Fire a big power bullet(always 3 power) then you know some avoider bot will try to avoid this bullet,then after 16 ticks,it's avoiding movement scheme was viewed,so then i fire a little power bullet(which will expect to reach him at the same time of last big power bullet),so if his aviod strategy is simple(eg:Cigaret
and some other mini-micro-bot),i will hit it very much.
Of course it is not effective when against some advance avioder eg:SandboxDT
however,in my testing it did very good against DuelistMini
You can try to use Spark or CigaretBH to try,some times you will see three bullets hit the enemy at the same time,it is beautiful.
but...some times one of my bullet will hit my another bullet,so...west two bullets.:(
This is a great idea. The question I would ask is: does it pay? I mean, do you hit often enough with a big enough bullet to make it better than firing a power 3 bullet with whatever hit percent your gun has? --tobe
Againt some bot it does pay.But ordinarily it does not,as Ash get 4# but Spark get 9# in ER.So,CigaretBH choose the better strategy with a certain bot.it did better than Cigaret.(CigaretBH is a bot just added a ChaseBullets strategy on Cigaret).
Did i get the point of the "pay"? -- iiley
Funny...i do pretty much the same thing. just a little different : TriggerBullets
No, TriggerBullets sounds more like AngularTargeting/Factored and AveragedBearingOffsetTargeting. The thing with ChaseBullets is that the second bullet has a velocity that will make it arrive at the target at the same time as the first, full-power, bullet. It's designed to collect hits agains bullet dodgers. It's more like Marshmallow/CowboyGuns but done the right way. -- PEZ
's current (better said: former;-) WhiteWhale
, et.predator 1.8, seems to use exactly this technique. When you don't expect these ChaseBullets
they can really mess up your beautiful movement algorithms :-) Vic
Correct. But back to the draing board on wverything else.
This idea seems to be similar to something I am working on to deal with wave surfers, which is essentially to drive them into a corner by firing lightweight bullets at it, to drive it into the corner. Once it is there, the idea is to limit the escape angles so that my enemy has to choose between driving into a bullet, and/or the wall, or sitting still and taking a hiding. It would take a reasonable level of sophistication to choose which to do, and several of the robots I have looked at would simply spin in a tight circle. I got the idea from a fluke win by my first robot against jp.Perpy, in which Perpy spawned in the corner and couldn't wave-surf because of the walls.