Scenario designing with AI

[Lt Bull]

Hi,

I have had reason to be inspired to create a SP scenario that will feature the player as the attacker.

It has occurred to me that there are some basic game concepts that I deal with as a player that I do not know how or if are they dealt/are made to be dealt with by the AI.

I know there are things like "AI plans" and "AI triggers" now in the game, but I'm just looking for a basic run down on how the following things are typically dealt with by scenario designers when managing/scripting the AI.

- Off map artillery: I am guessing the AI doesn't call in a fire mission whenever or where ever it thinks it is appropriate. If the scenario designer doesn't "code " it in somehow there will be no arty fire missions.  How then are AI arty fire missions controlled by the scenario designer? I can imagine the most basic types are just fixed "scripted" pre-planned fire missions designed to strike at one spot at a given time.  I'm guessing that perhaps these same kind of fire missions could instead be "triggered" to occur is and when an enemy unit enters some kind of predetermined "zone".  What else is possible? PS:  does an AI trigger get activated by a player unit moving in to a zone regardless of whether it is spotted or not?

- Covered Arcs/Hiding:  What exactly is/isn't possible as far as assigning unit orders like covered arcs and hiding to individual units?  I am pretty sure from experience playing vs AI that it is at least possible to set the two types of covered arcs.  I can kind of imagine assigning these orders at setup (including hide) but once they are assigned I am guessing that order will remain with the unit (unless it breaks/retreats/rallies).

[domfluff]

The AI will happily call in off-map assets by itself, give it the assets and a Forward Observer in a good position, and enough time to call it in.

You can also set pre-planned missions.

I'm not sure about triggering artillery missions, but you can certainly trigger the FO's movement into a spotting position.

Triggers get activated regardless of LOS or other issues, they're magic like that.

Orders like covered arcs can be assigned at setup, or can be pre-baked - you could assign the AI an entire plan just using the movement options (Move fast here, pause for 2 minutes, crawl to here, covered arc west, etc.) but this has to be a "pre-baked" scenario, so there are other limitations.

 

[Lt Bull]

Wow, kind of surprised that the AI will happily call in off-map assets by itself. Have you any idea what kind of fire mission it is/can be?

Point/line/area?  Harass/light/medium/heavy/emergency? etc?

I get that you can "pre-bake" at setup.  I'm guessing you could set the "pre baked" orders at start then potentially assign an "AI plan" that can be activated later in battle?

By activate I mean on a timer (after x time elapsed), or via the triggers?

Are triggers only activated by enemy units entering a specified zone?  Is it just one unit, regardless of type?

 

 

[Bulletpoint]

 

- Covered Arcs/Hiding:  What exactly is/isn't possible as far as assigning unit orders like covered arcs and hiding to individual units?  I am pretty sure from experience playing vs AI that it is at least possible to set the two types of covered arcs.  I can kind of imagine assigning these orders at setup (including hide) but once they are assigned I am guessing that order will remain with the unit (unless it breaks/retreats/rallies).

You don't set arcs like in the normal game. What you do is to change the "stance" of the unit group to for example "Engage 200 metres", either at setup or at the end of an order, and it means the unit group will only open fire at targets within 200 metres - effectively working like a circular target arc.

The downside of this is that you can't give such orders to individual units unless you assign just that unit to a specific AI group. In my small scenario "Crossroads at Pierrefitte-en-Cinglais", there are only few Germans, but a lot of various AI groups, allowing me flexibility of setup and orders.

[RockinHarry]

Wow, kind of surprised that the AI will happily call in off-map assets by itself. Have you any idea what kind of fire mission it is/can be?

Point/line/area?  Harass/light/medium/heavy/emergency? etc?

 

With regard to AI artillery and as already has been mentioned, FO´s are needed to be placed or moved to good spotting positions, preferably with looks to map areas where also AI TRP´s are placed. The more AI TRP, the better. The AI adheres to usual C2 restrictions, meaning that any spotter needs to be in chain of command AND is allowed to access a particular Arty or mortar asset. The AI is no different in this regard as a human player. Also counts for FO´s and artillery units soft factors, meaning higher quality units respond, spot and shoot better.

Beside that, AI TRP´s should be placed with security stand off ranges for various calibre Arty in mind, as otherwise the AI won´t risk targeting areas that are too close to friendly units.

Triggers are of little use for anything related to AI Arty, but I could imagine situations where map triggers (enemy, any) could be used to shift positions for a particular FO, so it gets LOS to an area where it formerly did not have any. Or retreat a FO and onmap mortars, when enemy units come too close. There´s generally more to think about for an attacking AIP, but once you get aquainted with triggers, you can become quite creative in its usage, also for a defending AIP.

There´s no guaranty that the AIP will use its Arty or use it oftenly, but the above mentioned methods definitely help a lot.

What works and what not, needs lots of playtesting, preferably in scenario author mode. There you can also see who spots for AI Arty/mortars, when they´re used and also used targeting patterns. From my observations I know the AIP uses point, area and line, but never smoke missions. One can´t see type of intensity mode used by the AIP, but from observations I think it mostly uses short duration destructive fires.

