BletchleyGeek

For something this specific it would probably be easier to figure out which units were in the geographical area that you are looking at and then try to find out what vehicles were in those units.  The 233rd Reserve Panzer 'Division' was located at Horsens in Denmark from August 1943 until the end of the war.  It only had 34 tanks and was used as a training unit so I doubt that any Panthers were present in the unit, but you never can be sure without finding more details about the unit.

That sounds like a good concept. The map size (given the general consensus of 1k frontage above) might get a bit cramped. Defense will likely be spread out over maybe double to triple that. It's hard to average these things, with so many variables- as it is a "proof of concept", I'd say 1k attack frontage, 2.5k defense. How that is dealt with also becomes a critical point- the 1k frontage could be the map. Or the 2.5k, with something like a "static" attacker side force along on the other side of the 1k phase-line. Lots to consider.
I dealt with this same fundamental question in a different way for the Aachen campaign for FB. Campaigns have a lot of life in them, and stuff we haven't mined, I'm sure. If I remember for Aachen- I ended up splitting the battalion depicted into component companies which fought over differing areas of a single city. The last fight recombined the entirety of the battalion, where the consequence of each battle leading up to the last one would be felt cumulatively. This allowed smaller forces for most of the battles, and also (if the player didn't get too bloodied along the way) allows for a familiarity with the entire force both as separate entities, and as a coherent unit. That's all baked into the original concept- yours sounds like something people would be interested in (I am).
That campaign was for the game's release, and I was trying to be reactive to the "smaller scenario size" crew (I'm one), and as a dry-run for my patience for a Berlin map- you can be free of such constraints, but just be aware that you may lock some people out by going huge-huge.
Combatintman's post is good advice. What else is new?

Not on the commercial engines, but Graviteam's engine does.

TBH you guys lost me at 'algorithm'. 

That was an awesome post. Thanks for the background info and the links to references. My AI knowledge is from listening to and reading podcasts of interviews with technology inventors and practitioners and interacting with the group in my company that are working on integrating some machine learning into our product line. It is very interesting to see how things are progressing from the expert systems that I started working on using Fortran back in my university days. Also, very predictable that the latest work builds upon the previous work - duh
As a side note there is a clear difference in reading your well thought out charitable arguments on the subject. References certainly help solidify your posts.

That, plus flares, and we would have a great night battles sim.

Revamp the way direct fire mortars work. Currently, they either fire very fast for a full minute, wasting many shells. Or they fire very slowly (target light), which takes forever to knock out the target.
Instead, they could work like in real life: The mortar is given a target order; it then then fires a couple of rounds to get the range; then it rapidly fires three shells on target. The officer evaluates the effect and can then order another salvo. This time without the spotting rounds, as the mortar already has the range.
Alternatively, it could work like this: Giving a Target Brief order, the timer would only start to count down once the mortar was done aiming and finding range, and had begun the fire for effect phase.

