Archived from groups: rec.games.corewar (More info?)
Hi all.
I promised that I would release the strategy behind my tiny scanners
when one of them would be pushed off the tiny hill. This happened
today. I used the name stone/scan, anyway I do not think this is
correct.
Usually, a scan instruction looks like this:
SEQ.I (or SNE.I) loc1, loc2
however, if I tune my scanner so that it never scans twice the same
locations, I can use the same scan instruction to leave some garbage
behind:
SNE.I <loc1, <loc2
leaving garbage behind is a good way to trigger the opponent's scan
(this works as a decoy maker; anyway, it keeps making the decoy while
it scans). In the meantime, if loc1 and loc2 are empty instructions,
it scans the locations before; if one or both are not empty, there is
a good chance that: either it triggers the clear, or it drops a
"color" bomb on the opponent's code.
The "mov }dest, >dest" instruction that I put before the scan
instruction is just a bombing part that I added to explain the
"stone/scan" name. The rest of the warrior is a classical bishot
clear.
Now you can guess how Milkshake IV work... 😉
I sent the warrior to both Sourceforge and Koenigstuhl, and I hope
that John Metcalf will soon enter it in his infinite tiny hill.
Best regards.
LAchi.
;redcode-tiny
;name Strawberry and Lemon
;author LAchi
;kill Strawberry and Lemon
;strategy The best original bishot stone/scan I could get.
;strategy Tuned by Fizmo.
;assert CORESIZE==800
STEP equ -7
FIRST equ 56
WIDTH equ 404
loop sub.f sPtr, dest
start mov.i }dest, >dest
dest sne.i <FIRST, <FIRST+WIDTH
djn.a loop, <loop-61
add.f dest, sPtr
djn.f sBomb, sPtr
for 2
dat 0, 0
rof
sPtr dat STEP, STEP
dat 10, 402
sBomb spl #13, -328
sLoop mov sBomb, }sPtr
mov sBomb, >sPtr
sBack djn.a sLoop, <sBomb
end start
Hi all.
I promised that I would release the strategy behind my tiny scanners
when one of them would be pushed off the tiny hill. This happened
today. I used the name stone/scan, anyway I do not think this is
correct.
Usually, a scan instruction looks like this:
SEQ.I (or SNE.I) loc1, loc2
however, if I tune my scanner so that it never scans twice the same
locations, I can use the same scan instruction to leave some garbage
behind:
SNE.I <loc1, <loc2
leaving garbage behind is a good way to trigger the opponent's scan
(this works as a decoy maker; anyway, it keeps making the decoy while
it scans). In the meantime, if loc1 and loc2 are empty instructions,
it scans the locations before; if one or both are not empty, there is
a good chance that: either it triggers the clear, or it drops a
"color" bomb on the opponent's code.
The "mov }dest, >dest" instruction that I put before the scan
instruction is just a bombing part that I added to explain the
"stone/scan" name. The rest of the warrior is a classical bishot
clear.
Now you can guess how Milkshake IV work... 😉
I sent the warrior to both Sourceforge and Koenigstuhl, and I hope
that John Metcalf will soon enter it in his infinite tiny hill.
Best regards.
LAchi.
;redcode-tiny
;name Strawberry and Lemon
;author LAchi
;kill Strawberry and Lemon
;strategy The best original bishot stone/scan I could get.
;strategy Tuned by Fizmo.
;assert CORESIZE==800
STEP equ -7
FIRST equ 56
WIDTH equ 404
loop sub.f sPtr, dest
start mov.i }dest, >dest
dest sne.i <FIRST, <FIRST+WIDTH
djn.a loop, <loop-61
add.f dest, sPtr
djn.f sBomb, sPtr
for 2
dat 0, 0
rof
sPtr dat STEP, STEP
dat 10, 402
sBomb spl #13, -328
sLoop mov sBomb, }sPtr
mov sBomb, >sPtr
sBack djn.a sLoop, <sBomb
end start
Facebook
Twitter
Instagram