Issue 25 April 15, 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
Newbies should check the stormking page for the FAQ, language specification,
guides, and tutorials. Post questions to rec.games.corewar. All new players
are infinitely welcome!
If ftp.csua.berkeley.edu is still out of service, you can download Pmars at
Terry's web page--http://www.infi.net/~wtnewton/corewar/
______________________________________________________________________________
Greetings.
Not many news from the hill this week, but the fall of the 1200+ old Evol
cap, but some interesting code: my Tornado 3.0 and Stepping Stone by Kurt
Franke; I hope mant other authors will publish their ones.
There has been some discussion in the newsgroup about self fights, and not
one spoke if favor for keeping them. Is there a good reason, apart the
historical one, to go on with self fights?
--Beppe Bezzi
______________________________________________________________________________
Current Status of the Internet Pizza Server ICWS '94 Draft Hill:
Hill Specs:
coresize: 8000
max. processes: 8000
duration: after 80,000 cycles, a tie is declared.
max. entry length: 100
minimum distance: 100
rounds fought: 200
instruction set: ICWS '94 Draft
# %W/ %L/ %T Name Author Score Age
1 47/ 33/ 20 Mirage 2 Anton Marsden 161 96
2 48/ 37/ 15 Scanny Boy David van Dam 158 160
3 44/ 33/ 23 Thermite II Robert Macrae 155 1330
4 39/ 24/ 37 Walk Like An Egyptian John K W 154 11
5 46/ 39/ 15 Stepping Stone Kurt Franke 154 143
6 43/ 36/ 21 Chameleon M R Bremer 150 801
7 43/ 35/ 22 Grilled Octopus v0.5 David Boeren 150 546
8 43/ 36/ 21 Gem of the Ocean P.Kline 150 199
9 31/ 15/ 53 Hazy Shade II John K W 147 289
10 37/ 28/ 35 Blue Funk 5 Steven Morrell 147 474
11 34/ 21/ 45 Impfinity v4g1 Planar 147 1529
12 41/ 38/ 21 Blanket Party J E Long 145 54
13 29/ 14/ 57 EvolCap XI John Wilkinson 144 39
14 44/ 45/ 11 Wind-up Toy v0.7 Ian Oversby 143 10
15 36/ 28/ 36 Barrage Anton Marsden 143 803
16 39/ 36/ 25 prova Maurizio Vittuari 141 306
17 36/ 32/ 32 Flurry Anton Marsden 141 291
18 41/ 42/ 17 Harmony P.Kline 141 15
19 37/ 33/ 31 Lithium John K Wilkinson 141 337
20 37/ 34/ 29 Tornado 3.0 Beppe 139 1131
21 41/ 43/ 16 Wind-up Toy v0.6 Ian Oversby 138 23
22 37/ 39/ 24 T.N.T. Maurizio Vittuari 134 1055
23 4/ 0/ 0 Test 3 Ian Oversby 12 2
24 4/ 0/ 0 Test 4 Ian Oversby 12 1
25 4/ 0/ 0 testest M R Bremer 12 3
Weekly age: 87
New warriors 9 Turnover/age rate 10%
Average age: 386 ( 397 last week, 401 the week before )
Average score: 130 ( 140 last week, 141 the week before ) There are 3 killed
warriors on hill bottom.
The top 25 warriors are represented by 14 authors, JKW and Oversby with 4,
Marsden, with 3.
After dominating the hill last week, Scanny boy loses king position at the
end of this week. There has been a lot of fight for the best spot: Scanny
boy, Mirage 2, Impfinity, Thermite, Blue Funk 5, Grilled octopus all have
been seen in the top.
Pspacer are losing a bit against standard warriors, maybe all new
submissions have some brainwash component to keep'em at bay; perhaps next
week we'll see if this is a new trend or but a fluctuation.
Scanners and qscans are still dominating and replicators are near extiction
having lost both Hector and Marcia, only Barrage remains to represent them.
Reading through ;strategy lines I noticed there are a lot of references to
previous versions, without telling what the warrior is doing; I know what
Blue Funk and Thermite are, as much as most of us, but a beginner maybe puzzled.