Just some notes on the above, as a member of the research community, I feel I need to comment on this briefly.
First of all, that stance of "anything rulebased is anything but AI" is a disingenous position that does not hold when contrasted with the state-of-the-art literature. Domain knowledge can be expressed in many ways: with if-then statements, or behaviour trees, all the way to neural network architectures engineered to "capture" very specific processes and signals. The successes that have been widely hyped up - like the Star Craft/DOTA players by Deepmind and OpenAI - are fundamentally hybrids of what you refer with scorn as "AI" and machine learning. Even the Alpha players rely on not an insignificant amount of handcrafted knowledge, from the basic features used to parametrize states to the particular selection of activation units and interconnection patterns. All those choices were made by humans seeking the best combination of parameters, architectures, initialization strategies and more. If you check the paper on AlphaGo on Nature you'll see that the section devoted to explain those details is actually longer than the main paper.
 Even more interesting is to see how former preachers of the "end-to-end learning" gospel are now turning to classics like early 1980s subsumption-like architectures to bootstrap and guide those neural networks training process. Suffices to say that all major companies working on self-driving vehicles have abandoned that gospel and are scrambling to snatch leading researchers of areas which two years ago were considered to be "not relevant any more". 
I have no idea what is Palantir trying to pitch but it sounds to me as pure bull**** tbh. This is for several reasons: tactics require to deal with partial information, considering processes that flow at different time scales, on environments which are complex and dealing with a wide variety of platforms that operate autonomously (e.g. single riflemen, AFVs, a drone and its controller). At the contrary than in games like Go, where the number of pieces is fixed and known, and the board is always the same, a contemporary, near-future or past tactical environment shares little or none of those features. The most fundamental issue - there are several, and there's plenty of fundamental issues to choose and work on - to me is that neural models are not composable.
That is, you can work out a neural network to say, steer a simulated squad of simulated robots broken into two teams just all right along a given line of advance and against a specific amount and direction of enemy fires. Here is a list of the dimensions such neural network has to generalize in order to be useful and interesting:
- Initial distance to target (assuming that the order is to Assault)
- Effective volume of fires on enemy positions as distance to them changes.
- Type of terrain the unit maneuvers.
- Obstacles obscuring LOS and LOF
- Equipment of unit
- Hypothetical equipment of the enemy
There is absolutely zero guarantee that a given NN that performs at a certain level, for any meaningful performance index, on a finite sample along these directions will generalize to any possible combination of the above. If you have an algorithm for that which you can use on any problem at hand, then congratulations, you probably solved too Hilbert's 10th problem.  This applies to everything, including Starcraft: how many possible Starcraft maps there are? Can you classify all possible tactical and strategic situations neatly into discrete homogenous categories? That's also why doing funny stuff to allegedly state of the art CV pre trained networks - like adding a 1-pixel wide black border to an image - catastrophically degrades the accuracy of object identification. Luckily, other than perhaps Russia and China I think, nobody even considers to deploy deep learning systems for target identification and acquisition. If somebody does, they're criminally insane or selling snake oil, or both.
Provably you haven't done any of the above, but you may have a quite decent closed loop control strategy that works well enough to make some nice videos to impress people, or even beats some hand coded controller that somebody put a decent amount of effort in designing exploiting knowledge about the laws of Physics or some other fundamental process.
That can be good enough, it all depends what you're comparing it against. Definitely you can't make any guarantees on suitability for any purpose other than that captured by your training set: YMMV.
The problem of composability is illustrated by the following question: can I use that neural network as a building block to coordinate the movements of a platoon? The answer, so far, has been a quite deafening no.
There is no known way to constrain back propagation to guarantee that the knowledge acquired by the neural network you are using as a building block is going to be obliterated or changed in a fundamental and undesirable fashion during training for the "composite" problem.
Composability also challenges the ability to train incrementally, as the capabilities of the unit change due to casualties or changes in equipment. There's again no guarantee that any knowledge will be preserved when re-training after changing those elements in the environment that generates the training data for the neural network.
Last, composability has to do with time: what is the minimum period of time to be considered? Is there a sensible upper bound on the number of such consecutive periods of time to consider? Taking off-the-shelf techniques used for Natural Language Processing has been shown to be pretty much like dancing about architecture, spoken and written word has a very definite temporal structure, for which we know its "laws" (because we invented grammar and rules of style!).
Another fundamental problem linked to this last observation is that whatever the neural network learns we cannot be sure that it is capturing the essential first principles that allow the behaviours which are to be mimicked. This is analogous to the fundamental issue with the classic research by T. N. Dupuy and the HERO institute - in the 1960s, one could overfit a model only by hand, in the 2020s you can use neural networks too!
Contemporary machine learning has a niche, like those "rule-based" approaches you disparaged in your post do. And I certainly appreciate the good things in deep learning, for instance, the dependability and efficiency, provided that the right conditions for the techniques involved to work properly are an invariant of the set of situations I need to deploy them.

