SITREP

Discussion in 'Command Ops Series' started by Dave 'Arjuna' O'Connor, Jan 15, 2015.

  1. john connor

    john connor Member

    Joined:
    Oct 22, 2014
    Messages:
    1,949
    Likes Received:
    104
    The main problem, SamuraiN, is not the HQs leading out front for the player (who can do as you do, or, at any rate, monitor movement and intervene) but for the AI controlled enemy. And whilst you might have expected some lower level soviet leaders to lead from the front (though most often we read about them pushing from the back, with threats...) in game we're talking about whole higher HQ formations leading off. It's when you're playing as Axis, for example, and you come across the XXX Corps HQ ahead of all other enemy troops, for instance. Obviously, that didn't happen as a practice in real life. Anyway, Dave seems very close to fixing it for HQs, fingers crossed. I hope the similar fix for arty follows quickly as one of the reasons there is sometimes less enemy arty (AI controlled enemy arty) is that the enemy AI controlled arty isn't basing sufficiently far back and hence is always busy moving instead of shooting...
     
  2. SamuraiN

    SamuraiN Member

    Joined:
    Oct 24, 2019
    Messages:
    54
    Likes Received:
    3
    I see, looking forward to the tests!

    I would not worry too much about the arty at the moment, because they do not appear to be all that effective in AI's hands anyway, especially for offensives, for various reasons as I see. Since in actual situations arty coordination was very difficult and the control of timing was a sophisticated art, the main AI code for arty actions can hardly be easy. Nonetheless, I think hard coding specific arty actions for each scenario can be a viable solution.

    As a side comment, personally in game I often encounter some HQ cowardice when it sits far behind the line when weakened front line troops needs one extra push to break through. The armoured HQ is particularly despicable, as they are often well equipped. But in St. Vith the SS HQs fight rather well.
     
    #742 SamuraiN, Nov 4, 2019
    Last edited: Nov 4, 2019
  3. Kensal

    Kensal Member

    Joined:
    Mar 25, 2018
    Messages:
    8
    Likes Received:
    0
    An issue I find when playing the AI is the ineffectiveness of AI controlled arty. It does tend to spend a lot of time wandering its arty around rather than using it for fire missions. It possible to balance this by limiting your player controlled arty to corps or division level arty, but even then in the bigger scenarios this can still amount to a lot of fire power. Its difficult also to keep your hands off the arty when you have a crucial attack going in or a wavering defence.
     
  4. SamuraiN

    SamuraiN Member

    Joined:
    Oct 24, 2019
    Messages:
    54
    Likes Received:
    3
    Constraining player control of artys would be unreasonable, because all units should be under chain of command. Can't just do it for balancing against AI. Plus, what about player vs. player?

    The key is arty mechanism. Currently arty support is 'called', by other units triggering certain conditions, I guess. So no arty preparation, suppression, etc. Defensive arty actions would be hard to code, I think, because of consideration of friendly fire, duration of fire, etc. And in fact the best action would be to focus on enemy reserves instead of first wave attacking troops, unless urgent. And communication with high level arty units in the rear is not that good in real life, anyway, with possible exception of US forces, though in those missions of US on the defence, they are always in disarray. So I would not worry about this part at the moment.

    On the other hand, I think AI arty can be made more effective when it is on the offence, or counterattacking. My opinion is that maybe hard coding specific actions for scenarios: identify breakthrough areas and related critical junctions, and automatically trigger arty fire once enemy force identified exceeds a certain level or some other threshold. The problem is, doing so for each scenario would be terribly painstaking. It would be viable if we have an open source project and players can contribute, but then the business model would be changed...
     
  5. SamuraiN

    SamuraiN Member

    Joined:
    Oct 24, 2019
    Messages:
    54
    Likes Received:
    3
     
  6. Kensal

    Kensal Member

    Joined:
    Mar 25, 2018
    Messages:
    8
    Likes Received:
    0
    sorry I wasn’t suggesting the game limiting players’ use of arty -it’s merely something I try to limit voluntarily when I’m playing agts the AI to provide more balance. Also sorry for hijacking this thread which is, after all, for updates on the game.
     
  7. john connor

    john connor Member

    Joined:
    Oct 22, 2014
    Messages:
    1,949
    Likes Received:
    104
    I very rarely call in manual arty. I think it's like cheating, the way things are presently coded. You can have a great game with plenty arty from both sides without calling it in manually. I like that as well, because I'm so lazy. Sometimes it's astonishing how much arty the enemy AI will drop on you, at just the wrong moment, wrecking all your plans. On the other hand, sometimes you can have a poor game where the enemy AI uses very little arty and your own side too. It varies. Overhauling the arty code is, I believe, on Dave's list and has been for a long while. I wouldn't hold my breath. The pace of updates, we will all have noticed, is constrained by the size of the team involved - basically, Dave, for these big things, I think. Changes seem to take ages at the moment, updates and and new DLC are long delayed. Short of someone funding Dave for three extra coders, nothing can be done about that. I believe some kind of change that will reduce the amount of moving around the enemy arty does (instead of firing) is, on the bright side, part of the present round of basing changes Dave is working on, so there may be improvement there much sooner.
     
