Author Topic: Pathway errors on NA maps  (Read 1913 times)

Offline ksl94

  • Jr. Member
  • **
  • Posts: 739
  • British Imperialist
    • View Profile
Pathway errors on NA maps
« on: 24-01-2012, 21:01:30 »
All right everybody, I noticed that there are many pathway errors still present on NA maps. This thread shall serve the purpose of collecting locations which are causing trouble for bots. Please feel free to post your observations here to aid the developers in updating the pathmaps (only if they want to, of course).
I will start: On Op. Hyacinth the bots get constantly stuck in/at the shed you encounter on the way to the first flag. On Fall of Tobruk bots often like to get stuck at walls (I will list the actual locations soon).

Offline Tankbuster

  • Jr. Member
  • **
  • Posts: 921
    • View Profile
Re: Pathway errors on NA maps
« Reply #1 on: 25-01-2012, 17:01:16 »
Tanks cannot go from the Church to the Brit mainbase in FoT.

Offline ksl94

  • Jr. Member
  • **
  • Posts: 739
  • British Imperialist
    • View Profile
Re: Pathway errors on NA maps
« Reply #2 on: 29-01-2012, 22:01:00 »
On Fall of Tobruk bots like to get stuck at a wide variety of places, especially at piles of rubble. The tanks that don't advance properly as you described it may have to do with the horrific AI driving behavior that was introduced with FH2 2.4.

Offline Void

  • Moderator
  • ***
  • Posts: 162
  • AI Dev
    • View Profile
    • ESAI at Moddb
Re: Pathway errors on NA maps
« Reply #3 on: 30-01-2012, 04:01:08 »
For sp development I am pretty much it right now. As far as this thread helping me out, here is an idea for you guys and some explanation:

Many of these issues being described are just due to outdated navmeshes needing to be regenerated. That is part of what I am doing these days already.

You need not tell me exact locations on sp maps where bots get stuck. Many existing sp maps are going to undergo changes in layout and flag positions under my hand anyway.

Discuss yourselves which maps have the worst pathfinding problems overall, then make a priority list. That will influence the order in which I generate new navmeshes.

Offline ballard44

  • Newbie
  • *
  • Posts: 30
    • View Profile
Re: Pathway errors on NA maps
« Reply #4 on: 31-01-2012, 12:01:42 »
The tanks that don't advance properly as you described it may have to do with the horrific AI driving behavior that was introduced with FH2 2.4.

No, there is no mesh there for bots to continue driving on in the 1st place.
Unfortunately, the space needed for bots to continue is just too narrow, so it was not generated.



Offline djinn

  • Masterspammer
  • ****
  • Posts: 5.723
    • View Profile
Re: Pathway errors on NA maps
« Reply #5 on: 31-01-2012, 12:01:31 »
A good many changes in the next patch will cause a long-term project of redoing alot of navmeshes, so dont worry about pathmaps. Many will be redone to be optimized.

With me as Beta tester and Void as Dev, you can rest assured, these things will be addressed.


Offline Tankbuster

  • Jr. Member
  • **
  • Posts: 921
    • View Profile
Re: Pathway errors on NA maps
« Reply #6 on: 01-02-2012, 06:02:25 »
Thanks

Offline sheikyerbouti

  • Full Member
  • ***
  • Posts: 1.402
  • Yay, Rep feature is dead
    • View Profile
Re: Pathway errors on NA maps
« Reply #7 on: 01-02-2012, 11:02:28 »
A good many changes in the next patch will cause a long-term project of redoing alot of navmeshes, so dont worry about pathmaps. Many will be redone to be optimized.

With me as Beta tester and Void as Dev, you can rest assured, these things will be addressed.

 You guys seriously need to consider working with the Veteran's gaming coop server community. hands down, they are the number one user of the COOP maps and would provide a lot of firsthand experience and advice with the maps.

 Better a whole community than a one man Beta test army.
My Quebec includes Canada

Offline Void

  • Moderator
  • ***
  • Posts: 162
  • AI Dev
    • View Profile
    • ESAI at Moddb
Re: Pathway errors on NA maps
« Reply #8 on: 01-02-2012, 17:02:10 »
Well, we have an army of beta testers who have access to what I am currently working on. Djinn is a BT who happens to be my main go-to guy with what I would call my "pre-beta" work, before it goes to the rest of the testers. This is largely to prevent embarrassing myself with silly mistakes I might otherwise put into the beta build.

VG is a great community and I am acquainted with some of its members. I value their input as much as anyone else's. Any VG member is welcome and encouraged to express their ideas or concerns in this forum, and yes, I will listen.

Thank you for expressing your opinion on this, just thought I would clarify that we have many beta testers, though input from the gaming community is always welcome.

