Need More Help tweaking 2 config values

Post by **hellcatv** » Thu Aug 23, 2007 4:49 am

Jacks and I spent all night tweaking spec--and I've spent some time today (just committed a new .exe file and new source).
We've got it down to a 3.5 minute transit time from dock port to dock port on planets....and have reduced some of the slide problems

however, the autopilot code, carefully tuned for the old values, now needs some tuneup--and if folks could help us that'd be great.

The variables that we're most interested in finding good values for a variety of bases are the ones in the
<section name="physics">
<var name="enough_warp_for_cruise" value="1000"/>
<var name="warp_perpendicular" value="80"/>
in vegastrike.config

the first one specifies how fast you need to be SPECing before you head directly for the target no matter what. This usually causes units to warp out into space right now... and orbit the sun a few times--which doesn't look cool.

The second value is how long it will go perpendicular to the surface of the planet to minimize spec values.... right now we have that set at 80...but maybe it should be lower too.... perhaps it should stay
If you folks could tweak those values a bit and post about your experience with the new faster autopilot let us know!

Please post to this thread discussions of these values and specific concrete problems you're having with this particular implementation of spec transit. (there's another thread if you want to talk about viability of spec in general http://vegastrike.sourceforge.net/forum ... php?t=9269 )

Post by **hellcatv** » Thu Aug 30, 2007 10:42 pm

Please someone, ANYONE help!!! we need this for the release--yet no one seems to want to be bothered by it

Post by **loki1950** » Thu Aug 30, 2007 10:54 pm

It's kinda hard to test when you can't get to the game yet

Enjoy the Choice

jere19 · Post by **jere19** » Fri Aug 31, 2007 5:38 pm

I'll try to help as soon as i can get the svn version running

(for the moment it just seg fault when i try to start the campaign)

charlieg · Post by **charlieg** » Fri Aug 31, 2007 7:30 pm

jere19 wrote:I'll try to help as soon as i can get the svn version running
(for the moment it just seg fault when i try to start the campaign)

Ditto

pheldens · Post by **pheldens** » Fri Aug 31, 2007 11:08 pm

use a working revision
like
Revision: 11333

(run svn update -r 11333 inside /data4.x and /vegastrike)

vegastrike is broken shortly after that

Post by **loki1950** » Fri Aug 31, 2007 11:14 pm

I guess i just need to run it in data4.x then as i still have all my binaries back to 11065

Edit: if you want to keep all on the stuff that was downloaded for up to svn11447 make a copy of your repo and run the svn update -r 11333 in that directory it will take awhile as all the stuff is downloaded again from the main repository.

Enjoy the Choice

Post by **hellcatv** » Sat Sep 01, 2007 5:50 pm

current SVN should work---I just committed one a few hours ago

rigelan · Post by **rigelan** » Sat Sep 01, 2007 10:16 pm

I'll help, is there a quick way to use new values, or should i completely close the program and open it again? Does it put it into memory or if I change it when vegastrike is running, will it read the new values?

Post by **ace123** » Sat Sep 01, 2007 10:49 pm

Unfortunately there's no easy way to do this at the moment.

Post by **loki1950** » Sat Sep 01, 2007 10:49 pm

@rigelan the values are in vegastrike.config so it's end game change values restart game just the nature of the beastie and yes helcatv svn11461 does load the head data set

Enjoy the Choice

rigelan · Post by **rigelan** » Sun Sep 02, 2007 2:10 am

I set up a route to go from Cephid to Everett to Enyo to Cephid, and went around it twice, stopping the clock when the next system was loading, and finding the next system and pressing A immediately. SVN 11452.

It seems that the multiplier doesn't multiply the momentum anymore, it seems to multiply your speed in the direction the ship is facing, and this has had a major impact on these values.

I ran the route twice, and my conclusion is that changing these values within this range, and with the new SPEC multiplier makes little significant difference.

my Values:

1000/80 = 17:25
800/80 = 17:17
500/80 = 17:16
500/50 = 17:17

If you think there is a better set of values for me to test, let me know. But it seems that for my set of jumps, they consistantly took 2:50 - 3:00 minutes per jump.

I would estimate an error of plus or minus 2 seconds because of my reaction time in pressing the stopwatch.

Post by **hellcatv** » Sun Sep 02, 2007 9:56 am

3 minutes from hotspot to hotspot seems reasonable--I'm happy you've been testing it in major systems

yes nowadays we changed spec to not multiply velocity but only multiply facing velocity--this makes it a ton easier to fly around...

we also made the autopilot extra smart about what it does when you're over 3 minutes into the planets atmosphere....

pheldens · Post by **pheldens** » Sun Sep 02, 2007 11:40 am

put the 2000 value to 10000 and don't notice much difference.

wirser · Post by **wirser** » Fri Sep 07, 2007 1:29 pm

well I just had svn 11483 and it was terribl e spec wise
will post back when 11497 compile and I can test it