Issue 23 April 1, 1996
______________________________________________________________________________
Greetings.
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.
Greetings to all and good first of April, I hope you enjoyed our little joke.
This was a quiet week, but 90 challenges and no important warrior pushed
off; even if the turnover has been high most charges have been new versions
of existing warriors. Just a bit of relax for the spring, or something
burning under the ashes.
Planar's corner is back this week, we hope he will continue working for us.
Anyone wanting to pupblish an article or write his own number is welcome.
--Beppe Bezzi
P.S. I changed my mail address; new one is: bezzi@nemo.it
I still have the old one, but it will be checked less often.
______________________________________________________________________________
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/ 32/ 21 Gem of the Ocean P.Kline 162 3
2 48/ 40/ 12 Naked dancer Beppe Bezzi 156 55
3 41/ 34/ 25 Thermite II Robert Macrae 147 1105
4 41/ 34/ 26 prova Maurizio Vittuari 147 81
5 41/ 35/ 24 Chameleon M R Bremer 147 576
6 41/ 37/ 22 Grilled Octopus v0.5 David Boeren 146 321
7 34/ 25/ 41 Barrage Anton Marsden 144 578
8 43/ 41/ 16 Stepping Stone Kurt Franke 144 446
9 43/ 43/ 14 Frontwards v2 Steven Morrell 143 1357
10 39/ 35/ 26 T.N.T. Maurizio Vittuari 143 830
11 37/ 32/ 31 A Hard Day's Night Maurizio Vittuari 142 67
12 37/ 32/ 31 Tornado 3.0 Beppe 142 906
13 36/ 30/ 35 Flurry Anton Marsden 142 66
14 27/ 15/ 58 Hazy Shade II John K W 140 64
15 41/ 42/ 17 Wind-up Toy v0.4 Ian Oversby 139 21
16 23/ 9/ 68 Evol Cap 6.6 John Wilkinson 138 1119
17 43/ 49/ 8 Memories Beppe Bezzi 137 498
18 34/ 32/ 34 Lithium John K Wilkinson 137 112
19 27/ 18/ 54 Impfinity v4g1 Planar 137 1304
20 40/ 43/ 17 Error 0 Planar 136 38
21 38/ 41/ 20 Sloth v0.4 Ian Oversby 135 94
22 30/ 25/ 46 Hector 3 Kurt Franke 134 109
23 38/ 42/ 20 myConfuser2.1 Paulsson 134 16
24 32/ 31/ 37 Blue Funk 5 Steven Morrell 134 249
25 29/ 34/ 38 tripod v1.3 Brian Haskin 124 1
Weekly age: 91
New warriors 10 Turnover/age rate 11%
Average age: 401 ( 350 last week, 333 the week before )
Average score: 141 ( 142 last week, 142 the week before )
The top 25 warriors are represented by 14 authors; welcome to Ian and Brian
The age of the Octopus seems to be over; Paul Kline, with his new version of
Gem of the Ocean, has managed to reach n. 1 spot, the only one making him
feel confortable; and this with but 33 submissions this week.
Will now David refurbish the Octopus, submitting v04 to claim again top
spot, or let it age toward the hall of fame?
Beppe's Naked Dancer, and Grilled Octopus, are the only other warrior to
have been in top position.
After past weeks carnage, no aged warrior, not even mid aged, left us; all
millenaries seem to be in good health, even if Thermite has been for a while
in the red zone, under 20th position, to rebounce quickly up.
Planar, after a long silence, has submitted a new warrior. Next week,
perhaps, we'll see in these columns what our archiver is doing.
Bombers and P-spacers are sharing the top positions, while scanners are
having bad times and replicator near extinction, with but Barrage an Hector
to represent them on the hill.
The scores are very balanced, there are but 13 points beetween n 3 and n.
24, should a new warrior enter in mid hill none will feel safe. Maybe next
week we'll see some movement again
______________________________________________________________________________
94 - What's New
# %W/ %L/ %T Name Author Score Age
7 37/ 36/ 27 prova Maurizio Vittuari 138 1
15 33/ 33/ 33 A Hard Day's Night Maurizio Vittuari 133 1
19 30/ 30/ 40 Flurry Anton Marsden 130 1
4 25/ 15/ 60 Hazy Shade II John K W 136 1
10 41/ 44/ 15 Naked dancer Beppe Bezzi 137 1
20 37/ 45/ 17 Error 0 Planar 129 1
23 36/ 44/ 20 myConfuser2.1 Paulsson 127 1
1 45/ 34/ 21 Gem of the Ocean P.Kline 156 1
25 27/ 35/ 38 tripod v1.3 Brian Haskin 119 1
New versions of: Gem of the Ocean, Naked dancer, prova, Flurry, A Hard day's
Night, Hazy Shade II.
