Issue 11 January 8, 1996
______________________________________________________________________________
Core_Warrior_ is a weekly newsletter promoting the game of corewar. Emphasis
is placed on the most active hills--currently the '94 draft hill and the
beginner hill. Coverage will follow where ever the action is. If you have
no clue what I'm talking about then check out these five-star internet locals
for more information:
FAQs are available by anonymous FTP from rtfm.mit.edu as
pub/usenet/news.answers/games/corewar-faq.Z
FTP site is: ftp.csua.berkeley.edu /pub/corewar
Web pages are at:
http://www.stormking.com/~koth ;Stormking
http://www.ecst.csuchico.edu/~pizza/koth ;Pizza
http://pauillac.inria.fr/~doligez/corewar/ ;Planar
______________________________________________________________________________
Hi, and happy new year.
Back home, after some days at the mountains one of my first thoughts
was checking mail to be sure my old Jack was still here, he is, you
have been kind with him, so I'll give you another issue of Corewarrior
instead of spending the time crying on my bed.:-)
Tuc is working hard to change the scripts of Stormking koth, soon
we'll have new features available, password, detailed status and the
chance to change ;redcode attribute of our warrior from verbose to
quiet, for example. Everyone is invited to submit some sample warrior
to kothtest@valhalla.stormking.com trying any oddity to crash it, this
avoid problems in future; any suggestion for enhancement is also
welcome, mail them to tuc@stormking.com
Stefan is still on vacations so we cannot yet see Steven's, Robert's,
and Paul's, tournament winner and medalists, faces on the web; I'm
sure all corewar girls are waiting.
And now stop rambling, let the dat bombs speak for me.
______________________________________________________________________________
Current Status of the Internet Pizza Server ICWS '94 Draft Hill:
# %W/ %L/ %T Name Author Score Age
1 41/ 38/ 20 quiz Schitzo 144 341
2 39/ 36/ 24 Harmony P.Kline 142 11
3 39/ 37/ 24 Frontwards v2 Steven Morrell 140 194
4 33/ 31/ 36 Mason 2.0 Robert Macrae 134 77
5 19/ 6/ 74 Evol Cap 6.2a John Wilkinson 133 8
6 23/ 13/ 64 Impfinity v4g1 Planar 132 141
7 25/ 19/ 56 juliet storm M R Bremer 132 31
8 26/ 20/ 54 La Bomba Beppe Bezzi 132 336
9 18/ 5/ 77 Evol Cap 6.2 John Wilkinson 132 25
10 23/ 14/ 63 Hector 2 Kurt Franke 131 256
11 28/ 24/ 49 Jack in the box Beppe Bezzi 131 708
12 24/ 17/ 59 juliet and paper M R Bremer, B. Bezzi 131 337
13 30/ 30/ 40 Torch t18 P.Kline 130 720
14 37/ 45/ 18 seventyfive Anders Ivner 130 50
15 27/ 24/ 48 endpoint . M R Bremer 130 13
16 35/ 41/ 24 Boombastic Maurizio Vittuari 129 76
17 36/ 44/ 20 Provascan 3.0 Beppe Bezzi 127 91
18 17/ 7/ 76 Night Train Karl Lewin 126 228
19 29/ 32/ 39 Tornado 2.0 h1 Beppe Bezzi 125 181
20 19/ 14/ 67 myOmy Paulsson 125 5
21 16/ 7/ 77 DoorMat v0.1 K Lewin 125 190
22 18/ 12/ 70 Cheap Hack M R Bremer 125 28
23 35/ 54/ 11 itest ai 116 4
24 21/ 36/ 43 test9 Maurizio Vittuari 105 1
25 28/ 55/ 17 John's Zizzor 2g John Wilkinson 100 7
The hill aged 74 this week with 11 new entries. quiz is always holding
the king position even if less strongly than in the short past; having
given the top position to Frontwards, Impfinity and Harmony for short
periods.
After Aeka we have a new resubmission of an old warrior, juliet storm,
this time by its author, she holds a good 7th position.
Hill scores are much lower, average 128 (136 last week). Average age
is 162 (168 last week)
______________________________________________________________________________
94 - What's New (in order of appearance)
5 43/ 45/ 12 seventyfive Anders Ivner 140 1
14 30/ 25/ 45 juliet storm M R Bremer 134 1
24 21/ 16/ 63 Cheap Hack M R Bremer 126 1
9 19/ 4/ 77 Evol Cap 6.2 John Wilkinson 134 1
24 29/ 31/ 40 endpoint . M R Bremer 127 1
2 40/ 40/ 20 Harmony P.Kline 141 1
5 19/ 5/ 77 Evol Cap 6.2a John Wilkinson 132 1
25 28/ 55/ 17 John's Zizzor 2g John Wilkinson 100 1
18 19/ 13/ 68 myOmy Paulsson 125 1
24 36/ 53/ 11 itest ai 118 1
24 21/ 36/ 43 test9 Maurizio Vittuari 105 1
Many new entries this week, above all new version of Harmony, after
some dissonant trial, by Paul Kline; other entries worth noting are
the fast bomber seventyfive by Anders Ivner, the two Evol Caps by John
Wilkinson and the evergreen juliet storm by Myer Bremer.
______________________________________________________________________________
94 - What's No More
26 0/ 1/ 3 testPW Maurizio Vittuari 3 9
26 4/ 0/ 0 Dynamo v.1a John K. Wilkinson 11 8
26 1/ 3/ 0 Provascan 2.0d B.Bezzi 3 4
26 2/ 1/ 1 Harmony P.Kline 7 22
26 35/ 47/ 19 Qwiksand Wayne Sheppard 122 99
26 34/ 43/ 22 myVamp v3.7 Paulsson 125 643
26 23/ 21/ 56 Aeka T.Hsu 124 75
26 36/ 47/ 18 Porch Swing 2 JKW 125 226
26 34/ 44/ 22 Persistence Kurt Franke 125 136
26 32/ 40/ 27 Derision M R Bremer 124 351
26 33/ 50/ 17 crow Karl Lewin 117 167
The leading trio is broken, myVamp is no more at the venerable age of
643, 9th in the 94 hall of fame.
Other losses include Derision at 351, Porch swing 2 at 226, two
centenaries crow and Persistance, a near centenary, Qwiksand, and a
bunch of warriors killed by their authors. The old Aeka, whose
resubmission caused a few discussions, is no more a problem, at least
in 94 hill.
______________________________________________________________________________
What's Old
13 30/ 30/ 40 Torch t18 P.Kline 130 720
11 28/ 24/ 49 Jack in the box Beppe Bezzi 131 708
1 41/ 38/ 20 quiz Schitzo 144 341
12 24/ 17/ 59 juliet and paper M R Bremer, B. Bezzi 131 337
8 26/ 20/ 54 La Bomba Beppe Bezzi 132 336
10 23/ 14/ 63 Hector 2 Kurt Franke 131 256
18 17/ 7/ 76 Night Train Karl Lewin 126 228
The warrior passing 200 age are still seven, but they are no more the
same. myVamp and Derision are not more, replaced by Hector and Night
Train.
quiz, juliet and paper and La Bomba are all very near entering hall of fame.
_____________________________________________________________________________
HALL OF FAME
* means the warrior is still running.
Pos Name Author Age Strategy
1 Iron Gate 1.5 Wayne Sheppard 926 CMP scanner
2 Agony II Stefan Strack 912 CMP scanner
3 Blue Funk Steven Morrell 869 Stone/ imp
4 Thermite 1.0 Robert Macrae 802 Qscan -> bomber
5 Blue Funk 3 Steven Morrell 766 Stone/ imp
6 Torch t18 P.Kline 720 * Bomber
7 Jack in the box Beppe Bezzi 708 * P-warrior
8 HeremScimitar A.Ivner,P.Kline 666 Bomber
9 myVamp v3.7 Paulsson 643 Vampire
10 Armory - A5 Wilkinson 609 P-warrior
11 Phq Maurizio Vittuari 589 Qscan -> replicator
12 B-Panama X Steven Morrell 518 Stone/ replicator
13 NC 94 Wayne Sheppard 387 Stone/ imp
14 Cannonade P.Kline 382 Stone/ imp
15 Torch t17 P.Kline 378 Bomber
16 Lucky 3 Stefan Strack 355 Stone/ imp
17 Derision M R Bremer 351 Scanner
18 Request v2.0 Brant D. Thomsen 347 Qvamp -> vampire
19 Dragon Spear c w blue 346 CMP scanner
20 Leprechaun on speed Anders Ivner 344 Qscan -> scanner/bomber
myVamp leave the venerable trio a little before establishing the
Swedish top score, Torch and Jack pass HeremScimitar and 700 in good
health, next target Blue Funk 3.
Bremer's Derision enters the hall in 17th position pushing juliet storm off.
