Down The Core And More - round 3 rules announcement
Nenad Tomasev
Round 3
"tWINs"
Well, it's time for the third round of the tournament. In the first two
rounds, you have had a lot of options, and haven't been restrained by
the rules. This time, it's going to be different. This will be a
thematic round. We'll be exploring a strategy that has mostly been
neglected and unexplored. It's the doubled scanners, of course. But,
we'll spice things up a little more, as you'll soon discover. The rules
are:
CORESIZE: 8000
MAXPROC: 8000
MAXCYCLES: 80000
MAXSIZE: 100
MINDIST: 100
PSPACE: not allowed
INSTRUCTION SET: ICWS '94 (RedCode)
(standard battle parameters)
no selffights
obligatory part of the code:
wlength equ 18
wdist equ 3643
org boot
wstart ...
... ; this is where your code will be
for 77
dat 0, 0
rof
tb dat.f (wstart+wlength), (wstart+wlength+wdist)
boot mov.i {tb, <tb
djn.b boot, #wlength
spl.b *tb, 0
jmp.b @tb, 0
end
tournament type: round robin, 400 rounds per matchup
all entries will be inserted into the round robin tourney, along with
Hullabaloo, Gargantuan, Mooncake, DanceOfFallenAngels and Frantic.
In other words, they will fight other entries, and they will also fight
the above mentioned warriors. Permitted warrior types for insertion
into designated space are CDS, SCN and CLR. In other words: clear
directing scanners, HSAish scanners, fixed carpet length scanners,
SPL/JMP and similar scanner types, oneshots, bishots, twoshots, even
coreclears! And there will be some limitations. To avoid people
inserting MOV.I bomb, wdist or something like that to turn the scanner
(or coreclear) pair into a single working piece of code, the following
restriction is included: in the first 300 cycles in melee, the number
of processes has to be >= 2.
You've probably noticed the WDIST being "ugly" - and that's the whole
point. It would be easy to synchronize the warrior modules if the WDIST
was, for instance, 4000. You'll have to put in a little bit more
effort. Also, you must have noticed that the bench warriors are
designed to crush scanners. Generally, double scanners should be more
resistant against their attacks. That is, if you do a good job. You'll
also have to calculate the fact that you'll be fighting a lot of other
scanners... So, you have to find the optimal warrior size, scan step,
type of attack, possibly implement airbag, synchronize the attack,
balance the speed and, above all, be original in your coding :).
Good luck!
The DEADLINE will be January 5th, 23:00 CET.
Send your entries to tomasev_at_nspoint.net
or pbrain_at_cincilator.com
inversed
If my scanner is smaller than 18 instructions, may I decrease wlength
to save some cycles during boot?
saye...@gmail.com
inversed wrote:
> If my scanner is smaller than 18 instructions, may I decrease wlength
> to save some cycles during boot?
Good point! My dclear is not that long too :)
Nenad Tomasev
In case your code is smaller than the mentioned wlength, you can use
its length as a parameter instead. But the length has to be less than
the original one.
Sorry for the late reply.
Happy Holidays!
Lukasz Adamowski
> The answer is: YES.
> In case your code is smaller than the mentioned wlength, you can use
> its length as a parameter instead. But the length has to be less than
> the original one.
Strange, I was expecting the answer would be negative... So shorter warriors will start faster, that makes difference when fighting with known opponents which all have Q-scan.
Another question is: how many warriors can we send for this round? Two or one?
Bye
Lukasz Adamowski
P.S. And Happy New Year to all of you! :)
-= "War is a problem, never a solution" =-
-= "Wojna jest problemem, a nie rozwiazaniem" =-
----------------------------------------------------
W wieku 15 lat została królową Francji, chociaż chciała być tylko
zwyczajną dziewczyną. Nowy film Sofii Coppoli - MARIA ANTONINA.
W kinach od 5 stycznia !!
http://klik.wp.pl/?adr=http%3A%2F%2Fadv.reklama.wp.pl%2Fas%2Fantonina.html&sid=974