|
|
Special Reference: Shiny Hunting
Due to the newly risen discussion about the so-called Shiny Hunting and the postulation
of some people to prove the validity of 'Shiny Hunting', I now post a reference about this concern.
Since all these doubting people want a 'technical statement', the following won't be
easy to understand if you don't like numbers, probabilities and computer-terms. So if you are just another
pokemon player who tries to find 'the final answer' to why-the-hell-don't-i-get-any-single-shiny-with-this-tactic
(which is btw 42, as you all know) then quit reading here.
Warning: After reading this document some people will have to accept a 'uncomfortable' truth. Don't
go on if you can't live with the existence of 'Shiny Hunting'. (Yes, I'm a little bit sarcastic atm. -.-)
Okay. And here it comes. This is the routine which is responsible for the decision if the pokemon is
a Shiny: (I added some comments for you...)
ROM:0205E40C SpecialShinyHandler ; CODE XREF: CheateChainSet+5Cp
ROM:0205E40C PUSH {R3-R5,LR} ; r0 == chainsize
ROM:0205E40E CMP R0, #0 ; if equal 0 then no special shiny handling
ROM:0205E40E - the normal 1 out of 8192 chance is used
ROM:0205E410 BNE loc_205E416
ROM:0205E412 MOV R0, #0
ROM:0205E414 POP {R3-R5,PC} ; exit in case of chainsize == 0
ROM:0205E416 ; ---------------------------------------------------------------------------
ROM:0205E416
ROM:0205E416 loc_205E416
ROM:0205E416 ADD R2, R0, #0 ; r2 = r0 = chainsize
ROM:0205E418 MOV R1, #0xC8 ; r1 = 0xC8 = 200
ROM:0205E41A LDR R0, =0x2008 ; r0 = 0x2008 = 8200
ROM:0205E41C MUL R2, R1 ; r2 = chainsize * 200
ROM:0205E41E SUB R5, R0, R2 ; r5 = 8200 - r2
ROM:0205E420 CMP R5, #0xC8
ROM:0205E422 BGE loc_205E426
ROM:0205E424 ADD R5, R1, #0 ; prevent negative value if chainsize
ROM:0205E424 is bigger than 40
ROM:0205E426
ROM:0205E426 loc_205E426
ROM:0205E426 LSL R0, R5, #0x10
ROM:0205E428 LSR R0, R0, #0x10 ; overflow secure
ROM:0205E42A BNE loc_205E430
ROM:0205E42C BL IgnoreMe ; never executed
ROM:0205E430
ROM:0205E430 loc_205E430
ROM:0205E430 LSL R0, R5, #0x10
ROM:0205E432 LSR R0, R0, #0x10 ; prevent overflow
ROM:0205E434 CMP R0, #1
ROM:0205E436 BHI loc_205E43C ; r0 is always > 1, so jump
ROM:0205E438 MOV R4, #0
ROM:0205E43A B loc_205E46A
ROM:0205E43C ; ---------------------------------------------------------------------------
ROM:0205E43C
ROM:0205E43C loc_205E43C
ROM:0205E43C BL get_random_val ; Get a randomized value between 0 and 65535
ROM:0205E440 ADD R4, R0, #0 ; r4 == rand
ROM:0205E442 LSL R1, R5, #0x10
ROM:0205E444 LDR R0, =0xFFFF ; r0 == 0xFFFF == 65535
ROM:0205E446 LSR R1, R1, #0x10
ROM:0205E448 BLX div_r0_r1 ; 65535 / (8200 - chainsize * 200)
ROM:0205E44C ADD R1, R0, #0
ROM:0205E44E ADD R1, R1, #1 ; (65535 / (8200 - chainsize * 200)) + 1
ROM:0205E450 LSL R1, R1, #0x10
ROM:0205E452 ADD R0, R4, #0 ; r0 = rand
ROM:0205E454 LSR R1, R1, #0x10
ROM:0205E456 BLX div_r0_r1 ; rand / (65535 / (8200 - chainsize * 200)) + 1
ROM:0205E45A LSL R0, R0, #0x10
ROM:0205E45C LSR R4, R0, #0x10 ; r4 = r0 (r0 == lastresult)
ROM:0205E45E LSL R0, R5, #0x10
ROM:0205E460 LSR R0, R0, #0x10 ; r0 = (65535 / (8200 - chainsize * 200)) + 1
ROM:0205E462 CMP R4, R0
ROM:0205E464 BCC loc_205E46A ; if r4 < r0 then jump
ROM:0205E466 BL IgnoreMe ; this time it is executed but it doesn't
ROM:0205E466 change r4 so ignore it once again
ROM:0205E46A
ROM:0205E46A loc_205E46A
ROM:0205E46A
ROM:0205E46A CMP R4, #0 ; last test. If r4 == 0 Then...
ROM:0205E46C BNE loc_205E472
ROM:0205E46E MOV R0, #1 ; ... make it SHINY!
ROM:0205E470 POP {R3-R5,PC} ; return
ROM:0205E472 ; ---------------------------------------------------------------------------
ROM:0205E472
ROM:0205E472 loc_205E472
ROM:0205E472 MOV R0, #0 ; ... else: no shiny
ROM:0205E474 POP {R3-R5,PC}
ROM:0205E474 ; End of function SpecialShinyHandler
ROM:0205E474
ROM:0205E474 ; ---------------------------------------------------------------------------
ROM:0205E476 ALIGN 4
ROM:0205E478 dword_205E478 DCD 0x2008 ; DATA XREF: SpecialShinyHandler+Er
ROM:0205E47C dword_205E47C DCD 0xFFFF ; DATA XREF: SpecialShinyHandler+38r
ROM:0205E480 ; ---------------------------------------------------------------------------
Okay. That's the technical part. Let's analyze it the mathematical way.
The basic formula is this one: (n is the chainsize)
It works for all cases of n between 0 and 40. If n is bigger than 40 the denominator is
always 200.
After calculating this, a function, which returns a random number between 0 and 65535, is called.
The random number (r) is divided by Z and afterwards the result has to be lower than 1. (It's a math.floor operation,
so anything < 1 returns 0.)
Because r can be anything between 0 and 65535 the probality P is:
So, for the last time: NO, shiny hunting isn't a bug as many people state and YES,
shiny hunting depends on the 'length' of your shiny chain.
Greetz
MX
|
|
