Minecraft Forums

The changelog is up which means the update should be live in 2 to 2 1/2 weeks. There are an enormous amount of bug fixes which is awesome, plus world generation is seeing an overhaul so those hours and hours of searching.... yeah about that. That is a tad bit depressing but hey, a ton of bugs are being fixed and if you need ocelots you might wanna get them before the update goes live provided you don't plan to just start a new world as soon as the update goes live.

I think I can look for yet another world after the update for this... Nice to see they fixed the "world changing slightly" issues as well as overhauling the biome generation algorithms.

It is very nice to see, yes. Just a tad unfortunate for those of us who invest countless hours trying to find good seeds. But heck that means perfect seeds should be a bit more common and that is what I'm looking forward to.

Just for clarity, I don't think they overhauled the generation algorithms for biomes. They just changed it so when you let it make a random seed, the seed it chooses has a lot of biomes.

But that is still awesome.

Quote from crasyleg73»

Just for clarity, I don't think they overhauled the generation algorithms for biomes. They just changed it so when you let it make a random seed, the seed it chooses has a lot of biomes.

But that is still awesome.

no, I think they may have in fact changed the algorithms due to size limitations. Current gen doesn't have that same problem so narrowing seeds would be easier but on pre-gen the algorithm would have to change, incorporating even smaller biomes and a bit more compact. I expect we will see a lot of mixed biome islands after the update comes. The only other thing is that several others as well as myself had spent hours analyzing seeds. If there were any better ones we would have found them. Several decent seeds were found but only a handful that were near perfect, obviously no perfect seed was ever found.

Quote from Wolfhusband»

It is very nice to see, yes. Just a tad unfortunate for those of us who invest countless hours trying to find good seeds. But heck that means perfect seeds should be a bit more common and that is what I'm looking forward to.

I wouldn't throw away your "near perfect" seed... the change just might be just enough to make it absolutely perfect!

Quote from crasyleg73»

Just for clarity, I don't think they overhauled the generation algorithms for biomes. They just changed it so when you let it make a random seed, the seed it chooses has a lot of biomes.

But that is still awesome.

A random seed should and is likely just that - a random number generator... it doesn't (and shouldn't) "choose" anything. Doing that would make a number of random numbers generated come up as "invalid." What 4J changed are the probabilities of certain things happening after that number is generated (and that means that they've changed the algorithms that the game uses to convert that random number into a representation of biomes, technical biomes, and structures within the seed). I suspect (and I'm purely speculating here) that what they changed will be that we'll see fewer repeat biomes scattered all over the map... So, to use my dice analogy, if you roll 2 or more dice of the same biome, the generator will likely now be forced to put those side by side somehow (i.e. combine them into one biome), rather than be allowed to scatter them in little blobs all over the map. As a result, the terrain on the Xbox 360 maps may not blend quite as nicely as on the PC. They may also have adjusted the technical biomes and structures algorithms to allow them to generate more frequently within smaller biomes... and, according to the change log, they've at least made them stable with the seed again.

How would this help seeds get more biomes on average - well, if we're only rolling 61 dice, we've got one chance at a perfect roll, but if we're rolling more (say 2x as many) dice, our chances improve of getting at least one of each number.

The only thing that annoys me really about the change log is the correction for the texture of the red sandstone stairs. I liked that smooth texture and have already made a few builds that use that particular block for that reason. Aside from that, that's quite a few bug fixes nonetheless.

Whats interesting though is that the generation currently looks similar to PC. Except it's four times smaller!(I think).

I don't think they are going to make the biomes any smaller.

and it doesn't say it changes the world generation, it says it changes the random seed generator.

Quote from Changelog

• Improved random seed generator to select worlds with a better spread of biomes.

If I am correct this also means that choosing your own seed will cause it to have a lower change of having diverse biomes, compared to generating a random one.

Quote from crasyleg73»

Whats interesting though is that the generation currently looks similar to PC. Except it's four times smaller!(I think).

I don't think they are going to make the biomes any smaller.

and it doesn't say it changes the world generation, it says it changes the random seed generator.

Quote from Changelog

• Improved random seed generator to select worlds with a better spread of biomes.

