From NetHackWiki
Jump to navigation Jump to search

Stack shock by dipping inventory?

I got a message about 300 rocks shuddering when I dipped them into a potion of polymorph. Is the following really true?

"Also note that if you polymorph an object in your inventory by dipping it in a potion of polymorph, it won't get a system shock, so if you have enough potions, it's safe to try turning a pile of 300 rocks into diamonds, though if you quaff it or throw it at a monster, system shock will still apply."

Tjr 03:13, 19 May 2009 (UTC)

wizards and books

", unless you are Wizard, in which case you are very likely to succeed in writing a book of your choice"

I prefer to first polypile duplicate / useless books, then write the remaining spells (possibly on the books blanked by polymorph). Very early in the game, when you do not yet have many books to polypile, you most likely also don't have the necessary luck to write ones. Anyway, this is strategy advice and belongs in that section.Tjr 12:53, 20 May 2009 (UTC)

  • I believe it is strictly correct for a wizard generated with a spellbook of cure blindness, a potion of polymorph and no ring of polymorph control to immediately polypile the spellbook; by the time cure blindness needs to be relearned, it will probably no longer be necessary; in the meantime, should the resulting spellbook be low-level, the player knows an extra spell - 16:28, October 15, 2010 (UTC)

alternating material piles

[rgrn thread] -- is it really true golems will be generated from the top? IIRC, that was randomized. If not, it would be an efficient (and tedious) way to protect you rings etc. -Tjr 14:59, 3 June 2009 (UTC)

Yes and no. Items are polymorphed from the top (the most recently dropped item) down. Once an item stack shudders and produces a golem, though more item stacks may shudder, nethack will NOT check for golem creation again. The first shuddering stack to produce a golem picks the type of the golem, and more importantly, the material the golem will consume. If you have a bunch of mithril items on top, and they shutter resulting in an iron golem (and since they are on top, they are likely to be the one's to pick the golem), then mithril items will be absorbed. Leather, iron (yes, iron), copper, stone, etc. will be safe. So the top materials pick which other materials are safe from absorption.
OTOH, golems absorb SO MANY items, that putting your stuff on the bottom will not keep it safe if it's made of what the golem wants to eat. (They consume their weight in items, forming iron golems attempt to consume 2000 items of whatever material spawned them, be that iron, mithril, or generic metal.) Blackcustard 20:06, 17 July 2011 (UTC)

polypiling artifacts

Is it really true artifacts cannot be polymorphed? I thought they just have a much higher chance of resisting, and will never turn into another artifact. -Tjr 10:59, February 4, 2010 (UTC)

I just tried it in wizmode, and it looks like you're right. After zapping my pile fore more than 100 times, the artifacts eventually polymorph. The wand, potion and spellbook of polymorph, on the other hand seem to be completely immune to polymorph. Galehar 15:19, February 4, 2010 (UTC)

Strategy section

The strategy section reads like a list. Which is fine, or at least it would be if it were actually a list. I was thinking of adding, later, a bunch of sections to break them off into like "When to polypile," "When not to polypile," "General tips," etc, so that it's far easier to see what's relevant to a player's situation. Feagradze 13:50, June 29, 2010 (UTC)

Feel free to flesh out a separate list "the perfect polypiling" or some such. Your plans look very reasonable. (I'm the original author.) Tjr 23:51, June 29, 2010 (UTC)
  • The rambler zaps a wand of nested formatting +4!
    Done. Can't help but think it's missing something, though. And I don't know what... « Feagradze 06:07, July 1, 2010 (UTC) »
  • Speak of the devil, though. I was hoping to eventually make a complicated spreadsheet that forks out the ideal way to set up a line, as per a few constraints (namely, number of wand charges, though also a sanity switch for how much time they feel like spending). Something like Clippy for polypiling, albeit in desktop form. Currently, that's beyond my ability. Conversely, not for long. More limited by finding the notes for such in my god-awfully huge un-indexed pile of notes. With luck, expect an external link on this page to that eventually. « Feagradze 06:51, July 1, 2010 (UTC) »

polypiling spellbooks

The article is flat wrong; polypiling spellbooks is an amazing way to learn dozens of new spells for nearly free. 23:32, 23 November 2010 (UTC)

