The Tavern

Polymorph is probably the most spoiler-heavy thing remaining in crawl. Learning it from experience does not work that well, there are multiple results of polymorphing the same monster and you can be lucky in your first game but unlucky in second and die.

I suggest to make polymorph result in any monster generated in current level.
So if you want to polymorph an Orc Priest on D3, you will get any monster who has been generated on D3, even if D3 has been almost cleared and the Orc Priest is last remaining monster on the level. Assuming D3 generated Orc, Orc Wizard, Adder, Ogre, Rat, Bat, Worm you will get one of those monsters with equal chances.

Pros
1) No need for spoilers, no need to even know that HD exists. You basically know what to expect assuming it is not your first monster found on the first floor of a new branch.
2) More "oh, sh.t" and "that's nice" moments. You didn't like Orc Sorcerer (HD 9) and polymorphed it? Get a Stone Giant (HD 16). You didn't like the Stone Giant and polymorphed it? Get a plain Orc (HD 1)
3) New tactics of luring a dangerous monsters to another floor before polymorphing

Cons
1) No more Titan/Acid Blob from Hydra etc.
2) Some low MR high HD monsters can become trivial. I believe it is not a big problem since you already can trivialize them via confuse/paralyze/hibernate/enslave.
3) Number of polymorph zaps/wands should probably be decreased for balance reasons, players are not going to use them on easy targets

Encouraging luring to another floor as well as keeping track of a floor's monsters sounds really annoying, I would much rather have the polymorph pool capped to some small, fixed list and have that information transmitted to the player.

Yes, that would be much better than current situation too.

Luring dangerous monster is not trivial, it has to be adjacent for some time after all and then your polymorph can fail multiple times in a row.

just always turn the target into a frog

with a 1/20 chance of it going wrong and turning into a (very)ugly thing

Agree with polymorph being a problem for essentially the reasons laid out in OP.



Agreed. One possibility would be to make polymorph of enemies work more or less as suggested in OP, except based off of the floor on which the target was generated, not the floor the target currently occupies.

(NB: I have no idea if Crawl currently keeps track of this information (floor on which enemies generated), nor how hard it would be to change, if it does not do so.)

EDIT: Some portals would have very small number of polymorph "destinations"; I think it would make sense for ice caves, wizlabs, labyrinths, etc. etc. to be special-cased so as to generate enemies based on the floor on which the portal generated. Also note this version of polymorph would be more useful in some branches than others, kind of like ?summoning or shadow creatures, but I think that's ok.

Nah but seriously the frog idea sounds good to me.

The problem with these suggestions is that they don't solve the original problem at all. First of all, just communicating it is nontrivial, and secondly, once you do, it still hasn't solved the spoiler problem because unspoiled players don't know monster generation rules.

The core problem is not really polymorph, but HD. HD affects a wide range of important monster properties, from accuracy to damage to hex power to resistance to mephitic clouds to polymorph targets, and is completely hidden from players. You can get rid of polymorph but you still have the same problem for everything else HD affects, which is a lot of things (probably more since that list was made).

Polymorph effects present a lot of issues, but for now, if you want to make monster polymorph safer, you can do it by basing it off of XP value instead of HD. XP is still a bad proxy for monster danger, but it's a better one than HD, and this is an easy change to make from a technical perspective. (RIP polymorphing uniques, though)

Yeah, polymorph should turn the monster into a random other monster worth between 70% and 99% of the xp of the original monster. And, yes, you would be able to use it multiple times to eventually turn the monster into a bat, though you'd only get bat xp for it.

How about just remove polymorph as a player-accessible ability? Maybe there's a niche as a monster spell, or better yet sequestered off in the Xom ghetto, but there's no job that a polymorph wand can do in player hands that isn't done as well or better by the other hex wands, which are mechanically similar. It's nice to make a humanoid monster stop using an unpleasant item, for instance, but confusion or paralysis would work just as well for that without as many clunky and obscure mechanics involved.

Polymorph is too much fun to be removed imo. How about displaying the monster HD in xv and explaining how polymorph works in the wand description?

Remove poly wand, add a God who has poly other as one active power.
Wait, remove also Transmutation school and add the best of that shit to that God too.

100% coherent to DCSS development plan and probably a good solution too
(this post is 50% serious)

It won't help IMHO. How is player supposed to know monsters with HD 9, for example? By checking all monsters in "?/m"?