 

[RockinHarry]

 

I get that you can "pre-bake" at setup.  I'm guessing you could set the "pre baked" orders at start then potentially assign an "AI plan" that can be activated later in battle?

By activate I mean on a timer (after x time elapsed), or via the triggers?

Are triggers only activated by enemy units entering a specified zone?  Is it just one unit, regardless of type?

Don´t think that you can pre bake a scenario in the mentioned way. You can´t start a 2 player game, save it as "baked" and then provide the saved game to be played in vs. AIP mode, so it then adheres to any given AI plans in existence. I think the "baked" format was in older CMSF version only and then removed later and not again reintroduced in any other of the CMX2 series.

The basics with regard to triggers is well covered in the game manual, but both type of triggers (orders and map based) only serve the one and only purpose to time an already given movement order for a units group in a preset AI plan. So it means always "start moving to next order zone, when trigger becomes active". The other movement order parameters (move after, move before, triggered by) determine the time frame for the trigger to be used, once it´s become active. There´s a number of possibilities here, that are covered in the game naual.

Triggers get activated only one time in the game (or never at all) and with regard to map zone triggers, it doesn´t matter how many units enter a trigger zone (one single unit activates the trigger). Only type of map trigger determines if a trigger is set to be active. They´re  friendly all, friendly armor, enemy all and enemy armor. It´s "silent" triggers and it doesn´t matter if friendly or enemy units can see any unit moving into the trigger map zone.

[Lt Bull]

Thanks for the good replies.

I think I may not really know what is meant by "baked" however. Perhaps I should ask in a different way.

Is it possible to initially setup a defensive AI unit with say a "hide" order, and have a 'enemy enters zone" type trigger initiate a move order for that unit to move to another location.  Can a string of triggers be assigned to a unit so that you could effectively simulate a succession of triggered move orders that in theory could result in the unit falling back sequentially to different positions?

On considering implementing these kinds of "triggered fallback move orders", how is end of move facing typically handled?  Can the triggered order include a facing/stance/covered arc etc order at the end?  Wouldn't want troops falling back to end up facing away from enemy. In the case of vehicles (especially tanks), is it possible to create a "triggered fallback movement order' that involves the unit falling back by reversing?

PS: RockinHarry...just looked at your system specs...time for an upgrade?? And still on WinXP 32 bit?

Edited January 6, 2016 by Lt Bull

[Bulletpoint]

Is it possible to initially setup a defensive AI unit with say a "hide" order, and have a 'enemy enters zone" type trigger initiate a move order for that unit to move to another location.  

Yes, they can be set up to hide. Yes, they can then be given a move order when an enemy enters a trigger area.

Can a string of triggers be assigned to a unit so that you could effectively simulate a succession of triggered move orders that in theory could result in the unit falling back sequentially to different positions?

Yes.

On considering implementing these kinds of "triggered fallback move orders", how is end of move facing typically handled?  Can the triggered order include a facing/stance/covered arc etc order at the end?  Wouldn't want troops falling back to end up facing away from enemy. In the case of vehicles (especially tanks), is it possible to create a "triggered fallback movement order' that involves the unit falling back by reversing?

No to all of these. The unit will just end up facing in the direction of travel. I wanted to do the same as you, a triggered fallback AI, but it's useless because the AI will not be able to take advantage of cover.

If you order a unit to fall back to a wall or a hedgerow, they will end up on the wrong side of the cover, facing away from the enemy.

You can, however, order infantry to fall back to foxholes or trenches. They will then sit there facing away from the enemy, but reorient themselves quite quickly once the enemy appears. Still looks dumb though.

There's a workaround possible where you give troops two orders - first an order to move one square behind the hedgerow, then another order to move up to it. Then they will end up on the right side with the right facing, but this is a bit bothersome to work with. Also it means the fallback takes longer to complete.

If you want a tank to fall back, it will always turn and expose its backside.

[Lt Bull]

 

No to all of these.

OMG. I don't know if I am more shocked or disappointed by what you are telling me for a number of reasons.

I recall BFC long ago clearly drawing a line and stating essentially this: "We don't really care so much for the needs/wants/features of the MP aspect (customers) of CM, they are just a minor part of the market. The MP players can even consider themselves lucky to even have PBEM in CMx2 (yes, PBEM was not in the original plan or release of CMx2)). The focus and priority of CM is and always will be the SP experience."

If BFCs position on this has changed, I have not seen it stated anywhere.  I for one am a MP-centric player and understand that catering for both the SP and MP aspects of the game is no easy thing to do.  I also see how even the introduction of seemingly simple/trivial features in to the game that can be handled by players no problems, would/could present themselves as "CPU opponent coding nightmares" when considered in the context of "if we can't code the CPU opponent to intelligently utilise the same set of unit orders/features a human player can easily assign their units, then we won't include it because it will disadvantage the CPU opponent too much" philosophy. I can also see how a developmental dilemma like this can lead to/has led to the "stunted/modest growth in CM game features" some have complained about on these forums.  With this perspective, I am not surprised by the "modest" growth of game features and rejection of otherwise seemingly "simple" feature requests/ideas we have seen.

