Topic: New movement calculation [as of 2014.06.10]

Some of you already noticed the yellow cube trail that your robot leaves now on the test server. The cubes basically serve a debugging purpose, they indicate your real position on the server.

We have changed the movement calculations to be much more server-side than before. This should alleviate many problems, including people getting stuck on top of mountains, jumping walls, and similar freaky things.

Please test it and report bugs if you find any.

Re: New movement calculation [as of 2014.06.10]

If you have two clients open at the same time and move robot A around and stop, it won't stop robot A at client B where it should. Looks about 7 seconds out of sync or some prediction error. When the robot finally stops at the other client it's in the wrong location.

Re: New movement calculation [as of 2014.06.10]

Voted you leave these in for X number of seconds as mech footprints.

Re: New movement calculation [as of 2014.06.10]

oh, after reading the topic title, i almost thought you would start to take terrain slope into consideration for robot speed

*Disclaimer: This post can contain strong sarcasm or cynical remarks. keep that in mind!
Whining - It's amazing how fast your trivial concerns will disappear

Re: New movement calculation [as of 2014.06.10]

Annihilator wrote:

oh, after reading the topic title, i almost thought you would start to take terrain slope into consideration for robot speed

this tricked me into thinking the same thing lol
ah well, we can always dream

+1
-Confucius

Re: New movement calculation [as of 2014.06.10]

Would be pretty cool if you introduced a new extension and module that allowed you to "track" other robots and use those boxes as "tracks" you can only see with the module running.

[18:20:30] <GLiMPSE> Chairman Of My Heart o/
CIR Complaint Form

The Imperial Grand Wizard of Justice

Re: New movement calculation [as of 2014.06.10]

Your calculation is introducing huge amounts of error.  Or De Sync.

The Gifter
Top  Killer 2013  - 01: 334 -- 17 -- 317  : Merkle
Top  Killer 2012  - 01: 027 -- 472 -- 445 : Merkle

Scarab Kill Count - 13

Re: New movement calculation [as of 2014.06.10]

The error burial described above happens consistently between clients--i.e., bots are shown as moving to new positions after stopping. None of the movement markers are being drawn, but the bot's still moving.

Re: New movement calculation [as of 2014.06.10]

Bots just seem to teleport around randomly and the position on the screen doesn't seem to have much to do with where they are with regard to being in/out of cover

+1
-Confucius

Re: New movement calculation [as of 2014.06.10]

Ozy wrote:

The error burial described above happens consistently between clients--i.e., bots are shown as moving to new positions after stopping. None of the movement markers are being drawn, but the bot's still moving.

Yup, that is an issue and we're currently fixing it.

[14:15:15] <Freya Sabbat> ...Dear god, the Devs are as bad as us

Re: New movement calculation [as of 2014.06.10]

I FEEL THAT MULTITILE ROBOTS WILL COME SOON big_smile:D:D:D:D:D

Re: New movement calculation [as of 2014.06.10]

Alexadar wrote:

I FEEL THAT MULTITILE ROBOTS WILL COME SOON big_smile:D:D:D:D:D

well, they could just increase the resolution of the ground tiles... then even the smallest bot would be multitile big_smile

*Disclaimer: This post can contain strong sarcasm or cynical remarks. keep that in mind!
Whining - It's amazing how fast your trivial concerns will disappear

Re: New movement calculation [as of 2014.06.10]

We're putting out the fix on the test server now; take a look.

[14:15:15] <Freya Sabbat> ...Dear god, the Devs are as bad as us

Re: New movement calculation [as of 2014.06.10]

I cannot get it to load it just hangs at connecting.

The Gifter
Top  Killer 2013  - 01: 334 -- 17 -- 317  : Merkle
Top  Killer 2012  - 01: 027 -- 472 -- 445 : Merkle

Scarab Kill Count - 13

Re: New movement calculation [as of 2014.06.10]

Fixed now.

[14:15:15] <Freya Sabbat> ...Dear god, the Devs are as bad as us

Re: New movement calculation [as of 2014.06.10]

Eh this seems to be live already, more testing time, next time, please smile

+1
-Confucius

Re: New movement calculation [as of 2014.06.10]

This needs to be rolled back.

People are rubber banding through narrow passes and anywhere near Plants.

It is a disaster.

forget PvP

DEV Zoom - "If you mean the NPC aggro, that's been like that for months already."

Re: New movement calculation [as of 2014.06.10]

We've deployed a new version, plz try.

[14:15:15] <Freya Sabbat> ...Dear god, the Devs are as bad as us

Re: New movement calculation [as of 2014.06.10]

Have been testing since you posted this and have not managed to break movement in any way.

No rubber banding
No odd things

No problems at all


+1

DEV Gargaj wrote:

We've deployed a new version, plz try.

DEV Zoom - "If you mean the NPC aggro, that's been like that for months already."

Re: New movement calculation [as of 2014.06.10]

This has to be deployed asap, so please do test it, otherwise we're just going to believe Rage smile

21 (edited by Twiz 2014-06-11 20:01:56)

Re: New movement calculation [as of 2014.06.10]

Was toying on the test server.  Seems  there is still some issues with the rubberbanding. While running around with just 1 account, you dont feel anything.  But as soon as you have 2 accounts, where you put one of your bots on follow on the other, then you still get the rubberbanding or the bot just warp back to a previous spot.

From what I see is, the "follow" function have issues.  If your Bot A use autorun/autopilot/manual control, then Bot B which is on "Follow" will either rubberband or instant port back to a previous point.  However it doesnt always seems to happen.  Sometimes it does it, othertimes it runs fine.

22 (edited by Syndic 2014-06-11 23:51:38)

Re: New movement calculation [as of 2014.06.10]

1 account was fine but as soon as I scaled it up to my usual 3 it became unreliable and I started experiencing the same thing as before.

[18:20:30] <GLiMPSE> Chairman Of My Heart o/
CIR Complaint Form

The Imperial Grand Wizard of Justice

Re: New movement calculation [as of 2014.06.10]

Syndic wrote:

1 account was fine but as soon as I scaled it up to my usual 3 it became unreliable and I started experiencing the same thing as before.

+1

Steam achievement Unlocked:  Being a Badass
http://www.perp-kill.net/kill/239407
Dev Zoom: I think its time to confess, Ville is my alt
Dev Zoom: Ville can be sometimes so sane it's scary.

Re: New movement calculation [as of 2014.06.10]

I ran two accounts no problems,  hmmmmm

did many types of tests


are most people having problems with follow and AP ??

DEV Zoom - "If you mean the NPC aggro, that's been like that for months already."

Re: New movement calculation [as of 2014.06.10]

If you shift click on the minimap to autopilot but do this clicking on an impassible terrain, your bot seems to walk like your on follow, and in particular if you hit a building on the way, it gets "stuck" there and you will rubberband. This can be reproduced easily. The small yellow " breadcrums" becomes a very bright circle and you can walk away from that point just to rubberband around for 2-3 minutes.

Also, if you again use autopilot to an impassible terrain a a lot further away on the map and keep changing your autopilot direction to another place, again into impassible terrain, your bot will do a little twitch movement and when your moving your camera around, it will be very slow/sluggish, cpu usage increases for a short period, 5-30 secs.