Issue 36 July 1, 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 unreachable, you can download pMARS at:
Terry's web page--http://www.infi.net/~wtnewton/corewar/
Planar fpt site--ftp://ftp.inria.fr/INRIA/Projects/para/doligez/cw/pmars
Fechter ftp site--ftp://members.aol.com/ofechner/corewar
______________________________________________________________________________
Greetings.
At least some of the hill changes have been implemented at Pizza. Check out
all the spiffy decimals on the hill reports for greater precision. Apologies
to our European players who prefer commas. It will definitely take some
getting used to. If you have any comments or requests for new features,
mail them to Thos at <sd@ecst.csuchico.edu>. In case you missed them, I have
kept the 'New Pizza Features' seen in the last issue.
Personally, I was hoping the ;test option could be used to fight individual
warriors in addition to challenging the entire hill. Otherwise, I fully
support all the changes proposed.
Just a reminder that the 'limited process hill' is open. Send your warriors
to <koth@wastedyouth.us.itd.umich.edu> with a ;redcode-94lp header. To get
more information, send an e-mail with a subject line of 'help' (no quotation)
marks to the above address. Thanks to J K Lewis and Andrew Fabbro for their
work on the new hill.
Next week, we will have a very special issue brought to you by Anton Marsden.
Check out some of his work in this week's hint.
Before anyone yells at me, I just noticed that the hill reports Bezzi has
been feeding me are _very_ out of order my apologies for any confusion. It's
not Beppe's fault. I need to get back on the hill soon.
--M R Bremer
______________________________________________________________________________
New Pizza features
o Remove self fights on all hills that use pspace.
o Add the "test" command to decrease artificial aging of the hill.
I think the best (and easiest) way to do this would be as an
argument to the ";redcode" command like "quiet" and "verbose" are
now. ";redcode test" would challenge all the warriors on the hill
(only 100 times?), but the results would be discarded. This way
you could see how your changes affect your performance against all
the warriors on the hill, but you get the results back twice as
fast.
o Add the ";password" command, to make it harder to forge ";kill"
commands.
o Make ";name" commands mandatory, and make blank ";kill" commands
illegal.
o Add the ";url" and ";show" commands for web access. ";url" is
pretty much self explanatory, but ";show" will be used to customize
how much info on your warrior you want to make public. "[no]source"
and "[no]score" are two possible arguments to add.
o Add the ";change [warrior name]" command to allow you to change info
in your warrior without having to kill it and resubmit it. If you
wanted to make it "quiet" while you go on vacation, or change the
";url", just send one piece of mail with ";change" in it. This would
do a substring match like ";kill", and any other commands in the mail
would be changed in the warrior. Does this make sense?
Any other ideas, or comments on the above proposals?
Thos
______________________________________________________________________________
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 42.8/ 39.9/ 17.3 Blur 2 Anton Marsden 145.7 2
2 41.7/ 41.8/ 16.5 myVamp5.3 Paulsson 141.5 11
3 37.6/ 34.3/ 28.1 Goliath David van Dam 140.9 3
4 38.8/ 38.0/ 23.3 Yogi Bear P.Kline 139.5 59
5 40.4/ 43.9/ 15.7 Goldfinch P.Kline 137.0 4
6 37.4/ 39.4/ 23.2 Twister Beppe Bezzi 135.5 309