So if the SP experience is BFC's favourite baby and where they really want to shine, then I am shocked at how "un-finished"/crude the very tools available to scenario designers to "create a SP experience" seem to be.  It is only through these tools along with the creativity of the scenario designer to work within the limits of these tools (aside from the omnipotent TacAI of course), that the whole SP experience is based upon, and upon which the concept of a CM SP CPU opponent will ultimately be judged on.

I just would have expected simple things like the triggered move orders I have been asking about to be much more complete, polished and useful to a scenario designer tasked with "scripting his own AI plans", which in reality is essentially "simulating" the kind of things a human player would do.  It seems so frustrating that the core/bones/intent of the mechanics is there but by omitting (overlooking?) some seemingly basic finishing touches (eg. end of movement facing orders), so much capability and potential for scenario designers to "design in" a much more "intelligent" CPU opponent behaviour is lost.  It's like it's just one order/finishing step away from potentially being a much more useful and intelligent tool for scenario designers to use.

In some ways, not being too much of a SP player (though I do really enjoy campaigns and the odd recommended SP-only scenarios), I shouldn't really care too much.  However, if "good for player, bad for CPU opponent/TacAI coding" features are being rejected/dismissed by BFC in favour of development of less "complicated"/messy , more compartmental features, like improving the suite of AI programming tools available to the scenario designer, I want to see them get the polish they deserve.

I really hope that if BFC are serious about their commitment to/focus on the SP experience, that the tools available to the scenario designers to "simulate a human opponent" get.  This area of the game definitely is not "held back" in the same way "good for player, bad for CPU opponent/TacAI coding" features are.

Edited January 6, 2016 by Lt Bull

[RockinHarry]

OMG. I don't know if I am more shocked or disappointed by what you are telling me for a number of reasons.

I recall BFC long ago clearly drawing a line and stating essentially this: "We don't really care so much for the needs/wants/features of the MP aspect (customers) of CM, they are just a minor part of the market. The MP players can even consider themselves lucky to even have PBEM in CMx2 (yes, PBEM was not in the original plan or release of CMx2)). The focus and priority of CM is and always will be the SP experience."

If BFCs position on this has changed, I have not seen it stated anywhere.  I for one am a MP-centric player and understand that catering for both the SP and MP aspects of the game is no easy thing to do.  I also see how even the introduction of seemingly simple/trivial features in to the game that can be handled by players no problems, would/could present themselves as "CPU opponent coding nightmares" when considered in the context of "if we can't code the CPU opponent to intelligently utilise the same set of unit orders/features a human player can easily assign their units, then we won't include it because it will disadvantage the CPU opponent too much" philosophy. I can also see how a developmental dilemma like this can lead to/has led to the "stunted/modest growth in CM game features" some have complained about on these forums.  With this perspective, I am not surprised by the "modest" growth of game features and rejection of otherwise seemingly "simple" feature requests/ideas we have seen.

So if the SP experience is BFC's favourite baby and where they really want to shine, then I am shocked at how "un-finished"/crude the very tools available to scenario designers to "create a SP experience" seem to be.  It is only through these tools along with the creativity of the scenario designer to work within the limits of these tools (aside from the omnipotent TacAI of course), that the whole SP experience is based upon, and upon which the concept of a CM SP CPU opponent will ultimately be judged on.

I just would have expected simple things like the triggered move orders I have been asking about to be much more complete, polished and useful to a scenario designer tasked with "scripting his own AI plans", which in reality is essentially "simulating" the kind of things a human player would do.  It seems so frustrating that the core/bones/intent of the mechanics is there but by omitting (overlooking?) some seemingly basic finishing touches (eg. end of movement facing orders), so much capability and potential for scenario designers to "design in" a much more "intelligent" CPU opponent behaviour is lost.  It's like it's just one order/finishing step away from potentially being a much more useful and intelligent tool for scenario designers to use.

In some ways, not being too much of a SP player (though I do really enjoy campaigns and the odd recommended SP-only scenarios), I shouldn't really care too much.  However, if "good for player, bad for CPU opponent/TacAI coding" features are being rejected/dismissed by BFC in favour of development of less "complicated"/messy , more compartmental features, like improving the suite of AI programming tools available to the scenario designer, I want to see them get the polish they deserve.

I really hope that if BFC are serious about their commitment to/focus on the SP experience, that the tools available to the scenario designers to "simulate a human opponent" get.  This area of the game definitely is not "held back" in the same way "good for player, bad for CPU opponent/TacAI coding" features are.

It´s true that we just have some "basic" features for the AI to act in certain ways with the current game engine and that we´re lightyears away from getting an AI opponent that has similar capabilities like a mediocre human opponent. Triggered moves aren´t perfect, also true, yet one can still achieve a number of things when one gets aquainted with the scripting capabilities generally. So I wouldn´t give up on it yet until you at least tried some AI scripting in practice.

