History/Concepts

• 2001 Sep - CoreBot? is the first bot to have an extendable framework that allows for the easy addition of new modes. CoreBot? has two of these "Cores," ramming in 1vs1 and wall-riding in melee.

• 2001 Sep - Graygoo invents bullet sensing/dodging. His first bot that used it was Grasshopper, but the strategy only looked so promising before he released Wolverine in November, which became the first undisputed king of 1vs1 by using it coupled with a pattern matching gun. Mat Nelson correctly predicted it would spawn a new generation of pattern-matching bullet-doding robots.

• 2001 Oct - Tobe's bot Relativity is the first to use AntiGravity movement.

• 2002 Mar - HCBot, the first human controlled bot, is released.

• 2002 Apr/May? - Qohnil introduces first bot using Neural Networks. (XBot)

• 2002 Jan/May? - The idea of classifying bots according to their code complexity spreads. Not sure how the idea appeared, but for sure Christian Schnell (who developed the CodeSize utility to measure the complexity of a bot), Tobe (who set up the MiniBotChallenge and defined the MiniBot, MicroBot and NanoBot categories) and M3thos (author of the Eva bots) deserve the credit. Nowdays these categories are standars in Robocode world.

• 2002 Jul - MogBot introduces a pattern matcher he claims is "simple", and becomes the first to publish an understandable article outlining the algorithm (http://realj.com/robots/MogBot.html).

• 2002 May/Jul? - Paul Evans introduces guessFactor/statistical aiming. The notable feature of this algorithm is that its the first designed to work against random movement.

• 2002 Jun/Jul? - The idea of a cooperative robot framework is born; Qhonil? and Vuen set out independantly to complete the framework, and both fail (Vuen because of a computer crash (I'll rebuild it, I promise! :D)).

• 2002 Dec - Nano introduces the Perceptual bots concept.

• 2002 Dec - With Cigaret, Iiley introduces the Wave concept to calculate the "right bearing to hit" without using VirtualBullets.
  • The real contribution with this is probably the descriptive name Wave. It was called Probe? by Dummy in his HummingBird? and lots of my bots used them under various names before I realized it was the same as Waves. Not that this contribution is small. Naming is hard to master art and important as well. -- PEZ

• 2003 Feb - NanoLauLectrik leads the way to a new generation of nanos introducing nanonized SymbolicPatternMatcher s, which take advantage of java built in functions to do the search.

• 2003 Mar - PEZ introduces the VirtualGuns targeting strategy by his main bot Marshmallow.
  • Was it in Mar? -- iiley
    • I don't really remember. I think it could have been Mars when I wrote the wiki page. Marshmallow was equipped with these from the very earliest publi versions. All the while the array has been tested and rated using Waves by the way. Which shows how versatile this concept is. -- PEZ
      • Isn't this another naming contribution? I seem to recall multi-gun bots based on accuracy data existing before this. -- Kuuran
        • Yes, it probably is. The idea is too simple to not have been explored earlier than 2003. -- PEZ
          • The oldest example of a bot using multiple guns I can find on my hard disk is okki.Robokki 0.4 by Ole-Martin Mørk. This last version was built for Robocode v 0.98.3, but I bet it wasn't the first bot to do this either. -- Kuuran
            • Just a check question. Was Robokki using VirtualGuns or was it just using multiple guns in a MultiMode type of way? -- PEZ
              • Definite answer: Ask the author :p My guess is, however, that is was, because it's debug output contains some pretty extensive data on all it's guns success after just one round, and they all seem to update quite frequently. Of course this could just be thrashing modes since none of it's guns are particularly amazing. -- Kuuran
                • I was using a kind of VirtualGuns, but as Kuuran pointed out none of the guns was really good. [Ole-Martin Mørk]?
                  • AH, But that goes for all my bots that has used VirtualGuns too. -- PEZ

• 2003 Mar - (Or there abouts) Kawigi sets the search light on the importance of CurveFlattening with his SandboxFlattener. By April, he released FloodMicro, with his first public attempt at this sort of movement, along with an adaptive element he termed "DynamicDistancing".

• 2003 Sep - PEZ hints that a lot of simple and effective inventions are still to be made. This hint was delivered with the rediculously simple movement of VertiLeach. A movement which made a very specialized and effective gun possible. I hereby name this gunning strategy ReducedDimensionalSegmentation. -- PEZ

• Who and when introduced WaveSurfing? -- Albert
  • I'd say that it was ABC introducing it. Even if he kept it a secret and even started to obfuscate his code, he left enough hints on the wiki to lead others of us to WaveSurfing. When... it was with the first Shadow that threatened DTs throne. Maybe it was in Feb this year. I'd also say that it was Axe's MusashiTrick that started the common thought process of fooling simple targeting that eventually got ABC to question the need for calls to random() for your movement. It takes a really open mind to question something like that. -- PEZ
  • I would even argue that iiley's Lacrimas may have been the first to put *most* of the concept into practice. It was the first bot that I know of that avoided the last guess factor that it was hit at. So you see it is a very difficult question indeed. -- jim