7 36.0/ 37.3/ 26.7 T.N.T. pro Maurizio Vittuari 134.8 573
8 36.0/ 37.5/ 26.5 test Anonymous 134.6 20
9 35.8/ 37.9/ 26.3 Thermite II Robert Macrae 133.6 2194
10 26.7/ 20.1/ 53.2 blue flame c2 bjoern guenzel 133.3 23
11 32.0/ 31.2/ 36.8 Simple v0.3 Ian Oversby 132.9 14
12 34.2/ 35.7/ 30.1 Jo Clark II Ian Oversby 132.8 213
13 30.0/ 27.2/ 42.8 Jack in the box II Beppe Bezzi 132.8 212
14 29.1/ 26.8/ 44.1 Armory II John K W 131.3 48
15 29.1/ 27.1/ 43.8 the historian bjoern guenzel 131.1 328
16 38.7/ 46.3/ 15.0 Stepping Stone Kurt Franke 131.1 1007
17 19.1/ 7.2/ 73.7 The Avalanche John Wilkinson 131.0 78
18 27.6/ 24.6/ 47.9 blue candle bjoern guenzel 130.6 442
19 25.5/ 20.6/ 53.9 Rosebud Beppe 130.5 528
20 32.7/ 35.5/ 31.8 test Maurizio 130.0 253
21 29.3/ 29.3/ 41.3 Pulp v0.1 Ian Oversby 129.3 12
22 22.4/ 16.0/ 61.6 test Maurizio Vittuari 128.7 7
23 36.7/ 46.1/ 17.2 HopScotch 1.6 Robert Macrae 127.3 322
24 19.4/ 11.7/ 68.9 ompega Steven Morrell 127.1 206
25 35.0/ 44.8/ 20.2 Yet 3c Justin Kao 125.1 1
Weekly age: 56 ( 108 last week, 67 the week before )
New warriors: 11 Turnover/age rate 20%
Average age: 275 ( 296 last week, 269 the week before )
Average score: 133 ( 127 last week, 133 the week before )
The top 25 warriors are represented by 13 authors: Vittuari, Bezzi, guenzel,
and Oversby with 3 (hogs), and Kline, Macrae, and JKW with 2.
King Report: Blur 2 has had a commanding hold on the hill, but there were
a few surprises this week. Miro by Ivner popped up into first place on its
second submission, and just as quickly, he killed it from the hill.
Interesting strategy. blue flame by guenzel also appeared briefly in the
king position. blue flame c2 is already out, but at the less prestigious
position of 10th place. Kline had both Goldfinch and Yogi Bear on top at
times, but that's no real surprise there. I would hazard a guess that Kline
has had more different warriors on top of the '94 hill than any current
author on the hill. I could be wrong though . . .
Speaking of Kline: his test retinA was seen on the hill tying over 80% when
paired with mirrored imps. One reason I find replicators so frustrating is
that they can't win massively against any other warrior. I believe Evol Cap
used separate bombers to help gain some wins. However, as a p-component,
papers do take losses and turn them into ties.
Heavy pspacer Armory II and Goliath are doing well. Goliath did have Scan
Man, Bomber Boy, and a move bomber in it, but a test with Goliath's brain
and three basic warriors called 'Paper, Scissors, and Stone' actually scored
a bit higher when it was submitted aginst the hill.
There are lots of pspacers and imp heavy programs on the hill. If you want
to make it on, you could utilize a quick scan (but not as effective with the
new qscan detection system) or be extremely imp hostile.
______________________________________________________________________________
94 - What's New
# %W / %L / %T Name Author Score Age
7 27.2/ 20.0/ 52.7 blue flame c2 bjoern guenzel 134.4 1
8 36.7/ 38.4/ 25.0 test Anonymous 134.9 1
13 25.8/ 22.2/ 52.0 Armory II John K W 129.3 1
1 41.7/ 40.6/ 17.7 Blur 2 Anton Marsden 142.7 1
3 41.7/ 42.6/ 15.6 Goldfinch P.Kline 140.8 1
4 37.0/ 34.9/ 28.1 Goliath David van Dam 139.2 1
4 41.7/ 41.7/ 16.7 myVamp5.3 Paulsson 141.7 1
24 28.6/ 29.9/ 41.5 Pulp v0.1 Ian Oversby 127.4 1
16 30.8/ 29.6/ 39.6 Simple v0.3 Ian Oversby 132.1 1
22 22.0/ 15.1/ 62.9 test Maurizio Vittuari 129.0 1
25 35.0/ 44.8/ 20.2 Yet 3c Justin Kao 125.1 1
Half of the top ten warriors are, again, new submissions. Although most are
slight tweaks on old versions.