______________________________________________________________________________
Current Status of the Internet Pizza Server Beginner's Hill:
# %W/ %L/ %T Name Author Score Age
1 61/ 26/ 12 Koolaid II: The Wrath of David Boeren 196 11
2 48/ 14/ 38 Spacehead1.1 Warpi 182 72
3 45/ 12/ 43 Spacehead Warpi 178 83
4 55/ 31/ 14 Our Vamp v2 R Bartolome & JS Pul 178 60
5 38/ 14/ 48 Thunder V 1.0 Andy Nevermind 163 7
6 42/ 31/ 27 Winter Werewolf W. Mintardjo 153 52
7 43/ 36/ 21 Paperstone Warpi 151 87
8 32/ 17/ 51 Spiral Again Calvin Loh 147 1
9 27/ 12/ 62 GCSpiral Example 142 53
10 26/ 14/ 59 SpiralTest2 Calvin Loh 139 10
11 36/ 36/ 28 Paperstone v1.2 Warpi 137 84
12 43/ 49/ 7 polyphemous Kafka 137 94
13 36/ 39/ 25 RedPixel.2 John Lewis 134 8
14 33/ 35/ 32 Everybody Must Get Stoned Nathan Summers 132 96
15 33/ 36/ 30 Everybody Must Get Vamped Nathan Summers 131 19
16 37/ 43/ 20 Slime 1.1 Anton Marsden 130 12
17 35/ 45/ 20 Test Stone #1 David Boeren 125 27
18 31/ 38/ 30 EveryoneMustGetStoned 2.5 Nathan Summers 125 18
19 33/ 43/ 24 test-A Raul Bartolome 124 34
20 35/ 45/ 20 Czar Master 3.0 Dark Scneider 124 21
21 33/ 43/ 24 test-b Raul Bartolome 122 35
22 30/ 38/ 32 Forked Lite Ning 4.02 Ansel Greenwood Serm 121 9
23 29/ 41/ 30 Forked Lite Ning 4.06 Ansel Greenwood Serm 117 3
24 29/ 42/ 30 Craw IIe Bach 116 2
25 25/ 39/ 35 Forked Lite Ning 4.03 Ansel Greenwood Serm 111 5
Aeka and Rave have been retired and now the hill is dominated by
Boren's Koolaid II that, if I'm not wrong, did the 94 hill for a while
too. Go on trying, if you made it once you are not so far.
______________________________________________________________________________
The hint Steps
Bombing and scanning steps are an important choice to make when coding
a warrior; the right choice may be the thing that change your score of
many positions in the hill, and require a bit of wisdom.
Having to bomb, for example, every fourth cell of core (modulo 4) we
can choose any number with 4 as greater common divisor with Coresize,
but results are greatly different if we choose a bad or a good one.
Using four as step against a twelve lines long opponent, makes us bomb
it thrice every time we bomb it once, a sort of overkill that's not
needed in most cases; our oponent instead, using a smarter step and
better spreading its bombs may bomb us but once three times as
often. Result: 75/25 for him :-(
Number theory help us in our choice, even if, as usual in corewar,
there is not a best choice but a sort of tradeoff.
Steven Morrell so speaks in its book:
----
One factor that could mean the difference between a top-rate stone and an
unsuccessful stone is the choice of step size. The program that manages to
bomb the enemy first has a decided advantage, and some bombing step sizes
are more efficient at scanning for the enemy than others. So what makes a
good step size?
Ideally, it ought to hit every location in 8000 bombs, every other location
in 8000/2=4000 bombs, every third location in 8000/3=2667 bombs, etc.
Unfortunately, this is impossible, especially with a single step size, but
it suggests a basic strategy -- go for the biggest programs first and then
fill in the gaps.
One way of rating the efficiency of a step size is to find the length of
the largest unbombed section of code after each bomb is dropped. By adding
up all of these lengths, we get a number that tells us how big an average
gap is. (Indeed, by dividing this number by the number of bombs dropped,
we get the average gap size.) If we minimize this number over all step
sizes, we get the "Optima Numbers." For a coresize of 8000, these optima
numbers are:
mod-1 3359/3039 under-100 -> 73
mod-2 3094/2234 under-100 -> 98
mod-4 3364/3044 under-100 -> 76
mod-5 3315/2365 under-100 -> 95
mod-8 2936/2376
mod-10 2930/2430
Another common rating is how closely to in half the new bomb subdivides the
old gap when it is dropped. By taking the differences between where the
bombs fall and the middle of each gap and adding these distances up, we get
an alternate method for testing efficiency.
Both of these methods are useful for finding general-purpose step sizes.
But suppose you wanted to find a step size optimized for killing other
stones. Since stones usually have four or five instructions, you would
want a step size that would bomb every 4th and 5th instruction quickly,
regardless of how it does in general.
Fortunately, there is a program in the public domain that calculates all
of these things quicky:
Corestep by Jay Han can be found as
misc/corestep.c, and calcutates optima numbers and optimal step sizes.
You will need a C compiler to use it, but it is otherwise self-contained.
For more infromation, FTP a copy and read through it. The classic formula
calculates optima numbers, the alternate formula calculates the sum of the
distances between bombs and midpoints, and find-X calculates optimal step
sizes against a specific program length.
