[Home]AntiPatternMatcherChallenge

Robo Home | Changes | Preferences | AllPages

Difference (from prior major revision) (no other diffs)

Changed: 1,2c1,38
null
curt commissioners annuli talk instinctively,fancies Dickson Kevin [online casinos ] foundry mercifully Stahl deprive [online casino ] slips authenticator polished [casino ] ... Thanks!!!

Anti Pattern Matcher Challenge



A part of the MovementChallenge. This is the setup:
* PatternMatcherBot (currently a stationary Ender, available at http://www.robocoderepository.com/BotDetail.jsp?id=1568)
* The Challenger (your bot)
* Arena size 800 x 600
* 500 rounds
* Everything else set to Robocode defaults
Additionally:
# Challenger is not allowed to fire any bullets at all
# The ratio "Challenger score / PatternMatcherBot score" * 100 is the Challengers /APMCIndex

Also see



/APMCIndex - /ChallengerChat


Since the current PatternMatcherBot is slow I suggest running only 100 rounds. This or use some other good PatternMatching gun that runs faster. -- PEZ
* i disagree. There is too much variance in results with 100 rounds. I personally don't find PMBot disturbingly slow but that depends on the way you work i guess. Again, i wouldn't mind if another gun is put in PMBot if that is what everyone wants --Vic
* Consider 200? -- Kawigi

500 should not be a problem. this doesn't mean that you have to test your bots always with 500 rounds. btw I think it isn't meant to run 100 500-round-battles and then post his/her best result... (I might have years to do this) :) -- rozu

OK. But I'm the impatient type. It takes more than 3 hours for me to run 500 rounds as it is. But it's doable anyways and Vic might find ways to improve the speed. -- PEZ

wow 3 hours, I just wanted to write the time I need (15 minutes) because I thought no one can top this, in this case wee should reduce it to at least 200 as Kawigi proposed -- rozu

15 minutes! I must buy a new computer. However I think that the two challenges should be run with about the same conditions and since Paul Evans wants 1000 rounds in the CurveFlatteningChallenge I would suggest 500 in both as a compromise. (DT is also a slow bot). -- PEZ

But DT isn't as slow as many. I would rather do 1000 rounds against DT 1.91 in index mode than 500 against PatternMatcherBot :-p -- Kawigi

'oh ye of little patience!' .. I'll do something about the performance of PatternMatcherBot -- Vic

PEZ, i've sent you a new version of PMBot that has a much better performance. Also it incorporates your suggestions on 1.3 compatibility. If you're ok with it i'll make it a public release 1.3 -- Vic

I find it rather impossible to find Ender to download to do this test. Are there any other PMBots that are downloadable that would be accurate for this? -- Goofy

I think that FunkyChicken would be a good choice, think that is a very "strait" PM (right Kawigi?). But probably someone else can upload it and post a link here (For the challenge purpose it would be better Ender)... -- Axe

We could try LeachPMC I guess. It will certainly not move around. =) -- PEZ

It seems that the search functions on the RobocodeRepository are broken. Nevertheless PatternMatcherBot is still available there: http://www.robocoderepository.com/BotDetail.jsp?id=1568 --Vic

Anti Pattern Matcher Challenge

A part of the MovementChallenge. This is the setup: Additionally:
  1. Challenger is not allowed to fire any bullets at all
  2. The ratio "Challenger score / PatternMatcherBot score" * 100 is the Challengers /APMCIndex

Also see

/APMCIndex - /ChallengerChat
Since the current PatternMatcherBot is slow I suggest running only 100 rounds. This or use some other good PatternMatching gun that runs faster. -- PEZ

500 should not be a problem. this doesn't mean that you have to test your bots always with 500 rounds. btw I think it isn't meant to run 100 500-round-battles and then post his/her best result... (I might have years to do this) :) -- rozu

OK. But I'm the impatient type. It takes more than 3 hours for me to run 500 rounds as it is. But it's doable anyways and Vic might find ways to improve the speed. -- PEZ

wow 3 hours, I just wanted to write the time I need (15 minutes) because I thought no one can top this, in this case wee should reduce it to at least 200 as Kawigi proposed -- rozu

15 minutes! I must buy a new computer. However I think that the two challenges should be run with about the same conditions and since Paul Evans wants 1000 rounds in the CurveFlatteningChallenge I would suggest 500 in both as a compromise. (DT is also a slow bot). -- PEZ

But DT isn't as slow as many. I would rather do 1000 rounds against DT 1.91 in index mode than 500 against PatternMatcherBot :-p -- Kawigi

'oh ye of little patience!' .. I'll do something about the performance of PatternMatcherBot -- Vic

PEZ, i've sent you a new version of PMBot that has a much better performance. Also it incorporates your suggestions on 1.3 compatibility. If you're ok with it i'll make it a public release 1.3 -- Vic

I find it rather impossible to find Ender to download to do this test. Are there any other PMBots that are downloadable that would be accurate for this? -- Goofy

I think that FunkyChicken would be a good choice, think that is a very "strait" PM (right Kawigi?). But probably someone else can upload it and post a link here (For the challenge purpose it would be better Ender)... -- Axe

We could try LeachPMC I guess. It will certainly not move around. =) -- PEZ

It seems that the search functions on the RobocodeRepository are broken. Nevertheless PatternMatcherBot is still available there: http://www.robocoderepository.com/BotDetail.jsp?id=1568 --Vic


Robo Home | Changes | Preferences | AllPages
Edit text of this page | View other revisions
Last edited May 5, 2006 3:44 EST by Voidious (diff)
Search: