Spawning Algorithm Broken - Mobs Don't Do Block Checks

In 1.11.0.8 Beta I had a wandering trader spawn on the wood slab roof of my base.

It also seems as though the mob cap is off, at least on realms. There are far more mobs, hostile and passive than there should be, especially since our realm has a lot of mob farms with members who are on all over the main area. I’ve never seen so many horses, sheep, pigs, skeletons, creepers, etc.

@unknown: That's most likely the "donut effect", which isn't related to this issue. When you spend a long time in the same small area, mobs spawn in a 'ring' around you (between 24 blocks and the edge of the player ticking area, which is determined by your Simulation Distance setting). Like all mobs, they move around randomly, but when they try to walk out of the ticking area they become suspended in the chunk they're entering and are no longer counted toward the mob cap. That leaves room for more mobs to spawn behind them, and some of those will likewise wander into the unticked chunks. Over time, a lot of mobs will accumulate in chunks just outside the ticking area this way. Later, when you move toward that direction, the outside chunks become ticked. Suddenly a whole bunch of suspended mobs become reactivated right at the chunk boundary, possibly exceeding mob caps. By the time you get near them they've spread out a little but haven't wandered very far, so you see clusters of them.

You can recognize that this is what is happening when there are too many mobs of all different types (hostile and passive) clustered at your simulation distance away from a chunk you spent a long time in. It's especially common outside the perimeter of a fully mob-proofed area, because the only places mobs can spawn are closer to the unticked chunks, so they're more likely to wander into them. You can reduce the donut effect by setting your simulation distance smaller; if it's set to less than the perimeter radius, you can even eliminate the donut effect entirely for that area.

The developers are aware of the donut effect and are always looking for ways to make it work better.

Updated this report to include all the 1.11 updates for bedrock.

@unknown: Updates to the affected versions automatically send an update notification. Separately commenting that you did so is redundant.

Is this issue fixed in the latest 1.11 version?

For some mobs, yes. Need to do additional testing to get a full list of whats broken🙂
Off the top of my head, i know llames can spawn under ground on stone. But thats just what i have randomly noticed.

I had a bear spawning on a glass block or a hopper yesterday

This is a major issue now that mob events are now a thing. Patrols have been reported to begin to spawn in caves and end up in Player's mob grinders which I know that is 100% not intentional and forces you to find a Pillager Outpost instead of random encounters with the Patrols. I think the best way to fix it, is for the patrols to always spawn in loaded chunks at the same y-level as the player. This could reduce the chances of the Patrols getting stuck on something like trees.

You raise a fair point, things like wondering traders and pillager patrols don't really do block checks either.

The patrols should purely be a surface spawn, that would solve the underground issue.
As for the block checks, stuff needs fixing😛

Zombies and pillager patrols do have an odd spawning behaviour.

They will spawn in mob grinder if some certain blocks are on top of the roof; stairs and bottom slabs.
Without it as in roof made out of regular blocks not covered by stairs or slabs, they will not spawn in there. All other hostile; witch, skeleton, creeper, and spider spawn normally.

Raid can spawn on a 1 deep water as big as 8 chunk radius. I don't mind but that's up to the community and developer. Raid won't work in the ocean biome/ 2 or deeper water. They can spawn on waterlogged bottom slabs.

@unknown: Are you sure you meant zombies and pillager patrols? I'm pretty sure zombies will spawn regardless of how the roof is constructed.

On the other hand, it sounds like patrols are working as they should. In Bedrock, the "surface" has different definitions for different purposes. For hostile mob spawns, the surface is normally the topmost solid opaque block, but if the topmost (not necessarily solid opaque) block (excepting glass and leaves) is a slab or stair, an exception is made: It and any number of consecutive solid opaque blocks below it are ignored for the purpose of deciding where the surface is. I imagine that this exception is made to allow surface spawns under roofs, which people usually make out of slabs or stairs to prevent hostile mobs spawning. Roofs are normally above ground level, so people intuitively understand that the "surface" is the ground below them, and this exception agrees with that intuition.

It might also have been intended as a way to improve mob grinders. To be dark inside, you have to make its roof out of solid opaque blocks, but then you have to add slabs or stairs above them so mobs won't spawn on the roof at night. Without this exception the top spawning floor would only qualify for cave spawns. By making the exception, it's possible to have surface spawns on the top floor and cave spawns on lower floors, which could mean more mobs will spawn inside.

The exception does have a big flaw, though: It doesn't matter how many consecutive solid opaque blocks below a slab or stair there are, they're all ignored when looking for the surface. I have seen cases where it was all solid blocks all the way down to lava level, so "surface" spawns were happening in extremely deep caves. See MCPE-41273 for a more thorough description.

In your case, your roof qualifies the top spawning floor in your grinder for surface spawns, which is why patrols can spawn inside (but not on any lower spawning floors if you have them).

People keep saying that patrols are spawning in caves, but I'm not convinced. I think patrols are limited to surface spawns, and what people are finding are patrols that spawned with a cave mouth between them and the village. They're not smart enough to understand that they can't reach it through the cave, so they merrily just walk inside, and then they get stuck there because getting out again would require them to walk away from the village, which their AI won't let them do.