    panzerpit and Seb3brv78 like this.
  8. Keydet

    Keydet Member

    Joined:
    Jun 7, 2015
    Messages:
    639
    Likes Received:
    12
    There are a number of things about fire support that sinks my pinball. [I too think it should be left to AI but I don't] The concerns are:
    1. There is no transparency to what the AI is doing with My fire support. I don't know nor can I set the priorities of fire.
    2. I have no ability to build and manage a fire support plan
    3. Final protective fires in the defense can't be set much less know which units have which FS units giving the FPF.
    4. FS unit logs don't have each fire request/fire mission described with Target description http://www.poeland.com/tanks/artillery/targets.html, mission (suppress, destroy, etc), sheaf used http://www.poeland.com/tanks/artillery/sheafs.html, and FO results report.
    5. the logs of requesting units don't report instances FS request made and result fired or denied (and why: no comm, ammo rationed, no tubes available, etc?).
    6. There is not a FS realism slider like orders delay: perfect world to realistic.
    7, In every game I see some battalion mortars never fire for any of its battalions engagements which is inexplicable while most other battalions have near 100% response..
    8. Some regiments never succeed in getting any of the attached artillery to fire while others get near 100% response.
    9. Artillery battalions are dragged around by the HQ's these are supporting rather than independently considering the missions and locations of supported units, ranges, lines of communications. and enemy actions and locations.
    10. Artillery battalions with ongoing support missions don't move by batteries.
     
    #748 Keydet, Nov 17, 2019
    Last edited: Nov 17, 2019
  9. Dave 'Arjuna' O'Connor

    Dave 'Arjuna' O'Connor Panther Games Designer

    Joined:
    Jul 31, 2014
    Messages:
    2,943
    Likes Received:
    288
    SITREP Wed 4 Dec 2019

    Hi all,

    Sorry for the delay in posting. I have been head down and tail up with the HQ Basing feature. This has been a mammoth coding exercise because it affects so much of the AI and in particular the planning code. I have finished coding up the doctrines to support this and have been testing and refining it over the last two weeks. I got it to run and process through a number of scenarios. But in the more complex scenarios it encounters an infinite recursion caused by a recursive call sequence inside the firebasing code - the code that bases artillery.

    I have opted to redo the firebasing code using the new HQBasing code as a template. So they will base along the main supply route or mission route of their own route depending on the circumstances. I am modifying this to handle the minBombardRange of arty units. HQs don't have a minimum command range. I hope in doing this that I'll also address feedback I've received about arty units following their HQ around, always moving and hence not firing and deploying into dangerous territory.

    I'm still implementing this.
     
    Keydet, Seb3brv78 and panzerpit like this.
  10. john connor

    john connor Member

    Joined:
    Oct 22, 2014
    Messages:
    1,949
    Likes Received:
    104
    Fantastic news, Dave - that you're looking at the arty too. Many thanks for that. Courage!
     
  11. Rick S.

    Rick S. Member

    Joined:
    Jun 22, 2019
    Messages:
    5
    Likes Received:
    0
    Keep at it, can't wait.
     
  12. Dave 'Arjuna' O'Connor

    Dave 'Arjuna' O'Connor Panther Games Designer

    Joined:
    Jul 31, 2014
    Messages:
    2,943
    Likes Received:
    288
    SITREP Friday 6 Dec 2019

    Hi all,

    I believe I have finished the firebase overhaul. I've eliminated all the recursions and tweaked the suitability calculations and added smarts to prevent long range arty from moving to an unsafe objective. Of course it needs some robust beta testing, but I plan to write up all my changes later today and check this code base in. Then we need to merge it with the master and put out a new build for our beta testers. If it's halfway decent, we'll then put out a public beta build for all to test.
     
    Seb3brv78, Keydet, panzerpit and 2 others like this.
  13. Dave 'Arjuna' O'Connor

    Dave 'Arjuna' O'Connor Panther Games Designer

    Joined:
    Jul 31, 2014
    Messages:
    2,943
    Likes Received:
    288
    Further to the above, I have just committed all my work and pushed it to our Github repository for Pavlo to integrate with his work. The change list is massive. I'll post that later. :)
     
    Seb3brv78 and Keydet like this.
  14. ironsight

    ironsight Member

    Joined:
    Oct 18, 2015
    Messages:
    25
    Likes Received:
    3
    Sounds like we may all get an early Christmas present soon! :happy:
     
  15. john connor

    john connor Member

    Joined:
    Oct 22, 2014
    Messages:
    1,949
    Likes Received:
    104
    Indeed! Great news!
     
  16. nelmsm

    nelmsm Member

    Joined:
    Feb 21, 2017
    Messages:
    6
    Likes Received:
    3
    Awesome news!
     
  17. SamuraiN

    SamuraiN Member

    Joined:
    Oct 24, 2019
    Messages:
    54
    Likes Received:
    3
    Would you mind posting the Github repository?
     
  18. panzerpit

    panzerpit Member

    Joined:
    Dec 12, 2014
    Messages:
    48
    Likes Received:
    6
    Don't you thing it's a bit ridiculous to expect that autor will publish code repository of a commercial game? :)
     
  19. SamuraiN

    SamuraiN Member

    Joined:
    Oct 24, 2019
    Messages:
    54
    Likes Received:
    3
    Well, since the current business model is that the engine is free, but scenarios need to be purchased, I do not see a terribly big issue there. And one doubts how many people will make use of them anyway. Though I do acknowledge there is the issue of tempering with the code without authors' consent.

    Personally I feel that charging for engine but let users create scenarios would make more sense for the community, or game ecosystem, But admittedly that would not be commercially feasible.
     
  20. john connor

    john connor Member

    Joined:
    Oct 22, 2014
    Messages:
    1,949
    Likes Received:
    104
    Users can - and do - still create scenarios, of course...
     

Share This Page