on how long it has been equipped.

Or temporarily suppressed after a controlled teleport. The -cTele status light would fire up in yellow, getting brown as it expires. This would avoid problems with cheap chained blinks. — KiloByte 2012-11-21 19:17

This is mentioned under “General Changes” — Galefury 2012-11-22 00:52

Some of the suggestions in the “General Changes” section would make semicontrolled blink from blink a lot weaker. And of course removal is always an option. It is called “cTele” after all, not “cTloc”.

Semicontrolled blink could be changed to use something similar to cloud targeting instead of the current direction-based targeting. Some kind of floodfilly fuzz. If not enough places are reachable from the target location, the current player location could be used as a point of origin for additional target locations, or some could be randomly distributed.

Semicontrolled blink should have some sort of targeting preview that shows you where you might land.

These are not specifically targeted at any of these problems, but might be good regardless.

The controlled teleport status could end after a single use (teleport or blink). Glow already does that to some extent. This would make semicontrolled blinks from blink a lot weaker. Spellpower could affect the chance for cTele to end on teleporting. This only works if there are no permanent sources of cTele (current ring and mutation).

Instead there could be a cooldown preventing you from controlling another translocation after any controlled translocation. Spellpower could affect cooldown length for the spell, power levels for the ring and mutation would be low. This would work even with permanent sources of cTele.

I agree with the analysis, and I think that it's worthwhile to try to keep cTele. Unchaining (as detailed by kilobyte and having controlled teleports kick in slower are good approaches: assuming that the latter has a message (Your controlled teleport takes longer to activate. whenever there is additional delay) means that the change is transparent. Also, these seem simple enough — not too much effort invested in case we decide to cut cTele anyway. — dpeg 2012-11-22 01:55