______________________________________________________________________________
94 - What's New
# %W/ %L/ %T Name Author Score Age
2 33/ 23/ 44 Walk Like An Egyptian John K W 144 1
16 34/ 38/ 28 Blanket Party J E Long 131 1
23 22/ 10/ 69 EvolCap XI John Wilkinson 134 1
14 44/ 45/ 11 Wind-up Toy v0.7 Ian Oversby 143 10
I haven't got this entry
8 39/ 39/ 22 Harmony P.Kline 139 1
18 38/ 43/ 19 Wind-up Toy v0.6 Ian Oversby 132 1
22 38/ 47/ 15 Test 3 Ian Oversby 130 1
25 21/ 78/ 1 Test 4 Ian Oversby 64 1
16 41/ 47/ 13 testest M R Bremer 134 1
Two entries in the top ten: Walk like an Egyptian, a pwarrior by JKW, and
Harmony by Paul Kline.
______________________________________________________________________________
94 - What's No More.
# %W/ %L/ %T Name Author Score Age
26 39/ 51/ 10 Memories Beppe Bezzi 126 159
26 32/ 38/ 31 A Hard Day's Night Maurizio Vittuari 125 289
26 31/ 48/ 22 No Prisoners John K W 114 2
26 21/ 14/ 65 Evol Cap 6.6 John Wilkinson 129 1299
26 0/ 0/ 4 Marcia Trionfale Beppe Bezzi 4 29
26 32/ 38/ 30 Eternal Flame John K W 126 128
26 30/ 34/ 35 test P.Kline 126 79
26 29/ 32/ 38 Hector 4 Kurt Franke 126 2
26 C I A Anders Ivner ???
I haven't got the message with C I A push off
Another millenary warrior leaves us this week: Evol cap 6.6 by JKW
______________________________________________________________________________
94 - What's Old
# %W/ %L/ %T Name Author Score Age
19 27/ 18/ 54 Impfinity v4g1 Planar 137 1529
3 41/ 34/ 25 Thermite II Robert Macrae 147 1330
12 37/ 32/ 31 Tornado 3.0 Beppe 142 1131
10 39/ 35/ 26 T.N.T. Maurizio Vittuari 143 1055
7 34/ 25/ 41 Barrage Anton Marsden 144 803
5 41/ 35/ 24 Chameleon M R Bremer 147 801
6 41/ 37/ 22 Grilled Octopus v0.5 David Boeren 146 546
24 32/ 31/ 37 Blue Funk 5 Steven Morrell 134 474
16 39/ 36/ 25 prova Maurizio Vittuari 141 306
T.N.T enters the millenaries club
______________________________________________________________________________
HALL OF FAME
* means the warrior is still active.
Pos Name Author Age Strategy
1 Jack in the box Beppe Bezzi 1620 P-warrior
2 Torch t18 P.Kline 1539 Bomber
3 Impfinity v4g1 Planar 1529 * Stone/ imp
4 Frontwards v2 Steven Morrell 1420 One shot scanner
5 Thermite II Robert Macrae 1330 * Qscan -> bomber
6 quiz Schitzo 1262 Scanner/ bomber
7 Evol Cap 6.6 John Wilkinson 1299 Imp / stone
8 Tornado 3.0 Beppe Bezzi 1131 * Bomber
9 T.N.T. Maurizio Vittuari 1055 * Bomber
10 Iron Gate 1.5 Wayne Sheppard 926 CMP scanner
11 Agony II Stefan Strack 912 CMP scanner
12 Blue Funk Steven Morrell 869 Stone/ imp
13 Barrage Anton Marsden 803 * Qscan -> replicator
14 Thermite 1.0 Robert Macrae 802 Qscan -> bomber
15 Chameleon Myer R Bremer 801 * P-warrior
16 Blue Funk 3 Steven Morrell 766 Stone/ imp
17 Night Train Karl Lewin 755 Replicator
18 Mirage 1.5 Anton Marsden 736 Scanner/ bomber
19 Blizzard Anton Marsden 713 Qscan -> replicator
20 HeremScimitar A.Ivner,P.Kline 666 Bomber
21 La Bomba Beppe Bezzi 650 Qscan -> replicator
22 myVamp v3.7 Paulsson 643 Vampire
23 Hazy Shade Of Winter John Wilkinson 616 P-warrior
24 Armory - A5 John Wilkinson 609 P-warrior
25 juliet and paper Bremer & Bezzi 607 P-warrior
Evol cap 6.6 end its run at 1299 in 7th position. Impfinity is now very
close to Torch and at reach of Jack in the box for the Kothof (King of the
hall of fame)
______________________________________________________________________________
Current Status of the Internet Pizza Server Beginner's Hill:
Hill Specs:
coresize: 8000
max. processes: 8000
duration: after 80,000 cycles, a tie is declared.