______________________________________________________________________________
94 - What's No More
# %W / %L / %T Name Author Score Age
26 31.2/ 35.4/ 33.5 * Satura * bjoern guenzel 127.0 12
26 22.2/ 17.7/ 60.3 Hazy Shade II John K W 126.8 1102
26 1.6/ 2.0/ 0.4 Blur 2 Anton Marsden 5.0 6
26 2.1/ 1.7/ 0.1 Goldfinch P.Kline 6.4 11
26 1.3/ 1.6/ 1.0 Goliath David van Dam 4.9 43
26 1.2/ 1.2/ 1.6 myVamp5.1 Paulsson 5.0 53
26 31.0/ 35.0/ 34.2 Thalamus mod Edgar 127.0 203
26 2.0/ 1.1/ 4.6 rhodium v1.3 Martin M. Pedersen 10.6 64
26 29.8/ 37.9/ 32.3 Scotch Broth 1.2 Robert Macrae 121.8 98
26 0.3/ 0.7/ 2.9 test E P.Kline 3.8 19
26 0.2/ 0.2/ 3.6 test jb21 Beppe Bezzi 4.1 15
Hazy Shade II is dead. JKW is in mourning, evidenced by his Armory
submissions . . .
A challenger has arrived on the ICWS '94 Draft hill! Vital statistics:
Program "Armory II" (length 100) by "John K W"
(contact address "jwilkinson@mail.utexas.edu"):
;strategy Alas, poor Hazy, I knew thee well...
The loss of Edgar's Thalamus mod pushed him completely off the hill.
Hopefully he will have a new one on the '94 hill soon.
______________________________________________________________________________
94 - What's Old
# %W / %L / %T Name Author Score Age
9 35.8/ 37.9/ 26.3 Thermite II Robert Macrae 133.6 2194
16 38.7/ 46.3/ 15.0 Stepping Stone Kurt Franke 131.1 1007
7 36.0/ 37.3/ 26.7 T.N.T. pro Maurizio Vittuari 134.8 573
19 25.5/ 20.6/ 53.9 Rosebud Beppe 130.5 528
18 27.6/ 24.6/ 47.9 blue candle bjoern guenzel 130.6 442
15 29.1/ 27.1/ 43.8 the historian bjoern guenzel 131.1 328
23 36.7/ 46.1/ 17.2 HopScotch 1.6 Robert Macrae 127.3 322
6 37.4/ 39.4/ 23.2 Twister Beppe Bezzi 135.5 309
Stepping Stone breaks the 1000 mark, the first vampire to do so.
______________________________________________________________________________
HALL OF FAME
* means the warrior is still active.
Pos Name Author Age Strategy
1 Thermite II Robert Macrae 2194 * Qscan -> bomber
2 Impfinity v4g1 Planar 1993 Stone/ imp
3 Jack in the box Beppe Bezzi 1620 P-warrior
4 Tornado 3.0 Beppe Bezzi 1567 Bomber
5 Torch t18 P.Kline 1539 Bomber
6 Chameleon Myer R Bremer 1437 P-warrior
7 Frontwards v2 Steven Morrell 1420 One shot scanner
8 Evol Cap 6.6 John Wilkinson 1299 Imp / stone
9 quiz Schitzo 1262 Scanner/ bomber
10 T.N.T. Maurizio Vittuari 1204 Bomber
11 Grilled Octopus v0.5 David Boeren 1154 P-warrior
12 Hazy Shade II John Wilkinson 1102 P-warrior
13 Stepping Stone Kurt Franke 1007 * Qscan -> Vampire
14 Iron Gate 1.5 Wayne Sheppard 926 CMP scanner
15 Agony II Stefan Strack 912 CMP scanner
16 Barrage Anton Marsden 876 Qscan -> replicator
17 Blue Funk Steven Morrell 869 Stone/ imp
18 Flurry Anton Marsden 835 Qscan -> pwarrior
19 Thermite 1.0 Robert Macrae 802 Qscan -> bomber
20 Blue Funk 3 Steven Morrell 766 Stone/ imp
21 Night Train Karl Lewin 755 Replicator
22 Mirage 1.5 Anton Marsden 736 Scanner/ bomber
23 Blizzard Anton Marsden 713 Qscan -> replicator
24 HeremScimitar A.Ivner,P.Kline 666 Bomber
25 La Bomba Beppe Bezzi 650 Qscan -> replicator
Thermite II continues on, yaddah, yaddah--the thing is like the Energizer
Bunny. Qscan detectors are a definite threat though and someone has a anti-
qscan qscan, whatever that is. But Thermite seems to be holding it's own,
even with the wave of new blood entering the hill.