Polymorph should turn an enemy into a Mutant Beast of the tier closest to its original HD. Give these beasts immunity to further polymorphs so you can't just spam it until you get one that matches your resists; but maybe the effect times out after a while.

At that point I'd be more inclined just to remove polymorph wand (could leave it on chaos brand and Xom). What is fun about pmorph is that you sometimes change things into monsters that are more dangerous — it is a risk.

The problem is that, while you can naturally (non-spoiler, just playing game) get a sense for good polymorph candidates — low enough MR, but dangerous ability — the level of risk is based on what other things they could polymorph into, and that information is very spoilery. So you can get a good sense of the benefit of polymorphing something, without resorting to spoilers, but not of the risk involved. Without spoilers you cannot take reasonable calculated risks with polymorph wand.

You can have it turn the enemy into a super-weak enemy whenever it succeeds, or what have you, but at that point it becomes "just another hex wand." We already have plenty of wands that wreck low-MR targets. What is fun about polymorph (potentially very strong but also risky) is directly related to its design problems.

Two ways to approach that issue. Either keep the "calculated risk" element (which differentiates polymorph from wand of paralysis), but somehow make those risks reasonably legible to the player in-game. Or remove that element, at which point (imo) polymorph is functionally indistinct from other hex wands and should just be removed. (I thought that changing polymorph so that it was no more spoiler-ish than shadow creatures was sufficient, as did the OP, but perhaps that is not the case.)

If I have time to try polymorphing the low MR monster into something unknown, it is usually not a good idea because I can get a more dangerous monster with high MR and ranged attack and die. It's a gamble where I can win little (I might escape instead of trying to polymorph) but can lose much.

Edit. I agree with the rest of your post.

Right — what I meant there by "candidate" was, "candidates that you might want to polymorph." In order to know whether it is actually a good idea to use polymorph, however, you have to know what it can morph into, and at what (at least roughly) likelihood. That part is spoilery, so very careful play recommends against using polymorph, unless it is a last resort, or unless you know a lot about how polymorph works. (Way more than I do, at least.)


Thinking about it a bit more, here might be a way around the problem, while still keeping what is cool about polymorph. (I understand the desire to keep it, by the way, as it is at least a distinctive form of attack compared to other MR-checking actions.) This is basically a version of what bananaken suggested up-thread.

Make some lists based on "type" and HD range.

Weak humanoids
Strong humanoids

Weak beasts
Strong beasts

Weak holy
Strong holy

Weak demon
Strong demon

Weak unholy non-demon
Strong unholy non-demon

[etc.]

Each category gets a short list of possible polymorph results. Each result has an equal chance if the polymorph beats the target's MR check. Display the list in xv. Each list of results should have a few really weak (i.e., highly beneficial) results for that category, mostly medium ones, and at least one strong one (the "risky" element). Lists should be short but not so short that the "risky" polymorph result makes it 100% a bad move.

Drastically reduces spoiler problem, keeps the flavor and "neato" factor, and retains an element of calculated risk.

NB: This could also be applied to chaos and Xom, which would be nice for consistency, while also perhaps helping to fine tune those features away from BS deaths...

EDIT — That is probably a lot of coding legwork, and therefore might not be worth it. If that is indeed the case, then I think polymorph wand should simply be removed.

vv I like that, too. "Easier to apply, but riskier" is a good niche. Yes, in that case I agree player should only have one bite at the apple.

If we want to keep polymorph, main polymorph problem is that it is always inferior to paralysis/confusion. Obvious solution is to make polymorph chance higher, much higher than chance for wands of paralysis/confusion.

Maybe even ignore MR completely (except MR immune and probably uniques) but allow polymophing a monster just once. You polymorph Foo into Bar and you no longer can polymorpth the Bar further. But we still need to solve problem of spoilers, player should be able to see possible Bar outcomes for every Foo in the game, without out-of-game resources or source diving.

Honestly, I think this is really close to right. In my opinion the ideal fix is for polymorph to always turn the target into one of the following, with the following probabilities (ballpark):

• 40% rat
• 40% giant frog
• 10% randomly selected from every monster in the game with the same holiness
• 10% randomly selected from every monster in the game with the same holiness, but very heavily biased toward things than can kill you

So in the early game polymorph would be really risky; as the game progressed it would become a respectable emergency option but retain a potential for hilarity. And there'd be no need for spoilers.

After reading the topic line, I somewhat expected the proposal to essentially be "restrict polymorph results to monster types you have already seen in the current game". Which seems kind of strange, but could maybe work if you considered all monsters on levels you've visited to qualify.