Jump to content

Need odd LOS issue looked at


Lt Bull

Recommended Posts

10 hours ago, Lt Bull said:

FWIW, in my "EVADE testing of random units during orders phase of various turns" in the scenario I am playing (the one where  the friendly directions are apparently 90deg out!) I have noticed that units in a particular building will EVADE in a different direction to the units located around the building. All the units around the building will tend to EVADE NW while the units in the building will instead EVADE SW.  I haven't really tested this thoroughly with other buildings etc, it's just an initial consistent observation.

I am surprised that you considered "friendly map edges" more in terms of potential issues with range and small calibre off-map arty than as a foundation to understand how you would expect units will react when evading/retreating/routing etc, especially in the context of understanding the EVADE mechanics you have been investigating.

I can tell you that when I edited the Allied/Axis Friendly Direction setting in the scenario editor and repeated the EVADE order testing on the units I had put on the map, they definitely did start EVADING in the newly set direction.  Quite easy to test.  Try it out. Was there a reason why you seem to have thought that the Friendly Direction setting may not have had an effect on influencing the EVADE order waypoint?

I haven't gone checking a bunch of scenarios myself (yet) but what you are saying certainly is not what I would expect to find (even when I consider all the scenarios I recall playing).  I would expect most scenarios to be designed around forces engaging north-south or east-west rather than diagonally across the rectangular map, and as a consequence the Friendly Directions set as N/S or E/W respectively.

Can you elaborate more on the reasons you have found to avoid scenarios with N/S and E/W map edge settings/confrontations as opposed to the scenarios based on diagonal set friendly directions?  This is most curious.

If a scenario has been set up so the main battle lines and deployments are N/S or E/W, I can only imagine that if the Friendly Directions are instead set at NW/SE or NE/SW then it should be expected that it would be more likely to experience "illogical" EVADE and/or retreat behaviour (liek we are discussing) as units would tend to "retreat" diagonally rather than directly away from the N/S or E/W front line.

I just made the obvious test. Reversing friendly and enemy map edges in my vs AIP test game. And see..... I was wrong. In fact the map edges do matter quite considerably for evade direction, although TacAI selected waypoints remain largely FUBAR. I likely had too many the odd evade/retreat towards the enemy occurrences to assume the map edges do not work generally. :wacko: Maybe too much focus on buildings ATM. Likely got to test other combat and terrain situations more, although buildings remain a most important part in my mission designs, for the AIP quite in particular. Fact remains that evade seeks to interrupt LOS/LOF from a particular enemy unit and too little taking into consideration other important matters like exposure (to other enemies), cover terrain and shortness of move to waypoint. Map edge setting is only part of the problem, considering you can actually make the known enemy frontline the friendly map edge when from the AIP and TacAI situational awareness it actually should know better. Need to mention I play in iron mode only, possibly putting my forces situational awareness to the minimum maybe. Just another guess of mine, but who knows.

Link to comment
Share on other sites

1 hour ago, RockinHarry said:

I just made the obvious test. Reversing friendly and enemy map edges in my vs AIP test game. And see..... I was wrong. In fact the map edges do matter quite considerably for evade direction, although TacAI selected waypoints remain largely FUBAR. I likely had too many the odd evade/retreat towards the enemy occurrences to assume the map edges do not work generally. :wacko: Maybe too much focus on buildings ATM. Likely got to test other combat and terrain situations more, although buildings remain a most important part in my mission designs, for the AIP quite in particular. Fact remains that evade seeks to interrupt LOS/LOF from a particular enemy unit and too little taking into consideration other important matters like exposure (to other enemies), cover terrain and shortness of move to waypoint. Map edge setting is only part of the problem, considering you can actually make the known enemy frontline the friendly map edge when from the AIP and TacAI situational awareness it actually should know better. Need to mention I play in iron mode only, possibly putting my forces situational awareness to the minimum maybe. Just another guess of mine, but who knows.

That's good info. Thanks, RockinHarry.

So, the Engine is taking into account faction-related map edge assignment in general and applying it to rout/evade. However, the undesirable suicide evade is resulting from the TacAI at the micro level being unable to properly assess its immediate environment in relation to current threats. So, for example, it doesn't value nearby building cover adequately, or, similarly, will take a deadly path, rather than a safe path, due to some unfortunate weighing of the criteria it uses for path choice.

Two first thoughts on this:

One: the problem doesn't happen all the time, which means the TacAI decision process must have something good in there. So, eliminate the wheat from the chaff? Also, although we haven't determined if it's really true or not, if higher-quality units are less prone to suicide rout, maybe find out why and apply that to all units. Then penalize lower-quality units in a different way. A good option I mentioned before might be extending the "cool off" period from the panic, so that it takes longer for lower-quality units to accept new orders. In other words, they rout/evade safely, but stay in their new safe spot longer - like bailed out AFV crews often do.

