Chess, but everything is random | The Chess Randomizer Ep. 2

[@bobbobert9379]

Underpromoting into forced mate is goated

[@mathmagician8191]

2:32 j4 is the only move that avoids mate, but it ends up needing to sacrifice a queen for a rook so it is still losing

[@Rafael-w4j]

The pawns should be random also, Pawn or Checker or Knight.

[@Rodricz123]

This is an extreme version of chess960

[@mathmagician8191]

This position has a somewhat decent white advantage - my engine evaluates the starting position as +0.86

[@theominouspigeon]

there should be a mode where both people move at the same time

[@nikro8275]

this is genius

[@hy7864]

Synchronous Chess by Vitaliy Korolev, 1991

[@purplenickname]

In order to make the king space completely random you should first use a d12 to see where the king goes and then do the other pieces. With the current method not all spaces have equal chance to be a king space.

[@GreenLemonGames]

The king square is the first thing decided in the code. It's just revealed left-to-right visually

[@nikro8275]

Will this mode be added to deckmate public?

[@tacodoodle]

This is kind of like Fischer Random

[@CheckmateSurvivor]

Scramble is better than Fisher Random Chess.

[@DanielGamer-wn1hy]

Hi

[@godowskygodowsky1155]

How are you randomizing the king location? It might not be uniform.

[@GreenLemonGames]

The game picks a random square for the king before the other pieces are randomized. It seems to be working perfectly fine - I did a handful of test runs before making the first video.

[@godowskygodowsky1155]

@@GreenLemonGames Oh, that's fine. I was worried that you were revealing the pieces as they were being randomized, which is still possible to make uniform, but it's harder. (I understand this is how pieces are revealed in Fischer random for the purpose of making it a spectacle.)

[@BinaryBolias]

For making it uniform while generating in real-time, the chance for each square to be a king (until a king is placed) should simply be 1 divided by the amount of remaining, _un-generated_ squares. In confirming this method, the first square quite obviously has the correct chance of containing a king: The first square's _"absolute king chance" (for it to contain a king on any given generation)_ is the same as its _"local king chance" (1 divided by remaining square quantity - except equal to 0 after a king has been generated)._ Every square subsequent can have its absolute king chance confirmed by multiplying its _local king chance_ by the chance of getting to the square without previously generating a king _(the sum of all previous absolute chances subtracted from 1; a value between 1 and 0)._ In performing this absolute king chance calculation, you would find that the absolute chance of a king appearing on any given square is the same for all squares _(and is equal to 1 divided by the total quantity of valid squares)._ This is the method I would prefer, as it's more fun, and more authentic to the revealing process. This can also be easily modified to allow for multiple kings, and to also provide restrictions for other piece types, as well. _(quick explanation: Per piece type, use the _*_remaining amount of the piece to place_*_ ("RAotPtP", I guess) in the fraction numerator of the local piece chance. Piece types can be done in arbitrary order, per square, but the RAotPtP values of all pieces must sum to the total amount of valid squares. If the sum of RAotPtP values is greater than the square total, then a single piece type can still be placed in its full desired amount if it is checked for first on each square)_

[@ScrapFrog]

Will this be a in the full game? You should add it if not

[@GreenLemonGames]

Yes! It will be part of Version 1.7, which is coming soon

[@TaisGamingChannel314]

10.5x10.5 when

[@Abdullah0091a]

Is it the same guy from last time?