In the meantime I´d prepared bits of info as I find this thread could develop to something with lots of potential, particularly for players who want to start on serious AI scripting as well.

Based on the mentioned, trigger driven retrograde movements I prepared some very basic example, with just two opposing units involved (1 US AI and 1 German human player Forward Observer).

Situation and Task:

A US AI FO unit starts the game sitting in its initial setup zone with the task of just observing towards the german line and to retreat move to a fallback position, once a german unit "touches" (means moving into) a preset Trigger Objective Zone (Trigger1). The procedure is repeated if the german player moves into a second Trigger objective zone (Trigger 2), with the US AI FO finally retreat moving from its fallback position to...somewhere making sense thereafter.

The US FO is the only AI unit on the map and the only member of a Group (Group1/A1).
Mission length, 30 minutes.

US Group 1 orders in AI editor are as follows:
(...) parameters are not important for the example.

"Trigger 1" is a US trigger objective zone with setting "enemy all".
"Trigger 2", the same.

Group1

Setup

(Mixed)
Hide
(No Dismount)
Exit Between: 00:00
...and             : 30:30
Wait for: Trigger 1

The US AI FO will sit and wait in his setup position, until a german unit moves into and thus activate "Trigger 1" zone, in this case the german FO. If no german unit activates "Trigger 1", the US FO will not move away at all during the whole game time (unless routed ect.). That means none the orders starting with Order 2 would be executed. "Exit between" - "..and" sets the time frame that the trigger is allowed to become active. In this case it can become active from game start until slightly beyond end of the game. "Hide" is actually not the best choice as it seriously limits spotting capabilities. An "Ambush 75m" would do the purpose better.

Order 2
(Not Trigger)

Quick
(Mixed)
Hide
(No Dismount)
Exit Between: 00:00
...and             : 00:30

Once the trigger got activated, the US FO quick moves to the fallback position. "Hide" stance only becomes active when the Order 2 fallback position is reached finally (soldiers go to prone). Otherwise the "Hide" should prevent the unit during movement to engage into any combat actions if possible. "Cautious" would be another alternative stance. "Exit between" - "..and" tells the US FO to not wait and immediately start execution of Order 3. Nonetheless, the AI includes a further game turn for the US FO to face toward the objective zone for Order 3, before it plots the movement order and starts moving the next turn. That needs to be taken into consideration for proper movement timings.

Order 3
(Not Trigger)

Quick
(Mixed)
Ambush 75m
(No Dismount)
Exit Between: 00:00
...and             : 30:30
Wait for: Trigger 2

The US FO gets "Ambush 75m" stance activated and quick moves to the "final" fallback position, which is 1 AS towards the german frontline again. This as Bulletpoint already explained, is to get the unit facing back toward german lines at the earliest opportunity. Now again the US FO will sit and wait until a german unit moves into and activates "Trigger 2" objective zone. Then Order 4 would be executed the next game turn. Just like during "Setup", the activation time frame for "Trigger 2" is set to last from games start until slightly beyond games end (30 minutes). If "Trigger 2" does not get activated, Order 4 will not get executed and the US FO never moves on from there. The Trigger activation time frame off course can be set to different "Exit between" - "..and" figures, but for the given example I´ve chosen the "whole game time" setting so no matter what, the US FO will only execute movement orders when the triggers get activated. The V3.11 game engine manual gives further info for the "Exit between" - "..and" settings when used with Triggers. 

Order 4
(Not Trigger)

Quick
(Mixed)
Normal
(No Dismount)

When "Trigger 2" gets activated, the US FO will move away from its fallback position to another map zone further back.

Example mission link attached below. Test play as german human player vs.US AI in scenario author mode! Move the german FO towards the US FO and touch the trigger objective zones in succession (ALT-J to show objectives). Click/activate the US AI FO and see what actions the AI takes and when.

https://www.dropbox.com/s/spkxe92hfht3in3/CMBN Trigger Example.btt?dl=0

Edited January 6, 2016 by RockinHarry

[A Canadian Cat]

Nice @RockinHarry that looks right.

[Lt Bull]

Hey RockinHarry,

Thanks for detailed reply and sample file.  That seems to be the best way to achieve what I was talking about.  Unfortunately this string of orders would most likely be inappropriate (useless) for tanks/vehicles because they would have to pivot on the spot 180deg, move back all the while exposing rear to enemy, then again pivot 180deg to face the enemy again.  Will try messing around with some other ideas.

It´s true that we just have some "basic" features for the AI to act in certain ways with the current game engine and that we´re lightyears away from getting an AI opponent that has similar capabilities like a mediocre human opponent. Triggered moves aren´t perfect, also true, yet one can still achieve a number of things when one gets aquainted with the scripting capabilities generally. So I wouldn´t give up on it yet until you at least tried some AI scripting in practice.

