/* GLOBAL SETTING */
MAX_FRAMES 120
--------------------------
/* GAME MONITORING */
/* A, X, Y, PC correspond to the values of the CPU registers. */
// D57D = FrameUpd
// C5AB = RandomFunc
// BF25 = CreateDrop
// A231 = HurtFunc
// C833 = GotItem
// C219 = KillFunc
// BEEC = EnemyKill
WHEN PC = 0xBF25
DEFINE RockmanBonus = Y /* Y identifies the item that was generated */
WHEN PC = 0xA231
DEFINE RockmanHurt = 1
WHEN PC = 0xC219
DEFINE RockmanHurt = 2
WHEN PC = 0xC833
DEFINE RockmanItem = A /* A identifies the item that was collected */
WHEN PC = 0xBEEC
DEFINE RockmanKill = A /* A identifies the enemy that was destroyed */
/* Bonus items in Rockman:
* 0x3B = no item
* 0x3c = score
* 0x3D, 0x3E =small refills (wpn,nrj)
* 0x3F, 0x40 = big refills (wpn,nrj)
* 0x41 = 1up
*/
--------------------------
/* How to restart */
CLEAR RockmanBonus
CLEAR RockmanHurt
CLEAR RockmanItem
CLEAR RockmanKill
--------------------------
/* RULES, HOW TO CONTINUE AT EACH FRAME */
WHEN RockmanHurt COUNT_AT_LEAST 1
OR RockmanItem COUNT_AT_LEAST 1
OR RockmanBonus IS_NOT 0x3F,0x3B
RESTART
WHEN RockmanBonus HAS 0x3F COUNT_AT_LEAST 2
/* When two refills have been found, accept this result */
ACCEPT
/* Set this as a new cap (don't try to find longer sequences) */
SET_LIMIT
/* And go find more solutions. */
RESTART
--------------------------
/* HOW TO CREATE INPUT */
/* These hex numbers are bitwise ORs of controller input
* bits. 0x01 = A, 0x40 = left, 0x80 = right, 0x02 = B
*/
INPUT_TABLE 0x00 0x01
0x40 0x41
0x80 0x81
0x01 0x41 0x81
0x01 0x41 0x81
0x01 0x41 0x81
// initialize input with zero.
INPUT 0x00
// 1:20 chance that B will be held this frame
IF_RANDOM 20 INPUT_OR 0x02
// 1:20 chance that LastInput will be changed
IF_RANDOM 20 RANDOM_ASSIGN LastInput INPUT_TABLE
// Include the LastInput in this frame
INPUT_OR LastInput
A more general way of setting the "goal" that the emulator has to find is to set it as being "earliest frame that has sprite #(insert address here) loaded on the graphics palette". That would be one way of knowing that the level has ended, for example, or that the enemy died, or that you got the power up, etc.
dont matter how you look at it, the user will have to know the relevant ram addresses.
dont matter how you look at it, the user will have to know the relevant ram addresses
The instructions didn't work, as in, there's a bug somewhere so it didn't save the same input it used to play with?
much better would be to have BisqBot optimize each hit individually, and only roll back all the way if it can't find any input that works at all for one of the later hits (or if it succeeds in killing the boss and wants to try again). That's what I had in mind when suggesting to try it in boss battles, that the segmented nature of the battle be taken advantage of to speed up the search. (I'm talking about boss battle optimizations in general so maybe I'm overlooking something specific to Gutsman here...)
for bigger events, the only potential problem with this method is that is would eliminate the possibilty of skipping the "goal" entirely. for instance if you were to try to make this play the wily stage of mm1 and one of the goals was to get to cutman as soon as possible it would never "discover" how to bypass the boss entirely.
I wonder if Generic Algorithms would be of any use here? Although calculating the fitness value for just *one* genotype could be quite expensive.
You probably mean genetic, not generic.
I've been thinking about it, but there isn't any "learning", or room for learning, in the current system. I don't see how it could work...