______________________________________________________________________________
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
# %W / %L / %T Name Author Score Age
1 52.4/ 29.1/ 18.5 Versatility 1.2 Ross Morgan-Linial 175.7 58
2 53.9/ 33.3/ 12.8 Syzygy 1.0 Philip Kendall 174.5 43
3 53.6/ 34.2/ 12.2 Saboteur 0.4k shar 173.0 1
4 52.7/ 34.3/ 13.0 Yet 3c Justin Kao 171.0 39
5 49.6/ 35.3/ 15.1 Phoenix Alpha Andy Nevermind 163.9 44
6 51.0/ 38.8/ 10.2 Saboteur v0.3p shar 163.3 65
7 50.8/ 39.3/ 10.0 Thalamus Edgar 162.2 91
8 48.8/ 36.1/ 15.1 Sandwich Bags Andy Nevermind 161.4 32
9 46.2/ 43.2/ 10.6 3-clear 0.6 Ilmari Karonen 149.1 66
10 45.6/ 44.2/ 10.2 Not Very Pretty 2.0 Ross Morgan-Linial 147.1 89
11 45.1/ 43.6/ 11.3 Yet 3a Justin Kao 146.6 40
12 41.6/ 37.5/ 20.9 Antivenin Ross 145.7 31
13 40.3/ 37.9/ 21.8 test Justin Kao 142.8 55
14 43.5/ 44.2/ 12.2 Switch Hitter 0.3t Ross Morgan-Linial 142.8 70
15 34.4/ 26.4/ 39.2 Cannon Fodder I. Karonen 142.3 45
16 31.6/ 20.9/ 47.5 Ties, Ties, Ties!+ Ross 142.2 10
17 25.4/ 9.5/ 65.1 Nematode v1.3c Jonathan Stott 141.3 25
18 41.0/ 41.0/ 18.0 3-clear 0.5 I. Karonen 141.0 80
19 42.3/ 46.2/ 11.5 hyper 1.0 Ross 138.5 63
20 28.8/ 20.0/ 51.2 Ties, Ties, Ties! Ross 137.7 28
21 30.8/ 27.3/ 41.9 Avenger I Oliver Fechner 134.2 62
22 32.2/ 42.4/ 25.4 Utility Knife Robert J. Street 122.1 2
23 19.8/ 60.4/ 19.8 WarHawk Robert J. Street 79.1 7
24 23.1/ 70.8/ 6.2 Kevorkian Robert J. Street 75.4 5
25 3.8/ 0.0/ 0.0 1234567890123456789012345 Anonymous 11.5 17
Versatility is still on top. With strong components from '94 hill warriors,
this code should live up to its name.
Syzygy 1.0 is a scan/spl carpet to spl/dat clear. Is the scan one shot or
repetitive. Check out this week's hint. Blur utilizes a scan/carpet -->
spl/dat clear.
Lots of pspace and core clear warriors. Make sure you check out the latest
and greatest core clear innovation. d-clear by guenzel is a dat/djn clear
wiping at 1.0c. It makes a great p-component, especially if you can get it
a spl clear tacked onto it without increasing the length (I'm trying).
______________________________________________________________________________
The Hint
I suppose this should be an Extra Extra and not a Hint, but if I had any good
hints, I would be on the hill right now.
------------------------------------------------------------------------------
The Mirage/Blur Scanner Series
by Anton Marsden
Mirage 1.5 was my first successful warrior. It was a 33% scanner with
continuous 33% carpet in the scanning loop. It did quite well against
imp-type warriors and papers but was thrashed by bombers. I wrote several
similar warriors similar to Mirage 1.5 and didn't want to publish them
immediately - I was still experimenting. But now I'm quite happy with my
latest version and have decided to release them all now, including the Hill
version. I haven't included the booting code and decoy but they're nothing
special. If there's enough demand for the full versions I'll post them in the
newsgroup at a later date.
Mirage 2 was based on the same technique as Mirage 1.5 but used a SEQ scan.
Here is Mirage 2:
;redcode-94
;name Mirage 2
;author Anton Marsden
;strategy Faster version, boot and decoy, no pspace
;assert CORESIZE==8000
;kill Mirage 2
ORG scan
step EQU 3024; mod-16
ptr: dat.f >0,$step
dat.f >0,$btm-ptr+3
cc: spl.a #1,$btm-ptr+4
top: mov.i $cc,>ptr
scan: seq.i $2*step+8,$2*step
mov.b $scan,$ptr
a: add.f $inc,$scan
jmn.b $top,$scan
inc: spl.a #step,>step
mov.i @1,>ptr
btm: djn.b $-1,{cc
This warrior had several weaknesses, the major one being that it was only
a mod-8 scan. I knew that mod-5 was achievable (and also mod-4) but had the
disadvantage that the locations scanned would be only 5 apart (not exactly
optimal!).
As Mirage 2 slowly made its way off the Hill I was wrote a mod-5 version and
named it Blur (two versions is enough for me!):
;redcode-94
;name Blur
;author Anton Marsden
;assert CORESIZE==8000
;strategy 0.5c scan, 0.25c carpet -> SPL/DAT core clear
;strategy Mirage 2 with a better step and some other changes.
;strategy This product is PSpace Free :-)
;kill Blur
ptr: dat.f $0,$step+5
dat.f $0,$0
dat.f $0,$0
dat.f >0,#btm-ptr+3
cc: spl.i #0,#btm-ptr+3 ; DJN stream protection
top: mov.i $cc,>ptr ; Putting this line here enhances protection
scan: seq.i $2*step,}2*step+5 ; A-indirect ensures B field is attacked with
mov.b $scan,$ptr ; no delay. Also leaves decoy for scanners
a: add.f $inc,$scan
jmn.a $top,$scan ; Falls through when all locations scanned
inc: spl.a #step,>step
mov.i @1,>ptr
btm: djn.b $inc+1,{cc
The A-indirect scan improved Blur's score against bombers but left "bombs" in
core that a decent dodger could take advantage of.
One of the things I had discovered about all these scanners was that if the
mod step in the scan was sufficiently good (ie. enough scans were done) then
there was no need for a SPL/DAT core clear against non-imp warriors. Then I
found the perfect clear for Blur - DAT/DJN (dclear?). I added this to Blur
and made a few more adjustments and came up with Blur 2:
;redcode-94
;name Blur 2
;author Anton Marsden
;assert CORESIZE==8000
;strategy 0.5c scan, 0.25c SPL carpet -> DAT/DJN clear
;strategy Final version (for a while)
;kill Blur
ORG scan
step EQU 70; mod-10
gate EQU top
ptr: mov.i $inc+1,>step ; New DJN protection... better than a SPL?
top: mov.i $inc,>ptr ; Or worse?
scan: seq.i $2*step+5,$2*step
chg: mov.b $scan,@top
a: add.f $inc,$scan
djn.b $top,#800
inc: spl.i #step,#step
mov.i $clr,>gate
btm: djn.f $-1,>gate
clr: dat.f <1,#clr-gate+2
Note on booting: The decoy and boot code was aligned such that Blur didn't
attack it's old code at all.
There are many variations on the basic engine I have tried and probably a
few that I haven't. Can you improve on it?
Things to try:
- different gap, eg. seq.i $2*step+15,$2*step, mod-5 scan
- different (ptr-scan) offsets
- psuedo mod-1 scan using A/B indirect scanning, eg.
top: mov.i $BOMB,>ptr ; bomb should probably be spl #0,0
scan: seq.F $2*step,}2*step+5 ; could also try seq.f >2*step,}2*step+5
mov.b $scan,$ptr
a: add.f $inc,$scan
djn.b $top,#0
inc: spl.i #step,#step
mov.i $clr-10,>gate; A field incremented 10 times in 8000 loops
______________________________________________________________________________
Questions? Concerns? Comments? Complaints? Mail them to people who care.
authors: Beppe Bezzi <bezzi@nemo.it> or Myer Bremer <bremermr@ecn.purdue.edu>
|