If I am correct this also means that choosing your own seed will cause it to have a lower change of having diverse biomes, compared to generating a random one.

symmantics. All I'm saying is that there really isn't a large number of even decent seeds. I've been through at least 1000 or so. Plus the thousands of others who were hunting. Generally speaking most seeds have a mesa or a mega taiga or an ice spike plains. Also about 85%of seeds that I have seen at least had 1 or 2 trees missing. Granted a perfect seed is not 4j's concern. But other biomes suffer too like the mooshroom biomes. Of the 1000ish seeds I searched mooshroom only appeared 4 or 5 times. I know it states that generation is simply being improved for random seed generation but that would severely hamper the amount of seeds available to about 15-20%. If that is the case then all the seeds deemed unfit would simply become useless.

I just think it would make more sense to adjust the entire generator rather than remove a large portion of it. But thats just me. According to the text that is what will happen, cutting out the excess. It just doesn't make sense to do so.

Tee-hee, one little exploit I found isn't on the list~

But sweet, this looks good!

I guess it depends on how it works.

Like does it examine 5 random seeds, and then pick the best 1?

Actually If it did that, a seed that got rejected one time, could be accepted in a different situation depending on what seeds get picked with it.

And even if it does limit the number of seeds, it will still be a vast number.

What I'm saying is that I think it will probably pick better seeds, but not actually have tight requirements.

This seems to make sense, because I don't think they would have the game examine hundreds of seeds at once.

Oh and here is more supporting evidence:

• Improved random seed generator to select worlds with a better spread of biomes.

Quote from crasyleg73»

I guess it depends on how it works.

Like does it examine 5 random seeds, and then pick the best 1?

Actually If it did that, a seed that got rejected one time, could be accepted in a different situation depending on what seeds get picked with it.

And even if it does limit the number of seeds, it will still be a vast number.

What I'm saying is that I think it will probably pick better seeds, but not actually have tight requirements.

This seems to make sense, because I don't think they would have the game examine hundreds of seeds at once.

Or... are we just going to be able to just make a choice to specify whether we want it to generate a random seed with either 1) a greater variety of biomes included in the map (acknowledging they will be smaller in size) vs. 2) larger biomes (acknowledging that there will likely be fewer biomes in the map and perhaps more "missing" blocks as a result). I think this this either/or choice is the most likely (albeit perhaps the least satisfying) option. Perhaps, with luck, we might also be able to specify that we want all basic biomes (so that all blocks are available) and on the other side of that coin, be able to specify that we want a single-biome world - giving us 4 options to choose from. This would be a broad way of telling the random seed generator what sort of seed were looking for (survival vs. creative themed) without actually dictating to it what biomes to generate. The generator would then just pull randomly from one of either two or four pools of seeds that fit the option selected. I think we'll still have to fly through them to see if they included the specific biomes we had in mind.

That way, it wouldn't change or invalidate any seeds that we've already generate... They would generate the same still. The random generator would just limit itself to generating seeds of a certain type if the player chooses that option. If the player chooses to instead type in a seed, it would work the same as it always has.

Well your suggestion makes sense, But based on the changelog does not look like it is happening right now. But quite possibly something similar will happen in the future.

In the future there may be an option for world type(It would probably be all the same seed just different type)

It might have options like this

Normal

Large Biomes

Custom - choose the biomes, so you could make a single biome world.

Ach... Now my head hurts from debating.

Quote from crasyleg73»

I guess it depends on how it works.

Like does it examine 5 random seeds, and then pick the best 1?

Actually If it did that, a seed that got rejected one time, could be accepted in a different situation depending on what seeds get picked with it.

And even if it does limit the number of seeds, it will still be a vast number.

What I'm saying is that I think it will probably pick better seeds, but not actually have tight requirements.

This seems to make sense, because I don't think they would have the game examine hundreds of seeds at once.

Oh and here is more supporting evidence:

• Improved random seed generator to select worlds with a better spread of biomes.

I read it last time. I understand the writing on the wall. Well guess I will give it a shot but definitely keeping my near perfect seed. Just seems odd. Better spread is pretty vague. So in other words just cut out the ocean worlds. I won't pretend to know how many seeds currently exist nor how many will be available through semi-random selection after the update

The Thing is that no seeds will cease to exist If it works like we think it will, Because poor seeds just need to get analyzed next to worse seeds, and different seeds are analyzed each time.

So it is not neccesarilly better seeds all the time, just a better chance for good seeds

The only seeds that could cease to ever possibly exist, under this theory, would be the worst 4(if it analyzed five) because they could never get in first place next to other seeds.

