Jump to content

AI plans and a more responsive AI


Recommended Posts

On 6/3/2020 at 9:26 PM, RepsolCBR said:

IntrestingĀ idea...šŸ™‚

And I agree...all possibilties are probably not explored yetĀ  !

RE self trigger evading/retreating AIP units. It works basically. Once the TacAI selects a heavily suppressed (pinned state) unit to evade and it touches the trigger zone, all of its group members getĀ to move to next (rearward) movement zone. This for units initially placed in a defense stance and meant to hold as long as possible from given position. In this case the trigger zone would be placed to the rear of this unit/group. Standoff can be anything from just one action spot to multiple, as long as the trigger zoneĀ is in the mostĀ likely retreat direction of the evading unit. That got to be figured out fom testing. A retreat trigger zone closer to original defense position gets all of the group to retreat sooner/faster than one placed further back.

Then thereĀ“s various possibilities forĀ move mode once the groups (retreat) movement got triggered. Assuming a Plt size group/formation, quick/dash will move more of that group at once, while the more attack oriented modes (advance/assault/max assault) moves in smaller pieces over longer time. Similar consideration for combat stances. Better move back at "cautious" (shoot little) or make a fighting withdrawal with "active"? Many (hard) choices here.

So at last this method might be quite useful for smaller outpost groupsĀ Ā containing just 1-2 units.Ā Once forced evading they "retreat" toward a switch postion as selected by mission designer, instead of leaving that to the TacAI. Sequential switch positions are also thinkable. In case the group survives long enough and noĀ other frindlies move on the trigger zone. So far so good.

Link to comment
Share on other sites

5 hours ago, RockinHarry said:

RE self trigger evading/retreating AIP units. It works basically. Once the TacAI selects a heavily suppressed (pinned state) unit to evade and it touches the trigger zone, all of its group members getĀ to move to next (rearward) movement zone.

This is a very good idea. Realistic, increases challenge, and it avoids the immersion-breaking situation where you'll force enemies back from hedgerows and then just find them camping out in fields later.

Link to comment
Share on other sites

7 hours ago, RockinHarry said:

RE self trigger evading/retreating AIP units. It works basically. Once the TacAI selects a heavily suppressed (pinned state) unit to evade and it touches the trigger zone, all of its group members getĀ to move to next (rearward) movement zone. This for units initially placed in a defense stance and meant to hold as long as possible from given position. In this case the trigger zone would be placed to the rear of this unit/group. Standoff can be anything from just one action spot to multiple, as long as the trigger zoneĀ is in the mostĀ likely retreat direction of the evading unit. That got to be figured out fom testing. A retreat trigger zone closer to original defense position gets all of the group to retreat sooner/faster than one placed further back.

Ā 

1 hour ago, Bulletpoint said:

This is a very good idea. Realistic, increases challenge, and it avoids the immersion-breaking situation where you'll force enemies back from hedgerows and then just find them camping out in fields later.

Agreed - that's a neat trick. I must try it :)

Link to comment
Share on other sites

In one scenario I made, I did it a bit differently - I painted trigger zones that would make enemy teams fall back if the player is close to occupying flanking positions, or if the player approaches the enemy positions to a certain distance. This means that enemy teams that have been thrown back will get a chance to retreat before the player takes those hedgerows and guns them down in the open.

Edited by Bulletpoint
Link to comment
Share on other sites

12 hours ago, Bulletpoint said:

This is a very good idea. Realistic, increases challenge, and it avoids the immersion-breaking situation where you'll force enemies back from hedgerows and then just find them camping out in fields later.

thatĀ“s a good example, yes.

10 hours ago, Bulletpoint said:

In one scenario I made, I did it a bit differently - I painted trigger zones that would make enemy teams fall back if the player is close to occupying flanking positions, or if the player approaches the enemy positions to a certain distance. This means that enemy teams that have been thrown back will get a chance to retreat before the player takes those hedgerows and guns them down in the open.

another good possibility. But as said above you donĀ“t want to get a larger friendly group fall back just because a tiny enemy scout unit moves onto the trigger.Or in case of V4.x, an enemy straggler evading forward. So again a number of "what-ifĀ“s" would come handy. WeĀ also just have 15 terrain objectives and when becoming too creative with them, one runs out of them just like with AI groups. So while everybody asks for more AI goups (me too), IĀ“d say we also need more terrain objectives. IĀ“m fine with using them for that purpose (trigger) exclusively as I donĀ“t need Occupy types for scoring. Preserve and Destroy is also not that much usedĀ I think.Ā At least I havenĀ“t had a use for them yet. Touch and Exit leaves more interesting possibilities though.

Would also be nice if terrain triggers can be tied to a certain AI group only. Generic friendly/enemy is not quite sufficient here. Another one would be to have terrain objective triggers activate or deactivate just for certain time frames, or time generally. In fact thatĀ“s another What-if option as it needs to be tied with the orders sequences which can become a bit complicated. Multiple trigger dependencies would be interesting too. I.e have an AI group start any actions/orders, when more than one trigger gets activated and such. Much food for thought....:wacko:

Ā 

Link to comment
Share on other sites

4 hours ago, RockinHarry said:

another good possibility. But as said above you donĀ“t want to get a larger friendly group fall back just because a tiny enemy scout unit moves onto the trigger.

Ā 

If a scout team is able to get into the trigger zone, that means the AI group is not longer able to control that ground, so it's good that they fall back. In my scenario, the AI groups were very small, down to individual teams at times.

Link to comment
Share on other sites

On 6/13/2020 at 6:25 AM, RockinHarry said:

Once the TacAI selects a heavily suppressed (pinned state) unit to evade and it touches the trigger zone, all of its group members getĀ to move to next (rearward) movement zone.

+1Ā  Cool idea.Ā  I did something similar in the CMBS scenario: Tactical Operations Center.Ā  Ā But instead of the evading unit causing other units to fall back it caused a different AI unit to come forward to reinforce.Ā Ā 

Ā 

7 hours ago, RockinHarry said:

So while everybody asks for more AI groups (me too), IĀ“d say we also need more terrain objectives.

Yes.Ā  More AI Groups and terrain objectives.Ā  Especially since terrain objectives are also used as exits and triggers.Ā 

Link to comment
Share on other sites

3 hours ago, MOS:96B2P said:

+1Ā  Cool idea.Ā  I did something similar in the CMBS scenario: Tactical Operations Center.Ā  Ā But instead of the evading unit causing other units to fall back it caused a different AI unit to come forward to reinforce.Ā Ā 

Ā 

Yes.Ā  More AI Groups and terrain objectives.Ā  Especially since terrain objectives are also used as exits and triggers.Ā 

another nice variation! B)Ā Hope we could start some "Scripting AIP solution center" sort of thread anytime soon. After the next patch release maybe. Too many good ideas and solutions get lost scattered in too many different threads.

Link to comment
Share on other sites

11 hours ago, MOS:96B2P said:

+1Ā  Cool idea.Ā  I did something similar in the CMBS scenario: Tactical Operations Center.Ā  Ā But instead of the evading unit causing other units to fall back it caused a different AI unit to come forward to reinforce.Ā Ā 

Ā 

Yes.Ā  More AI Groups and terrain objectives.Ā  Especially since terrain objectives are also used as exits and triggers.Ā 

How come I keep hearing the song from Oliverā€¦.

Ā 

Link to comment
Share on other sites

On 6/14/2020 at 7:14 AM, RockinHarry said:

But as said above you donĀ“t want to get a larger friendly group fall back just because a tiny enemy scout unit moves onto the trigger.Or in case of V4.x, an enemy straggler evading forward. So again a number of "what-ifĀ“s" would come handy.

If there was a more than or less than option to enemies triggering that line it should also work.

Maybe something like:

Red triggers Line Yellow

If Red = > 50

Ā  then move back

If Red = 50 or < 50

Ā  then stay put

Edited by BornGinger
Link to comment
Share on other sites

Exits are a cool idea I think just need to be used differently than they usually are. One of the ways i'd really like to use them is to recoup some of the points invested in my force by allowing broken units or vehicle crews with no vehicle to simply exit the play area and perhaps make up a little of the difference for some other KIA tags.Ā 

Link to comment
Share on other sites

46 minutes ago, BornGinger said:

If there was a more than or less than option to enemies triggering that line it should also work.

This is something i want also šŸ˜Ž...

