Entry Summaries
The attached mail is the collection of write-ups I received. As usual
they are mostly unedited (only personal POTM-master insults were removed).
Write-ups are provided in order of finish (more or less).
SHOWDOWN Program entry descriptions
Beware of long mail .... but there ARE gems buried!
Even if you don't study the whole thing, you might
want to check out the first four or five writeups!
The attached mail is the collection of write-ups I received. As usual
they are mostly unedited (only personal POTM-master insults were removed).
Write-ups are provided in order of finish (more or less).
There will also be selected code listings available on the websites.
If you need even more detail - contact the authors - not ME!!!
=Fred
---()--()--()--()--()--()--()--()--()--()--()--()--()--()--()--()--()--()---
-()--()--()--()--()--()--()--()--()--()--()--()--()--()--()--()--()--()--()-
---()--()--()--()--()--()--()--()--()--()--()--()--()--()--()--()--()--()---
======= WINNER =========== Dilbot ===
========CODE WILL BE POSTED ON WEBSITES =========================
=================
Dilbot was submitted by Mike Anderson at mranderson@att.com
=================
++ What strategies did Dilbot use?
Dilbot is primarily a Sleeper with a little bit of Adjuster and Zapper
mixed in. Dilbot's basic strategy is to deny his opponent a WIN or TIE (the
really BIG points!) by sitting on his home square and throwing WALLs in
front of his opponent to collect OUCH points while tracking his position to
figure out the best place for the next wall while also avoid getting ZAPped
himself should his opponent get too close to Dilbot's home row or colunm.
If his opponent does get "close", Dilbot goes into ZAPping mode, shooting
down the row or column that his opponent is likely to stumble into and
hopefully HIT him then repeatedly ZAP him for more points until the match
is over. Otherwise, Dilbot repeatedly WALLs his opponent one or two move
squares in front of his opponent along the diagonal across the path to
Dilbot's home square. If Dilbot detects that his opponent has stopped to
deal with a wall, he WALLs him in closer to try to replace the WALLs that
may have just been ZAPped. This can result in some OUCHes before his
opponent tries to break this replacement cycle, or a lot of OUCHes if he's
stuck and keeps trying to MOVE. Because Dilbot doesn't move, he can't give
OUCH points to his opponent (although Dilbot won't get MOVE points either
until/unless he make his run; see below). And because he stays in the
corner, his opponent can't go past Dilbot's home row or column which
increases the chance to ZAP him as he has to move into that row or column
to get to the home square.
However, should Dilbot's opponent sit still for a period after the turn
passes where it's no longer possible for his opponent to reach Dilbot's
home square, Dilbot morphs into a Zapper. He will move on a randomly
selected "diagonal" path towards his opponent's last location to collect
MOVE points as well as trying to ZAP him and collect ZAP points (this is
intended to also defeat a Sleeper opponent by racking up MOVE points as the
Sleeper sits on his home). Each MOVE is proceeded by a ZAP to clear any
wall dropped in front, or to hit his opponent they crossed paths. In case
his opponent WALLs in a way that makes Dilbot give an OUCH after clearing a
wall, Dilbot will change course and try a ZAP/MOVE again. Another such OUCH
puts Dilbot in a random ZAP/MOVE mode to get out of a WALL replacement cycle.
There are weaknesses in Dilbot's strategy though. If his opponent does get
in close to Dilbot while he's still sitting on his home square, Dilbot
won't be able to best guess whether his opponent will try to ZAP him on the
home row or column (if his opponent moves in at all). Also, Dilbot exposes
himself to getting ZAPped and to OUCHes when he's doing his Zapper run.
And there are other weaknesses that can be exploited, such as the opponent
could ZAP Dilbot right when he takes a break to RADAR his opponent.
Oh, well, no strategy in this Showdown can seem to guarantee a win, but
maybe only improve the chances of a win; there will always be holes that an
opponent could exploit, either on purpose, or by dumb luck. But I hope the
above will work well for Dilbot. (A Runner type opponent will be in big
trouble against Dilbot).
++ Did Dilbot do anything differently if you went first or second?
Yes, but only slightly. Since he's primarily a sleeper, he attempts to WALL
in his opponent right from the start. If Dilbot goes first, he knows that
his opponent must still be on his home square and therefore will alternate
building a wall on the two squares immediately adjacent his home until
either he gets an OOPS (opponent managed to zap a wall just after it was
built and successfully moved), or it's not longer possible for the opponent
to reach Dilbot's home square in 33 moves. It's possible for the opponent
to get stuck in a cycle of ZAPs and MOVEs that result only in a bunch of
OUCH points for Dilbot until either the opponent breaks the WALLing cycle
(OOPS) or Dilbot makes his end game ZAPping run.
If Dilbot goes second, his opponent may have moved. So he uses his standard
WALLing algorithm to collect OUCH points and slow down his opponent.
++ This POTM could have been better if only you had ....
Perhaps provided the mediator program to save the time to write and test one.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
AT&T Labs / AT&T WorldNet Service,
Lincroft, NJ,
software developer,
the POTM,
I'm Bill Gates' illegitimate half-brother,
more "bot" programming with different goals, constraints, layouts,
commands, physics, etc.
=================================================================
======= WINNER =========== 6.5_KillerBytes ===
========CODE WILL BE POSTED ON WEBSITES =========================
=================
6.5_KillerBytes was submitted by Don Kirkby at DKirkby@Sierrasys.com
=================
++ What strategies did 6.5_KillerBytes use?
Combination Sleeper and Zapper. It tried to block with walls and when
the opponent came in range, fired blindly until it hit or was hit.
++ Did 6.5_KillerBytes do anything differently if you went first or second?
Not much, if it went second, it used a RADAR to see which way the
opponent had moved, otherwise it knew where the opponent was and dropped
a wall in front of him.
++ This POTM could have been better if only you had ....
Put more time into it. When I started, I had ideas to run an artificial
evolution system to breed competitors. I didn't quite get around to
that. Oh well.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
Consulting company in Vancouver, B.C., Canada. I play Ultimate Frisbee
and commute by bicycle.
I was thinking about a contest involving balancing objects on a plane
with a fulcrum in the middle. Either solo problem solving or head to
head competition. I have to flesh out the idea a bit more. I really
liked the sound of the previous competition where you had to wander
through a building.
=================================================================
======= FINAL_FOUR =========== SOS ===
=================================================================
=================
SOS was submitted by Aivars Zogla at fix.lv!uvsk@kcig2.att.att.com
=================
++ What strategies did SOS use?
I would call strategy of SOS as oscilator: to go to the opposite corner
of the short side then back home with tripple horizontal zapping after each
move. If something on the road - zap it and continue. No radars, no walls.
++ Did SOS do anything differently if you went first or second?
No.
++ This POTM could have been better if only you had ....
Millions of strategies. Lotery. Of course, attempt to prepare good adjuster
could be winning. But it's too much to do it for fun. I can't say that this
was the best POTM I took part. But it's only my point of view. I know many
of POTMers like such kind of competition. I wish them good luck this time.
++ COMPANY? LOCATION? JOB?
Teacher in Ugale Secondary School, Latvia.
DO FOR FUN?
I've noticed that my hobbies changes in a period of 5 years. This is second
year while taking a lot of fun in solving computer problems.
INNERMOST SECRET?
"The darkest night is before sunrise." I don't know exactly where I pick up
this, it could be from POTM's, but if you are trying to figure out something
in some area, this can realy help you.
POTM IDEAS?
Go on, Fred, go on!
=================================================================
======= FINAL_FOUR =========== DeputyDawg ===
=================================================================
================
DeputyDawg was submitted by Sam Wilson at samw@crss.esy.com
================
+ What strategies did DeputyDawg use?
The original submission was basically a Runner which would throw out walls here
and there. Toward the end of the game it went into a mode where it would
alternate moving with throwing a wall into the opponent's goal. I thought this
would keep the opponent from ever entering the goal. My 14 year-old daughter,
Beth, demonstrated a flaw in my reasoning, though, so I knew I had to try
something different.
I also realized that although the Runner type programs were scoring more points
against The_Marshall and were going to be seeded higher, the programs most to
be feared were those that could catch me in an unending sequence of ZAPs. I
wrote a simple minded Zapper and it beat the original DeputyDawg pretty badly.
(BTW, I used the TCL mediator provided by Gerald Williams, but I had to apply
several fixes before it worked properly.)
