Thread history

From Talk:Dynamic Clustering
Viewing a history listing
Jump to navigation Jump to search
Time User Activity Comment
18:54, 16 December 2013 Voidious (talk | contribs) New reply created (Reply to Density Calculations)
21:55, 15 December 2013 Straw (talk | contribs) New reply created (Reply to Density Calculations)
21:22, 15 December 2013 Voidious (talk | contribs) Comment text edited (er)
21:20, 15 December 2013 Voidious (talk | contribs) New reply created (Reply to Density Calculations)
21:10, 15 December 2013 Straw (talk | contribs) New reply created (Reply to Density Calculations)
20:53, 15 December 2013 Voidious (talk | contribs) New reply created (Reply to Density Calculations)
20:33, 15 December 2013 Straw (talk | contribs) New reply created (Reply to Density Calculations)
17:40, 15 December 2013 Voidious (talk | contribs) New reply created (Reply to Density Calculations)
15:28, 15 December 2013 MN (talk | contribs) New reply created (Reply to Density Calculations)
05:05, 15 December 2013 Straw (talk | contribs) New thread created  

Density Calculations

I have been working on a pure anti surfer gun (Not for use against flatteners!) and I am having some trouble implementing firing angle density calculations. I've mostly been testing vs DrussGT with flattener and shooting disabled. Right now I just fire at the nearest neighbor to the current situation, because every time I try to look at more neighbors and use some highest density calculation, it significantly reduces my hitrate. I've tried doing simple things like counting points near each point, and some more advanced stuff with gaussian rolloff for both distance (in kNN) and how near the point is on the GF scale [-1,1]. Help?

Straw (talk)05:05, 15 December 2013

Surfers are designed to screw up highest density calculation on purpose.

You are better off trying to shoot everywhere, except where they were in the past and were hit by your bullets (the highest peak). Peaks which formed only after your last bullet hit still works though, and the easiest way to do this is using data decay. Using less neighbors has the same effect to a lesser extent.

MN (talk)15:28, 15 December 2013

Indeed, I use a pretty low K in my anti-surfer gun. Higher than 1 though.

Also, make sure you're running enough battles to get an accurate measurement. It might be more battles than you think. More than a few bad Robocode conclusions have come from not gathering enough data.

Voidious (talk)17:40, 15 December 2013

How low is your k? I've only run a few battles, but when my hitrate goes under 8% vs a bot which can't shoot and can't flatten, and normally its above 12%, it makes me think there is something wrong. Ill run a few more to make sure.

Straw (talk)20:33, 15 December 2013

My Anti-Surfer gun uses 4 data sets. They are k=3 of last {125 | 400 | 1500 | 4000} data points. So the least-decaying one has ~5 rounds of data, and many of the selected data points will overlap.

8% vs 12% of raw hit rate over a few battles doesn't could still be just fluctuation. If it's a normalized hit rate that might be significant... I generally run a few hundred to a few thousand battles to test a change, for what it's worth.

Voidious (talk)20:53, 15 December 2013

Its normalized calculated by DrussGT. I just did 15 battles, it seems unlikely due to random fluctuation. Is your anti-surfer gun a composite of multiple other guns? How does it work?

Straw (talk)21:10, 15 December 2013