Being able to set some kind of treashold would be useful not only for triggers but also for objectives...

As of now...a single sniper is all that is needed to hold an objective (if no enemy is present)..

Many times it would be kind of cool if those requirements could be somewhat higher...

Lets say...a platoon-sized force is needed (atleast) holding a certain objective to considder the requirement met...and score points for it...

We already have a nice unit-value/cost system in the game...the QB-battle purchase costs...

Perhaps those values could be used to calculate the appropriate treashold-values for triggers and objectives šŸ˜Ž..

Ā 

Link to comment
Share on other sites

Quite a few years ago I wrote about having more possibilities to the AI editor in the game and was told that Battlefront won't make any changes.

Ā 

But if military personnel are using this game to improve their tactical approach on the real world battle fields, as visible here, wouldn't Battlefront really want to improve their Tac AI?

Ā 

*I wonder why the text to the video says "Enjoy our work with Slitherine Games" when they are showing Combat Mission Shock Force or Black Sea.

Edited by BornGinger
Link to comment
Share on other sites

11 minutes ago, BornGinger said:

Quite a few years ago I wrote about having more possibilities to the AI editor in the game and was told that Battlefront won't make any changes.

That was not an official statement from Battlefront i guess...

If it was...that would be a strange thing to say...

After all...the areafire, face and withdraw commands are fairly new...

Link to comment
Share on other sites

19 minutes ago, BornGinger said:

Quite a few years ago I wrote about having more possibilities to the AI editor in the game and was told that Battlefront won't make any changes.

Ā 

But if military personnel are using this game to improve their tactical approach on the real world battle fields, as visible here, wouldn't Battlefront really want to improve their Tac AI?

Ā 

*I wonder why the text to the video says "Enjoy our work with Slitherine Games" when they are showing Combat Mission Shock Force or Black Sea.

Answer from Steve himself is in this threadĀ 

Ā 

Link to comment
Share on other sites

I recall long ago Steve commenting on the topic of newĀ IU commands. He said individual proposed special commands may sound good but you've got to factor in 'player overload'. Clutter up the UI with too many new orders and players will start tuning them out, gameplay will become burdensome. I think he has the same philosophy with theĀ AI editor. I recall when they were first discussing adding triggers Steve turned down all suggestions where the scenario designer would be obliged to write code to get it to work. Keep it reasonably basic and reasonably intuitive for most users. A sleek sportscar may be more desirable than a suburban SUV but if you don't know how to drive stick the sportscar's not going to get much use.

Even longtime scenario designers can be unaware of the AI's capability. It was just recently that we learned (to our great surprise) you can navigate through groups and orders using the keyboard arrow keys. How many scenarios have I made and I didn't have a clue about that! One of our merry band recently learned how the 'retreat' AI map marker works. Another recently learned about directing on-map mortars to fire outside of LOS using the AI. The CM AI isn't just the AI, it a collaboration between the AI and the scenario designer. You need an AI the designer can work with and the AI needs a designer that knows what he's doing. its a balancing act.

Link to comment
Share on other sites

10 minutes ago, MikeyD said:

It was just recently that we learned (to our great surprise) you can navigate through groups and orders using the keyboard arrow keys. How many scenarios have I made and I didn't have a clue about that! One of our merry band recently learned how the 'retreat' AI map marker works.

How do you actually do that? Is it to click on one choice and after that use the arrow keys to go up and down? If there are functions with the AI which many don't know anything about, I believe it's about time someone makes a pdf with them so people can use the AI tools better.

Link to comment
Share on other sites

38 minutes ago, MikeyD said:

I recall when they were first discussing adding triggers Steve turned down all suggestions where the scenario designer would be obliged to write code to get it to work.Ā 

This makes sence but hopefully he agrees with the fact that things like...

- more AI groups

- more objective/trigger slots

- reinforcements by trigger

- timed objectives

- a new trigger option for AI group casualtylevel higher then XXX

- A new timing option for waypoints that is based on tlme on location rather then the game clock...

- A remake of the dreadful/very unintuetive AI artillery program interface...

- A way to set some kind of value in zone limitation on triggers and objectives...

