I don't get the point of multiple spawning floors... Every single floor you add adds one more chance in 128 of bailing out before placing any mobs (bad) and also adds one more chance in 128 of placing a mob in that chunk before moving to the next (good).
The good and the bad should cancel each other out EXACTLY, no matter how large your mob trap is, shouldn't they?
A point is chosen from the whole of the chunk, not just the surface area in it. That's 1 point in 32,768. A single floor in a chunk is 16x16=256 squares of that 32,768, and each floor adds another 256 legal spawning locations. Since all spaces other than the first floor were non-viable before the second floor was added, you've reduced, not increased, the blocked spaces.
More floors means more places they can spawn and fewer places they can't.
256 / 32,768 = 1 / 128... That's exactly what I said.
I don't think you understand my argument. Yes, you are adding 256 more legal spawning places, BUT you are also adding 256 more places where the algorithm will select solid rock and bail out of the script (no mobs).
Imagine rolling a 6-sided die. One floor in your tower would be the equivalent of winning on a 1 and losing on a 6. 2-5 is a re-roll. Adding a floor is the equivalent of winning on a 1 OR a 2 but also losing on a 5 OR a 6 (3-4 are re-roll). The number of wins in the long run is exactly the same! Similarly, the number of mobs will be the same no matter what.
I just coded and ran a monte carlo simulation in matlab for this, in fact, and the number of floors does not change the average mob spawning over five hundred thousand trials. Adding solid rock with no air space above it, however, does change mob spawning (e.g. building a trap at high elevation)
No. 2-5 would not be re-rolls. 2-5 would be lose. You lose if the spot chosen is solid, and you lose if the spot chosen is in mid-air. A mid-air test may succeed, but if the follow-up spawn locations that get chosen are not at a level where the mob spawns on ground, they fail. That means a successful spawn of 0 mobs.
If you have a perfectly even chunk with no hills and no caves, the only place that will be successful is at sea level, and any mob that tries to spawn higher or lower than that will fail. All those higher and lower places are possible choices, and they are fails, not re-rolls.
Adding a level, in the air or underground, removes 256 fails and replaces them with successes. The place where the floor itself is added was already 256 fails, you do not gain fails.
Oh okay. We are thinking about different scales is the problem, I believe.
I am thinking of the entire spawn range of 144 chunks, and you seem to be talking about a tower just at one chunk (which I had thought people gave up on after Notch changed the code).
If your trap is only one chunk, then mid-air is just as much of a fail as solid rock, because even if it moves on, the mob won't be in your trap anyway. But if your trap is many chunks large, then mid-air is NOT as bad as solid rock, because if it's mid-air, you then get more chances for spawns in the other chunks of your trap, whereas if it is solid rock, that cycle is finished for your whole trap.
So if your trap is say, 2 chunks large, and you have X floors, then a solid rock in the first chunk means 0% chance of spawns. But mid-air in the first chunk means moving onward, and then you have an X/128 chance of spawns in the last chunk.
There would be a SLIGHT advantage in a 144 chunk trap to have many floors, because whatever the very last chunk is, it would be more likely to spawn there, and mid-air would = solid rock in that last chunk. But that is only if the other 143 chunks chose mid-air... the extra construction effort would absolutely not be worth it. I'd say if your trap is smaller than about 20 chunks or so, making extra floors is worth it, probably. There is no hard rule, though, because it just depends how much you value your digging time.
Rollback Post to RevisionRollBack
Przerwap, upon looking at some code I had just written: "...Gav... That's not how programming works."
I have been under the impression that, in fixing the "0,0" exploit, Notch caused the full spawn cycle to not bail any more on encountering a fail, but just move on to the next chunk....
I have been under the impression that, in fixing the "0,0" exploit, Notch caused the full spawn cycle to not bail any more on encountering a fail, but just move on to the next chunk....
Source?
Rollback Post to RevisionRollBack
Przerwap, upon looking at some code I had just written: "...Gav... That's not how programming works."
From page 12 of this thread there's this post. Not sure if it's been verified.
Quote from Glurak »
What I got so far but I am not sure about. Tell me if I pointed something wrong.
- There is no more chunk0,0 exploit
- There is no more "bail out" algorithm (it haven't got put back after revert)
- There is code for slimes and giants spawning but blocked completely
- Nothing spawns within 24blocks "circle" around you/spawnpoint from bedrock to sky. (mob spawners are exception)
- Nothing spawns on water(source/current)/lava/halfrock/glass/midair/ice
- I have sometimes seen not moving creeper. That haven't moved/jumped until you got too close. Bug? Feature?