______________________________________________________________________________
94 - What's No More.
# %W/ %L/ %T Name Author Score Age
26 10/ 12/ 5 Gem of the Ocean P.Kline 35 66
26 37/ 48/ 15 Pagan John K W 127 27
26 1/ 1/ 2 Flurry Anton Marsden 6 13
26 3/ 1/ 0 Gem of the Ocean P.Kline 9 16
26 2/ 1/ 1 prova Maurizio Vittuari 6 24
26 2/ 1/ 1 A Hard Day's Night Maurizio Vittuari 6 24
26 0/ 0/ 3 Hazy Shade II John K W 4 64
26 29/ 37/ 34 Clueless v0.4 Ian Oversby 122 34
26 2/ 2/ 0 Naked dancer Beppe Bezzi 6 85
26 30/ 43/ 27 Spook v0.1 Ian Oversby 116 16
26 3/ 0/ 0 myConfuser2.0 Paulsson 9 42
______________________________________________________________________________
94 - What's Old
# %W/ %L/ %T Name Author Score Age
9 43/ 43/ 14 Frontwards v2 Steven Morrell 143 1357
19 27/ 18/ 54 Impfinity v4g1 Planar 137 1304
16 23/ 9/ 68 Evol Cap 6.6 John Wilkinson 138 1119
3 41/ 34/ 25 Thermite II Robert Macrae 147 1105
12 37/ 32/ 31 Tornado 3.0 Beppe 142 906
10 39/ 35/ 26 T.N.T. Maurizio Vittuari 143 830
7 34/ 25/ 41 Barrage Anton Marsden 144 578
5 41/ 35/ 24 Chameleon M R Bremer 147 576
17 43/ 49/ 8 Memories Beppe Bezzi 137 498
8 43/ 41/ 16 Stepping Stone Kurt Franke 144 446
6 41/ 37/ 22 Grilled Octopus v0.5 David Boeren 146 321
24 32/ 31/ 37 Blue Funk 5 Steven Morrell 134 249
______________________________________________________________________________
HALL OF FAME
* means the warrior is still running.
Pos Name Author Age Strategy
1 Jack in the box Beppe Bezzi 1620 P-warrior
2 Torch t18 P.Kline 1539 Bomber
3 Frontwards v2 Steven Morrell 1357 * One shot scanner
4 Impfinity v4g1 Planar 1304 * Stone/ imp
5 quiz Schitzo 1262 Scanner/ bomber
6 Evol Cap 6.6 John Wilkinson 1119 * Imp / stone
7 Thermite II Robert Macrae 1105 * Qscan -> bomber
8 Iron Gate 1.5 Wayne Sheppard 926 CMP scanner
9 Agony II Stefan Strack 912 CMP scanner
10 Tornado 3.0 Beppe Bezzi 906 * Bomber
11 Blue Funk Steven Morrell 869 Stone/ imp
12 T.N.T. Maurizio Vittuari 830 * Bomber
13 Thermite 1.0 Robert Macrae 802 Qscan -> bomber
14 Blue Funk 3 Steven Morrell 766 Stone/ imp
14 Night Train Karl Lewin 755 Replicator
16 Mirage 1.5 Anton Marsden 736 Scanner/ bomber
17 Blizzard Anton Marsden 713 Qscan -> replicator
18 HeremScimitar A.Ivner,P.Kline 666 Bomber
19 La Bomba Beppe Bezzi 650 Qscan -> replicator
20 myVamp v3.7 Paulsson 643 Vampire
21 Hazy Shade Of Winter John Wilkinson 616 P-warrior
22 Armory - A5 John Wilkinson 609 P-warrior
23 juliet and paper Bremer & Bezzi 607 P-warrior
24 Phq Maurizio Vittuari 589 Qscan -> replicator
25 Barrage Anton Marsden 578 Qscan -> replicator
Frontwards and Impfinity are quickly approaching Torch, both passing 1300
milestone; Tornado should enter the over 1000 club next week if, crossing
fingers, it survives.