Second thought:

In terms of CM being what it should be (the best, that is), units really need to have better situational awareness regarding buildings/cover and how to use it versus threats. That sounds like a beefy coding job, but I think the game needs to go there, at least eventually.

2 hours ago, RockinHarry said:

Think once the next patch got released we should start a dedicated "Mission design and how to set up an AI player " thread sort of. There´s quite a lot of interesting things to share and learn alike.B) 

I was planning to start a "QB AI Plan Design: Best Practices" thread when I get to that point with my map, as I want to offer a number of QB slices with it. I'm still hoping for some surprise Engine improvements before then, since my map won't be finished until after the CMRT module is released. Maybe not looking so good on that, but there is always hope.

The current state of QB AI Plan functionality is a bit of a head-scratcher. We know from Steve that most players play CM solo. That means the QB system must be heavily used by players when they can't find a scenario/campaign they want to play. Yet, so many obvious design tools are still missing that would allow designers to really enhance the solo QB experience.

For example, I'd really like to be able to funnel specific unit types into specific groups in a QB AI Plan. That seems extremely basic to me -- at least this many years in.

But yeah, it would be good to have some fresh threads for both Scenario and QB AI Plan design. Can't get into that yet, though. No time right now.

Edited by Macisle
Link to comment
Share on other sites

3 hours ago, RockinHarry said:

I was wrong. In fact the map edges do matter quite considerably for evade direction....

Yes, and I hope no one is too surprised to see that this actually is the case.  It shows at least the manual EVADE waypoint placement mechanics are working as expected/as they should ie. EVADE direction influenced by the scenario defined parameter "Friendly Direction".  It should bring some relief to know (not the least Battlefront!) at least this aspect of the game mechanics is working as advertised.

Importantly, I have just heard back from the scenario designer of "A Nasty Surprise". He thanked me for pointing it out. It appears that he was oblivious to this scenario setting and had never set it in the first place, so the scenario retained whatever default setting it happened to be (which is Allies west, Axis east).

So my casual observations of "odd" behaviour did not deceive me. I do feel somewhat vindicated now for originally even suggesting that "something may not be quite right" with the way the unit in the video behaved when the TacAI made it evade "towards the enemy". Under similar circumstances, I would feel confident that the infantry unit in question would NOT have behaved in the same "suicidal" way had the scenario defined Allied Friendly Direction instead been set to north, rather than west as it was.

3 hours ago, RockinHarry said:

... although TacAI selected waypoints remain largely FUBAR. I likely had too many the odd evade/retreat towards the enemy occurrences to assume the map edges do not work generally. 

RockinHarry, in saying this, are you of the belief that the waypoints assigned by the TacAI during "auto-evade" are calculated differently to the waypoints assigned when a player presses the evade button ("manual-evade")? As suggested before, I would think they are essentially one of the same same thing, with the only superficial difference being in whether the evade routine was being initiated by the player directly (by pressing the EVADE order button) or by the TacAI automatically (during RT or mid-WEGO replay action).  Do you have any examples that might support your view?

3 hours ago, RockinHarry said:

I likely had too many the odd evade/retreat towards the enemy occurrences to assume the map edges do not work generally. :wacko: Maybe too much focus on buildings ATM. Likely got to test other combat and terrain situations more, although buildings remain a most important part in my mission designs, for the AIP quite in particular. Fact remains that evade seeks to interrupt LOS/LOF from a particular enemy unit and too little taking into consideration other important matters like exposure (to other enemies), cover terrain and shortness of move to waypoint. Map edge setting is only part of the problem, considering you can actually make the known enemy frontline the friendly map edge when from the AIP and TacAI situational awareness it actually should know better. Need to mention I play in iron mode only, possibly putting my forces situational awareness to the minimum maybe. Just another guess of mine, but who knows.

Given you had indicated that you had never really considered the Friendly Direction settings with respect to evade direction, it now makes me wonder what the Friendly Direction settings actually were in all the cases you have mentioned (and for that matter, ever other case that has been discussed).

