User:Poochy/Konoha Mode Analysis

Based on the video from the first location test (2009), the randomizer appears to be history of length 3. Until some hidden condition is met (5 Bravos?), it will only pick from pieces not in history, so any 4 consecutive pieces dealt are all different until that point.

初ロケテス(2009年)の動画では、全消し5回(?)完成時まで、NEXT画面が必ず4種とは限らない。

Case notation: +X-Y means X held, Y missing from next piece preview.

「+X-Y」が「Xホールドしている、NEXT画面の4種にYが無い」。

Case #		Missing
		-I	-L	-O	-T	-J
Held	+I	1	2	3	4	2
	+L	5	1	6	7	8
	+O	3	6	1	9	6
	+T	10	11	12	1	11
	+Z	5	8	6	7	1

Case 1: One of each piece (5手5種ツモ)

	→
	→
	→
	→
	→
	→

Can be mirrored.

//TODO: Permutations that don't work with this build or its mirror image.

Case 2: +I-L, +I-J

Diagram for +I-L, mirror for +I-J. Use held I first (horizontal), then hold L if necessary.

→

Case 3: +I-O, +O-I

Use held I or O first. Make a 4x2 with I's or O's. (Note: For +O-I in 20G, build the 4x2 on the right to ensure the second O can reach its destination.) With remaining L, J, T, the following can be done in any order:

+

+

Special case: If I comes last in +I-O, hold the first of L or J and use one of the following builds (if J comes first, mirror if L comes first):

T comes first or second:

→

T comes third:

→

Case 4: +I-T

Use held I first, and hold the second I. Build depending on relative order of L, O, and J, then finish with I for a Tetris Bravo:

OLJ, OJL	LOJ, JOL	JLO, LJO

Case 5: +L-I, +J-I

Strategy for +L-I, mirror for +J-I.

Build depending on relative order of L, O, and J:

OJL, OLJ, LOJ: Use held L first, then hold the second L.

→

JOL: Same as above. Hold J and use the held L, then hold the second L and use the held J. Alternate no-hold strategy:

→

LJO, JLO: Use held L first, then hold J.

→

Case 6: +L-O, +J-O, +O-L, +O-J

(Note: For +O-* in 20G, build the 4x2 on the right to ensure the second O can reach its destination.)

Diagram for +L-O, mirror for +J-O. Hold J.

→

Diagram for +O-L, mirror for +O-J. Hold J.

→

Case 7: +L-T, +J-T

Diagram for +L-T, mirror for +J-T.

//TODO

Case 8: +L-J, +J-L

Diagram for +L-J, mirror for +J-L.

If I comes first or second, use the held L first and hold the second L:

→

Else if O comes first or second and I comes third:

→

Another build, requires T to come first and I to come second or third:

→

→

//TODO: Permutations. In particular none of these work if I comes last.

Case 9: +O-T

(Note: In 20G, build the 4x2 on the right to ensure the second O can reach its destination.)

→

Hold L.

Case 10: +T-I

//TODO

Case 11: +T-L, +T-J

Diagram for +T-L, mirror for +T-J.

If I comes first, hold J for Triple Bravo:

Otherwise, use held T first, hold I:

J before O	O before J

The above will not work in 20G for TIOJ or TIJO, in which case save the held T for last and use the following alternate builds:

	TIOJ →		→
	→ TIJO		→

Case 12: +T-O

→

(Note: It is possible to make a similar build with the J on the left and the L on the right, but this is less safe in 20G since failing to IRS the first of L or J will get it stuck in a position where it can't be moved to the desired position. With the above build, IRS is not strictly necessary.)

//TODO: Permutations