So if this is how it works, the number of available seeds after the update will be:

(current number of seeds)-(number of seeds analyzed by seed generator - 1)

Quote from Wolfhusband»

I read it last time. I understand the writing on the wall. Well guess I will give it a shot but definitely keeping my near perfect seed. Just seems odd. Better spread is pretty vague. So in other words just cut out the ocean worlds. I won't pretend to know how many seeds currently exist nor how many will be available through semi-random selection after the update

"Better" is really a personal thing for each player (e.g better for me is a single-biome ocean world for treeless survival, which would probably be the worst seed that Crazylegs would eliminate entirely). Therefore, I don't think the game is going to analyze "better" or "worse" - the player is going to select what they consider to be better or worse out of, probably, only 2 alternatives - larger and less vs. smaller and more. Once the player toggles their choice, the game will use one of 2 algorithms to generate a random seed string (out of the pool for that algorithm). So, if the choices are 2, then there will be 2 pools of seeds encompassing 1/2 of all valid seeds based on whether they fit the parameters for "larger and less" or "smaller and more." (say, more than 1/2 of all biomes or less than 1/2). No current random seeds become impossible, they'd just be in one or the other pool only. With "smaller and more" the odds are just doubled on that particular roll that you'll get variety since you won't have any chance in that generation of getting a seed out of the "larger and less" pool. Typed in seeds would probably default to a third, more general algorithm (probably the current one) that might generate anywhere from 1 biome to all of them. If opting to type in a string, the "selection" would be disabled.

All three algorithms could be essentially the same as far as how they place the biomes, etc. (could be the same as the PC), just the ones activated by the "selection" the player makes would have an additional limiting formula so, for example, one won't generate a seed with fewer than 1/2 the available biomes and the won't generate a seed with more than 1/2 of the available biomes. It could world the same with 4 options - one generating only seeds with less than 1/4 of the available biomes, one generating only seeds with 1/4 to 1/2 of the available biomes, etc. Also, the divisions don't have to be equal. The "smaller and more" pool could be everything with more than 90% of the available biomes and the "larger and less" pool could be everything else with less than 90% of the available biomes.

Thank you for your comment, as it has cleared up my mind.

The thing is, the choice your describing, choosing between a seed with lots of biomes, and choosing one with a few is pretty much already there in this update(tu32). and adding just a plain old "Large Biomes" option, would complete it.

I have turned my prediction into a---

SCENARIO FOR DUMMIES. No offense intended

pretend tu32 has come out and then later they added a Large Biomes option

Warning: some of the details in the below scenario are not verified but just there to show the concept

Alice wants lots of biomes in her world, so she loads a world leaving the seed blank.

The game picks 5 random seeds. It then analyzes them and sees which one has "the best biome spread"

She gets a world with a decent amount of biomes.

Bob wants a normal world without too many biomes. He types in something random, "hilariousmonkey", or "567899954".

He gets an average seed like how they were in tu31.

Carl isn't wants just a few biomes. he either chooses his own seed or chooses a random one. But then he chooses the "Large Biomes". He gets a world with bioes 4 times as large(PC size).

That would be great, yes. But going verbatum... "Seeds with a better spread" seems to be the phrase of the game. Better is subjective only until it is given a constant value. So say "better" currently has a minimal of 15 biomes of the available 61 with the upgrade increasing it to 40/61.

This will make seed hunting slightly better for those looking for multi-biome seeds but a lot worse for survival island type seeds and 1 of a kind biome seeds.

This doesn't guarantee that the next seed generated will have mesa, mega taiga, ice spikes or mooshroom island which are the only 4 biomes that really matter for a horder like myself but rather guarantees that there will just be an overall mix of biomes. The parameters set for the current engine are unknown because it really does seem random. But I suspect the new generation engine will be easier to exploit after some crazed mathmatician decides to calculate how many biomes actually generate in any given string or set of strings.

Tied up nice and neat, survival island will be restricted to seed input and variety will be better on semi-random. It would be extraordinary to select what we are looking for but that ssems a bit defeating to the entire generation engine. Even if it pulls a number of seeds that match or closely match your criteria. Sure would beat wasting hours trying to find one seed though. Less search = more play. Quick! Someone insert a "I like my facebook games on facebook" gag

We will have to wait untill the update comes out, to see If I am right, and you are wrong, or vice versa.

Until then....