I would not be so presumptuous to even think that the determination of the evade waypoint by in CM (under any circumstances) involves what I can only imagine would have to be very complex coding based on battlefield geometry, LOS/LOF interrupt calculations, unit SA/FoW considerations etc even though interrupting LOS/LOF and reducing exposure IS a core objective of the evade behaviour in the first place.  If there is an expectation that the game does (or can) explicitly consider LOS/LOF when determining evade waypoints, then I can understand the expectations one may have on HOW they expect a unit to evade.  I don't think we really should be expecting too much "intelligence" based on the "local" LOF/LOS, battlefield geometry etc. A simple "calculate evade waypoint" algorithm that works through a list of possible alternatives should be sufficient enough to satisfy in most case our need for a "realistic"/"intelligent" waypoint solution in the majority of circumstances.  I would be happy to conceded that perhaps just the relative direction of "known" enemy units may play some part in perhaps tweaking the waypoint determination in secondary way.

In thinking about all the possible ways and factors the TacAI coding might consider in it's calculation to determine a suitable "destination" action space for the "evade" behaviour of any given "rattled" infantry unit, I realised that I had never considered one very simple "solution" option: the location of nearby unaffected infantry units.  For example, in the instance I highlighted in the video, you can see in the side view screenshots that 20m to the rear left of the infantry unit in question is the 1st Plt HQ/G Co located in a safe position in total defilade. By virtue of the fact that a friendly unit was actually at a location that was:

  1. not under fire
  2. out of enemy LOS
  3. further away from known enemy locations/source of enemy fire/enemy map edge
  4. not too far away
  5. accessible via a path that would not increase the units current exposure to enemy fire (probably the hardest thing to calculate/code for),

by definition this identifies/qualifies that action square as a totally valid/legitimate destination for the infantry unit to 'evade" to. Essentially most (if not all) of the checkboxes boxes for "good choice for an evade destination" are ticked. It certainly is much safer destination to move to than where it actually did move to!

So rather than just simply relying on dealing with/modifying existing complex coding algorithms (involving actual analysis considering actual local terrain geometry, LOS/LOF scenarios from various enemy units and all the complexities related to FoW, as well as other factors like other local incoming fire etc) to calculate/determine (from essentially first principles) a suitable "safe" destination for an "evading" rattled TacAI driven infantry unit to move to, perhaps if the algorithm also considered the action squares occupied by nearby unaffected friendly infantry units as valid qualifying destinations, and compared it against other alternatives using a quick "checklist" like the 5 points I listed above,  perhaps many of the complex "poorly calculated" outcomes chosen by the game as "a safe destination" (like we are seeing) could be avoided. If such a thing was considered in the case of A Team/1st Sqd/1st Plt/G Co, no one would have an issue with the TacAI choosing the action square of the the nearby 1st Plt HQ/G Co unit as an "evade" move destination.  At least we know that the action square was (is) safe.

I know this easy, quick and dirty "cheats" way of circumventing complex algorithms to determine a "safe nearby action square" as a possible destination for evading infantry units may not be applicable in all situations, primarily in instances where there are no nearby friendly units, or that it somehow may not even provide "the best" destination in all cases, but I think it would help in a vast majority of the situations where the existing TacAI would instead pick questionable destinations like we are seeing.

Edited by Lt Bull
Link to comment
Share on other sites

PS: I've gone looking for the threads I read months ago (and indirectly referenced in this thread) about the "evade towards enemy through gaps in hedgerow" issue.  I remember seeing a video of some German infantry breaking cover and evade directly across open ground towards the enemy fire and another of US infantry lined up along hedgerow with a gap in it.  If you know what I am talking about, please provide me the links as I can't seem to find them.

Link to comment
Share on other sites

On a positive note about the protection factor of complex modular buildings...

I had time for a couple of turns today in my playtest. I'm still working on the one that I posted the suicide rout example shots from. Well, the northern threat was beaten into hiding, so I sent in a tank to deal with the east alley threat. Due to LOS, it had to target the ground in front of the enemy wall, rather than the wall itself. However, in my experience, that can give good results, too. -Sometimes better, as the HE splash can knock out walls across multiple adjacent building sections, along with perhaps causing casualties. Anyway, I put in a few shots and knocked out their wall.

CMRT_092219_01.thumb.jpg.2dbebc90bf20e4306b952cbc8d5be4f3.jpg

 

Here's what happened as soon as the smoke cleared:

CMRT_092219_02.thumb.jpg.d6521ccb884fb8ed667e61cfef876cfe.jpg

No survivors.

Link to comment
Share on other sites

6 hours ago, Sgt.Squarehead said:

Man I want that map!  B)

I just finished the core building all the way through. Now, I'm in the polishing phase. I need to fine tune the elevations and knock out internal walls in a lot of buildings I did before I caught on to the technique. Then, it's on to filling out the flavor objects, working on a texture pack and getting the scenario/campaign/QB slice work done. I need the CMRT module for textures and flavor objects, though.  Oh, and I've got some funky church constructions in the German alamo area that I need to test for pathfinding a bit more.