max. entry length: 100
minimum distance: 100
maximum age: At age 100, warriors are retired.
rounds fought: 200
instruction set: ICWS '94 Draft
Last challenge: Mon Apr 15 06:36:53 PDT 1996
# %W/ %L/ %T Name Author Score Age
1 47/ 16/ 37 Vectorfantasy v0.3 basehead 177 51
2 54/ 34/ 12 Wind-up Toy v0.5 Ian Oversby 174 26
3 48/ 24/ 29 Fork v0.2-9p/51b Christoph C. Birk 171 54
4 51/ 38/ 11 Violent Micro v0.3 basehead 164 39
5 50/ 39/ 12 C-Seagal II 1.02 jShelton 161 22
6 49/ 39/ 12 look-see Anonymous 160 84
7 44/ 34/ 21 Ripped v0.1 Ian Oversby 154 10
8 40/ 26/ 34 Pro-Bomber v0.3 David van Dam 154 1
9 45/ 42/ 12 Gunslinger 2 Julian 148 9
10 37/ 28/ 34 Hyakutake Perihelion Scott Manley 146 91
11 39/ 35/ 26 Rabbit v0.2a Christoph C. Birk 143 57
12 41/ 41/ 18 AK-47 Julian 140 14
13 44/ 48/ 9 Uni Charm Password jShelton 140 30
14 40/ 42/ 17 AK-47.1 Julian 138 13
15 39/ 41/ 20 why? V 0.3 bjoern guenzel 138 2
16 39/ 49/ 12 Optma Dwarf v5.3 Phoenix 129 37
17 30/ 30/ 40 Toxin v0.7 Edgar 129 35
18 30/ 45/ 26 Antibodies Mk II Andrew Fabbro 115 61
19 30/ 46/ 24 JuJu 7 Andrew Fabbro 114 18
20 30/ 48/ 22 Napalm Iain Hogg 111 33
21 32/ 61/ 8 Whoops. Didn't mean to se JKW 102 15
22 9/ 4/ 2 Anti Imp V1.1 bjoern guenzel 30 34
23 6/ 5/ 4 smoju 1 (another pun ...) bjoern guenzel 24 6
24 2/ 4/ 2 why? V 0.4 bjoern guenzel 7 4
25 1/ 1/ 2 Pro-Bomber v0.2 David van Dam 5 17
basehead's Vectorfantasy is the leader, followed by Oversby's Wind up toy,
whose brother v07 is in the pro hill.,Corewarrior list
Not a lot of movement here this week, but 37 challenges; just a lack of new
players or the b-hill is too hard to make?
______________________________________________________________________________
The Hint
Tornado 3.0 by Beppe Bezzi
As announced here is Tornado 3.0a, exacly as is on the 94 hill, after all
the small flaw I was afraid of isn't that easy to spot, after all I spent
more time on it than anyone else.
It's a 60%c using, OTOH, tornado engine. Main differences from Jack's
version, even if it's perhaps easier to find resemblances, :-) are:
- It's self splitting, for more robustness.
- Uses djn stream for further protection against one shot scanners.
