Difference between revisions of "User talk:Zyx/Radar Widest Area Angle"

From Robowiki
Jump to navigation Jump to search
(I think you are right)
(almost sure is the same thing, one very rare corner case to account for)
 
Line 6: Line 6:
  
 
: I will add small check to see if turning to the center gives the same result as this code, but I think you are both right, intuitively it seems to have the same output and code is so much simpler :-/. About not adjusting, I call the adjust method in the onScannedRobot event and at the beginning I set to turn radar, gun and bot. But that is for 1v1 and not melee. --[[User:Zyx|zyx]] 05:08, 17 October 2009 (UTC)
 
: I will add small check to see if turning to the center gives the same result as this code, but I think you are both right, intuitively it seems to have the same output and code is so much simpler :-/. About not adjusting, I call the adjust method in the onScannedRobot event and at the beginning I set to turn radar, gun and bot. But that is for 1v1 and not melee. --[[User:Zyx|zyx]] 05:08, 17 October 2009 (UTC)
 +
 +
My code, and turning to the center gave the exact same answer for over 2000 tests. That's not a formal proof, but is enough for me to think that [[Starrynte]]'s method gives the exact same result using roughly 1/90 lines of code :-).
 +
 +
The only case I can think of were turning to the center fails, is if the bot is spawn facing exactly (to Java's double precision) to the center, so the angle is an exact 0 (very very rare). But it shouldn't be hard to check for this very unlikely, but somewhat possible, situation and turn whatever side you prefer in such a case as the probability is 1/2 for each side.
 +
 +
But thanks to [[Starrynte]] for simplifying my radar code to almost non existent again --[[User:Zyx|zyx]] 05:37, 17 October 2009 (UTC)

Latest revision as of 06:37, 17 October 2009

In one of my melee bots (now retired), I set the radar, gun, and robot to turn towards the center, and not to adjust gun and radar for robot turn until after 3 ticks. Not sure if this would have a similar effect --Starrynte 01:10, 17 October 2009 (UTC)

Is disabling 'adjust xxx for yyy turn' really needed? I wouldn't expect it to be needed... hmm --Rednaxela 02:43, 17 October 2009 (UTC)

Hmm... well, I may be mistaken... but thinking this over, I'm pretty sure turning the radar towards the center will always do the same as this 'widest area angle' indeed... --Rednaxela 01:53, 17 October 2009 (UTC)

I will add small check to see if turning to the center gives the same result as this code, but I think you are both right, intuitively it seems to have the same output and code is so much simpler :-/. About not adjusting, I call the adjust method in the onScannedRobot event and at the beginning I set to turn radar, gun and bot. But that is for 1v1 and not melee. --zyx 05:08, 17 October 2009 (UTC)

My code, and turning to the center gave the exact same answer for over 2000 tests. That's not a formal proof, but is enough for me to think that Starrynte's method gives the exact same result using roughly 1/90 lines of code :-).

The only case I can think of were turning to the center fails, is if the bot is spawn facing exactly (to Java's double precision) to the center, so the angle is an exact 0 (very very rare). But it shouldn't be hard to check for this very unlikely, but somewhat possible, situation and turn whatever side you prefer in such a case as the probability is 1/2 for each side.

But thanks to Starrynte for simplifying my radar code to almost non existent again --zyx 05:37, 17 October 2009 (UTC)