The current version is mostly a Zapper. It maintains a variable that keeps
track of how close the opponent can have gotten to my current row and column. I
do a RADAR at intervals to reset this variable. As long as I know I can safely
move (not move into a ZAP) I do so. If I think the opponent is in position to
be ZAPped, I try one. If I catch him in a ZAP, I ZAP forever. If he gets past
me somehow, then I go ahead and run for my goal.
I have come to the conclusion that there is no perfect strategy - which makes
this a very interesting problem. If I knew my opponent's strategy before a
contest, I think I could always produce a counter-strategy that would beat his.
But then a counter-strategy tuned to do well against that particular strategy
would be very likely to do poorly against many others. So I think it is quite
likely for there to be three strategies A, B, & C such that A beats B, B beats
C, and C beats A. I would be interested to see if a round-robin tournament
would produce different results from a single-elimination tournament.
+ Did DeputyDawg do anything differently if you went first or second?
Not really. Some of the variables are given different initial values, is all.
+ This POTM could have been better if only you had ....
The one change I would be most in favor of would be to have it take a turn for
your "gun" to recharge after a ZAP. This would, of course, take away the
ability to ZAP forever, and put more emphasis on actually reaching the goal
first.
An alternate to this might have been to put a limit on the number of WALLS and
ZAPS you could use in the game.
COMPANY? Raytheon E-Systems
LOCATION? Garland, TX
JOB? Software Engineer
DO FOR FUN? Coach and judge Odyssey of the Mind, Read, Play games
INNERMOST SECRET? I'd have to ZAP you if I told you.
=================================================================
======= ELITE_EIGHT =========== Renegade ===
=================================================================
=================
Renegade was submitted by John Engels at jengels@att.com
=================
++ What strategies did Renegade use?
Renegade follows a conservative hunt and destroy strategy. I say
conservative because Renegade will avoid getting into the line of fire
of the opponent. Instead, Renegade will get close and then wait in
ambush until the opponent crosses his line of fire. Of course, if the
opponent is following the same strategy, the result will be a deadlock.
If a deadlock situation occurs, Renegade will sense the deadlock,
determine who would win the game if the stalemate were to continue, and
attack only if he is behind on points.
++ Did Renegade do anything differently if you went first or second?
No
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
Company AT&T
Location Piscataway, NJ
Job Programmer/Analyst
Fun Skateboarding/Skiing/Wakeboarding
Innermost Secret I am hopelessly in love with a girl who
works in the next building.
=================================================================
======= ELITE_EIGHT =========== CVGA ===
=================================================================
=================
CVGA was submitted by Pedro Demasi at demasi@centroin.com.br
=================
++ What strategies did CVGA use?
Zapper: tried to find and shoot your opponent
Adjuster: tried to figure out what opponent did and adjust strategy fittingly
(it uses this one to decide where to build walls)
Defender: at some point, it build at least 6 walls in its base to make it
harder to the oponent to get there. And when it finds the opponent
is to close to the base, it keeps on bulding walls to avoid the
opponent's victory.
++ Did CVGA do anything differently if you went first or second?
No, it always start building walls in front of the opponent.
++ This POTM could have been better if only you had ....
... more time to elaborate my program, since I entered the
tourney too late!
... AND if I could write my program in Pascal (I LOVE PASCAL!) :))))
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I'm from Rio de Janeiro, Brasil. So, sorry for the bad English! ;))))
I'm a student of Computer Science at UFRJ
(Rio de Janeiro Federal University).
Yes, I did for fun! ;)
=================================================================
======= ELITE_EIGHT =========== Zappa ===
========CODE WILL BE POSTED ON WEBSITES =========================
=================
++ What strategies did Zappa use?
Zappa has a probabilistic strategy that varies according to the situation,
sort of a modified "Adjuster". If Zappa finishes ahead, its almost always
because it has succeeded in zapping the opponent repeatedly.
The probabilities of the strategy depend on:
1) the closest (either row or column) the opponent was to Zappa
at its last known position.
2) the closest the opponent is likely to be, based on study of
opponent's recent motion
3) a comparison of the distance Zappa and opponent have remaining
to go to reach their respective goals
Each combination of distances has its own set of probabilities for moving on a
row, moving on a column, zapping, using radar, and/or building a wall.
At the beginning, distances are large, so Zappa does a "Runner" strategy
until the opponent could get close. If the opponent is near the same
row or same column, Zappa doggedly follows a "Zapper" strategy.
(If there is a HIT, Zappa ZAPs forever.) If the opponent gets really
close to its goal, Zappa lapses into the "Autistic" strategy, repeatedly
building either WALL A 1 or WALL I 26.
Zappa calculates the current trajectory of the opponent, as well as the
direction the opponent had to go to get to the goal, then arrives at
a consensus direction and speed. The opponent's direction and speed
determine the direction for zapping as well as probabilities and
direction for moving.
++ Did Zappa do anything differently if you went first or second?
Only as a secondary results of the probabilistic strategy. For instance,
Zappa is more likely to build walls at the beginning when it goes first,
because of its exact knowledge of opponents position.
++ This POTM could have been better if only you had ....
Had more than 2 matches between opponents, or done a round-robin tourney.
With a probabilistic strategy, there may be a lot of luck involved on
any one game.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
Company: AT&T
Location: Florham Park
Job Title: Division Manager of the Speech Processing Software
and Technology Research Department
Innermost secret: My job title is too long to fit on a business card.
=================================================================
======= ELITE_EIGHT =========== Anti_Marshall ===
=================================================================
==================
Anti_Marshall was submitted by Ramkumar Srinivasan at rsrini@iiap.ernet.in
==================
++ What strategies did Anti_Marshall use?
Zapper : Just stay put at your initial position and keep Zapping till you get
your opponent tranquilized.
++ Did Anti_Marshall do anything differently if it went first or second?
Nope
++ This POTM could have been better if only you had ....
No zapping.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
Iam form INDIA. Iam doing my undergraduate studies in Electronics &
Communication.
For fun I write system programs (Disk utilities etc), play games etc.
POTM ideas : A TIC-TAC-TOE game.
Thank you Fred for a great contest. It was fun :-)
=================================================================
======= SWEET_SIXTEEN =========== Russian ===
=================================================================
=================
Russian was submitted by Paul Banta at Paul_Banta@NOTES.YMP.GOV
=================
++ What strategies did Russian use?
Russian
An Adjuster
What made this problem fun was the fact that you almost never had
enough information to know what the best move was. You could ZAP to the
right, but that didn't guarantee that you could MOVE right and not bump
into a WALL. Or, if you did a RADAR, that only told you where your
opponent USED to be, not where he currently was. I found two exceptions
to not having enough information. These are what my entry was built
around.
Basically Russian is a sleeper. The name "Russian" was meant to
disguise his true nature. Russian - "Rush In" - "Rush'n". I decided that
a pure sleeper could be beaten too easily, so I taught the old dawg some
new tricks. The main strategy (Player 1 Primary Strategy below) has been
the heart of Russian from the beginning and is the only thing that I'm
proud of. Everything else is just tweaks and percentages. Judging from
all the other scores against The Marshall, it looks like I am the only
one who stumbled across this strategy. I think it's fundamentally sound.
Player 1:
Primary Strategy: REPEATEDLY build WALLs on I 25 and H 26!! This
guarantees that my opponent cannot move from his initial spot without me
getting an OOPS. Think about it! So basically I provide my opponent with
ample opportunity to bump into WALLs until I see that he got out by
receiving an OOPS (The Marshall couldn't get out). This is an example of
actually having enough information - you know your opponent is on I 26 as
long as you're able to build WALLs without getting an OOPS.
Secondary Strategy: Take the BYE. Along the way, I found a way to beat
The Marshall with a bogus program (1460+ points). Then I incorporated
that strategy into my real program. In essence, I found a way to
reasonably detect when I was playing against The Marshall and alter my
strategy to beat HIM (Nothing personal Fred). That gave me a BYE in the
first-round. And YES, it WAS legal.
Secondary Strategy: More WALLs followed by ZAPping. If my opponent is
able to move from I 26 (not impossible, just detectable), I build a
series of WALLs that he will almost certainly bump into. Then, I ZAP
repeatedly across the A row for a while followed by alternating ZAPs down
the 1 column and across the A row.
Secondary Strategy: Guard against a Sleeper. If I'm still building
WALLs on I 25 and H 26 after 95 moves, I try to move right two squares.
This is just in case I'm playing against a sleeper who doesn't move at
all.
