The odds of you generating a random special world?

Probably [1 / (2^31 - 1)]. Random seeds will always be an integer from [0] - [2^31 - 2] as a string - rand.Next().ToString(), where rand.Next returns a number from [0] to [2^31 - 2] : [2^31 - 1] values total.

Both the Bees seed and FTW are checked against the text of the seed, so they're out. That leaves only the Drunk world, which is checked right as the world begins generating. Thus, [1 / (2^31 - 1)].
There's a chance that this is wrong, but I don't see any issues with it.

Edit: Changed some instances of [2^31 - 2] to [2^31 - 1]. rand.Next() does generate a value from [0] to that, but the number of values is still [2^31 - 1]. Also added brackets to all the numbers.
 
Last edited:
Probably 1/(2^31 - 2). Random seeds will always be an integer from 0 - (2^31 - 2) as a string - rand.Next().ToString(), where rand.Next returns a number from 0 to 2^31 - 2.

Both the Bees seed and FTW are checked against the text of the seed, so they're out. That leaves only the Drunk world, which is checked right as the world begins generating. Thus, 1 / (2^31 - 2).
There's a chance that this is wrong, but I don't see any issues with it.

Yes now that I think about it, it looks like there are not any negative integer values in seed creation. So for the max 32-bit integer (2147483647) it would come out to 1:2147483647 chance of generating the one of three special world seeds.

random.PNG


The game does not look like it has a check within the generation settings that process the random function that would prevent the game from selecting the drunk worlds seed of 5162020. From the traceroute of when the game creates a random seed, it does not alter the int after it passes the seed generation.

done01.PNG


This is a very interesting concept being you could in theory generate the custom drunk world seed by complete random. The chances of you getting this one lucky seed are greater in equivalence to you winning the match 5 jackpot.
 
I never thought that number of possible world seeds is so limited. At first, over 2 billions might be impressive, but further investigation proves that such low number of seeds produces quite a lot of collisions (i.e. twin worlds with the same seed).
A rough example may give an idea of how many twin worlds might be played presently. Imagine 30 millions Terraria players decided to generate 1 world each with default settings and random seed. A simple simulation model which counts frequencies of 30 mln random integers uniformly distributed in [0; 2^31-1] interval gave us approximate number of collisions:
~208000 twin world pairs
~950 twin world triplets
~4 twin world quadruplets

All of these don't appear to impact normal players' experience in any harmful way. However, I suppose it is not a kind of behaviour you would usually expect or prefer to have.
 
Back
Top Bottom