Going back to the games briefly. Regarding Graviteam, I learnt through a weird interaction with Andrey on the Steam forums a few months ago that he's pretty ignorant on any of these topics. Which is totally all right, he's not expected to be an expert on that. So my educated guess is that what you see animating those pixel truppen in Graviteam games are not too different from the techniques used in 99% of video games and 80% (?) of robotics: good old hand-designed controllers via behaviour trees, A*/D* and PIDs/SQP/Non Linear Programming.
Last, I want to address the comment which I read is blasting BFC (and video game developers in general) because of not using deep learning technologies. I have zero idea of what is the operating budget of BFC, but say, the cost in $$$ to say develop and train an Alpha-like system for one of the countless drills possible in CMx2 would probably be somewhere north of 1 million USDs (that counts salaries, on boarding of staff and compute for like 40 days with a similar amount of computation power as the one wielded by Deepmind to ensure you can beat Bil Hardenberger like 90% of the time). Indeed, they would probably amortize salaries and onboarding over time, but the cost of computation is what it is, and changes in the game mechanics, or even bug fixes, etc. would require retraining (or training new networks for that special case).
Indeed there are opportunities for more modest applications rather than end-to-end tactical battle management, but I am skeptical than they are cost effective for the return on investment Battlefront will get. I am pretty sure they're already doing this pretty much for the sake of the arts, and unless they get patronage, I can't see why should they spend tens of thousands of dollars per month on EC2 just to replace their code for animations, drills etc by neural networks. Or maybe you could work pro bono for Battlefront developing those
 

https://spark.adobe.com/page/Sk6iKJAIvCQ1H/

Hm alright that could be an explanation for it.

No causality implied but also good to keep in mind that the affected object is the most numerous flavor object on the map.
Was your map "just" large or did it also feature a ton of flavor objects either of a specific given type or in overall quantity?

https://www.bbc.co.uk/news/resources/idt-sh/Iraq_legacy_of_war
In 2007 the BBC’s Mark Urban was “embedded” with a platoon of soldiers in one of Baghdad’s most violent areas.
Ten years later he tracked down four of the men from the unit.
What effect did the war have on them?

I just finished “Player of Games“. Awesome, indeed! And a very unusual Science Fiction. Had it‘s length, but also many fantastic twists and turns. The only thing I missed, was a more detailed description of Azad. Hm, that’s probably too much of a gamers view.
Looking forward to “Use of Weapons” and thereafter, I might grab an Ian Banks audio book for free. But then it’ll be enough Ian Banks for a while.

Who knows, maybe there's a huge untapped market of Brazilians wanting to play Brazilians in a computer wargame.

Since, at this point of my life, I finally start to consider playing video games a waste of time, while, at the same time, I think that making wargames is totally awesome, this thread makes me particularly happy.
Best regards
Thomm

That, and similar occurrences, have knocked that game dead in the water for me.
Great post by @Thewood1 on those two white elephant games... and @Mord's coda was quite fun too.

You could say the same thing about Norm Kroger and Jim Rose.  They made some great games for their rimes.  But after Distant Guns and Jutland, no one will touch them.  And just like them, Scott will be remembered for POA2.
btw, The Tigers on the Prowl and Panthers in the Shadows engine has as much to do with POA2 as CM1 did with CM2.  It was completely rebuilt and only some of the most basic concepts carried over.

Good summary @Thewood1 - but here I would say that they put their mind to it, but their mind goes to places ours does not. I can't read their Russian-speaking forums, I would love to hear if the feeling is similar across the language divide. Maybe  @DMS can chime in?

Yeah, it probably should be in the General Discussion forum.  But let's face it, it is probably a good advertisement for them.  The only games getting significant mention are:
A game that has been rehashed and relaunched a dozen time (Close Combat)
A game that, at its heart is a military tank procedural trainer (Steel Beasts)
A game that ended up being a rip off after 16 years of development (POA2)
A game built by a Ukrainian team that could have the potential if they put their mind to it, but most likely won't. (Graviteam)
and a few odds and ends.
And none of them really compare directly, feature for feature.

I thought you liked seeing heads roll?

All I can think of is Lee looking for Stuart at Gettysburg.  There is a point when a game becomes more like work than fun.  Friendly FOW to the extreme that the entire playing experience is about getting info from your peer and subordinate units starts to feel like work.

Ummm....don't think your going to get an in-depth discussion of Battlefront competitors on thee Battlefront FORUM.  You should ask your question on a non-publisher forum.

My number one wish is to have working and realistic pathfinding routines or at least being able to give "follow the road" order to a group of vehicles. 

In my current game I've received a company size tank reinforcements. Now I have to give a detailed movement orders to every single vehicle, just to make sure that they arrive ok to the frontline. It's  completely mundane, boring and unproductive task. It often makes me quit the scenario that I was excited about, because I feel that I'm better off reading a book or watching a series than repeating over and over this exercise.

Was the ability to control individual soldiers in a FPS mode already mentioned?
Gotta think out of the box for CM3!
That, and Fulda Gap.
Best regards,
Thomm