Player 2:
Primary Strategy: WALLs followed by ZAPping. Because player 2 goes
second, there is no corresponding strategy to player 1's WALLs at I 25
and H 26. So, I build a series of WALLs that he will almost certainly
bump into. Then, I ZAP repeatedly across the I row for a while followed
by alternating ZAPs up the 26 column and across the I row.
Secondary Strategy: Still on A 1? I tried to be somewhat intelligent
in detecting and reacting to another sleeper like myself. After building
the WALLs, I do a RADAR (approximately my 8th move). If I see that my
opponent is still on A 1, I try to beat him by either moving 1 spot away
from my home or by winning the RADAR tie-breaker. If I don't think I can
move, I just RADAR the rest of the game (~ 85 RADARs). OK, so I'm
grasping at straws trying to win on RADARs.
Both Players:
Secondary Strategy: Response == HIT? I'm sure most entries reacted the
same when they received a HIT. I repeat whatever my previous action was.
This was the other case in which you DID have enough information in order
to reasonably decide what your next action should be.
++ Did Russian do anything differently if you went first or second?
Yes! Player 2 has no guarantee of detecting when player 1 moves from
A 1. That is, Russian player 1 builds walls on I 25 and H 26 and is
GUARANTEED to get an OOPS when the opponent moves from I 26. If player
2 built walls on A 2 and B 1, it's possible for player 1 to get off of
A 1 without player 2 getting an OOPS. So, I can't use this strategy as
player 2. Instead, I go straight to building some walls and ZAPping. I
guess the strategies are very similar, but not identical. If I had a way
of guaranteeing an OOPS, I'd play that strategy as player 2 also.
++ This POTM could have been better if only you had ....
1) It would have been even better if The Marshall played as both player
2 AND player 1.
2) It might have been better if ZAPping didn't put your opponent to
sleep. This would have closed the loop-hole of repeated ZAPping.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
Company: TRW
Location: Las Vegas, Nevada, USA
Job: Database Programmer
Do For Fun: POTM, Volleyball, Hicking, Camping, Fishing
(Necessarily in that order).
Innermost Secret: Bozo
POTM Ideas: Rubick's Cube
: Solitare
=================================================================
======= SWEET_SIXTEEN =========== WileECoyote ===
=================================================================
=================
WileECoyote was submitted by Andrew Gauld at Andrew.G.Gauld@att.com
=================
++ What strategies did WileECoyote use?
Runner: basic strategy is to go for the opposite home base quickly
- curved path, leave wall on starting point, shoot at ending pt.
++ Did WileECoyote do anything differently if you went first or second?
No.
++ This POTM could have been better if only you had ....
Needed pre-run evaluation to better reflect final run. Was almost a
different game. I was tempted to build a solution that looked at the
date and used different strategies for the initial runs than for the
final.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
AT&T. Lincroft. PTSM. Music. I'll never tell. Nope.
=================================================================
======= SWEET_SIXTEEN =========== Wanderer ===
=================================================================
=================
Wanderer was submitted by Rich Taylor at taylorrs@cs.rose-hulman.edu
=================
++ What strategies did Wanderer use?
Runner with weighted random walk (stayed away from walls).
++ Did Wanderer do anything differently if you went first or second?
Nope
++ This POTM could have been better if only you had ....
More time. :)
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
Student, Rose-Hulman Institute of Technology
=================================================================
======= SWEET_SIXTEEN =========== Runaway ===
=================================================================
=================
Runaway was submitted by Hank Turowski at turowski@accessone.com
=================
++ What strategies did Runaway use?
Runaway is a Runner. It puts up a minimal defense by building walls at A1
A2 and B1 for an A1 start or I26 I25 H26 for an I26 start. Then it runs
for the goal. On the INIT command, Runaway generates a random path through
the room, and writes its whole sequence of moves out to the temp file.
On results other than HELLO, OUCH, and ERROR it executes the next command
in the file.
On HELLO and OUCH, it will zap and on ERROR it will HELP and generate a new
movement file.
++ Did Runaway do anything differently if you went first or second?
Yes, when I moved first, I build walls at H26 and H25. With random
movement, there is a 75% chance that my opponent will run into at least one
of these walls.
++ This POTM could have been better if only you had ....
Maybe a round robin instead of a single elimination tourney... I know there
were a heck of a lot of entrys for this, but it looks pretty easy for a
weak program to get lucky and beat a very good program. Maybe groups of 5
or 10 programs in a round robin with the top one or two advancing to the
next group, not unlike World Cup matches.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I work for the Edmark Corporation in Redmond Washington. I spend most of
my days fixing computers and doing network administration. For fun, I play
soccer (lots of soccer) and I volunteer locally in the Deaf Community.
Innermost Secret: I can never remember that damn Quadratic Equation. I
had to cheat in High school Algebra and write it inside my calculator. But
that was just the beginning. I don't want to go into the details here, but
I will tell you that I have Avogadro's number tattooed on my left arm. I'm
just glad my C++ compiler has on-line help, or I'd be in big trouble.
=================================================================
======= SWEET_SIXTEEN =========== stalker ===
=================================================================
================
stalker was submitted by Bernard Hatt at bmh@arkady.demon.co.uk
=================
++ What strategies did stalker use?
The basic strategy was runner, the only modification was to
put a wall on the opponents home square and zap at it's own
finishing square on the assumtion that the opponent had done the
same.
[An unfinished version was attempting to detect the opponent and
synchronise putting walls on it's final square, so that it could be
doing (MOVE(OUCH), ZAP(BOOM), MOVE(OUCH) etc., whilst stalker
was doing WALL, MOVE, WALL etc.]
++ Did stalker do anything differently if you went first or second?
If it went second, it reversed the coords printed on input
and output, so that it always thought it was going A1->I26
++ This POTM could have been better if only you had ....
More free time ....
++ COMPANY?
I had been made redundant just before the POTM was announced,
so I managed to get a prototype entry and mediator going
fairly quickly. Fortunatly for my bank balance, and unfortunatly
for my POTM entry I got a job at a company called UniSoft
shortly afterwards.
++ LOCATION?
Hemel Hempstead (~20 miles NW of London, England)
++ JOB?
Programmer for UniSoft (Based in central London)
=================================================================
======= SWEET_SIXTEEN =========== sneaky_devil ===
=================================================================
=================
sneaky_devil was submitted by Kurt Van_den_Branden at kurtvdb@belgonet.be
=================
++ What strategies did sneaky_devil use?
Sleeper: stayed pretty much on your own home and built walls, etc.
In the beginning I thought my strategy was pretty devious.
I started randomly walking towards my goal, but when I think my opponent
is pretty close to his goal I start putting wall's at his goal-point.
So, if his only strategy is to walk to his destination, he will keep
hitting his head against the wall, and so I would win.
++ Did sneaky_devil do anything differently if you went first or second?
no
++ This POTM could have been better if only you had ....
a better strategy?
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I'm from Belgium, and I love the POTM. although I don't always have
time to compete
=================================================================
======= SWEET_SIXTEEN =========== showdown ===
=================================================================
=================
showdown was submitted by Sameer Mudaye at sbm@godrejnet.com
=================
++ What strategies did showdown use?
Runner: basic strategy is to go for the opposite home base quickly
and zap any walls encountered on the way
++ Did showdown do anything differently if you went first or second?
YES : If it starts at A1 it runs for I26 and vice versa .
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
It was great fun to participate and accept the challenge .
=================================================================
======= SWEET_SIXTEEN =========== spoof ===
=================================================================
=================
spoof was submitted by Edward Shipley at enshipley@attmail.com
=================
++ What strategies did spoof use?
Basically, spoof is a runner but with a set (1 or 2) of trigger points that
cause the program to take offensive or defensive action.
++ Did spoof do anything differently if you went first or second?
No
++ This POTM could have been better if only you had ....
I assume you is Fred. I think this is the most enjoyable POTM so far, and I
wish that it had been run as a series of tournaments in which users got more
detailed data (list of action for each turn). This would be sort of like the
pro tennis tour, with matches every one or two months, and time between to
repair defects and concoct sinister strategies. I'm sorry that this POTM is
ending.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
AT&T Middletown System Engineer For fun Yes The head to head competition
is a great idea.
=================================================================
======= BELOW_TOP_16 =========== tumbleweed ===
=================================================================
=================
tumbleweed was submitted by Dave Lynch at dfl@esun.ho.att.com
=================
++ What strategies did tumbleweed use?