Barrage enters in the last spot, pushing Hector 2 off. Chameleon is
scheduled for next week being at 576 age.
______________________________________________________________________________
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 55/ 34/ 11 Violent Micro v0.2 basehead 177 58
2 45/ 16/ 39 Vectorfantasy v0.2 basehead 173 59
3 48/ 37/ 14 Living Let Die III Julian 159 3
4 50/ 41/ 9 C-Seagal II Test jShelton 158 97
5 50/ 42/ 8 GPT 15l David van Dam 157 96
6 42/ 27/ 31 Obvious Thief.2 John K. Lewis 157 84
7 49/ 41/ 10 Machido jShelton 157 33
8 48/ 43/ 9 Goliath 0.3 David van Dam 154 53
9 41/ 29/ 30 Hyakutake Zenith Scott Manley 152 70
10 41/ 31/ 28 Bag of Tricks II Andy Nevermind 152 30
11 47/ 42/ 11 Revision Three Andy Nevermind 151 93
12 47/ 44/ 9 GPT 18a David van Dam 150 14
13 43/ 36/ 21 Ineffectual Ian Oversby 149 5
14 43/ 41/ 16 GPT 09k David van Dam 145 89
15 44/ 47/ 8 bloodstone V2.0 bjoern guenzel 142 15
16 42/ 44/ 14 Extreme Prejudice Scott Manley 141 91
17 33/ 28/ 38 AB Scanner 2.9.6 Chris Hodson 138 38
18 40/ 47/ 13 Optma Dwarf v5.1 Phoenix 133 19
19 36/ 40/ 23 Three Core Monty Andrew Fabbro 132 50
20 36/ 42/ 23 Testing V1.21 bjoern guenzel 130 4
21 29/ 30/ 41 Unknown John K. Lewis 127 83
22 35/ 53/ 12 Optma Dwarf v5.0 Phoenix 118 21
23 27/ 62/ 11 MissleCommand_v2 harleyQ2 91 2
24 14/ 56/ 30 djn monster bjoern guenzel 73 7
25 8/ 67/ 25 babyQ2 harleyQ2 50 1
Jack in the box is dead of old age today, this means that the beginners hill
is aged exactly 100 in last two weeks; don't worry, you won't see Jack anymore.
Basehead is dominating the hill with a couple of warriors seen in the pro
hill for a while, Julian and jShelton too did the 94 hill for short periods.
______________________________________________________________________________
Stormking 94x hill standings.
We had some movement in this hill in the past weeks, so we report hill
standings.
Hill Specs:
redcode 94x
coresize: 55440
max. processes: 55440
duration: after 500,000 cycles, a tie is declared.
max. entry length: 200
minimum distance: 200
rounds fought: 250
instruction set: ICWS '94 Draft
Last battle concluded at : Mon Mar 25 21:13:40 EST 1996
# %W/ %L/ %T Name Author Score Age
1 35/ 3/ 62 Evol Cap 4 X John Wilkinson 166 19
2 48/ 40/ 12 Illusion-94/55 Randy Graham 156 1
3 47/ 39/ 13 Frontwards v2 Steven Morrell 156 49
4 46/ 38/ 15 Stepping Stone 94x Kurt Franke 154 5
5 42/ 32/ 27 Lithium X 8 John K Wilkinson 152 10
6 46/ 41/ 13 Memories Beppe Bezzi 151 26
7 43/ 38/ 18 Derision M R Bremer 148 36
8 41/ 35/ 24 BigBoy Robert Macrae 146 44
9 43/ 44/ 13 Pagan John K W 143 4
10 38/ 34/ 27 Lithium X John K Wilkinson 142 17
11 39/ 35/ 26 Tornado 2.0 x Beppe Bezzi 142 43
12 39/ 39/ 21 Fire Master Xv1 JS Pulido 140 41
13 31/ 26/ 44 Hector 2 Kurt Franke 136 39
14 42/ 48/ 10 S.E.T.I. 4-X JKW 135 20
15 30/ 26/ 44 Aleph 1 Jay Han 135 9
16 38/ 47/ 16 Watcher-h Kurt Franke 128 38
17 36/ 44/ 20 impthingX Kurt Franke 128 3
18 37/ 46/ 17 Watcher Kurt Franke 127 42
19 31/ 39/ 30 Nice Try M R Bremer 122 48
20 37/ 53/ 10 Withershins Thrice P.Kline 120 2
The big hill uses the same redcode as 94 hill. Strategies are a bit
different, seems that imps are more powerful and qscan are obviously less
dangerous, but there is a lot to discover here.