Offline Tankbuster

  • Jr. Member
  • **
  • Posts: 921
    • View Profile
Re: Pathway errors on NA maps
« Reply #9 on: 17-02-2012, 07:02:38 »
The German Panzers have trouble breaching the 1st line of defence in SoT

Offline djinn

  • Masterspammer
  • ****
  • Posts: 5.723
    • View Profile
Re: Pathway errors on NA maps
« Reply #10 on: 17-02-2012, 18:02:23 »
its less the panzer and more the infantry. The way that map goes,
(similar to Gazala), it requires boots on the ground to take out those antitank guns and mortar positions.

The Stuka might do something with say, AA and mg, but the real ability to cap is the infantry.

They do eventually though, and once one falls, you'd notice that the allies put up less of a resistance at the front and fall back to the bunker at the rear.

Some tweaks need to be done on this map though, both sides need to be aggressive and yet winning should be possible by both sides, by some slightly weakened position falling or a German offensive being halted and pushed back.


The trick  with SP, especially since it is also COOP, is to make it such that, a human player wont matter, and YET, will.

We used to use the see-saw effect, where the attackers will push in all the way and be pushed back by the defenders, but issue there is, it makes it too easy for an attacking human to prevent a counter-attack, and a defending human player from getting any support as defense.

be patient with us, and we will get these tweaked.
If this conforts you, I should tell you, my main priority in playing FH2 is playing singleplayer (COOP, solo)
even to an unhealthy degree. So we ARE on it.


The singleplayer communuity has always been an open one - and its not closing any time soon.
Just pop over to the google BUGTRACKER and post errors you encounter so we know where to look as well
as suggest novel strategies for dealing with AI in general or in specific aspects in the 'Singleplayer fan project'

« Last Edit: 18-02-2012, 11:02:21 by djinn »

Offline Michael Z Freeman

  • Jr. Member
  • **
  • Posts: 746
  • The Few are many;) (Used to be known as DJ Barney)
    • View Profile
    • Homepage
Re: Pathway errors on NA maps
« Reply #11 on: 17-02-2012, 20:02:16 »
Many of these issues being described are just due to outdated navmeshes needing to be regenerated. That is part of what I am doing these days already.

So most of these kinda errors get automatically fixed by the navmesh generator ?

Quote
You need not tell me exact locations on sp maps where bots get stuck. Many existing sp maps are going to undergo changes in layout and flag positions under my hand anyway.

The errors are worth putting in the issue tracker. The issues can always be marked fixed when the new maps come out and allow proper tracking of all this (did issue abc really get fixed, etc, by the new navmesh).

Quote
Discuss yourselves which maps have the worst pathfinding problems overall, then make a priority list. That will influence the order in which I generate new navmeshes.

I'd suggest committing an issue for each map as there is a tag you can add for each map in the tracker.

All right everybody, I noticed that there are many pathway errors still present on NA maps. This thread shall serve the purpose of collecting locations which are causing trouble for bots. Please feel free to post your observations here to aid the developers in updating the pathmaps (only if they want to, of course).
I will start: On Op. Hyacinth the bots get constantly stuck in/at the shed you encounter on the way to the first flag. On Fall of Tobruk bots often like to get stuck at walls (I will list the actual locations soon).

Yeah, discuss them here even if ya don't commit an issue someone will probably gather them up and put them in there.

Offline Void

  • Moderator
  • ***
  • Posts: 162
  • AI Dev
    • View Profile
    • ESAI at Moddb
Re: Pathway errors on NA maps
« Reply #12 on: 18-02-2012, 00:02:49 »
Quote
So most of these kinda errors get automatically fixed by the navmesh generator ?

Most, yes. The ones that don't get fixed that way are fixed by manually navmeshing.

Putting these bugs in the tracker sounds fine, so long as they are clearly grouped by map name.

Offline gavrant

  • (Almost) retired dev
  • Developer
  • ******
  • Posts: 598
    • View Profile
Re: Pathway errors on NA maps
« Reply #13 on: 18-02-2012, 16:02:17 »
...
I'd suggest committing an issue for each map as there is a tag you can add for each map in the tracker.
...

Right now tagging is not accessible to non-members of the project. Common users just get 'subject', 'description' and 'attach a file' for new issues, and 'comment' and 'attach a file' for existing ones.

EDIT: if by tag you meant 'ComponentType' and 'ComponentMap' fields or Labels...
« Last Edit: 18-02-2012, 16:02:45 by gavrant »

Offline Michael Z Freeman

  • Jr. Member
  • **
  • Posts: 746
  • The Few are many;) (Used to be known as DJ Barney)
    • View Profile
    • Homepage
Re: Pathway errors on NA maps
« Reply #14 on: 18-02-2012, 16:02:48 »
Thanks  ::) ... I should have caught that. On it. I need to setup templates or a link for each map for the user (non-project member) to select when they report an issue. That will set the right ComponentMap label.
« Last Edit: 18-02-2012, 17:02:23 by DJ Barney »