Difference between revisions of "Pattern Matcher Challenge/Results"

From Robowiki
Jump to navigation Jump to search
(my result + some cleanup)
(finally)
 
Line 27: Line 27:
 
|101.38 %
 
|101.38 %
 
|Oct 16 2009 / 4 seasons. [[Sequential Prediction]] on rotation, velocity and acceleration, 16, 16 and 3 symbols respectively.
 
|Oct 16 2009 / 4 seasons. [[Sequential Prediction]] on rotation, velocity and acceleration, 16, 16 and 3 symbols respectively.
 +
|-
 +
|[[HoldenPMC]] 1.0
 +
|[[Darkcanuck]]
 +
|101.38 %
 +
|Oct 18 2009 / single season. [[Sequential Prediction]] using 17 velocity and 17 delta-heading values.
 
|-
 
|-
 
|[[LeachPMC]] 1.1.10
 
|[[LeachPMC]] 1.1.10

Latest revision as of 18:48, 18 October 2009

PMC Sub-pages:
Pattern Matcher ChallengePattern Bot - Results
Bot Author Index Date of Victory / Comment
EnderPMC 1.3 Vic Stewart 101.75 % jun 17 2003 / Using deltaheading and deltavelocity for matching. Improved heading prediction and initial damage.
WeeklongObsession Simonton 101.61 % jan 4 2007 / Using the pattern matcher described on my page. Cheated a little by increasing my bot's delta-heading segments for the challenge :). (It scored 101.07 before that, though)
Sila pmc.1.0 Zyx 101.38 % Oct 16 2009 / 4 seasons. Sequential Prediction on rotation, velocity and acceleration, 16, 16 and 3 symbols respectively.
HoldenPMC 1.0 Darkcanuck 101.38 % Oct 18 2009 / single season. Sequential Prediction using 17 velocity and 17 delta-heading values.
LeachPMC 1.1.10 PEZ 101.15 % Jun 19 2003 / Variable length search vector using heading delta and velocity for matching and the same for projection (see its page for more info and progress)
GlowBlowPMC Rozu 101.08 % jun 15 2003 / Using headingchange and velocity for matching (does account for discontinuity between two rounds)
Nat's PM Nat 100.98% Oct 16, 2009; 5 seasons (best 101.11%); average win 67% of rounds. No optimise for PMC. Match on velocity/deltaHeading, 17 and 10 symbols. Sequential Prediction.
Nibbler/NibblerPMC Crippa 100.87 % jun 17 2003 / Using deltaheading and velocity for matching. Tweeked to get less skipped events.
Toa dev User:KID 100.80 % 20091019 / Sequential Prediction and Folded Pattern Matcher
MicroAspidPMC Albert 100.50 % jun 17 2003 / Variable window lenght pattern matcher using velocity & heading (without low pass filter). Considers beginning and end of rounds
ParakeetPMC Dummy 100.14 % Using sampled displacement for matching (sampletime=4 ticks,search vector consists of 8 samples)
Pytko Chase 100.08% I was bored enough to test it..
RooseveltPMC Brainfade 100.04% Just found a bug....

Chase-san/PatternBot Chase-san 099.32% Just your ol' fashioned pattern matcher (500:100000:PI/128:.25)
TeancumPMC Kawigi 098.55% Using lateral and advancing velocity for matching and projection, fixed length of 50 or 1/4 of the current round.
WaylanderMicro Skilgannon 095.98% Using variable key length and decoupled delta heading and velocity to decrease learning time. No data saving between games. No PMC specific optimisations.
Aryary's Pattern Matching (Releasing Soon in a MicroBot) Aryary 095.89% Matching Velocity and Delta Heading. (Fixed Some Bugs.)
Recrimpo's future gun Tango 083.97% Lots of work to do, yet.
mini.FenrirPMC 0.33i Loki 079.33 % Using lateral velocity and acceleration for matching, variable length pm
HeadOnTargeting 034.22% Just for comparison (3248/9492)

Guns above the line have completed the challenge and have beaten Pattern Matcher Challenge/Pattern Bot.