And a number of other suggestions that have been mentioned with regnard to the AI and objectives, triggers and reinforcement optionsĀ 

Would not require the scenario designer to 'write code'...but rather be pretty signifikant quality of lite improvements that would simplyfy scenariodesign and make them better...

A lot of frustration could be avoided imo šŸ˜Ž

Link to comment
Share on other sites

5 hours ago, MikeyD said:

I recall long ago Steve commenting on the topic of newĀ IU commands. He said individual proposed special commands may sound good but you've got to factor in 'player overload'. Clutter up the UI with too many new orders and players will start tuning them out, gameplay will become burdensome.

I agree with this passionately!

I would much rather see the AI do things automatically, so you map designers could set up a map with simpler orders, than have more and more detailed control.

Edited by Freyberg
Link to comment
Share on other sites

5 hours ago, Freyberg said:

I agree with this passionately!

I would much rather see the AI do things automatically, so you map designers could set up a map with simpler orders, than have more and more detailed control.

I guess we all would like to see that....but is it realistic ?

Remember BFC have struggled bigtime with such 'simple' things as geting the AI to chose the right door when entering and exiting buildings...the hedgerow-bug etc...

It would require a massive amount of work to improve the situational awareness and tactical skill of the AI i belive.

Compare this with what would be required to add some of the suggestions mentioned in this (and other threads)....

Suggestions that pretty much only would require some redesign of the editor UI and have pretty much no impact on game performance or put any higher demand on the current AI....

With the added benefit of actually making scenariodesign EASIER ! šŸ˜

Edited by RepsolCBR
Link to comment
Share on other sites

11 hours ago, MikeyD said:

Ā 

Even longtime scenario designers can be unaware of the AI's capability. It was just recently that we learned (to our great surprise) you can navigate through groups and orders using the keyboard arrow keys. How many scenarios have I made and I didn't have a clue about that! One of our merry band recently learned how the 'retreat' AI map marker works. Another recently learned about directing on-map mortars to fire outside of LOS using the AI. The CM AI isn't just the AI, it a collaboration between the AI and the scenario designer. You need an AI the designer can work with and the AI needs a designer that knows what he's doing. its a balancing act.

I actually belive that the demand on the scenariodesigner to know about ALL the 'tools' avaliable to him is higher if we only have a few tools (as of now)...compared to having a lager toolbox šŸ˜‰...

With a limited number of options avaliable it might be more important to take advantage of all of them when designing a scenario in order to produce good result...

The benefit with a lager toolbox is that you...use whar you need...you use what you need to solve the current problem...what you don't need you don't use...

The intention is not that you HAVE TO USE every option avaliable in every scenario...

You simply use what you need šŸ˜Š

How goes that oldĀ saying ?

"Better to have it and not need it then to need it and not have it"

Ā 

Link to comment
Share on other sites

On 6/21/2020 at 8:03 PM, RepsolCBR said:

reinforcements by trigger

That is something I have wanted for some time. The enemy enters an area, trigger line, and after a couple of minutes reinforcements arrive.

I think that option often would be better than having reinforcements arrive by how much time since the beginning of the game.

Reinforcements by a trigger shouldn't be allowed to arrive too early. Could be nice if there were different time spans to them arriving. The first could maybe be arriving within 15 to 30 minutes from when the triggering occured, the second time span 30 to 45 minutes from the time of the triggering and so on.

Link to comment
Share on other sites

  • 6 months later...

Time to wake this thread up with a question.

How is it with AI triggers and quick battles? Don't they have any function in quick battles?Ā 

I'm making a quick battle map and painted an "AI Trigger (enemy)" line and after that did a test in Author Mode with my troops reaching and thus activating that trigger just to see that the AI-group didn't react. I made my troops run over that trigger zone three times without the AI-group reacting.

If I'm doing this correct the order I give to the AI-group should be to wait for the trigger area to be activated before they move and I should also put a time span (Exit between... and) for movement which is a bit beyond the time when I suspect that the trigger area will be activated. This should give the AI the instruction to move the AI-group forward when the trigger area gets activated. But if the trigger area isn't activated before the time span (Exit between... and), which I have put as a reserve, the AI-group will start moving when that time span is reached.

Edited by BornGinger
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Unfortunately, your content contains terms that we do not allow. Please edit your content to remove the highlighted words below.
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...