Tuc has implemented many interesting features, send a mail to
koth@stormking.com with body of ' help ' (no quotes) to get them.
______________________________________________________________________________
The Hint
by Beppe Bezzi
---
A collection of my hints in previous numbers of CW is available at
ftp.csua.berkeley.edu/pub/corewar/documents/bbhints.zip
---
Wilkinson's beginners benchmark.
J K Wilkinson posted to the newsgroup his beginner's benchmark, a collection
of warriors ,good for beginners to test their warriors against. Steve Bailey
asked me for some comments on them and here they are.
;name Blue Funk 3
;author Steven Morrell
;strategy Fixed another in-memory/in-register bug
The strategy line doesn't explain a lo about what the warriors does, unless
you know Blue Funk and Kline's Emerald, but a little testing can easily show
that Blue Funk 3 is a classic stone with a binary launched imp spiral.
The most interesting property of the stone is the self modifying bombing
pattern; in practice BF3, after completing its modulo 4 bomb run shifts the
pattern and repeat it displaced, going to bomb the whole core. To better see
it, try running BF3 after changing the 'spl i' line in a dat, so
deactivating the spiral.
;name Cannonade
;author P.Kline, Paul Kline, pk6811s@acad.drake.edu
;strategy break gates, score points
;strategy stone and gate-busting imp-spiral(s)
Cannonade is worth mention mainly for the stone structure, well described by
S. Morrell in its book, and for the gate busting imps, descibed in Morrell's
manual too. Cannonade relies more on the imps for attack than on the stone,
using it, with a wide mod 10 pattern, to kill scanners and provide a core
clear. Unfortunaltely, for it, 94 redcode allows forward clears and
increment gates, and Cannonade's imps are unable to crash them; it's still a
powerful opponent in 88 redcode.
;name Fire Storm v1.1
;author W. Mintardjo
;strategy Stone with anti-imp core clear
Firestorm is another stone, worth mention as a way to implement a two pass
clear in 88 redcode. To see it trace the 'ptr' line and look how the clear
pointers self modify.
;name Marcia Trionfale 1.3
;author Beppe Bezzi
;strategy attemping a more aggressive behaviour
That's the version of Marcia Trionfale I used in last winter tournament,
together with Tornado, but using more processes, it's half Jack in the Box.
Marcia Trionfale is a very aggressive silk replicator, very good against
stone imps like Blue Funk and Juliet storm, but vulnerable to scanners.
;name nobody special
;author Mike Nonemacher
;strategy Paper like TimeScape, but with lots more
;strategy core-trashing, anti-imping, etc.
nobody special is another silk replicator, relative of Timescape. It's less
aggressive than Marcia Trionfale, but more than Timescape. Interesting how
the add instruction, together with the bombing mov in the other body,
creates a non-carpet bombing.
;name Paperone
;author Beppe Bezzi
;strategy Silk replicator
Paperone is my first warrior to make 94 hill, it's nothing more than Jippo's
Silk Warrior with different constants.
;name TimeScape (1.0)
;author J. Pohjalainen
;assert CORESIZE==8000
;strategy I'm stuck with replicators! Here is _The Latest_ one!
;strategy \---------------------------\ ----------------------
;strategy / ts1 spl @ts1, }STEP1 / Phoenix/Cell warrior
;strategy \ mov.i }ts1, >ts1 \ body, 6+ processes to
;strategy / ts2 spl @ts2, }STEP2 / keep That Thing alive,
;strategy \ mov.i }ts2, >ts2 \ two of them working
;strategy / mov.i {ts2, <ts3 / together with proper
;strategy \ ts3 jmp @ts3, }STEP3 \ constants and you have
;strategy /___________________________/ found >>--> TimeScape!
;strategy \T I M E T O E S C A P E\ ----------------------
;strategy v1.0: added more havoc in above code (or I hope so!)
