Talk:Targeting Challenge RM/Results

From Robowiki
Jump to navigation Jump to search

Hmmm. So it seems my movement in CunobelinDC is pretty good =) --Skilgannon 15:04, 2 May 2009 (UTC)

Triphammer

Great work! I can't wait to see if these latest changes show any big improvement in performance in the rumble! --Skilgannon 16:24, 19 September 2009 (UTC)

Thanks. =) I actually used it in Diamond 1.39. Unfortunately, but not surprisingly, it didn't show much improvement over the old KNN gun, but I'm glad it at least performed on par. [1]. Maybe the points I lose to the VG are the bigger obstacle now, and of course I know there are just not nearly as many points left in the gun as in the movement. --Voidious 16:57, 19 September 2009 (UTC)

Removed TripHammer, as it's Diamond's "Main Gun" these days anyway. Btw, without the VG it gets 91.35. --Voidious 22:28, 17 March 2010 (UTC)

Collecting results for TCRM

I dusted off the RoboLeague installation I had lying around, and ran the TCRM for 50 rounds. Then I had to manually add (well Excel did the addition) all 50 results for 15 bots. I thought there was a way (xml-file?) that would give the correct and complete result without any extra action. Am I wrong, or are you all using RoboResearch, as I can not imagine anyone crazy enough (except myself) to actually spend an hour Excelling to get the results. --GrubbmGait 20:47, 30 September 2009 (UTC)

Yes, I'm using RoboResearch. Apparently we don't have the TCCalc page up in a working state any more, but I'll get that setup for you if you want to stick with RoboLeague. At least then you could just re-arrange the final scores and compute the subtotals, or I could add TCRM support like I did for TC2K6. --Voidious 21:00, 30 September 2009 (UTC)

I'll try RoboResearch first, no need to resurrect TCCalc (yet). --GrubbmGait 08:40, 1 October 2009 (UTC)

I used to use RoboLeague too, and I hacked a lot of XSLT to provide the result (via some extra calculation too) RoboResearch is better, anyway. Nat Pavasant 11:18, 1 October 2009 (UTC)

DrussGT 1.9.9

Really nice work Skilgannon! Is this an improvement to the main gun, the VG, or both? (I remember hitting 91.3 sans VG...) --Voidious 14:28, 8 June 2011 (UTC)

This is with the latest precise min/max GF calculations added, as well as removing all but 1 antisurfer gun (my interpolated VCS gun) from the virtual gun system. I found that removing my DC antisurfer gun as well as my PM gun from the virtual gun system boosted by score by a good 0.5% for this challenge, and I'm interested to see if it will show in the rumble. I'm still working on tuning that antisurfer gun but it should be ready soon. Currently I'm sitting at around 72.0 % on the TCAS. --Skilgannon 12:35, 12 June 2011 (UTC)

Contents

Thread titleRepliesLast modified
Problems with TCRM017:15, 25 March 2013

Problems with TCRM

I recently tried out the TCRM challenge and found the following issues that may need to be addressed:

1) TCRM page on the wiki doesn't state the rules. You have to follow another link to an older TC challenge to find the rules.

2) Reference bots download does not include the challenge file. It is unclear where to get the challenge file from. The reference bots and challenge file should probably be packaged together in a single download.

3) Multiple inconsistencies between stated robot version numbers on the wiki page and the actual specified version numbers (whether or not "TC" is shown, mostly).

4) SandboxDT.properties inside of pe.SandboxDT_1.91TC specifies the version as "1.91". To follow proper convention, it should specify version "1.91TC".

Skotty17:14, 25 March 2013