- Has a multipass external clear
- Uses a decoymaker instead of the boot
- Uses a mov {step,1 bomb instead of a simple dat.
Tornado enters clear self hitting with a spl #0 bomb and goes further on
bombing, slower, until the the clear starves the bomber, a protection
against Clisson, even if Clisson were far from coming at Tornado birth.
The step may look non optimal, but it's the best scoring one I found
beetween all 'not too large' mod 5 numbers, apart that I lost my faith in
optimal number when a version of Memories with a step optimal for coresize
55440, an error using Corestep, scored better than any other optimal for
coresize 8000 on the 94 hill. Maybe Steven, our optima number theorist,
knows the reason.
----
;redcode-94
;name Tornado 3.0
;author Beppe
;assert CORESIZE==8000
;strategy Fast 60% c bomber
;strategy v 3.0a new clear, paper and imp protection
;strategy decoymaker instead of boot
;assert CORESIZE == 8000
;kill Tornado
dms equ 25
dmd equ -2500
step equ -45
djnoff equ -2205
;---
A equ 43
B equ 24
a for 13
mov >dmd-dms*a,{dmd-dms*(a+15)
rof
;-break
JMP.B bombs, >2975
for MAXLENGTH-CURLINE-A
spl #0,0
rof
gate dat -100, 100
dat <-10, <2667 ;anti imp bit
dat -4000, djmp-gate+2
stclr spl #-3000, djmp-gate+2
clear mov @djmp, >gate
mov @djmp, >gate
djmp djn.b clear, {stclr
dat 0,0
dat 0,0
for B
dat 0,0
rof
bombs spl #step, -step ;hit spl
start1 sub incr, @b1
stone mov (0*step)+jump,*(1*step)+jump
b2 mov bombs, @stone
b1 mov bombm, *stone
jump djn.f start1, {djnoff ;hit by spl
jmp stclr, 0
incr dat >-3*step,>-3*step
last
bombm mov {step, 1
______________________________________________________________________________
Extra Extra
Stepping Stone by Kurt Franke
Stepping Stone is the result of ideas I had while writing System Trap.
The main warrior is a 2/3 c vamp/scan in the spirit of myVamp. The pit is
primarily not for stunning, but to notify my program processes have been
trapped, to brainwash, and to self destruct. The first line of the pit
causes Stepping Stone to jump to the core clear.
The quick scan has an interesting attack. It is a 3/5 c vamp using
tornado style bombing. Only one instruction is a jump back to the trap.
All the other jmp's go indirectly back to the original one pointing
to the trap. This isn't practical for a general attack, but seems to
work ok for the qscan attack. The quick scan vamp has its own pit.
The first line of the qscan pit, when executed, causes Stepping Stone to
go directly to the core clear without booting. Maybe this saves a few
rounds where the boot code or vampire have been hit by a dat, I didn't
really test. What does seem to be true is adding the quickscan
increased my score nearly 10 points. This version scored about 3 or
4 points higher than the handshaking version.
It seems I could eliminate a line by combining the increment
value with the add line to get
add.f #-INCR, $INCR
and then putting jbomb INCR places forward. I tried this but it didn't
work well (don't you hate it when scores don't improve after making what
really seems to be an improvement?). I'll look at it again later perhaps.
I have made other minor "improvements", but this is the one on the hill.
Oh, except I changed the qscan places slightly so you can't murder me by
reversing them.