Fixed. BTW, you can edit yourself, you don't need anybody's permission. --Tjr 01:27, 24 November 2010 (UTC)


The page states that upgrading an object will "fix" hazy, meaning it wont revert back. Either I dont understand this correctly or the info is wrong. I polypiled and got a chest, which I tinkered into an ice box. Then I stored my stash in it, left, and came back to a single unicorn horn. Just a friendly warning, upgrading has no effect on an object reverting back.

Looks like whoever added that misread the opening paragraph on Upgrading objects - it says that upgraded objects do not need to be fixed. That is true, but upgraded objects don't lose the "hazy" status, they just don't gain it. Why would you polypile for a chest, by the way? They're quite common. -Ion frigate 13:45, 4 December 2010 (UTC)

how much polyfodder

Somebody should add a section how large the piles should be. Say a table of expected number of morphs per item if the pile is x large, or such. --Tjr 03:17, 30 July 2011 (UTC)

Item generation - athames

If I'm reading the linked spoiler correctly, polmorphing will never generate items that aren't randomly generated, such as Athames. Can anyone confirm that? Maybe a note should be added to the section about when not to polypile. 02:27, 30 September 2011 (UTC)

True. --Tjr 03:05, 30 September 2011 (UTC)

golem type

coud there be a list on the page of which golems get created by what? paper=paper and iron=iron, sure, but corpses/gems/wood/gold? and in SLASH'EM, will polypiling candles generate a wax golem? -- 13:02, 9 December 2011 (UTC)

Material Golem Material description
Iron metal
Clay or stone (equal chance) lithic
Food ("Fleshy" or not) Flesh organic
Wax Wax wax
Wood Wood wood
Leather Leather leather
Cloth Rope cloth
Plastic Plastic plastic
Bone Skeleton bony
Gold Gold gold
Glass Glass glassy
Paper Paper paper
"if all else fails" Straw

That's everything from SLASH'EM. Source: zap.c in SLASH'EM 0.0.7E7F2, line 1186. It could probably use some cleanup before being placed on the main page. -- Qazmlpok 13:29, 9 December 2011 (UTC)


Someone changed the range from 3-6 to 4-7. I believe that's half right. The range is 3-7 piles.

bhit is called with range rn1(8, 6) which is 6 to 13 inclusive. ( ,

#include <iostream>

void test(signed int range) {
    while (range-- > 0) {
        std::cout << range << std::endl;

int main() {
    std::cout << "Go!" << std::endl;
    std::cout << 6 << ":" << std::endl;
    std::cout << 13 << ":" << std::endl;
    char c = '\0';
    std::cin >> c;
    return 0;


That also agrees with wiztesting. The wand can definitly hit 7 piles; it's harder to be sure with this one, but I'm pretty sure it's possible for the wand to also hit only 3 piles (it's always possible it hit a 4th but had no effect).

Blackcustard 15:14, 27 December 2011 (UTC)

One Large Pile

This technique can be improved if you know the Stone to Flesh spell, so that you can turn the Rocks to Meatballs:

  1. You can pick up a lot more meatballs at a time than rocks, so that you can curse them with a single Potion of unholy water.
  2. There are fewer desirable Flesh items than Mineral items, so instead of the polymorph consuming stone rings, wands, and Luckstones, it consumes Eggs and Tripe rations, which are less valuable.
  3. Flesh golem corpses convey great Intrinsics. --Elestan (talk) 18:34, 25 November 2012 (UTC)
Another Idea: As a warrior-role, drop Spellbooks on top. The arising Paper Golem will drop scrolls of blank paper, that can be used to write scrolls (if you have a magic marker). This is usefull because casting spells from books requires good physical abilities (INT,WIS,POW) and non-metal armour, while casting spells from scrolls will always suceed (exept you are confused, blind or the scroll is cursed). If you play a Barbarian, Caveman, Tourist or Walkyrie, you will have to eat a lot of Mindflayers to be able to learn a spell at all, so converting spellbooks to scrolls is a considerable idea. -- 15:30, 1 April 2013 (UTC)
I stumbled upon this situation by accident in the late midgame, while doing some Altar work. I got tired of waiting for large monsters to finish sorting through the piles of dropped gear (one Ice Troll tried on three helmets and three different shields) before they finally walked over to attack (hence standing upon the altar on which I had planned their sacrifice). So I decided to take advantage of golem generation and shuddering to eliminate as much of the junkpile as possible (I wasn't trying to obtain anything in particular from the loot). Thus, I used the One Large Pile technique without having read your note, above. The bit about what you put on top is interesting when the pile is large (25 or more items). Putting stones on top always generated a stone golem (when any golem was generated at all). Also the bit in the article about luck is really significant. My luck was maximized at that point, so I obtained very few golems, which was frustrating. Stupidly forgetting that golems do not die a Nethack-normal death, I had hoped for golems to wander over the altar to become the next sacrifice. When I finally did obtain a golem or two, they were *always* stone golems, which--like all non-flesh golems--do not leave a proper corpse (leaving statues instead, which led to more rocks--after my pickaxe--and more stone golems, until I read this spoiler and stopped putting the rocks on top). Bottom line: max luck = min golems, but doesn't really make much diff in the shuddering (which has more to do with item type, count and B/U/C) and therefore loss of items. (Once I gave up on the golems and stopped putting rocks on top, I reduced the large pile to five items or so in about eight or ten attempts, and ended up with a nice Hawaiian Shirt for my trouble). This indicates a useful time-minimizing technique for the huge piles you get from Fort Ludios. --Clothos

Behavior of the Golem

Are Golems produced by polypiling tame or hostile ? Since the player is their creator it would make sense beeing tame, but a golem is considered to be something bad so haow doe they behave ? - 17:58, 11 March 2013 (UTC)

About that n/1000 chance of merging

I see it in the code, but experimenting with a stack of 38 blessed arrows, splitting them off into separate, named stacks, I found that three out of four stacks in the pile would merge, whether the stacks I dropped were of quantity 4 or 3. This remained the case until I used up all the arrows. So for four stacks of three arrows, in this particular session, over several attempts, I saw only one stack remain un-merged, on each attempt. Alas, I did not keep copious notes on the tries.

Does the compiled version use a different denominator than 1000? Because this series of outcomes makes that number look doubtful. (Yes, I do know the odds are simply odds, and not actual likelihood, nor does one instance influence succeeding instances.) --Clothos