In the meantime I´d prepared bits of info as I find this thread could develop to something with lots of potential, particularly for players who want to start on serious AI scripting as well.

Well said and be my guest to turn this thread as a bit of a info/how-to thread for players interested in exploring the much under appreciated art  of AI scripting in CM.

But why are there just some "basic" features for the AI to act in certain ways? Is not the SP experience BFC's focus/ favourite baby? Lets explore this further.

It is worth noting and recognizing that apart from the omnipresent (lol I called it omnipotent in my last post!) TacAI, which is critical for the game but only relates to "local unit intelligence", everything else in the game we consider "the AI" is purely a result of the scenario designers efforts and creativity at utilizing what I think are an under developed and limited suite of tools to "simulate"/emulate the "traditional" kind of AI we typically think of when we think of the AI in a computer game.

However if you think about it, CMx2 actually really has no real proprietary "CPU Opponent" concept to it, which is strange for a game that is primarily targeted at SP.  In that respect, BFC have dodged the responsibility of creating a/the "CPU opponent" and instead totally deflected/placed that responsibility in the hands of the actual scenario designer who is left to tinker with these seemingly underdone set of tools.  It's like being handed just some wood and nails and expecting to build a robot.

It is the scenario designers (Not BFC) that breathe the "virtual opponent" life in to the game. BFC's role is to just provide the tools.  I would not necessarily say the range and quality of these tools are near the level at which I would have expected them to be after all these years of development.

Kind of a strange situation, but then again sandboxy games like the ArmA series fall in to the same category.  The huge difference however is that the guys who coded ArmA poured a huge amount of effort and support in to the suite of tools that scenario designers can use to create all sorts of "credible" CPU opponents/missions/scenarios etc, definitely surprising even the game developers themselves as to what is possible when you give the community the right tools to explore and create.  And rightly so of course, because they understand just how important it is. I would put CM and ArmA at opposite ends on the "how well designed/considered/implemented/developed are these tools" scale of things.  I do not think it would hurt BFC to at least compare their "scenario tools" to those in ArmA to at least provide inspiration/idea of what is possible, if inspiration and ideas is what is lacking.

Does it not seem odd that when you see talk about making CM a better game by both BFC and players, improvements to the tools available to the scenario designer don't really figure too much in the discussion?  In some ways yes, given BFC tell us SP CM gaming makes up the vast majority of player interests. But then again probably not. The actual tools that make CM SP gaming even possible are only ever looked at, let alone even considered by a minute fraction of the CM community.  The vast majority of people have remained clueless about the very capabilities/limitations of the tools that are even used to try and create a challenging/"realistic" SP experience "against the AI".

Would really be good to hear comments from BFC about any plans for enhancing/improving/polishing these tools used to create the SP experience.  Give the scenario designers the tools, and they will use them, even if the majority of us remain clueless.

Edited January 7, 2016 by Lt Bull

[Bulletpoint]

The editor would really benefit from a general overhaul. AI plans are just one of many convoluted aspects of it.

It would make me personally more interested in doing more scenarios, and I suspect I'm not the only one who might be interested in doing more, if it were made more accessible.

But one thing is learning curve and accessibility. Another is basic features that make better and more realistic scenarios possible. Such as a facing command.

The important thing is not so much facing itself, it's that troops end up on the wrong side of cover.

I'm surprised a "take cover towards" command at the end of a movement order hasn't been implemented yet, surely some of the volunteer official scenario designers must have needed it, especially now that we have a simple AI-trigger system that really begs to be used for fallback plans. 

It wouldn't need to be a 360 degree command like in the game, a simple drop-down menu to select a general compass direction would be just fine. (north, northeast, east, southeast... etc). 

[RockinHarry]

Thanks for detailed reply and sample file.  That seems to be the best way to achieve what I was talking about.  Unfortunately this string of orders would most likely be inappropriate (useless) for tanks/vehicles because they would have to pivot on the spot 180deg, move back all the while exposing rear to enemy, then again pivot 180deg to face the enemy again.  Will try messing around with some other ideas.

Yep, the TacAI just knows a single movement mode which is "Quick" for almost all cases and types of units.  Not sure what BFC´s idea is all about the limited choice of move modes for the scripted AI, but my guess would be that a number of choices are left to the dynamic TacAI intentionally, since lots of actions and movements would very much depend on combat situations and unit soft factors. Vehicle reverse movement is not a tactical choice that fits into a larger framework of a Combat Plan, which what the AI Plans are basically about to provide. That could be a "possible" descripton of the overall AI scripting concept of the game, although I wouldn´t take that as excuse to not provide more choices to a scenario maker and to the AI generally.

Vehicle reverse movement: One could argue that reverse movement is normally just very short distances (few meters up to..?) and that the decision to use it belongs to the commander of the lowest level, the vehicle commander. We know that this is implemented this way, as (Tac)AI vehicle commanders decide to reverse at high threat levels automatically (dependent on soft factors). Moving a vehicle to rearward positions as part of a larger battle plan is more something that the higher command is about to decide, not the micro management that BFC leaves to the low level TacAI commanders. So from taking it from that POV and beside hoping that BFC refines the system, one is only left to work with what is available now.