- Kurt
/-----------------
;redcode-94
;assert CORESIZE == 8000
;author Kurt Franke
;name Stepping Stone
;kill Stepping Stone
;strategy Quick scan -> Vampire (version 9h)
;components (to help Planar) :
; quickscan, boot, vamp, scan, stun, clear, pspace(brainwash)
;; - - - - - - - - - - - boot parameters - - - - - - - - - - - - - - - -
;; reference address for boot distances
REF equ (vamp-3)
;; first define the places things will go
BVAMP equ (3+282*10+REF)
BSTEP equ (6+295*10+REF)
BJUMP equ (1+298*10+REF)
BCLEAR equ (4+367*10+REF)
BPIT equ (5+67*10+REF)
;; ^
;; +---- don't change this column of numbers
;; they are magic to avoid self-bombing
;; now define constants for use in the code
;; (these are to be used when refering to a
;; line within that block of code)
CVAMP equ (BVAMP-vamp)
CCLEAR equ (BCLEAR-dbomb2+space)
CTARGET equ (BVAMP-vamp+target)
ADJUST equ (CCLEAR-CTARGET)
FIRST equ 0
INC equ 110 ;; mod 10 step for mod 5 vamp/scan
COUNT equ 8*5
;; - - - - - - - - - - - the vampire code - - - - - - - - - - - - - - -
vamp add.f $vamp+6, $(BJUMP-CVAMP) ;; incr will be at vamp+6
mov.i @0, @(BJUMP-CVAMP)
jmz.a *target, *(BJUMP-CVAMP)
mov.i @0, *(BJUMP-CVAMP) ;; a-field needs to be 0
target djn vamp, #COUNT
jmp (CCLEAR-CVAMP)
step jmp @0, BPIT-BSTEP ;; a-field needs to be 0
jbomb jmp @BSTEP-BJUMP-FIRST, $FIRST ;; indirect jump to pit
CPTR equ (dbomb2-1)
dbomb2 dat <2667, <2*2667+1
dbomb dat <-15, #10
space spl #0, #10 ;; a-field needs to be 0
clear mov @cloop, >CPTR
mov @cloop, >CPTR
cloop djn.b $clear, {space
incr dat $-INC, $INC
pit mov.a #ADJUST, $CTARGET-BPIT
spl #0, $0 ;; a-field needs to be 0
mov 2, >1
stp #0, $BCLEAR-BPIT+36
;; - - - - - - - - - - - - - boot code - - - - - - - - - - - - - - - - - -
pboot mov $pit+3, BPIT+3
for 3
mov {pboot, <pboot
rof
cboot mov $dbomb2+5, BCLEAR+5
for 5
mov {cboot, <cboot
rof
mov $step, BSTEP
mov $jbomb, BJUMP
mov $incr, BVAMP+6
vboot mov $vamp+5, BVAMP+5
for 5
mov {vboot, <vboot
rof
mov $0, $cboot
spl @vboot
mov $0, $vboot ;; erase important pointers
dat $0, $0
;; - - - - - - - - - - - - - quick scan - - - - - - - - - - - - - - - - - -
QINC equ 6
QCOUNT equ 7
qjump jmp $qpit-found, #0
qstep1 jmp *QINC, #-2*QINC
qstep2 jmp *-2*QINC, #-QINC
start ;'94 scan (from FAQ)
s1 for 4
sne.x start+400*(s1+12), start+400*(s1+12)+100
seq.x start+400*(s1+12)+200, start+400*(s1+12)+300
mov #start+400*(s1+12)-found, $found
rof
jmn which, $found
s2 for 4
sne.x start+400*(s2+6), start+400*(s2+6)+100
seq.x start+400*(s2+6)+200, start+400*(s2+6)+300
mov #start+400*(s2+6)-found-100, $found
rof
jmn which, $found
s3 for 4
sne.x start+400*(s3+2), start+400*(s3+2)+100
seq.x start+400*(s3+2)+200, start+400*(s3+2)+300
mov #start+400*(s3+2)-found-100, $found
rof
jmz pboot, $found
add #100, $found
which jmz.f -1, @found
sub.ba $found, $qjump
mov $qjump, @found
mov.ba $found, $found
qloop sub.f $qstep1, $found
mov $qstep2, @found
mov $qstep1, *found
found mov $0, @0
djn qloop, #QCOUNT
launch jmp pboot, #space
qpit mov.ba $launch, $launch
qerase spl #0, #start
mov $10, >qerase
jmp BPIT
;; - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
end start
______________________________________________________________________________
Questions? Concerns? Comments? Complaints? Mail them to people who care.
authors: Beppe Bezzi <bezzi@nemo.it> or Myer Bremer <bremermr@ecn.purdue.edu>
|
