Talk:Darkcanuck/VelocityTest

Scratch that last posted result -- was using the method from 1.7.1.2 Beta by accident. Here's the corrected test result:

Starting velocity=0.0 distance=6.0
  1 velocity=1.0 remain=5.0
  2 velocity=2.0 remain=3.0
  3 velocity=3.0 remain=0.0
  4 velocity=1.0 remain=-1.0
  5 velocity=-0.75 remain=-0.25
  6 velocity=-0.25 remain=0.0
  7 velocity=0.0 remain=0.0

Still a problem here:

• at tick 3 the bot should have stayed at velocity 2 then decelerated to 1 to hit the target
• the bot then overshoots and takes its time to get on target
• this should have been a 4-tick move...

--Darkcanuck 05:22, 15 July 2009 (UTC)

For another (worse) example of the above, try starting at 0 with a target of 10.

Starting velocity=4.0 distance=0.0
  1 velocity=2.0 remain=-2.0
  2 velocity=0.0 remain=-2.0
  3 velocity=-1.0 remain=-1.0
  4 velocity=-2.0 remain=1.0
  5 velocity=-0.0 remain=1.0
  6 velocity=1.0 remain=0.0
  7 velocity=0.0 remain=0.0

Starting from an overshoot condition, the velocity formula causes a second, unnecessary overshoot. --Darkcanuck 05:45, 15 July 2009 (UTC)

And here's a sample of what odd negative velocities can produce:

Starting velocity=-1.9 distance=10.0
  1 velocity=1.8074999999999999 remain=8.1925
  2 velocity=2.8075 remain=5.385000000000001
  3 velocity=3.8075 remain=1.5775000000000006
  4 velocity=1.8075 remain=-0.22999999999999954
  5 velocity=-0.22999999999999954 remain=0.0
  6 velocity=0.0 remain=0.0

--Darkcanuck 05:49, 15 July 2009 (UTC)