Let us know about your ideas when messing around and if they work!

Some thoughts on the situation (vehicle retrograde movements): Possibly they work best if they start from a position that offers both, hull down and key hole. Vehicle turn radius needs also some considerations, particularly for those vehicles that can not turn on the spot. So bits of space forward and sidewards is required for the vehicle to turn away from its position towards the next movement zone.

An idea: Dependent upon situation, one can place or move a vehicle into a hull down position and let it turn around on the spot (auto face due to non activated next order or trigger), so that just the turret (unless it´s a Stug) would show to a potential enemy unit. Off course It would start with the rear turret showing toward the enemy as well, but maybe with some "optimized" info sharing (the vehicle is under command of an infantry HQ spotter nearby), the vehicle commander could swing around the turret towards an enemy unit and engage with the reversed hull staying in hull down position. Possibly the TacAI would interfere with attempting to turn around the hull towards the enemy threat, if one exists. If not, the vehicle would be in best position to retrograde move quickly in forward gears. Even if it just has some limitied application it´s worth to make some tests with it. My task for today.

[RockinHarry]

 

Kind of a strange situation, but then again sandboxy games like the ArmA series fall in to the same category.  The huge difference however is that the guys who coded ArmA poured a huge amount of effort and support in to the suite of tools that scenario designers can use to create all sorts of "credible" CPU opponents/missions/scenarios etc, definitely surprising even the game developers themselves as to what is possible when you give the community the right tools to explore and create.  And rightly so of course, because they understand just how important it is. I would put CM and ArmA at opposite ends on the "how well designed/considered/implemented/developed are these tools" scale of things.  I do not think it would hurt BFC to at least compare their "scenario tools" to those in ArmA to at least provide inspiration/idea of what is possible, if inspiration and ideas is what is lacking.

I´d messed some with the ARMA tools and editors (ARMA 1+2) some time ago and while you can do lots of interesting stuff with them, it partly requires very good programming knowledge to get more advanced things to work. I also know of some hard coded limitations that you can´t work around. I remember that for some the WW2 mods (Invasion 44 and Iron Front 44), you can´t some the vehicle/tank behaviors to change to more realistic and better ones. Some the more important AI and game core elements are hardcoded, but bits of stuff can be programmed via the very interesting FSM ( https://en.wikipedia.org/wiki/Finite-state_machine ) files and scripting languages. However, a number of modding and programmer groups created outstanding content for the series and keep doing so. That includes work and expansions of the internal scripting/coding language, something that is left to real professional people at last. All the private ARMA improvement undertakings come at a price. Demands for computer hardware rise significantly.

Hard to tell how well both games can be compared, with ARMA beeing a tactical 1st person shooter and CM in its own tactical wargame niche. Opening up some the games core elements to be tweaked and improved by 3rd party groups might be a promising way to relieve the BFC staff to what they figure to be more important to work on with their limited resources. There´s lots what can be improved or added for CMX2 mission design purposes and this thread is a good place to discuss what is doable with what we have now. If anything can´t get to work, no matter what and if there´s a demand from a majority of players and mission designers, BFC will surely keep an open mind for further improvements in the near future. Personally I do not expect major improvements before V4.0 of the game engine though.

[RockinHarry]

The editor would really benefit from a general overhaul. AI plans are just one of many convoluted aspects of it.

It would make me personally more interested in doing more scenarios, and I suspect I'm not the only one who might be interested in doing more, if it were made more accessible.

But one thing is learning curve and accessibility. Another is basic features that make better and more realistic scenarios possible. Such as a facing command.

The important thing is not so much facing itself, it's that troops end up on the wrong side of cover.

I'm surprised a "take cover towards" command at the end of a movement order hasn't been implemented yet, surely some of the volunteer official scenario designers must have needed it, especially now that we have a simple AI-trigger system that really begs to be used for fallback plans. 

It wouldn't need to be a 360 degree command like in the game, a simple drop-down menu to select a general compass direction would be just fine. (north, northeast, east, southeast... etc). 

Hm...personally I do not find accessability of the editors that difficult, but admittedly one needs to spend lots of time with them due to the high learning curve. And there´s always something new to learn, almost every day and with every new mission design attempt. The mission design pros will certainly agree and I hope some of them will join here and share some their "secrets".

Personally I´d like to see some more user control on facing for AI driven forces as well. If one accepts that the micro management is mostly left to the TacAI, one can only attempt to work around and approach a scenario design more from the bigger tactical picture. My given scripting example (basically the same to be found in the V3.0 game manual) to handle retrograde movements and facing, is just one way to handle a particular situation for an AI unit. While AI infantry units can handle themselves well enough when it comes to facing, some problems remain with vehicle units that demand more attention from a mission designer, when it comes to movement zones, particular terrain types and path finding between zones. It´s a good place here for further discussions and possible solutions.