Adjuster: tried to figure out what opponent did and adjust strategy fittingly
Use radar to determine location of opponent.
Try to slip by without getting zapped.
Then run for home.
++ Did tumbleweed do anything differently if you went first or second?
Not really.
++ This POTM could have been better if only you had ....
been able to work on it full time from 9 to 5, Monday through Friday.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I am an AT&T human resource temporarily located in Holmdel,
New Jersey. Nobody above my immediate supervisor has the
slightest idea about what I do, but they don't realize this.
Fortunately, I do have a very happy personal life.
=================================================================
======= BELOW_TOP_16 =========== robo2000 ===
=================================================================
=================
robo2000 was submitted by Carl in_'t_Veld at cveld@digitaal.nl
=================
++ What strategies did robo2000 use?
It just moved stupidly to the other corner. It only used the MOVE command.
++ Did robo2000 do anything differently if you went first or second?
Nothing special.
++ This POTM could have been better if only you had ....
More players in the field.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
Filled it out some time ago already.
=================================================================
======= BELOW_TOP_16 =========== runcible ===
=================================================================
=================
runcible was submitted by Brett Coley at brett.coley@tivoli.com
=================
++ What strategies did runcible use?
Runner, working on adjuster version. Hope to send in by
the end of the week...
++ Did runcible do anything differently if you went first or second?
No.
++ This POTM could have been better if only you had ....
More time. Facing project deadline, and have had little time
since mid-May. Still have ideas though.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM
IDEAS?
Tivoli Systems, Austin TX. Developer/designer of network
management software products.
Enjoy old cars, woodworking, music.
Like to read code (that is a twisted secret to have).
=================================================================
======= BELOW_TOP_16 =========== blast ===
=================================================================
=================
blast was submitted by Mark Studebaker at mds@eng.paradyne.com
=================
++ What strategies did blast use?
Runner: basic strategy is to go for the opposite home base quickly
Also Radar'ed every 5 turns or so.
When opponent is about to win, Blast stops moving and just
Wall's the opponent's home square every turn. This ensures that
Blast will never lose by opponent getting home.
++ Did blast do anything differently if you went first or second?
no
++ This POTM could have been better if only you had ....
more time to test.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
Paradyne Corp., Largo, FL
=================================================================
======= BELOW_TOP_16 =========== Walleye ===
=================================================================
=================
Walleye was submitted by Carl Burke at cburke@mitre.org
=================
++ What strategies did Walleye use?
Walleye was pretty simple. I borrowed some ideas from the Hamsterdam work
at the MIT Media Lab, giving Walleye multiple drives from which to select
at any given time. This allowed me to balance 'moving to goal', 'protecting
my goal', and 'know where your enemy is'. A similar technique was used for
deciding whether to zap or move if move was selected. For blocking, I tried
to put walls around my starting point, then walls along a direct path between
the enemy and my starting point. Walleye kept an internal map of the game
board for deciding where walls were, in addition to the value of decision
variables, current location, enemy location, and expected unertainty about
enemy location.
I never added any code to target the enemy directly, and a modified version
that was a little better at wall placement never got mailed to Fred.
I was not at all pleased at how poorly I did against The Marshal, but that's
what I get for not testing enough. :(
++ Did Walleye do anything differently if you went first or second?
No; the only adjustments were to the math, to make sure he headed in
the right direction.
++ This POTM could have been better if only you had ....
Tested, tested, tested! Some really simple strategies were almost
impossible for me to defeat (like just moving along the outer edge
of the board). The weights in the decision functions could have
been balanced better to have Walleye waste less time, etc. etc.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I work for the MITRE Corporation in Reston, Virginia, as a 'simulation
and modelling engineer' (fancy name for a programmer). I like to play
around with alife ideas and behavioral simulations; obviously, I need
to do some more playing around.
=================================================================
======= BELOW_TOP_16 =========== lurker ===
=================================================================
Sorry ... no description available for lurker
=================================================================
======= BELOW_TOP_16 =========== Showdownmaster ===
=================================================================
=================
Showdownmaster was submitted by M.Sabadello L.Heinz at
Markus.Sabadello@blackbox.at
=================
++ What strategies did Showdownmaster use?
It uses a quite passive strategy that doesn't depend on the home
position or on what the opponent does (no RADAR used). Showdownmaster
follows a pattern that is the same every time: The first thing it does
is a WALL next to the opponent's starting point (I 24 or A 3). This is
intended for rather stupid programs that do not avoid or zap blocks.
After that, Showdownmaster does a few MOVEs (max 3), ZAPing walls
which are in its way. Then, another WALL goes to the own starting
position. From then on, Showdownmaster uses horizontal ZAPs, in the
direction where the opponent started. These ZAPs go on until the end
of the game.
I know that this is very simple strategy, and I think it only works
well on programs which aim to get to the opposite home, but I beat
The_Marshall, so who cares...
++ Did Showdownmaster do anything differently if you went first or
second?
No. I thought about this possibility, but I couldn't find any reason
for different behaviour.
++ This POTM could have been better if only you had ....
...had a good mediator and The_Marshal's source code ;-)
++ COMPANY? LOCATION? JOB?
Just passed my finals at high school... Now I want to get my driving
license for cars, motorcycles and lorries. Where I live (Austria),
this much more difficult than in the US and you aren't allowed to
drive until you are 18.
INNERMOST SECRET?
Tell my innermost secret to 'bout 700 people I don't even know?
No way!
POTM IDEAS?
Nope
=================================================================
======= BELOW_TOP_16 =========== extrospector ===
=================================================================
=================
extrospector was submitted by John Sichi at jsichi@broadbase.com
=================
++ What strategies did extrospector use?
Sleeper, but Runner when the opponent is sleeping and Zapper when the
opponent is nearby. The wall building is based on the opponent's
current direction.
++ Did extrospector do anything differently if you went first or
second?
not really
++ This POTM could have been better if only you had ....
more free time...I wrote the first and only version while I was between
jobs, and then I began my new job at a startup, which is very intense,
and that was the end of POTM for me, at least for this round.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM
IDEAS?
Broadbase Information Systems; San Mateo, CA; database kernel engineer;
Jedi mind tricks on my daughter; ...
=================================================================
======= BELOW_TOP_16 =========== sailor ===
=================================================================
=================
sailor was submitted by Wieland Eckert at eckert@research.att.com
=================
++ What strategies did sailor use?
Initially, it was just a runner. I hesitated to call the program ``runner''
thus it was named sailor :-))
However, I changed the strategies a little bit towards a adjuster: I try to
find the opponents postition and if it is sitting at one of the borders I
assume it is a zapper! Other than that I just build a few obstacles at
the opponents goal since I assume that he wants to win (as I do): winning
gives most points, much more than zapping alone.
I thought of writing a zapper. But if the zapper would miss the opponent
he is doomed. Compare a zapper and a runner: If one of them is just a little
bit intelligent (i.e. not a pure runner/zapper) then the other one is done.
So my strategy had to be something inbetween. But the primary strategy
is to be quite quick.
I hope that both the runners and the zappers will be killed quite early
in the tournament :-)))
++ Did sailor do anything differently if you went first or second?
Yes, when I'm second I have to build the goal blocking walls earlier.
But no strategy differences.
++ This POTM could have been better if only you had ....
Hmm, I think this is about my ingenuity against yours. It's about strategy,
not programming art. And with strategies there is always the problem of
missing just a tiny case ...
How about running the tournament with different rules: everyone against
everyone and summing up all points???
But I start to be blasphemous against the god of all POTMs...
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
AT&T Labs Research, recently moved from Murray Hill, NJ, to Florham Park, NJ.
Working on speech dialogue systems.
Fun? Who has fun???? OK, I shoot some Photos (the old fashioned chemical
way, not bit-shuffling)
Secrets? Find out yourself!
=================================================================
======= BELOW_TOP_16 =========== Tie_Fighter ===
=================================================================
Sorry ... no description available for Tie_Fighter
=================================================================
======= BELOW_TOP_16 =========== Trevlix ===
=================================================================
=================
Trevlix was ssubmitted by Tyler Hudak at tyler@uakron.edu
=================
++ What strategies did Trevlix use?