Wait, that can't be right. So I continued the experiment with three more big stacks of items (spears, daggers, and darts). It didn't seem to matter if I used stacks of 2, 3, or 4, they almost always merged. It looks more like there's a n/1000 chance the stack WON'T merge. --Clothos
Take a look at zap.c, line 1174 (and how it's called from zap.c, line 1514). --Tjr (talk) 07:35, 12 May 2013 (UTC)
I saw that when looking at the code earlier. This separate accounting for items (in those lines you reference) appears to only happen when it's been determined that system shock happened. Or am I incorrect?
In gameplay, I saw the merging behavior I'm referring to when no "shudder" message occurred, and (with the exception of merging within stacks), the object count seemed to remain the same (suggesting to me that system shock code was not triggered). In fact, I continued playing with it by throwing everything in Ludios into one huge pile (14 pages worth of items) and zapping away, merely out of curiosity (my high luck kept golems from forming around 95% of the time). Though I did get shuddering (expected in such a huge pile), the more significant factor is that within two zaps, every single stack of knives, spears, daggers and stilettos had merged into one item. (>>> B/U/C was unknown, of course, which could account for all of it <<<) I may be misreading the code, but it looks like each individual stack in the pile should get the same chance to remain unmerged, so having them all disappear in two passes seems unlikely--unless the actual chance for a merge is much higher than reported in this document. Could be my real-world, "RNG-luck" is vastly different than my character's status-luck. But it sure looks like stacks will merge to one or two items the vast majority of the times they are zapped with the wand of polymoprh.
The bottom line is that this (apparent) limitation (again, I could be misreading the code), coupled with the no-more-than-four stacking guidelines, makes it incredibly hard to get a stack of silver projectiles (daggers, arrows, spears) through polypiling. If the odds of merging were really n/1000, it should always be possible to have a useful quantity (ie, > 3) of silver ranged weapons after some work at Ludios ("some" meaning, "you shouldn't have to polypile the entire floor's huge complement of weapons into dust" to get three or more rare, non-artifact weapons). But hey, if that's the way it works, then play on. But at least we'll know.
For anyone following this, polypiling definitely stuck fairly close to the percentage of generation likelihood per item as detailed in the weapons page. E.G., over a massive session of dangerous, experimental polypiling at Ludios (time wasting outside of Wizard Mode isn't recommended), I saw more silver arrows than tridents, more tridents than silver sabers, and more silver sabers than silver daggers. (No lances this time, though. In another game, with a much more typical bit of polypiling of about ten weapons, I ended up seeing four lances I couldn't use. RNG, baby!) --Clothos
I just thought of something that's hard to tell from the code: I'm operating under the assumption that stackables always poly into other (or the same kind of) stackables unless merged. If stacks turn into random collections, then it would seem that the entire business of worrying over "stacks" in polypiles is far too esoteric to be worth worrying over. I.E., you're better off just unstacking everything via naming before zapping with polymorph. --Clothos
Right, stackable weapons and tools don't usually turn into stackables. In that case, the stack size is set to one (see zap.c, line 1234, zap.c, line 1312, objclass.h, line 15). So it's a good idea to unstack those before you polypile (and after you bless, rustproof, overenchant them).
But scrolls, food, potions, stones always stack stackable. --Tjr (talk) 06:31, 13 May 2013 (UTC)
Thanks! That clears up a lot! Quite a few missed/misused opportunities have come from not understanding that. --Clothos

simple version for beginners

The topic is complex, but maybe we can add a dumbed-down lead paragraph for beginners. How about this:

"Polypiling is like recycling. If you polymorph a pile of junk objects, they turn into other objects. But you cannot control into what other objects, so you have to repeat until you get something useful. For example, unicorn horns can become magic markers. For game balance reasons, a share of the input objects will vanish without trace. The optimal strategy depends on what you have (how much junk), what you need (what type), and how many polymorph shots you can afford. Polypiling takes patience to do and patience to learn its complexities." --Tjr (talk) 11:48, 30 May 2013 (UTC)

quibbles with one big pile

a footrice corpse is always useful if you're standing still, agreed. why is it notably useful if you're trying for stone golems?

if you're "going for optimality" (aka minimaxing) why wouldn't you polypile meatballs?

  1. rocks are plentiful enough that you wouldn't really need the stone golem drops
  2. the resulting flesh golems are, at least, useful for intrinsics
  3. you'd not be polypiling food items so you wouldn't lose anything through absorption -- 12:32, 18 September 2013 (UTC)
rule 1: RTFP. -- 12:35, 18 September 2013 (UTC)
The answer is that you may not have the stone to flesh spell to make meatballs. -- 08:36, 19 September 2013 (UTC)

Probability of shuddering for cursed stacks of >4

If I'm reading the code correctly, cursed stack of 5 items or more will always shudder unless it resists (with 5% probability), because 3/2 == 1 in C.

Gold golem

Since for the "one pile" strategy there is always at least a 5% chance of the stack resisting, it would be helpful to mention that dropping gold can help minimizing golem losses as well - this works well for either pile strategy as you have some gold left over (it's mentioned on the zorkmid page but not here) -PQ (talk) 14:56, 21 February 2018 (UTC)