"Take cover towards" is actually available in the game through one the instant commands (avoid?), yet not within the larger framework of AI plans. Obviosuly BFC considers this to be part of lowest level decision making, left to the TacAI and not as part of larger (AI) plan. BFC could possibly implement something that takes enemy/friendly map edges more into consideration, when it comes to facing. Currently map edges ARE taken into consideration when it comes to friendly/enemy side of cover terrain in an action spot. I´d good successes with experimenting with map edges when I noticed certain problems. Oftenly things work better if map edges are not just a single cardinal direction (N,S...), but 2 instead (NW, SE...). Map edges are used for off map Arty purposes and routing units and there could possibly some slight weighting of map edges for facing purposes, although hardly noticable. Some more weighting would be desirable, maybe some toggle that can be used by a mission designer (Use map edges for facing purposes, or leave it at default).

The 360° TA is probably implemented for the scripted AI so that it can react on enemy threats in more flexible ways. A human player can react and correct TA´s if the enemy approaches from unexpected directions, while the AI can´t. But your idea with allowing AI units at least some weighting towards a particular compass direction might be something for BFC to think about. Map edge weighting and a 180° TA might help to face selected AI units towards the most likely threats in a mission.

[RockinHarry]

I´d also like to point to JonS excellent Sheriff of Oosterbeek scenario design AAR, which contains a wealth of expert info for scenario makers. Of particular interest for the scripting AI topic is both these chapters:

http://community.battlefront.com/topic/109190-the-sheriff-of-oosterbeek-–-a-scenario-design-daraar/?do=findComment&comment=1448749

http://community.battlefront.com/topic/109190-the-sheriff-of-oosterbeek-–-a-scenario-design-daraar/?do=findComment&comment=1449559

Market Garden module owners will have the full DAAR already as PDF file in their inventory.

[JonS]

Thanks @RockinHarry

There are a couple of points I made there that are still really relevant to this thread, I think.

2) StratAI –

the Strategic Artificial Intelligence is programmed by the scenario designer, which provides the high-level scheme of manoeuvre that AI-controlled units will attempt to execute. Human-controlled forces have no StratAI

... it’s crucial to try and get ‘inside the head’ of potential players, and guesstimate they different ways they’re likely to go about tackling the tactical problems your scenario presents them. Also, I think that at a high level AI plans need to be fairly simple and generic – broad sweeping movements to seize or hold ‘big’ objectives, rather than fine scale movements from one piece of cover to another.

Basically, the AI can't really be programmed to move carefully from one hedgerow to the next ... so don't try to do that. Create an overall high level plan (i.e., a company- or battalion-level plan, depending on the size of the scenario) and set the AI groups moving within that framework. And don't worry too much about exactly where the AI d00dz end up, or which way they're facing. If your overall plan for them is simple and robust, then it will present a coherent and plausible challenge for the player, even if it isn't 'perfect.'

Keep it simple, and use the editor that's actually in front of you.

[RockinHarry]

Thanks @RockinHarry

There are a couple of points I made there that are still really relevant to this thread, I think.

 

Basically, the AI can't really be programmed to move carefully from one hedgerow to the next ... so don't try to do that. Create an overall high level plan (i.e., a company- or battalion-level plan, depending on the size of the scenario) and set the AI groups moving within that framework. And don't worry too much about exactly where the AI d00dz end up, or which way they're facing. If your overall plan for them is simple and robust, then it will present a coherent and plausible challenge for the player, even if it isn't 'perfect.'

Keep it simple, and use the editor that's actually in front of you.

Yep, really depends much on size of AI forces to be moved across a map. If it´s yet out of sight from the human player, one can keep things more generic when moving from a possible setup/deployment zone towards a zone just before likely contact is to be made with the opponent. With the inability of the (Op-) AI to keep forces seperated and move in halfway coherent "formations" toward a movement zone, one does not like to catch the AI with lots of units intermingled, when they all try to reach the next zone on more or less the same (shortest and fastest) path. If one can distribute the AI forces in more or less platoon sized groups among the available 16 groups, lots of that timing and movement trouble can be avoided. This becomes more and more necessary if the AI gets into the actual combat zone, where bunching up and possible seperation from formations to their HQ´s could hurt the AI efforts badly.

There´s a number of ways, particularly with regard to movement zones that makes the AI tasks somewhat easier and also helping with some the mentioned facing issues.  Assuming a Plt sized group, I´d call it the "T" setup, where the "I" (column) serves as an intermediate assembly zone and the following "--" zone ( a line) beeing the final deployment zone. Reason for this is that the next action spot within a movement zone is rather chosen randomly for single units, as long as it´s a yet unoccupied AS nearest to the enemy map edge and nearest to the next movement zone, if one exists. (Map edge effect on Zone deployment needs confirmation). Moving from the column deployment zone towards the line deployment zone gets less units intermingled and thwarted, thus also diminishing movement times between zones and additional tiring. Off course intermediate terrain needs to be halfway suited for that maneuvre.

Here´s a quick sketch with green squares showing the AS´s (from the movement zone), each containing a squad or team as part of a Plt size group. The black arrows show the possible paths (worst case selection) between zones and final facing (blue arrows). The "T" setup gives the idea that it generally provides less opportunities for single units to get into each others way, while moving towards the next zone line. It´s not a one fits all setup, but can be generally adapted to the terrain and possible situation in an area. The basic idea is that while all units from a group start movements simultaneaously, to avoid at least some the bunching up and timing troubles during movements, when not yet under enemy observation.

[RockinHarry]

 

An idea: Dependent upon situation, one can place or move a vehicle into a hull down position and let it turn around on the spot (auto face due to non activated next order or trigger), so that just the turret (unless it´s a Stug) would show to a potential enemy unit. Off course It would start with the rear turret showing toward the enemy as well, but maybe with some "optimized" info sharing (the vehicle is under command of an infantry HQ spotter nearby), the vehicle commander could swing around the turret towards an enemy unit and engage with the reversed hull staying in hull down position. Possibly the TacAI would interfere with attempting to turn around the hull towards the enemy threat, if one exists. If not, the vehicle would be in best position to retrograde move quickly in forward gears. Even if it just has some limitied application it´s worth to make some tests with it. My task for today.

A quick test revealed that it basically works, as long as the AI turreted vehicle is in perfect hull down position. It´s at a disadvantage though, with the turret also facing rearward. Unbuttoned commander and info sharing from nearby friendlies help to spot enemies and get the turret rotated to the threat, while the hull is kept facing to the rear.

So as a method to get a vehicle "reverse" from a position via AI Orders and triggers, it basically works. I think this to be a more worthy tactic for a human player, as TA and facing can be set independently. 

[BornGinger]

This thread has been hibernating for more than two years now, so it's maybe time to wake it up again.

I have quite recently finished my scenario map and am about to deal with the AI. Before I start, I have a few questions that I'm sure someone knows the answers to:

When I go about to do the AI plans, is there in the Scenario Editor the options to use a trigger like "if-then" and similar orders to the AI?

Example 1: If there are two areas where reinforcements are arriving to and there is a possibility that one of those areas, which contains a setup zone for reinforcements, would be occupied by the opponent, is there an option to tell the AI "Send reinforcements to setup zone 1 but if that one is occupied by enemies then send reinforcements to setup zone 2"?

Example 2: If A company has been given the order to take Position 1 but fails, is there a possibility to tell B company to "Go and take Position 2, but if Position 1 isn't taken then go and take position 1"?
 
In the Unit section of the editor there are two columns. One is a column with available troops and the other one is a column with activated troops. When I expand a battalion in the activated troops column, I have problems to check which troops I have in each battalion or company, and to remove those I don't want to use, because I can't scroll the list of the activated troops. Is it supposed to be that way or am I doing something wrong?

The picture below shows the 3rd Section of the 3rd Platoon expanded. But I can't see what the section contains because I can't scroll.

If I'd like to make my scenario map available for H2H as well as for a single player scenario, should I aquire troops for each side and deploy them or should the scenario map be empty of troops?

 

Edit: What is the problem with uploading a picture to show what I mean? First I uploaded a picture which was accepted. Later on when I wanted to crop the picture I was informed that I only could upload 160kb. So I made the picture smaller but of some reason I wanted to change something with the picture again. The third time I wanted to upload it I was told that I only could upload 20kb. So now there is no picture to help the reader to understand what I mean with the problem in activated troops.

 

 

 

 

Edited July 23, 2018 by BornGinger

[Mord]

4 hours ago, BornGinger said:

If I'd like to make my scenario map available for H2H as well as for a single player scenario, should I aquire troops for each side and deploy them or should the scenario map be empty of troops? 

I can't answer the scenario design questions but I can answer this. You could add the forces you wish to have players use, and/or you could make it a quick battle map. If you do make it a quick battle map just make sure you say it's HTH only.

 

Mord.

[BornGinger]

This picture should explain a little bit more what I mean by the text above when I say that I can't see what units a platoon or squad contains. The 3rd section in the 3Rd Platoon is expanded but there's no way for me to see what is below the edge of the Activated Troops window because I can't scroll in it

Edited July 23, 2018 by BornGinger

[BornGinger]

5 hours ago, Mord said:

I can't answer the scenario design questions but I can answer this. You could add the forces you wish to have players use, and/or you could make it a quick battle map. If you do make it a quick battle map just make sure you say it's HTH only.

Isn't HTH (H2H) something different than a quick battle map? I thought HTH (H2H) was the type of battle you play against another person either realtime or via pbem and quick battle an engagement against the computer where you choose your troops based on different points?

[George MC]

36 minutes ago, BornGinger said:

This picture should explain a little bit more what I mean by the text above when I say that I can't see what units a platoon or squad contains. The 3rd section in the 3Rd Platoon is expanded but there's no way for me to see what is below the edge of the Activated Troops window because I can't scroll in it

Are they on-map or off-map mortars?