Timescape is the father of modern replicators. With very few exceptions, all
use its structure. For a better discussion on silk replicators see Core
Warrior 1 and 3
;name Iron Gate
;author Wayne Sheppard
;strategy cmp scanner-SPL/JMP-Gate
Iron gate is a classic 88 redcode cmp scanner, using spl 0 /jmp -1 stun
bombs. Worth notice are the compact code and the decrement protection. For a
better discussion on Iron Gate and cmp scanners see my hint in Core Warrior 19
;name Porch Swing
;Author Randy Graham
;strategy Swing with a little wider range.
;strategy Now 80% bomb/scan with djn-stream once-thru
Porch Swing is the warrior I saw more dominate; last summer it scored more
than 170 pts in the 94 hill, to disappear as soon as people began to take
counter measures like boot and use of bombs with b-field of 1. It's a
classic example of once through scanner bomber with a deadly multipass clear.
;name Rave
;author Stefan Strack
;strategy Carpet-bombing scanner based on Agony and Medusa's
Rave is another classic cmp scanner, attacking with a carpet of spl 0. Its
relative Agony II, similar but using boot, is one of the longest living
warriors in the 94 hill.
;name Thermite 1.0
;author Robert Macrae
;strategy Quick-scan -> incendiary bomber.
An example of quick scanner, using vampire attacks, followed by an
incendiary bomber. For a discussion on Thermite see Core Warrior 5
;name Tornado
;author Beppe Bezzi
;strategy the original one
;strategy Fast 60% c bomber
Tornado is the fastest pure bomber appeared till now (the 75% and 80% are
mixed scan bombers) It's interesting the use of bombs as pointers for next
one. Tornado is discussed in Core Warrior 7
The version in the benchmark never stood in the hill for long, being too
vulnerable to paper, but was a good complement for Marcia Trionfale in the
p-warrior Jack in the box (that was in the hill for very long :-)
Exercise for the reader: make it self splitting to increase resistance, add
a djn stream, a multipass clear and a bit of paper protection; if you can do
it without bloating Tornado you'll make the hill with success. Yes, Tornado 3.0
Overall the benchmark is, IMO of course, a little unbalanced. There are too
many replicators, and less scanners than in the 94 hill, stone bombers are
outdated, modern bombers are different from Cannonade and Firestorm, a Die
Hard style warrior could have found its place as like as a simple pwarrior
like Jack or juliet and paper. Anyway we have to thank JKW fo doing the job,
the warrior choice is sometimes personal taste, everybody has his preferred
enemies, and trades off some overall point just to beat'em :-)
I tested my warriors, and a few other famous guys on the 94 hill, using JKW
benchmark; the results:
Warrior Wilkstones (or whatever you want to call them, score for
example :-)
Frontwards v2 170.33
Naked Dancer 168.67
quiz 159.58
Tornado 3.0 156.33
Jack in the box 155.83
Impfinity v4g1 149.41
Memories 146.92
Torch t18 140.33
These are some other results, courtesy of Steve Bayley, other scores have
been posted in r.g.c.
The warriors of the benchmark itelf:
Porch Swing 167.9583
Marcia t. 1.3 150.4167
Timescape 149.4167
nobody special 145.75
Blue Funk 3 142.25
Thermite 1.0 139.5
Paperone 138.875
Iron Gate 132.5
Cannonade 123.5417
Rave 120.125
Firestorm 107.5833
Tornado 84.375
And two classics
Mice 87.08334
Dwarf 49.375
______________________________________________________________________________
Planar's corner
A true story
Last week, I had a great idea for a new strategy. I wrote a first
prototype warrior and tried it against my personal benchmark. This is
8 warriors taken from my own Web page, section "Warriors on the
hill". They are:
Clisson Lite
Frontwards v2
Impfinity v4g1
Lithium
Memories
Blue Funk 3
Thermite 1.0
Tornado
Some of these are old versions of current hill warriors.
