;redcode-94 ;name Maybe ;author Ilmari Karonen ;strategy Very simple table-based pseudorandom switcher ;strategy Weights optimized using an "ecosystem" approach ;strategy since I have no idea what I'll be against. ;assert CORESIZE == 8000 ;planar pspace, other, boot, scan, stun, bomber, clear, paper ;>These particular values give a rather good overall performance. ;>This tends to do rather badly against TimeScape, though. Since ;>other versions do that too, I can only assume my test population ;>is biased. I just wish I knew what I'm facing. ; Switcher constants PSTRAT equ 375 PRAND equ 7 rmagic equ 6441 ; TimeScape constants TSTEP equ 1800 CSTEP equ 3740 NSTEP equ -1922 FSTEP equ 1870 ; Mini HSA constants hstep equ 9 hptr equ (hbomb-5) org think ; ** TimeScape strat2 stp.a #2, think ; I hope it's OK to leave these around.. stp.a rand, guess spl 1, <-200 spl 1, <-300 mov.i -1, 0 tim1 spl @tim1, }TSTEP mov.i }tim1, >tim1 cel1 spl @cel1, }CSTEP mov.i }cel1, >cel1 mov.i {-FSTEP, NSTEP ; ** Switching table in dat 0, loss-strat strat dat 0, win-strat rand dat 0, tie-strat ; ** Carbonite & boot strat1 stp.a #1, think stp.a rand, guess mov.i dbomb, }cboot cboot spl datz+4, }969 ; silk-style boot z for 4 mov.i dbomb+z, }cboot rof dbomb dat >-1, >1 spl #0, <-100 mov dbomb, tar-197*3500 tar add #197, -1 ; gets bombed to start coreclear djn.f -2, <-1151 ; ** Mini HSA & boot strat0 stp.a #0, think stp.a rand, guess mov.i hbtm, @hboot z for 8 mov.i hbtm-z, hptr, >hptr jmn.f hkill, >hptr ha add.ab #hstep+1, @hmptr hscan jmz.f ha,