9Auldrick

Yes, if zombies can't spawn in there, pillager patrols can't also.

I built 3 towers with each 4 spawning floors. 1 of the tower has its roof(opaque) covered with slabs allow zombie and patrol to spawn in there.

Yes, I did noticed that patrols spawned at the topmost spawning floor oftentimes and the zombies were spawning on all floors. I had night vision and watched the hole where they fall off from.

Very few patrols did spawned bottommost and middle.

Edit: In additional to patrol spawning on topmost floor where I often found was the captain with its banner sticking out the roof wandering constantly on one side against the wall as if it is stuck in pathfinding.

Each spawning floor do not have access to skyblock, all sides are covered with only one exiting hole in the middle bottom tower. Extended wall down to darken bottommost spawn floor.

No village within.

2 towers with both having opaque roof without cover never spawn 1 zombie or 1 pillager after having 100+ patrols in the other tower. 🙂

To clarify: 100+ patrols doesn't necessary mean I have witnessed all which floor they fell off. Banners were collected and sent to chest via hoppers. Each tower has its own item collection.

@unknown: To sum up, this ticket is about mobs spawning on blocks they shouldn't spawn on. How what's above them affects spawning isn't part of it, and belongs in another ticket if there's an issue there. Please search for a report that describes your issue, or create one if you can't find one.

...or they should have spawn underneath regardless what is sitting on the opaque roof. Block check might have other flaw. I will leave my comments above as additional information.

All kinds except stray don't spawn on snow biome. They still can spawn on ice on frozen river, pond, and lake. They can still spawn on patch of grass but very rare due to an abundance of snow that makes stray the dominance.

I doesn't know if it is a real bug, but llamas in savannah are spawning in cobblestone blocks

When updating a report, please try to do as much updating as possible in one transaction. Each update sends notifications to every user watching.

Oof, was afraid it might have done that 😬 Sorry for the spam all!

Can confirm in 1.11.4 | Wandering traders spawn in a cave and leave there llamas behind when they despawn.

Can confirm on 1.11.4 | Cats spawn in beds in my iron farm

Can confirm on beta 1.12.0.12: I have just experienced several llamas spawn on a platform with stone brick blocks (savanna biome).

The issues originally reported were fixed in the 1.11.0.4 Beta and the 1.11.0 release, and the ticket should be been closed at that time. Consequently, I have closing it now.

@unknown: You've made comments indicating that all the originally reported issues were fixed, but then you continued to update the description with additional issues. The only element they have in common is that they relate to mob spawning not working as you expect it to. That's not enough similarity to suggest that they're the same bug, especially since fixing the original issues didn't fix them. When you find new issues, please report them separately. Once we've forwarded an issue to the developers, it becomes a bug management problem to add anything to it unless it's very closely related in terms of game mechanics.

.... Really auldrick? This report has to do with mobs not spawning on the correct blocks, both passive and hostile. This was an issue, and is still and issue, passive mobs, and hostile mobs, spawn on things they should not. Just because it different mobs throughout the updates, doesn't mean its not the same underlying issue.

I even put this in the report, stating it was fixed, but is now back again..
"Most of this bug was fixed in 1.10, passive animals needed grass to spawn, drowns needed water, and so did guardians. Now in 1.11 passive animal spawning is broken again. But the drown/guardian fix persists."

So no, all original issues are not fixed, and you just closed an important report improperly. Have a great day

From the history, here is your original description:

What this bug basically allows for, is passive mobs to spawn on none grass blocks, such as stone, concrete, beacons, iron blocks etc etc. The only thing the game is really taking into account for where to spawn things, is biome. 

So for instance, in super flat river world, with absolutely no water at all, you can have drowns spawn on land (as long as its part of the river biome). As another example, guardians can spawn in air. As you can see from these couple examples, the block check mobs should be doing is clearly missing in recent updates. 

As a side note on drown spawns, they are the ONLY hostile thing that will spawn in river biomes, water or not. Just an interesting thing i noticed while doing some testing:)

You originally reported three issues:

1. Passive mobs spawning on non-grass blocks,

2. Drowned spawning on land, and

3. Guardians spawning in air.

The next day, MegaSpud entered a comment explaining that #1 and #3 were duplicates of other tickets (MCPE-39727, which is still open, and MCPE-39941, now resolved), and that they were therefore being separated from this report. He also stated that #2 would be partially fixed in the next Beta release. At this point, #2 was the only issue being tracked on this ticket.

On 29 March you reported (mistakenly) that 1.10 fixed the passive mob spawning. You also reported abnormal spawning on slabs and at high light levels on Realms, but that wasn't related to this issue and was ignored. (People mention irrelevant stuff all the time, just trying to be helpful. We don't want to discourage that, because sometimes it is helpful, so if something seems irrelevant we just disregard it.)

On 23 April the fix for the drowned issue was released in 1.11.0. At that time this ticket should have been closed, but it was overlooked.

Yesterday, nearly 2 months after the drowned issue was fixed, you completely changed the description, in effect erasing all the issues you originally reported and replacing them with several new issues that have nothing to do with the drowned issue. You're attempting to justify this by lumping everything together in a generic "mobs spawn on the wrong block" category. But bugs are not generic beasts; the developers can't do anything without specific information that lets them reproduce the issue, so if you change the specifics, chances are you're talking about a different bug. And besides that, the drowned bug was already fixed and nobody, including you, was challenging that fact, so this ticket was no longer active.

Furthermore, the new issues you're trying to report are additional duplicates of MCPE-39727. If they offer new information, it should be added there, not here.

As a side note, drowned are not hostile mobs. They (and guardians and elder guardians) are water mobs with hostile behaviors, but the two have different spawn rules.

I get your point, everything you have stated is true and fair, and thank you for bringing it all up. However, this report isn't for any one specific mob or creature, the mobs listed are simply a description/example of the bug at hand. Mobs are spawning on things they should not, in places they should not. If this report is being tracked as otherwise, for just one specific issue, that fine. Should i make another report more or less the same as this one, to have it tracked correctly? Help me out here. This is a big issue, that the community clearly cares a lot about.

 @@unknown Different mobs have different spawning conditions, and even if they all have a similar issue, they don't necessarily share the same code for spawning and thus should be tracked with separate tickets.

It's possible that a new ticket would be the best idea, since MCPE-39727 seems to have gone a bit stale, but it depends on exactly how MegaSpud reported it to the devs and what they're doing with it these days. I don't have access to that information so I'll discuss it with him in the morning and let you know.

If we do make a new ticket(s), here's some advice on how to structure it/them:
The truth is, neither you nor we can tell whether something that looks like the same bug in different circumstances or at different times is really the same bug or not. We all become more intuitive about it with experience, but the only people who can find out for sure are the developers, because only they have access to the code the bugs live in. So the best approach is to describe precisely what you did, in specific detail with as much surrounding context as seems like it might be relevant, and then say precisely what happened as a result and why you think it's a bug. This is almost literally what the Issue Guidelines tell you to do, which is why we always try to get people to read and use them.

Naturally, if you know some element of what you're describing is variable (such as the same thing happening with different kind of mobs), we want you to tell us that too, but keep it specific about which things you've actually observed. Because even if it happens to sheep, cows, turtles, and parrots, you can't really know that it happens to baby mules and cats unless you've actually seen it, and if you tell us it happens to them when it really doesn't, you'll send the developers on a wild goose chase, which just wastes everybody's time. (If it really does happen to them but you didn't tell us because you didn't know, that's no problem. The developers will know when they find the bug, because it'll be in code that baby mules and cats use, too. So you don't have to worry that if you only tell us about zombies, they'll only fix it for zombies and skeletons will still have the bug. That can happen, but it's not a common case.)

In short, what I'm saying is don't generalize, at least not without providing the specifics along with it, I know it seems like generalizing a problem makes it easier to think about and therefore do something about, but that's only in the human domain. It does the opposite for the developers, because they have to work in the computer's domain, and computers always and only work with the specifics.

Well i'll wait till tomorrow to make new reports then, likely just make one for the issues with raid spawning behavior, and then when it comes to passive mobs.. Do you want one for every individual passive mob or just one report including information on all the different ones that have issues? Or should i abandon a new report and just put all the info i have into MCPE-39727 ? There's a lot to be desired when it comes to making reports on this site lol.

If the same specific thing happens with different kinds of mobs, it's fine to just list them together in the same report. My point wasn't that every trivial difference calls for a new report, it was that discarding details by over-generalizing so you can cram them all into one report is the wrong approach. Where the balance point is between those two extremes is a gray area though, and we're not that fussy about it, so trust your judgment.

I'm definitely seeing all the raid issues as being a new bug, so feel free to work on that when you're ready. (I'm looking forward to that one, actually; I already have some questions in mind for you.) Whether to make a new one for the passive mobs or just add info to 39727 is what I want to talk with MegaSpud about. To be honest, though, I'm not sure what those issues are. The world you tested in wasn't vanilla, so I have to figure out which of those issues can be reproduced in a vanilla world.

Fair points, i likey. Am interested to see mega spuds take on all of this:3

Not having custom super flats on bedrock, coupled with extremely slow spawning rates, makes it tedious to test in a "vanilla world" hence why i use world edit to change the biome type. It works perfectly from every test that has been thrown at it.. But i suppose a full perimeter could be made, specifically to test each issue >.<

I have seen all of these issues in "vanilla worlds", but when it comes to showing them off, or testing them, using a modified world is the only reasonable way to do so.

One thing that probably should be considered are non-surface villages. Villages don't depend on sky access anymore and thus can be created underground in the overworld, and raids should be able to work with them as well. In the nether there are only cave spawns, and the raids on villages there would no longer work if they were entirely surface spawns.

(going to move that comment over to the bug report devoted to underground raid spawns)