So I ran my benchmark and got something like this:
opponent Win Los Tie Win Los Tie Win Los Tie Win Los Tie points
0ClissonLite 110 85 5 / 116 81 3 / 74 114 12 / 121 73 6 / 161
1Frontwards 115 69 16 / 112 62 26 / 119 63 18 / 121 67 12 / 184
2Impfinity 112 55 33 / 107 43 50 / 116 48 36 / 122 50 28 / 189
3Lithium 80 99 21 / 125 48 27 / 101 62 37 / 111 74 15 / 168
4Memories 74 124 2 / 86 114 0 / 113 82 5 / 85 115 0 / 135
9BlueFunk 121 44 35 / 79 84 37 / 107 65 28 / 113 51 36 / 174
9Thermite 19 167 14 / 22 173 5 / 19 171 10 / 15 176 9 / 32
Tornado.. 83 89 28 / 75 89 36 / 103 74 23 / 109 59 32 / 153
self 72 75 53 / 84 66 50 / 78 82 40 / 76 84 40 / 139
This looked promising, except for a definite weakness against
quick-scanners, so I started working on the constants (and fixed a
couple of bugs in the process). A number of hours later, and after
more than 1000 battles (at 200 rounds each, this is a lot of
computation), I had this:
0ClissonLite 125 73 2 / 121 77 2 / 112 87 1 / 34 39 127 / 163
1Frontwards 141 54 5 / 151 43 6 / 133 58 9 / 124 70 6 / 209
2Impfinity 113 53 34 / 141 44 15 / 114 55 31 / 120 43 37 / 197
3Lithium 103 56 41 / 107 63 30 / 94 76 30 / 87 75 38 / 164
4Memories 86 111 3 / 105 89 6 / 122 74 4 / 127 72 1 / 166
5BlueFunk 133 38 29 / 133 34 33 / 133 35 32 / 134 40 26 / 214
6Thermite 63 128 9 / 49 132 19 / 77 108 15 / 51 136 13 / 97
7Tornado. 116 84 0 / 121 79 0 / 136 64 0 / 119 81 0 / 184
self 99 88 13 / 92 89 19 / 130 49 21 / 93 91 16 / 163
"Good, I said, that should land me right on top of the hill, with at
least 160 points." So I sent my program to the hill and eagerly
awaited the results. Here they are:
points
name %W/ %L/ %T taken/given
-----
Gem of the Ocean 52/ 38/ 10 166 / 124
Grilled Octopus v0.5 53/ 41/ 5 165 / 129
Frontwards v2 66/ 28/ 5 204 / 90
Impfinity v4g1 80/ 10/ 9 249 / 41
Chameleon 19/ 65/ 15 72 / 212
Barrage 21/ 57/ 21 86 / 192
Hazy Shade II 14/ 17/ 68 112 / 119
Naked dancer 29/ 42/ 28 117 / 154
prova 32/ 55/ 12 109 / 178
Evol Cap 6.6 9/ 24/ 66 95 / 138
Blue Funk 5 64/ 22/ 13 207 / 79
Tornado 3.0 27/ 53/ 20 101 / 179
myConfuser2.0 34/ 40/ 25 129 / 145
Memories 58/ 41/ 0 175 / 124
A Hard Day's Night 41/ 48/ 10 135 / 154
T.N.T. 31/ 64/ 4 99 / 196
Sloth v0.4 37/ 61/ 1 114 / 184
Flurry 15/ 70/ 14 60 / 225
Stepping Stone 42/ 49/ 8 134 / 157
Hector 3 28/ 52/ 20 104 / 176
Clisson 16/ 21/ 62 112 / 125
Thermite II 16/ 70/ 13 61 / 225
Lithium 39/ 41/ 20 137 / 143
C Seagal II 37/ 61/ 1 114 / 184
Living Let Die v2.6 27/ 66/ 6 88 / 205
self 44/ 44/ 11 143 / 145
What the hell ?! This gives me 127 points, 24th position on the hill.
Needless to say, I'm not happy with these results. It's time to
analyze.
Note that the results are as good as expected against Frontwards, Blue
Funk, and Lithium, and even better against Impfinity (can you say
"suicide" ?). Note also the good results against Gem of the Ocean and
Grilled Octopus. But my warrior fails miserably against a lot of
others.
The problem is described by one word: bias. My test suite is strongly
biased. I won't tell you toward what because I don't want you to
guess my new strategy.
I think this story show the importance of a good, balanced test suite,
and the necessity of live testing (i.e. sending test warriors to the
hill), as opposed to off-line testing.
Now I'll go back to the drawing board, but I feel just a little bit
depressed. Maybe I can make it work, but it won't be as great as it
should be.
-- Planar
______________________________________________________________________________
Questions? Concerns? Comments? Complaints? Mail them to people who care.
authors: Beppe Bezzi <bezzi@nemo.it> or Myer R Bremer <bremermr@ecn.purdue.edu>
|