If you don't have access to a C compiler or want this for some other
reason, P. Kline has compiled a list of all 8000 step-sizes with their mod,
find-4, find-5, find-10, and find-13 numbers, along with imp-killing
constants and imp-numbers. This table is designed for use in spreadsheets
or databases. It is available in the misc/ directory under the name num8000.txt
---
If you look at the steps of published warriors you can notice that the
optima numbers are not always used and smaller steps are usually
preferred; the reason of this is mainly because there is no nedd to
quickly find a 500 lines long target and a small step can be used as a
gate, or as pointer in the clear, saving some space.
How to choose, beetween some good number, the one to use is often a
matter of choice. The modulo has to be chosen according to the size of
our warrior, in such a way as not to bomb ourself; once chosen it we
have to select a target size: tiny (5 lines), small (around 10 lines),
average (around 13/15) and select a number with a good find for our
target, and an average one for the others. In our choice we have to
consider the mod 1 numbers with a good find for our selected
modulo. They have bombing patterns approximating closely the one we
want, and usually end bombing the cell immediately before or after the
one we start with.
_____________________________________________________________________________
Planar's corner
This is Impfinity v4g1, the latest (and first truly successful)
version of my stone/imp.
The general organization is the same as v3i: a pair of stones and a
pair of continuous-launching spirals on top of each other. The stones
haven't changed either.
What has changed is the pumps: I've implemented the idea of
self-priming pumps that I got while writing for Core Warrior 7. This
replaces a 3-instruction pump and 4-instruction primer with a
4-instruction pump.
But the most important change is the length of the fuse. In
Impfinity v3i, I start the imp-launcher, wait for a while, then start
the stones. The "wait for a while" part is what I call the fuse. I
experimented with the fuse length. A shorter fuse will make the
spiral weaker, thus giving up more losses, but it will also make the
bombing more aggressive, thus bringing in more wins. Against my test
suite, the shorter fuse was always better.
Finally, instead of having the pump start the stone after a while, I
made the boot code start the stone right after the pump, jumping to
the loop body instead of the SPL itself: the closer to the JMP, the
earlier the SPL will start accumulating processes (the shorter the
fuse). Again, the shortest fuse is the best one.
Anyway, this is Impfinity v4g1, seen as low as 11th and as high as 1st
on the hill. There was a lot of movement on the hill recently. Now
that the papers seem to be back, I don't think Impfinity will fare so
well. This is the exact code that I sent to the hill. Try and abuse
the boot constants if you dare !
-- Planar
;redcode-94
;name Impfinity v4g1
;author Planar
;strategy boot,imp,stone,clear
;strategy With self-priming pump and very short fuse.
;assert CORESIZE == 8000
;kill Impfinity
istep equ 1143
bstep01 equ 2214
bstep02 equ 3285
magic01 equ (2)
magic02 equ (-1)
trash equ (Z-285)
impoff equ (Z+437)
pmpof01 equ (impoff+1*istep-577)
pmpof02 equ (impoff+4*istep-496)
stnof01 equ (impoff+3*istep-215)
stnof02 equ (impoff+6*istep-158)
org boot
Z
boot spl boot02
i FOR 2
boot&i
j FOR 4
mov.i <psrc&i, <pdst&i
mov.i }psrc&i, }pdst&i
ROF
mov.i *psrc&i, *pdst&i
mov.i instr&i, impoff+(i-1)*istep
spl @pdst&i, >trash-15-i*2
jmp *pdst&i, >trash-i*2
psrc&i dat bomb&i, pend&i
pdst&i dat stnof&i, pmpof&i+pend&i-pump&i
point&i equ (pump&i-pmpof&i+impoff+(i-1)*istep)
pump&i spl #1, >prime&i
ptr&i spl pump&i-pmpof&i+impoff-istep-1, {335+i
add.f #istep+1, ptr&i
prime&i mov.i point&i, point&i-2
pend&i
instr&i mov.i #1, istep
bomb&i dat <2667, <5334
stone&i spl #stone&i+bstep&i+magic&i, {-650
mov.i bomb&i, }stone&i
add.f #bstep&i-1, stone&i
j&i djn.f stone&i, <i-50
send&i
j FOR i==1
k FOR 17
mov.i #1, 1
mov.i #1, @1
spl #1, 1
ROF
ROF
ROF
end
______________________________________________________________________________
Questions? Concerns? Comments? Complaints? Mail them to:
Beppe Bezzi <bezzi@iol.it>, Myer R Bremer <bremermr@ecn.purdue.edu> or
Maurizio Vittuari <pan0178@iperbole.bologna.it>
|
