A midweek review of Corewar
May 13, 1993
-------------------------------------------------------------------------------
I. The Standings:
# %W/ %L/ %T Name Author Score Age
1 35/ 21/ 43 Night Crawler Wayne Sheppard 149 338
2 37/ 26/ 37 ImpsAreMyFriend J.Layland 148 12
3 36/ 28/ 36 FlyPaper 2.0 J.Layland 145 43
4 44/ 44/ 13 Dragon Spear c w blue 144 440
5 33/ 23/ 44 Imprimis 6 P.Kline 143 738
6 33/ 25/ 42 Sphinx v2.8 W. Mintardjo 140 1336
7 42/ 45/ 14 Iron Gate 1.01 Wayne Sheppard 139 166
8 30/ 23/ 47 +0 Stormbringer Dan Nabutovsky 138 1768
9 41/ 44/ 15 Sucker 6 Stefan Strack 138 360
10 38/ 39/ 22 Leprechaun 1b Anders Ivner 137 1250
11 29/ 22/ 49 ttest nandor sieben 136 214
12 42/ 49/ 9 Agony 5.1 Stefan Strack 136 492
13 29/ 22/ 49 Snake Wayne Sheppard 136 98
14 42/ 50/ 9 Medusa's v7 Mintardjo & Strack 134 631
15 38/ 43/ 20 Herem II Anders Ivner 133 2
16 41/ 50/ 9 Sting II Arno Fuhlendorf 132 156
17 27/ 22/ 51 ittt nandor sieben 132 215
18 40/ 50/ 10 Backstabber Anders Ivner 130 258
19 38/ 48/ 14 ScanTest J.Layland 127 1
20 26/ 39/ 35 sub-type-b+r c w blue 114 5
21 2/ 2/ 1 ScanTest J.Layland 6 7
-------------------------------------------------------------------------------
II. The Basics:
-Core War Archives are available via anonymous FTP at
soda.berkeley.edu in pub/corewar...
-FAQ for this newsgroup is available via anonymous FTP at
rtfm.mit.edu as pub/usenet/news.answers/games/corewar-faq.z
-------------------------------------------------------------------------------
III. The Scoop:
Round 1 of S. Strack's tournament has been completed. All human particpants
breathed a sigh of relief as those who had non-human opponents ('Automatic
Player') won :-) As a help for the second round, Stefan mailed out the
source code for all the submittals. Just to make things interesting,
he doubly-encrypted the source (Zip and uuencode) which gave some authors
a bit of a puzzle.
Just when things at the top of the Hill seemed to have settled down to
a back-and-forth struggle between imps and scanners, along came J. Layland's
FlyPaper 2.0 which is firmly locked in the top 5. Not only that, but
his ImpsAreMyFriend and ScanTest are also showing some muscle.
ImpsAreMyFriend is especially interesting as it is challenging Night Crawler
as the highest-ranked imp-program. King of the Hill competitors are
anxiously waiting some published code (WE'LL SETTLE FOR HINTS!) to see
if J.L has uncovered some new principle or what :-)
The Hill weathered a late shower of Dwarf's, Tests, and (everybody's
favorite) 'Unknown by Unknown'. The forecast is for a steady drizzle
of new participants - WELCOME! And an occasional blast from an old program
just shows that the Hill continues to get tougher.
-------------------------------------------------------------------------------
IV. The Outlook:
5 44/ 44/ 12 ScanTest J.Layland 144 1
8 28/ 20/ 52 Sphinx v4.3 W. Mintardjo 137 1
3 34/ 24/ 42 FlyPaper 2.0 J.Layland 145 1
3 46/ 44/ 10 sub-type-cmp c w blue 148 1
1 38/ 24/ 38 ImpsAreMyFriend J.Layland 151 1
-------------------------------------------------------------------------------
V. The Quick Look:
20 40/ 52/ 8 A2 Stefan Strack 127 1
21 33/ 57/ 10 BS J.Layland 108 0
20 1/ 45/ 54 BTB Fredrik 56 1
21 26/ 54/ 20 Tsp W. Mintardjo 98 0
20 26/ 52/ 22 Tsp W. Mintardjo 101 1
20 36/ 58/ 6 XBS Fredrik Ohrstrom 114 1
20 27/ 61/ 12 a11 S. Halvorsen 93 1
21 29/ 69/ 3 rs2 Shade 89 0