I try to get to the home base asap, but if i see my opponent,
i zap 'em to death
++ Did Trevlix do anything differently if you went first or second?
nope
++ This POTM could have been better if only you had ....
women in bikinis!!!!! actually, i think it was pretty good/
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
im a systems administrator in akron oh. my innermost
secret is i'm actually bill gates love child. :)
=================================================================
======= BELOW_TOP_16 =========== Sundance_Kid ===
=================================================================
Sorry ... no description available for Sundance_Kid
=================================================================
======= BELOW_TOP_16 =========== BlindFury ===
=================================================================
=================
BlindFury was submitted by Brian Schladen at schladen@itginc.com
=================
++ What strategies did BlindFury use?
Runner: basic strategy is to go for the opposite home base quickly.
Blaster: Assume that most strategies are going to put walls around
home base and blast them before arriving to save time.
Will lose every time to a Zapper if he is able to catch
me. Try to zig zag around so that the Zapper has less
chance of catching me.
I think that I had a pretty weak strategy, the mediator beat me
every time....
++ Did BlindFury do anything differently if you went first or second?
No.
++ This POTM could have been better if only you had ....
More time to work on it. I stopped after about the second week.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM
IDEAS?
I write software for ITG Inc., a S/W factory in Culver City,
California. Play Guitar for fun. I really hate these puzzles, but
cant keep away from them.
=================================================================
======= BELOW_TOP_16 =========== Potter ===
=================================================================
=================
Potter was submitted by Steve Robbins at steve@nyongwa.montreal.qc.ca
=================
++ What strategies did Potter use?
The strategy is basically a runner. It follows a fixed, non-backtracking
course to the goal location. I chose the paths by hand, sort of randomly,
but making sure I didn't follow the "obvious" ones: along the edges or the
diagonal.
In addition, it drops exactly three walls to hinder the opponent. One on
the opponent's goal, and one on each connecting square. The opponent must
encounter two of them.
If potter happens to bump into (HELLO) or zap (HIT) the opponent, it will
continue zapping until time runs out.
The adaptive bit is just this: if we run into a wall (OUCH) we switch to
zap/move mode, doing a zap before each move. When a zap misses, we switch
back to run-mode.
Finally, in case someone drops walls only on my goal and adjacent square,
(like potter does) we make sure we go into zap/move mode just before the
penultimate move.
++ Did Potter do anything differently if you went first or second?
Yes, the path is different; i.e. it is not the inverse.
++ This POTM could have been better if only you had ....
Run the tournament in round-robin fashion, rather than single knockout.
Also, in case some people use randomized algorithms, it might be
interesting to run "best of X" kind of matches (for each starting
location).
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
What is the question? I did it just for fun. I like these types of
program-a-strategy games.
=================================================================
======= BELOW_TOP_16 =========== Wanderlust ===
=================================================================
=================
Wanderlust was submitted by Graham McDermott at gmcd@lucent.com
=================
++ What strategies did Wanderlust use?
A sort of combination of these three.....
Sleeper: stayed pretty much on your own home and built walls, etc.
I stayed on my home base, using radar, walls and zaps.
Zapper: tried to find and shoot your opponent
I tried to detect (& anticipate) when the opponent was on or
about to enter the last row or column in order to send a zap.
Once I successfully zapped, I continued to zap until the end.
Adjuster: tried to figure out what opponent did and adjust strategy fittingly
While my opponent was wandering around the centre of the arena I
tried to build walls in places they would bump into, or if they
stayed still long enough I would build walls all round them
Not a very complicated strategy, just looked at what direction they
were travelling and built a wall in front of them.
++ Did Wanderlust do anything differently if you went first or second?
Only one thing, If I went first I built a wall on H25, which they
would hopefully bump into if they tried the zig-zag approach across
the board.
++ This POTM could have been better if only you had ....
Time. Time to come up with a better solution, time to research the
problem, but unfortunately I'm trying to arrange my wedding and
manage learning about a new job.
++ COMPANY?
Currently Lucent Technologies, soon to be Catapult Communications.
LOCATION?
Wiltshire, in the UK (specifically Malmesbury).
JOB?
Software Engineer.
DO FOR FUN?
Cinema, Bowling, Softball (Badly),
INNERMOST SECRET?
Its a secret so I'll rot-13 it and then you can look at it only if
you're interested:-
Abar bs lbhe ohfvarff - ohg V znqr lbh ybbx naljnl!
POTM IDEAS?
=================================================================
======= BELOW_TOP_16 =========== I_Still_Have_a_Day ===
=================================================================
=================
I_Still_Have_a_Day was submitted by Darren Davis at jpdavis@webspan.net
=================
++ What strategies did I_Still_Have_a_Day use?
My strategy was to try to box my opponent in (for the 1 in a
billion chance that they don't know how to knock down walls), to move
out of my starting space vertically, and then put walls around my
starting space. Then my main strategy began. There was a 50% chance of
building a wall somewhere on the board, and a 50% chance that I would
try to move. To move it would go randomly toward the destination by
moving in the row if:
1000*y-distance-to-go/(y-distance-to-go+x-distance-to-go)<=a random
number between 0 and 999.
It recorded its move to the temp file, and on the next turn, if
the response wasn't OK, it would zap the position it thought it was in,
and would change its internal variables to its real position.
The temp file contained the following information:
current row (or what it thought it was)
current column (or what it thought it was)
number of turns of strategy used - the initial 8 turns are the
same in my strategy, but if the enemy puts a wall in one of the moves, I
need to zap it, and redo my move. (But I never actually TESTED that it
worked)
if I was player 1 or 2
a flag to determine the direction of the last turn if it was a
MOVE.
++ Did I_Still_Have_a_Day do anything differently if you went first or
second?
No, except to move up and to the left instead of down and to the right.
++ This POTM could have been better if only you had ....
Heard about it when it was announced instead of the MIDDLE OF JUNE!!!
Some advice: Don't start POTM 5 days before it is due -- Even if you
had to study for finals!
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM
IDEAS?
Job: I just finished my sophomore year of High School, so my job is a
STUDENT!
IDEAS: Allowing programs to be submitted in ANY language would be nice!
=================================================================
======= BELOW_TOP_16 =========== ninja ===
=================================================================
Sorry ... no description available for ninja
=================================================================
======= BELOW_TOP_16 =========== Timage ===
=================================================================
=================
Timage was submitted by Matt Peters at mrp@acpub.duke.edu
=================
++ What strategies did Timage use?
Basically, Timage was a runner. Every so often (and in fact, exactly as
often as necessary, in the worst case), it did a radar, to find out if
the opponent was in a threatening position. If not, it just kept moving,
but if so, it switched to a "close encounters" strategy. In this mode,
Timage tried to get by without getting hit, as a first priority, and
maybe zapping the opponent on the way. I must admit, I goofed on my
strategy: if the opponent just fired continuously, it would get me. But
I was too busy (lazy) to fix it.
++ Did Timage do anything differently if you went first or second?
Nope! Except for figuring the possible positions, which depended on who
went first, Timage was a happy-go-lucky fatalist.
++ This POTM could have been better if only you had ....
I've never been good at these questions. It would have been better if
you gave a graphical Java program on the web wherein we could see our
matches played, but other than that kind of unreasonable wish, I can't
think of anything.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I'm an undergrad at Duke, probably a physics and comp sci major. I play
frisbee, and I'm a big fan of rpg's, paper or electronic. Right now, I'm
trying to learn Java.
=================================================================
======= BELOW_TOP_16 =========== LinearStochasticIdiot ===
=================================================================
=================
LinearStochasticIdiot was submitted by Bob Lied at lied@lucent.com
=================
++ What strategies did LinearStochasticIdiot use?
Runner: basic strategy is to go for the opposite home base quickly.
Randomly, it drops a wall near the opponent's goal, and infrequently
it zaps, just in case it might get lucky. Movement is also random,
although always in the direction of the goal.
++ Did LinearStochasticIdiot do anything differently if you
went first or second?
No. Making it play symmetrically was part of the fun. And way easier
to test.
++ This POTM could have been better if only you had ....
The POTM is a most excellent diversion, and you are to be commended, sir.
++ COMPANY?
Lucent Technologies
LOCATION?
The Planet Bob, a small cubicle rotating in a temporary modular
building around an office complex in Naperville, Illinois, USA
JOB?
software dilettante
DO FOR FUN?
basketball, classical guitar, processes for software reuse
INNERMOST SECRET?
Live by the words of the Tao: The master has no goals, and
so everything he does is a success. (Very timely for those
mid-year performance reviews.)
=================================================================
======= BELOW_TOP_16 =========== New_kid_in_town ===
=================================================================
=================
New_kid_in_town was submitted by Daylan Darby at daylan@baker.ds.boeing.com
=================
++ What strategies did New_kid_in_town use?
