From Robowiki
< Talk:Gilgalad
Revision as of 18:27, 13 February 2012 by AW (talk | contribs) (typo)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Figure 1: GilgaladTargetingDiscussionFigure 1.jpg

Figure 2: GilgaladTargetingDiscussionFigure 2.jpg

Figure 3: GilgaladTargetingDiscussionFigure 3.jpg

Figure 1 would be normal guessfactors when there is no wall. Basically going forwards 0.25, then back 0.25, then forwards 0.5 is the same as going forwards 0.75 and back 0.25. They aren't exactly the same because you can't instantaneosly change your velocity but they will be similar.

Figure 2 shows precise MEA's near a wall. The problem is that when you go forwards 0.75 GF and then back 0.25, you are not suing the same path both times because you only sue wall smoothing for the wall you are heading towards. This means that 0.25 - 0.25 + 0.5 is not always going to be similar to 0.75 - 0.25. The black lines show the angle the correct angle to shoot at.

Figure 3 shows my idea of having non-linearly scaled bins. This might work a little better because "denser" areas would be more likely to be shot at. The orange line is GF 0.5 with non-linear bins. The purple line is GF0.5 with linear bins. Both with precise MEA.