So CMSF2 Quick Battles vs. them in other CM games

[kraze]

I've noticed one thing is that in all QBs that I play - AI will counter-attack the objective if it loses it with forces it decides to keep in "reserve" (usually at the starting point). Yeah I know it's all smoke and mirrors like triggers and stuff - but apparently this is something I've never noticed in CMSF1/CMBS/CMBfN.

So are there any plans for QB upgrade in Black Sea/Battle for Normandy/etc? I now find QBs truly enjoyable in CMSF2, almost like a typical "Battle" vs. AI minus the detailed briefing and more complicated scripting.

Edited September 3, 2020 by kraze

[Bulletpoint]

It's probably just that scenario designers have become better at using the tools available.

It's very easy to make the enemies counterattack towards the end of the battle, even if the counterattack just consists of all enemy forces suddenly starting to run towards the objective area. Also some QBs in CMBN did the same trick.

[A Canadian Cat]

@Bulletpoint is correct. @MarkEzra (and now @benpark too) have been getting better and better at creating AI for the QBs. What they usually do is any new content comes with QB maps that they create / add / modify up to their new standards. They don't go back and update older maps - as a rule. I'll let them talk about any exceptions to that. So, when a new module comes out for CMBS it will likely contain new QB maps and those will likely be made using the latest and greatest lessons from their previous work. I use the term "likely" because there is no official policy that I'm aware of and I'm not allowed make declarations but so far that is how it has been done since the beginning of CM2.

[benpark]

I think we are just getting better at remotely understanding Charles' and Steve's brains.

The 4.0 AI additions help. "Withdraw" and "Area Fire" are very useful. Just remember the plan is somewhat general, so it may not be beneficial to overdue any behavior, and I expect you will do well.

It helps to do as much AI testing as possible. It's always an unknown as to the forces, so consider that a given plan may be foot/wheeled/tracked. Lots of forest movement might get complicated for tanks, etc. I don't fight the engine, I try to adapt to it and see if you can come at your idea from another angle if stuck.

The AI testing covers the understanding of the likely behavior patterns vs. objective zone (even down to what shapes seem to work best) and on the testing of the QB under a variety of circumstances.

QB's do not use Triggers- They only use one objective type- "Occupy", if I am recalling correctly. Triggers don't work with those.

A lot is based around the map itself, which I grab from the larger maps I make. I'll look for an area that seems suited to any of the various types of QB- "meeting engagement" maps have different concerns than an "attack" one, and so on. I defer to the terrain to set the agenda, unless there is a reason to set the map in a specific area (historical area of fighting).

There are a bunch of boxes to check for creating these (friendly side, etc)- make sure you get all the settings right for each side. It's easy to miss a setting for Blue after spending a day from  the Red plan vantage- it's best to be methodical, and do the steps in an order that suits, but is consistent.

They take a lot of time to do by the dozens, but making a few initial ones can set you on the right track.

Edited September 4, 2020 by benpark

[kraze]

2 hours ago, benpark said:

QB's do not use Triggers- They only use one objective type- "Occupy", if I am recalling correctly. Triggers don't work with those.

This is interesting (if slightly removes the magic) - so it means that these are timed objectives as it was pre Red Thunder? Because I've played 5 QBs in CMSF2 thus far and somehow AI counterattacks just as I'm entering/occupying the objective... and then as I'm going for the next one - AI is attacking that one too.

Am I that exceptionally lucky with timing?

[benpark]

Probably good timing by Mark in the plan as well if CMSF2, and lots of testing.

Watching the flow of the QB play out in testing over and over helps. Scenario Author, Turn-Based. Repeat until the plan is working (adjusting timings and AI locations to suit).

The 4.0 AI additions are a big help with doing the AI, regardless of other controls. "Withdraw" alone keeps behaviors looking like something humans might do, which ranks well in my estimation for a "must have". Add in some "area fire", and it can be (somewhat- it's a QB, thus completely open) controlled chaos. And "Facing"- that is a huge one without having to do the odd work-rounds we once did.

Hopefully all of this work translates to effective plans in such an open system.

Edited September 4, 2020 by benpark