Does anybody remember the APPLE ][ game "RobotWars"?
Can anybody still crank up their APPLE ][ and play it?
My strategy for New_kid_in_town (my first ever POTM entry) was half
runner/half zapper.
Basically try to make it home, but use RADAR (ALOT!) to keep track of
the opponent
and ZAP when appropriate. And if necessary track down the opponent and
ZAP it to death.
++ Did New_kid_in_town do anything differently if you went first or second?
The 99th move of the second player was always a move (A move towards
home is worth more than a ZAP?).
++ This POTM could have been better if only you had ....
Intermediate runs against other oppenents. I went crazy building a
strategy to beat the RUNNER, then a strategy to beat the previous strategy...
Eventually I ended up with a worthless strategy (against a RUNNER).
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
BOEING
KENT (10 miles south of Seattle), WA
Computer Security Research
For fun I play with my wife and kids, design and build stuff (an
addition, a deck...)
My innermost secret is "Computer Bum", for which I am sorry. (If you
don't understand it didn't happen to you).
I like competing directly (not timewise) against other opponents.
Variations on this same theme might include: 4 players (corners)
at a time. Random unzappable system walls. ZAP = x engery points,
MOVE = y energy points, only have Z energy points to
start. A bigger board with all contestants (randomly placed) at once
(no homebase radar only works in one direction at a time). 3 dimensions.
Other ideas include:
Number game. You give a series of numbers (or letters or words) and our
program must be the fastest to determine the next in the series.
Number game 2. You give a series of number (or letters or words) and
our program must be the fastest to determine which number doesn't belong.
Rock, paper, scissors. 3 oppenents try to win the most number of times.
We each get a file to store history. This can be expanded to 0..n
where x beats x-1 (0 beats n).
(Similiar in strategy to the current game "what will my opponents do next?").
Limited Core war. I get randomly assigned a number between 0..n, so
does my opponent. I can either move to one of 0..n or erase 0..n
If I move onto or erase my opponent I win!
=================================================================
======= BELOW_TOP_16 =========== FlyingSnake ===
=================================================================
Sorry ... no description available for FlyingSnake
=================================================================
======= BELOW_TOP_16 =========== BountyHunter ===
=================================================================
Sorry ... no description available for BountyHunter
=================================================================
======= BELOW_TOP_16 =========== RAMPOtm ===
=================================================================
Sorry ... no description available for RAMPOtm
=================================================================
======= BELOW_TOP_16 =========== ghost ===
=================================================================
=================
ghost was submitted by Fan Xiaobo at fxb@csnet1.cs.tsinghua.edu.cn
=================
++ What strategies did ghost use?
All the decision of which action should be taken next are based on the
probability, which is relevant to the weight of every action, which is
decided by many factors as well as their relations, such as the self
"surplus" (the additional action which can be taken except "MOVE") and
current position, the estimated opponent "surplus" and its position in
history, moving direction and so on. The details are followed:
1. According to the history and last action recorded in the tmp file,
and the reponse from the mediator, the newest info (surplus, position) is
acquired. Because of the special effect of "ZAP", if hit, continuous
"ZAP"; otherwise, goto 2.
2. Compute the weight of "MOVE", "RADAR", "WALL" and "ZAP". Here's
not so much definite explaination, but the result of sense and
ever-adjusting. Some relations are linear, some are not. Goto 3.
3. Generate a random number, see where it locates and take the
corresponding action.
++ Did ghost do anything differently if you went first or second?
Since the opponent's "MOVE I 25" is very probable, I think, if ghost goes
first, it'll "WALL I 25". If ghost goes second, it'll "MOVE H 26".
++ This POTM could have been better if only you had ....
The source of The_Marshall.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
Tsinghua Univ.
Beijing, China.
Graduate.
Yes.
I plan to make some more complex mathematical models, for instance, the
Petri Net, stochastic process and linear programming, but have no time to
fulfill.
May be the renju game (making five in a line).
=================================================================
======= BELOW_TOP_16 =========== Prophet ===
=================================================================
=================
Prophet was submitted by Matthew Jensen at who2@sfsu.edu
=================
++ What strategies did Prophet use?
Sleeper: stayed pretty much on your own home and built walls, etc.
At first, Prophet just built barriers around itself and sat in his
corner waiting for someone to zap.
I am trying to make last minute adjustments that ALMOST completely
change that strategy.
My only true stroke of genius working on this contest was my last
defense. I'm sure someone else thought of it, but my last defense was,
if the opponent was just a few spaces from winning, the Prophet would
just keep sticking walls in the goal. No one wins, but Prophet racks up
points as the opponent's program thinks that it has almost won. The
adjustments that I've been trying to make are to facilitate the opponent
getting to
this last defense faster without the Prophet getting zapped at all in
the proccess.
Adjuster: tried to figure out what opponent did and adjust strategy fittingly
If the opponent sleeps to long, the Prophet wants to move and at
least make a few points.
++ Did Prophet do anything differently if you went first or second?
Nope.
++ This POTM could have been better if only you had ....
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I am a Computer Science student at San Francisco State and working
this summer full time for Counsel
Connect which provides services for lawyers on the interenet. My
innermost secret is that I worked on this
contest for fun and spent WAY too much time on it.
=================================================================
======= BELOW_TOP_16 =========== XyBot ===
=================================================================
=================
XyBot was submitted by JT_Traub TA_Popiel at jtraub@dragoncat.net
=================
++ What strategies did XyBot use?
XyBot was meant to be an adjuster. Unfortunately, Alex and I ran out of
time in our busy schedules, and never completed all of the functionality.
Other than that it radared every so often, and if we were within some
distance of the other robot we did attempt to zap it. If we got past the
other robot without shooting it, we stopped radaring (since we were 'safe')
and just became a runner.
++ Did XyBot do anything differently if you went first or second?
Yes, if we went first we put a wall randomly to one side of the opponent in
hopes of a quick few points.
++ This POTM could have been better if only you had ....
The only thing I would have liked to see would have been the seeding trials
being run the same as the competition. IE: run us against the Marshall
twice, once going first and once going second.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
JT Traub works for Monolith Productions a computer game design company
in Kirkland WA. For work and for fun he writes computer games, and tinkers
with code of all sorts, shapes and sizes.
Alex Popiel works for Pensare Inc. an educational software company in
Menlo Park CA. For work and for fun he writes computer games, and tinkers
with code of all sorts, shapes and sizes.
And yes, we're very much alike.
=================================================================
======= BELOW_TOP_16 =========== Shot-in-the-Dark ===
=================================================================
=================
Shot-in-the-Dark was submitted by Pat Marcell at Patrick_Marcell@isp.csg.mot.com
=================
++ What strategies did Shot-in-the-Dark use?
Shot-in-the-Dark (SITD) is primarily a "Runner". It has a set sequence of
states which take it from its home base to the opponent's home base as well
as building walls for protection and to slow down the opponent. If SITD
runs into a wall, he zaps it and proceeds according the state machine -- SITD
is not smart enough to find an alternate route. At certain strategic
positions on the board SITG performs "RADAR" commands. If it is determined
that the enemy is closer to its objective than SITD, then SITD goes into
a defense mode in which it builds a wall on its home base (the opponents
goal) every other turn, as it continues to move toward its goal. If the
opponent is not smart enough, it will continue a sequence of zapping and
moving into the wall, giving SITD additional points and time to reach its
goal.
++ Did Shot-in-the-Dark do anything differently if you went first or second?
The state machines are different for going first or second, so the strategy
can easily be modified to make each behave differently. However, right now
there isn't a significant difference in the way I've got the state machines
set up.
++ This POTM could have been better if only you had ....
1.) Started with a random wall configuration.
2.) Set a limit on the number of consecutive times you could zap your
opponent before he/she could move again.
3.) Perhaps change the shape of the playing field or add some fixed barriers.
4.) Change RADAR to return list of objects (i.e., walls, opponent) within
a certain range of current position.
5.) Make running into a wall send the person back some number of spaces.
++ COMPANY? Motorola
LOCATION? Libertyville, IL
JOB? Software Engineer for Iridium Cellular Infrastructure
Group
DO FOR FUN? Golf, fish.
INNERMOST SECRET? If I told, it wouldn't be a secret.
POTM IDEAS? How about a board game like chess, checkers, or
othello?
