Your Robot's Way Far! Errr...

     Hey, it's robot's again! Have you been perfecting your's since my last post? I guess you did since you're here. Well, I too have been working on making my robot better. By utilizing some of the things I learned from the katas, I experimented ways to make a robot that could compete with some of the included sample robots of robocode. In my opinion, it's not the best robot, but it's decent and I call it WayFarer! (No, not sunglasses, but thought it sounded to cool to name it after it.)

     For my robot, I took some note from the code katas from the other blog. The design is basically my robot moving back and fourth along an axis. When the robot moves, it scans for enemies, turn its gun towards the enemy and fires every other turn. When the robot hits another robot, it moves away and when the robot gets hit by a bullet, it turns 90 degrees and traverses that axis. So how does this robot perform? Let's see!

     Moving on to testing. I put my robot against 8 of the sample robots:(Corners, Crazy, Fire, RamFire, SittingDuck, SpinBot, Tracker and Walls.) For this test, I ran 100 battles each and recorded the scores. Here are the outcomes:

Enemy Challenged	Rank Against Enemy	Total Score	Survival	Survival Bonus	Bullet Damage	Bullet Bonus	Ram Damage (x2)	Ram Bonus	1sts	2nds
Corners	1st	15493 (87%)	4850	970	8033	1579	31	0	97	3
Crazy	1st	9380 (69%)	2800	560	5196	700	124	0	60	40
Fire	1st	13177 (92%)	4900	980	9417	1852	29	0	98	2
RamFire	1st	15027 (53%)	3100	620	9492	1201	450	163	62	38
SittingDuck	1st	18034 (100%)	5000	1000	9966	2000	68	0	100	0
SpinBot	1st	9988 (61%)	2900	580	5668	791	49	0	59	41
Tracker	1st	17209 (71%)	4050	810	10366	1742	208	34	81	19
Walls	2nd	3060 (22%)	100	20	2815	41	86	0	2	98

     As you can see, WayFarer is a decent robot, but definitely not the best. It can only defeat one sample robot 100% of the time, but majority, it comes in first place. It;s not perfect, but good. WayFarer is a good robot because it moves constantly and changes direction to dodge incoming bullets. This might be a good strategy, but one robot isn't affected by it, meaning the walls robot as you see on the table. WayFarer doesn't go well against walls because it travels along the "walls" of the battlefield while tracking and shoot. Walls also moves in constant motion, so many bullets shot by WayFarer miss the target. What could be implemented to make WayFarer better is to predict the future location of the enemy to ensure that bullets will hit the target all the time.

     Building more on testing is some JUnit tests I created to tests some robot behaviors. My 4 tests include testing if WayFarer hits a wall, changing direction after being hit, turning its radar to scan enemies and seeing if bullets shot hit the enemy. When WayFarer hits the wall, it confirms that the robot traverses an axis. When WayFarer get's hit by a bullet, it should change directions of the traverse. When WayFarer turns its radar, it ensures that it is scanning and when a shot bullet hits an enemy, it ensures that the gun is pointed at the enemy.

     Overall, designing the perfect robot is not easy. There are some robots that are easier to defeat while others require more thinking. During this development using the Ant build system, inning testing and checking for errors are simple to do and faster than running them manually. Creating projects are now easier to transport for other users to use since the build system automatically installs and downloads required files. I can now see why build systems are viral to many projects and I intend to use it for future projects.