- The spawning algorithm iterate fast. If there is nothing except for one bigger room withing 144x144 blocks for mobs to spawn in, that room will get full of mobs within ? minutes.
- Notch is going to rewrite spawning algorithm completely
- You guys that nags into decompiled java bytecode have no life. :smile.gif:
Given these information I come out with this theoretical looting strategy:
- Build small house on the bottom of the world in the middle of some ocean.
- Find all caves within 144x144 blocks and light them up / fill them with water
- Choose one chunk 24 blocks away from you and build trap tower there.
- All mobs will spawn there, will spawn there fast and will give you effect nearly as good as chunk0,0 exploit.
I would trust that, except it came from a guy who in the same post says that decompiling is lame, implying that he didn't actually find that information himself in the code.
It is easy-ish to test though... You could time how long it takes to spawn stuff in cleanroom, and then time how long it takes to spawn stuff in a cleanroom at 120 altitude, and see if there is a difference. Should be pretty dramatic if there is still bailout. Should make zero difference if there is no bailout.
Rollback Post to RevisionRollBack
Przerwap, upon looking at some code I had just written: "...Gav... That's not how programming works."
I would trust that, except it came from a guy who in the same post says that decompiling is lame, implying that he didn't actually find that information himself in the code.
It is easy-ish to test though... You could time how long it takes to spawn stuff in cleanroom, and then time how long it takes to spawn stuff in a cleanroom at 120 altitude, and see if there is a difference. Should be pretty dramatic if there is still bailout. Should make zero difference if there is no bailout.
Yes, that particular post isn't a valid source, it includes the theory that the no-spawn radius extends the full map height, which would have completely shut down my trap when it was 38x38 with the collection point at the center of it but down 28 meters.
hey can you please start writing the updated version? I need real information on the current update (and include numbers now) to maximise the efficiency of my trap. Or tell me a way to get the spawn code. PM me please!
Rollback Post to RevisionRollBack
I have a new signature as that one was a year (or more?) outdated!
No. 2-5 would not be re-rolls. 2-5 would be lose. You lose if the spot chosen is solid, and you lose if the spot chosen is in mid-air. A mid-air test may succeed, but if the follow-up spawn locations that get chosen are not at a level where the mob spawns on ground, they fail. That means a successful spawn of 0 mobs.
If you have a perfectly even chunk with no hills and no caves, the only place that will be successful is at sea level, and any mob that tries to spawn higher or lower than that will fail. All those higher and lower places are possible choices, and they are fails, not re-rolls.
Adding a level, in the air or underground, removes 256 fails and replaces them with successes. The place where the floor itself is added was already 256 fails, you do not gain fails.
I am thinking of the entire spawn range of 144 chunks, and you seem to be talking about a tower just at one chunk (which I had thought people gave up on after Notch changed the code).
If your trap is only one chunk, then mid-air is just as much of a fail as solid rock, because even if it moves on, the mob won't be in your trap anyway. But if your trap is many chunks large, then mid-air is NOT as bad as solid rock, because if it's mid-air, you then get more chances for spawns in the other chunks of your trap, whereas if it is solid rock, that cycle is finished for your whole trap.
So if your trap is say, 2 chunks large, and you have X floors, then a solid rock in the first chunk means 0% chance of spawns. But mid-air in the first chunk means moving onward, and then you have an X/128 chance of spawns in the last chunk.
There would be a SLIGHT advantage in a 144 chunk trap to have many floors, because whatever the very last chunk is, it would be more likely to spawn there, and mid-air would = solid rock in that last chunk. But that is only if the other 143 chunks chose mid-air... the extra construction effort would absolutely not be worth it. I'd say if your trap is smaller than about 20 chunks or so, making extra floors is worth it, probably. There is no hard rule, though, because it just depends how much you value your digging time.
Source?
I would trust that, except it came from a guy who in the same post says that decompiling is lame, implying that he didn't actually find that information himself in the code.
It is easy-ish to test though... You could time how long it takes to spawn stuff in cleanroom, and then time how long it takes to spawn stuff in a cleanroom at 120 altitude, and see if there is a difference. Should be pretty dramatic if there is still bailout. Should make zero difference if there is no bailout.
Yes, that particular post isn't a valid source, it includes the theory that the no-spawn radius extends the full map height, which would have completely shut down my trap when it was 38x38 with the collection point at the center of it but down 28 meters.
I have a new signature as that one was a year (or more?) outdated!