=================================================================
======= BELOW_TOP_16 =========== Blind_Mans_Bluff ===
=================================================================
Sorry ... no description available for Blind_Mans_Bluff
=================================================================
======= BELOW_TOP_16 =========== levinb ===
=================================================================
Sorry ... no description available for levinb
=================================================================
======= BELOW_TOP_16 =========== zigzag ===
=================================================================
Sorry ... no description available for zigzag
=================================================================
======= BELOW_TOP_16 =========== Lazy_RISC ===
=================================================================
Sorry ... no description available for Lazy_RISC
=================================================================
======= BELOW_TOP_16 =========== StateWarrior ===
=================================================================
=================
StateWarrior was submitted by Rick Dickinson at rtd@notesguy.com
=================
++ What strategies did StateWarrior use?
StateWarrior was basically a runner, but also did some wall-building and
radars. If it thought that the opponent was close enough, it would try to
shoot.
I implemented a six-stage state machine, thus the name. The sequence of
actions was as follows:
1. Move
2. Move
3. Wall
4. Move
5. Radar
6. Move
At each step, if all was "OK", I would first check to see if I should try
to zap, and, if not, I would take action according to the current state.
Moving was done with a 25% chance of going in the X direction (A-I), and a
75% chance of moving in the Y direction (1-26). If one way was blocked,
the other direction would always be picked. If both were blocked, I would
pick a direction to zap according to the same percentages.
Wall building was completely random -- I just picked an X and Y and dumped
a wall there. One out of every 6 turns was spent "RADAR"-ing, to update my
opponent's position.
To check if I should Zap, I added together the minimum (X or Y) distance to
the opponent (last time I knew where he was) and the number of turns since
I saw him, picked a random fraction, multiplied by the total, and, if it
was < 1, I tried to Zap. If the difference in X coordinates was smaller, I
tried to zap in the Y direction, and vice versa. Looking back on my code,
I see that I always tried to zap towards the destination, even if the
opponent was behind me.... I should change that (and probably will before
the finals).
++ Did StateWarrior do anything differently if you went first or second?
The destination was different, but the algorithm was identical.
++ This POTM could have been better if only you had ....
I have never programmed in Java before, so I had a number of syntax issues
to overcome. I spent quite a bit of time getting my program to compile
that I could have been using to improve the strategy.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I work for Enterprise ArchiTechs, a Lotus Premium Business Partner
specializing in messaging and groupware solutions. We're located in
California, from San Diego to the San Francisco bay area. I, personally,
am a Senior Consultant, and work and live in Burbank. I've written several
articles for the Lotus Notes Advisor magazine, and really love my work.
For fun, I like to program.... I also have a love for folklore and urban
legends, and am a member of the San Fernando Valley Folklore Society, and
try to keep up with the usenet newsgroup alt.folklore.urban. I'm currently
teaching myself Java (about my 10th or 12th programming language), and plan
on becoming Sun certified by the end of the summer.
My innermost secret? Hmmm.... I don't know if it's a secret (my wife
would argue that everyone knows it), but I tend to go overboard when I
start learning something. It consumes all of my free time until I master
it, and then I start looking for something new. I tend to not get much
sleep when I start on something new....
As far as ideas for POTM contests, I like the idea of actual head-to-head
competition. Projects like the current one, where our programs can duke it
out against each other, are very good. Maybe a competition where we play
checkers against each other or connect-four or Pente or something would be
fun. Or, how about writing programs to play poker or blackjack? The
moderator program could deal the cards, and our programs could decide which
cards to keep or discard, how much to bet or raise, or, for blackjack, how
much to bet, whether to hit or stand, and whether to split or double down.
Blackjack might be a good one -- we each could start with a fixed dollar
amount, say $1000. The moderator could play the dealer, using Vegas rules.
Up to 6 players would be at each "table", and we could play 100 hands.
The top however many at each table would move on to the next round, and
play the winners from other tables....
- Rick "Or how about playing Gin?" - Rick
=================================================================
======= BELOW_TOP_16 =========== Iyriasi ===
=================================================================
Sorry ... no description available for Iyriasi
=================================================================
======= BELOW_TOP_16 =========== Terminator ===
=================================================================
=================
Terminator was submitted by Bill White at cosc002d@frank.mtsu.edu
=================
++ What strategies did Terminator use?
Runner: basic strategy is to go for the opposite home base quickly.
Terminator ran for the target, blasting anything it bumped into.
++ Did Terminator do anything differently if you went first or second?
No, he's quite a lame strategist.
++ This POTM could have been better if only you had ....
Spent more time on it! Had to abondon it after the first attempt.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I'm a computer science graduate research assistant for the University of
Tennessee Space Institute. I most enjoy discovering new aspects of computer
science; consequently, much of my time is spent programming or reading
about programming. Currently, my focus is on scientific visualization of
multivariate datasets. Any remaining time is spent at Far East Chinese
Restaurant!
=================================================================
======= BELOW_TOP_16 =========== bababooey ===
=================================================================
=================
bababooey was submitted by John Saalwaechter at saalwae@host1.in-motion.net
=================
++ What strategies did bababooey use?
Bababooey is a runner. Hopefully before the finals I'll submit
the changes so that bababooey starts watching his opponent and
and building walls if the opponent gets too close to his objective.
Otherwise bababooey simply runs for his target, zapping walls that
he finds in his way.
++ Did bababooey do anything differently if you went first or second?
No.
++ This POTM could have been better if only you had ....
More elaborate commands...
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I'm a unix analyst/admin in the Indianapolis area.
I'm surprised that I haven't ever seen a potm entry done
in LISP. I think I try to convince Fred to provide it for
the next contest. Actually, perl is my first love, but I
wouldn't mind lisping a little.
=================================================================
======= BELOW_TOP_16 =========== Sandman ===
=================================================================
=================
Sandman was submitted by William Stubbs at wstubbs@uow.edu.au
=================
++ What strategies did Sandman use?
Zapper: tried to find and shoot your opponent
Assuming that the other player actually tried to make it to the goal zone
then you have to cross paths at some time in teh x or y axis. Sandman's
strategy was to wander forward toward the goal and scan any time it was
possible that on teh next turn the opponent could lie on my x or y.
If this was the case then id zap in that direction and if i made a hit
zap until the game ended. If the opponent wasnt in the way then continue
moving forward toward the goal and mark the least scan time before they
could be in my x or y again
++ Did Sandman do anything differently if you went first or second?
Hehehe, Since i didnt really get much time after the initial testing sandman
doesnt even KNOW there is a first and second :( He just happily thinks he
always starts at A 1. See next point for an explanation.
++ This POTM could have been better if only you had ....
If only i had not been studying my final session for a Bachelor of Comp Sci
and also trying to hold down a full time job. Free time vanished after the
initial evening that i wrote sandman.. His total development time was around
90 minutes.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
Job: Systems Operator (Uni of Wollongong Australia) (ex student (YES!))
Fun: Taking on the RSA with the RC5 challenge (http://rc5.distributed.net)
Secret: Yep its exactly that
potm: Im mostly new to all of this but i like the idea of competition between
potm'ers than the solve xyz puzzle. Just imho of course.. but competiton
always gets the participation rates soaring =)
Grab some board game and make us do the AI, battleship, connect 4? etc
shrug
Looking forward to the next potm as i dont have that study thing eating
into my free time =)
=================================================================
======= BELOW_TOP_16 =========== Elmo ===
=================================================================
=================
Elmo was submitted by Gerald Williams at gsw@lucent.com
=================
++ What strategies did Elmo use?
Elmo is more an exercise in style than in substance. The first incarnation
was as a simple runner. What made Elmo unique is that it was implemented
in just a few minutes using about 20 lines of Tcl code. In fact, here is
an 11 line (rather unreadable) version:
proc p {c} {switch $c B {return A} C {return B} D {return C} E {return D}\
F {return E} G {return F} H {return G} I {return H} default {incr c -1}}
proc n {c} {switch $c A {return B} B {return C} C {return D} D {return E}\
E {return F} F {return G} G {return H} H {return I} default {incr c}}
gets stdin inp; if {[lindex $inp 0] == "INIT"} {set m [lrange $inp 1 2]
if {[lindex $m 0]=="A"} {set t {I 26}; set d n} else {set t {A 1}; set d p}
} else {source $argv; if {[lindex $inp 0] == "OK"} {eval $try}}
if {[lindex $m 1]==[lindex $t 1]} {set to [lreplace $m 0 0 [$d [lindex $m 0]]]
} else {set to [lreplace $m 1 1 [$d [lindex $m 1]]]}
set try "set m {$to}"; puts "MOVE $to"; set f [open $argv "w"]
puts $f "set m {$m}\nset t {$t}\nset d $d\nset try {$try}"
The current incarnation of Elmo is what I'll call a "scared runner"--it
heads straight for the opposite corner until the opponent MIGHT be getting
close. If RADAR determines that the opponent might be able to get a shot
off soon, some walls are built, then Elmo attempts to move past. Whenever
anything unexpected happens, Elmo starts blasting away. This version is
actually quite vulnerable to a persistent runner or ZAPPER...c'est la vie.
