Nathanael, and the rest of the Sigterm team (except the Documentation Manager). Names supressed for privacy issues. |
Nathanael, and the rest of the Sigterm team (except the Documentation Manager). Names supressed for privacy issues. |
Ok. It's designed as a melee bot, but does ok in one on one. One of its competitors in the uni competition was Seth. |
Ok. It's designed as a melee bot, but does ok in one on one. One of its competitors in the uni competition was Seth. |
When a small (0.1<=x<=3.0) energy drop is detected in an enemy designated as one of the best targets, a bullet path using a linear algorithm is projected, and a moving gravity well assigned. If all goes according to theory, the robot will avoid the gravity wells assigned to bullets, and hence never be hit by linear targeting. Depending on which build you're using, Sigterm scores about 50 in the WaveSurfingChallenge without a wave surfing movement. |
When a small (0.1<=x<=3.0) energy drop is detected in an enemy designated as one of the best targets, a bullet path using a linear algorithm is projected, and a moving gravity well assigned. If all goes according to theory, the robot will avoid the gravity wells assigned to bullets, and hence never be hit by linear targeting. Depending on which build you're using, Sigterm scores about 50 in the WaveSurfingChallenge without a wave surfing movement. |
No difference. |
No difference in strategy, except for scanning. There is a bug in one-on-one where the robot will scan and fire at a non-existent target, ie where an enemy robot isn't. But this doesn't occur in melee because the radar sweeps in melee to cover the whole battlefield, but only scans the enemy target in one-on-one. Sigterm's one-on-one score in RoboRumble would be better if this bug was fixed. |
Comments, questions, feedback: |
Comments, questions, feedback: |