20 28/ 38/ 34 Test Wayne Sheppard 118 1
21 18/ 74/ 8 Test T. M. Bezenek (Dewdn 63 0
21 0/ 89/ 11 imp4 Shade 11 0
21 26/ 55/ 19 test P.Kline 97 0
21 28/ 61/ 11 Dwarf George Gorman 95 0
21 4/ 72/ 24 Splat Shade 35 0
17 38/ 46/ 15 TTp12 W. Mintardjo 130 1
21 28/ 61/ 11 Dwarf30 George Gorman 94 0
21 15/ 53/ 31 Frobozz Fredrik Ohrstrom 77 0
21 0/ 46/ 54 Unknown Eric 55 0
21 12/ 76/ 12 impsimp Harvey 49 0
19 36/ 43/ 20 Herem II Anders Ivner 130 1
21 18/ 72/ 10 re-entry ECP 64 0
21 20/ 78/ 2 test v0.6 Raistlin Majere 63 0
21 20/ 63/ 17 Early Bird c w blue 77 0
18 26/ 25/ 49 Nurgle III c w blue 128 1
18 39/ 49/ 12 sub-type-s c w blue 128 1
21 25/ 75/ 1 test v0.6g Raistlin Majere 75 0
21 23/ 74/ 3 Dwarfold103 George Gorman 72 0
20 26/ 35/ 38 FlashGordon J.Layland 117 1
18 35/ 42/ 22 Night Shade Wayne Sheppard 128 1
10 26/ 22/ 52 Sphinx v2.9 W. Mintardjo 130 1
20 38/ 47/ 15 Sunburst 31 Jay Han 129 1
21 2/ 76/ 22 The Curtain Rudi Cilibrasi 27 0
21 15/ 84/ 1 sub-type-av c w blue 47 0
18 35/ 45/ 20 EarSplitting P.Kline 125 1
21 30/ 54/ 16 RoadRunner I S. Halvorsen 106 0
20 34/ 50/ 16 Sunburst 30b Jay Han 118 1
20 21/ 27/ 52 sub-type-aix c w blue 115 1
20 24/ 27/ 49 sub-type-b+r c w blue 121 1
21 37/ 52/ 11 Bomb The Base Fredrik Ohrstrom 122 0
18 36/ 44/ 20 Distance v6.1 Brant D. Thomsen 128 1
21 0/ 45/ 55 George Gorman George Gorman 56 0
21 22/ 65/ 13 Inchworm 1.3a Brant D. Thomsen 80 0
21 28/ 62/ 11 Dwarf15nocolor George Gorman 94 0
21 8/ 49/ 44 Improved Dwarf Ben Burch 67 0
21 6/ 72/ 22 Parasite v 6.1 Roderick Easton 40 0
20 31/ 51/ 18 Am I Evil 10.0d Jarkko Lindblad 112 1
20 38/ 50/ 12 Bomb The Base 2 Fredrik Ohrstrom 127 1
21 27/ 42/ 32 Desperate I (b) Brant D. Thomsen 111 0
17 19/ 23/ 59 FlasherTest 0.1 J.Layland 115 1
20 37/ 53/ 10 Glass House 2.1a J.Layland 122 1
21 4/ 45/ 51 Scars, Four Eyes Joshua Houk 62 0
21 26/ 74/ 1 test v0.6h again Raistlin Majere 78 0
21 9/ 33/ 58 Bubble-scrape v3.1 Joshua Houk 85 0
21 6/ 51/ 44 Improved Dwarf III Ben Burch 61 0
21 16/ 39/ 45 Trident [Version 2] W. Mintardjo 93 0
21 5/ 95/ 0 shirley's neighbourhood gg and is 15 0
-------------------------------------------------------------------------------
VI. The Hint:
Part of the discussion related to the new draft of the redcode (Corewars)
syntax, concerns the difference between 'in-register' and 'in-memory'
evaluation of operands.
Some of the discussion is available in the digests in pub/corewar at
soda.berkeley.edu, but this difference was made clear to me (at last) this
week when a program I was running successfully on Core! for the Mac
got slaughtered on KotH.
I was using this construct:
mov splbm,@tgt
tgt mov @0,<100
jmn tgt,@tgt
...
splbm spl 0,5
In Core!, this lays down a carpet like:
tgt+100 spl 0,0
spl 0,1
spl 0,2
spl 0,3
spl 0,4
but at KotH, this goes into an endless loop, after laying down:
tgt+100 spl 0,1
spl 0,1
spl 0,2
spl 0,3
spl 0,5
Core! uses in-memory evaluation, so 'spl 0,1' decrements to 'spl 0,0' and
is overlaid on top of itself; then the 'jmn' sees a zero and falls through.
Because KotH uses in-register evaluation, the 'spl 0,1' instruction keeps
getting overlaid on tgt+100, so the 'jmn' never sees a zero and keeps
jumping back.
So beware when using a line like:
mov @0,<xxx
or
mov <0,<xxx
-------------------------------------------------------------------------------
VII. The End:
Paul Kline
pk6811s@acad.drake.edu