If Elmo ever does manage to ZAP an opponent, it just keeps ZAPPING, rather
than possibly giving the opponent a chance to ZAP back.
++ Did Elmo do anything differently if you went first or second?
Elmo is far too simple-minded for that.
++ This POTM could have been better if only you had ....
More time to try out strategies, I guess.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I work at Lucent Technologies (Bell Laboratories) in Allentown, PA. I am
currently developing assemblers and similar tools for programming fixed-
point DSP's (borrowing on our slogan, I make the things that let YOU make
the things that make communication work). Between working on my master's
degree, various sports, traveling, and owning a house, I have very little
time for "recreational programming" like this.
=================================================================
======= BELOW_TOP_16 =========== tihSbmuD ===
=================================================================
Sorry ... no description available for tihSbmuD
=================================================================
======= BELOW_TOP_16 =========== Bozo_II_The_Headache ===
=================================================================
=================
"Bozo II - The Headache" was submitted by Scott McCord at
scotti@cs.unlv.edu
=================
++ What strategies did "Bozo II - The Headache" use?
E) Other
Eighteen lines of the best code I've ever written - and that includes
comments! No, it doesn't use any files! Yes, it plays equally well as
either player 1 or player 2.
Speed is the key here. If you can make moves very quickly, your
opponent won't have time to think about his current position and will
be more likely to make mistakes. It worked for Deep Blue against Garry
Kasparov. I chose not to use files in order to reduce both my code
size and execution time. As we all know, I/O (especially to disk) is
EXTREMELY slow. Considering that a minimum of 1 disk read and 1 disk
write would be required per move, I found this to be an unacceptable
performance hit. Instead, I chose to build a bullet-proof state-
machine. Since HELP reports back your current location, I use HELP
in order to decide where to move. This is much (MUCH) faster than
retrieving that information from disk. I estimate that my execution
time is orders of magnitude faster than the typical POTM entry. I
have every confidence that this will translate to eventual victory
in the tournament.
In addition, I have the added edge of having the smallest source
and executable codes as well as the quickest running time. If I get
into a tie-breaker situation, these could be a factor.
I realize that Bozo hasn't fared well against The_Marshal.
Although I have been able to make some improvements (my original loss
by more than 2,000 has been trimmed to under 1,500 - a 25% gain!), I
fully expect that Bozo will clean up against opponents that aren't
prepared for the fast-paced tempo that Bozo delivers. Making it into
the final 4 is fully within the realm of possibility. On the other
hand, I've studied the results that Fred has sent me and I'm
convinced The_Marshall is doing something that is fundamentally
brilliant. Fred has made numerous claims to be not much of a coder,
but I know differently - Bozo vs. The_Marshal proves that Fred's
coding borders on genius at the very least.
Since my code is so small, I've included it for your benefit.
Notice that my array, s[], is of size 192. This allows it to
fit perfectly within word boundaries on machines with word sizes
of 1, 2, 4, 6, 8, 12, 16, 24, 32, 48, and 64 bytes - a further
optimization and portability feature that is often overlooked.
- - - - - - - - - -
/* POTM Entry: Bozo_II_The_Headache */
/* 18 lines of the best code I've ever written. */
/* Scott McCord (scotti@cs.unlv.edu */
#include
void main(int argc,char*argv[]){
char s[192];gets(s);
if(s[0]!='I'&&s[0]!='L')printf("HELP\n");
if(s[0]=='I'&&s[5]=='A')printf("WALL B 3\n");
if(s[0]=='I'&&s[5]=='I')printf("WALL H 24\n");
if(s[0]=='L'&&s[4]=='A')printf("MOVE B 1\n");
if(s[0]=='L'&&s[4]=='B'&&s[6]=='1')printf("MOVE B 2\n");
if(s[0]=='L'&&s[4]=='B'&&s[6]=='2')printf("MOVE B 3\n");
if(s[0]=='L'&&s[4]=='B'&&s[6]=='3')printf("MOVE B 2\n");
if(s[0]=='L'&&s[4]=='H'&&s[7]=='4')printf("MOVE H 25\n");
if(s[0]=='L'&&s[4]=='H'&&s[7]=='5')printf("MOVE H 24\n");
if(s[0]=='L'&&s[4]=='H'&&s[7]=='6')printf("MOVE H 25\n");
if(s[0]=='L'&&s[4]=='I')printf("MOVE H 26\n");
}
++ Did Bozo II - The Headache do anything differently if you went first
or second?
No. It plays both sides identically.
++ This POTM could have been better if only you had ....
This POTM could not have been better.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM
IDEAS?
Company: Nevada Power, Las Vegas NV. Software - Lotus Notes, Oracle.
Fun: Get out of town.
Inntermost Secret: I'm really a lesbian.
POTM Ideas: None
=================================================================
======= BELOW_TOP_16 =========== BrainDead ===
=================================================================
Sorry ... no description available for BrainDead
=================================================================
======= BELOW_TOP_16 =========== Road_Runner ===
=================================================================
=============================================
Road_Runner was submitted by Robert Morrison at rhmorrison@compuserve.com
=============================================
++ What strategies did Road_Runner use?
Runner: basic strategy is to go for the opposite home base quickly.
Specifically, Road_Runner goes directly to the exit square using
a straight line (yet random) path to reach the square. This means
that the random function is so weighted such that 17/25 of the time
it will move horizontally, the other 8/25 of the time vertically, and
always towards the goal.
When enough moves have been made such that the opponent could be
close to his exit three walls are thrown up (two directly next to
the exit and one right on the exit).
Should Road_Runner crash into a wall he attempts to force his way
thru by zapping the wall and trying again. Road_Runner vs Road_Runner
will always result in a TIE with each player receiving exactly the
same number of points.
++ Did Road_Runner do anything differently if you went first or second?
No, this was a FIRST ATTEMPT and after I realized the winner is
going to be a zapper program I just lost interest...
++ This POTM could have been better if only you had ....
More powerful commands. A scan command should not cause the
program to give up his turn. The ability to scan at any angle,
and more info. The ability to shoot at any angle, etc. I feel that
the capabilities of the robots were just too simplified to make for
any interesting algothims that have any chance of winning.
Take a look at C-Robots or P-Robots.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM IDEAS?
I work for Hitex Systementwicklung GmbH in Karlsruhe, GERMANY. I enjoy
playing Chess, Tennis, working on the computer, POV RAY, watching movies
and reading Science Fiction. I am an avid Star Trek fan and just
can't wait to see "The Men in Black".
=================================================================
======= BELOW_TOP_16 =========== lance_alot ===
=================================================================
Sorry ... no description available for lance_alot
=================================================================
======= BELOW_TOP_16 =========== HighNoon ===
=================================================================
=================
HighNoon was submitted by J.Eddings V.Groh at jeddings@ix.netcom.com
=================
++ What strategies did HighNoon use?
-- Two-pronged strategy: If I started at A1, Runner. If I started at
I26, Zapper. I figured I would kill two birds with one stone.
++ Did HighNoon do anything differently if you went first or second?
-- See above.
++ This POTM could have been better if only you had ....
-- Me personally? Built more "smarts" into my program. The program I
submitted was quite straight-forward. I wanted to get a working model
to a point where the smarts could be dropped in easily. But my job at
work ramped up, and became difficult to devote much time to.
++ COMPANY? LOCATION? JOB? DO FOR FUN? INNERMOST SECRET? POTM
IDEAS?
-- I work for Silicon Valley Internet Partners (Foster City, CA). I'm
an Internet consultant, and for fun, well, I do this. Program, play
around with my Linux box, read. My innermost secret? One word:
BOSCO. 8) POTM Ideas: Rat-in-a-maze learning program (set up an ASCII
grid maze, and work through maze.)
Thanks for the great fun! Keep up the good work!
=================================================================