I actually could start having folks battle test most areas now. The main problem there is that I've got a texture glitch that seems to block using more than eight textures unless you have my exact mod folder installed and I don't want testers seeing a texture vomit. Never could figure out why. Probably somthing incredibly simple. In addition to scavenging the new module for textures, I'm hoping installing it will fix the texture weirdness since I think it will expand the stock number to more than eight.

The last area of the map I built was the most dense area and it turned out to have some really funky constructions. Lots of variety. There is what looks like a university campus that yielded some pretty neat stuff.

I am so, so, soooo looking forward to having folks go weapons free on ma' baby! 🙂

Link to comment
Share on other sites

On 9/22/2019 at 7:10 AM, Lt Bull said:

Yes, and I hope no one is too surprised to see that this actually is the case.  It shows at least the manual EVADE waypoint placement mechanics are working as expected/as they should ie. EVADE direction influenced by the scenario defined parameter "Friendly Direction".  It should bring some relief to know (not the least Battlefront!) at least this aspect of the game mechanics is working as advertised.

IIRC map edges weren´t always working as advertised as I can´t otherwise explain my previous conviction they do not! It´s years back when I went through all of that testing and I just can guess it either got fixed or implemented with some the patches released between ~2015 and present. Odd. :unsure:

On 9/22/2019 at 7:10 AM, Lt Bull said:

RockinHarry, in saying this, are you of the belief that the waypoints assigned by the TacAI during "auto-evade" are calculated differently to the waypoints assigned when a player presses the evade button ("manual-evade")? As suggested before, I would think they are essentially one of the same same thing, with the only superficial difference being in whether the evade routine was being initiated by the player directly (by pressing the EVADE order button) or by the TacAI automatically (during RT or mid-WEGO replay action).  Do you have any examples that might support your view?

If circumstances remain the same generally then auto evade and player initiated are the same. You can also initiate direct command evade without having any enemy forces (and threats) on the map btw. I´ve various such test missions, mainly for friendly player map movement and pathing tests. So don´t need enemy forces on map then. Guess the evade command then works in relation to last time used face command of concerned unit. Evading from a map spot specified by face command so to say. Got to test that some more.

My regular, normal morale and in C2 german infantry realiably triggers self evade when morale state becomes nervous or cautious and suppression meter going up to full tilt (= pinned). When in buildings, they fast move down to basement, otherwise leaving the building, moving diagonally 4-5 AS (30m) toward the friendly map edge at average. But never into another building and always in very bad places with no cover in AS (unless wall or hedges maybe). What counts as said is breaking LOS/LOF to the threatening enemy unit so cover is any cover terrain placed between that enemy unit and target evade AS. 

Just could repeat the gapped bocage bug again. With map edges set to friendly = S and enemy = N I had an evading german team moving toward enemy map edge (N) on enemy side of gapped low bocage. That german team was already behind low bocage (starting position when evade kicked in) and to reach the new intended place it first got to cross another gapped bocage horizontally and then forward another, final one. 2 AS altogether.Evade alternative would´ve either been a stone building straight to the rear (1 AS away) or better yet, remain in place heads down. On one occasion that same german team moved exactly in that gapped bocage AS where in next turn an enemy mortar shell came down. If I hadn´t relocated the unfinished evade waypoint in orders phase they´d all got killed instantly. So I could make them avoid that direct mortar shell hit, but nonetheless they didn´t make it to safety as further mortar rounds coming down nearby killed all of them anyway. FUBAR. :P Glad it was just a test mission, as otherwise I´d not taken it as that funny.

Link to comment
Share on other sites

On 9/22/2019 at 4:52 AM, Macisle said:

In terms of CM being what it should be (the best, that is), units really need to have better situational awareness regarding buildings/cover and how to use it versus threats. That sounds like a beefy coding job, but I think the game needs to go there, at least eventually.

On 9/22/2019 at 2:45 AM, RockinHarry said:

Think once the next patch got released we should start a dedicated "Mission design and how to set up an AI player " thread sort of. There´s quite a lot of interesting things to share and learn alike.B) 

I was planning to start a "QB AI Plan Design: Best Practices" thread when I get to that point with my map, as I want to offer a number of QB slices with it. I'm still hoping for some surprise Engine improvements before then, since my map won't be finished until after the CMRT module is released. Maybe not looking so good on that, but there is always hope.

The current state of QB AI Plan functionality is a bit of a head-scratcher. We know from Steve that most players play CM solo. That means the QB system must be heavily used by players when they can't find a scenario/campaign they want to play. Yet, so many obvious design tools are still missing that would allow designers to really enhance the solo QB experience.

For example, I'd really like to be able to funnel specific unit types into specific groups in a QB AI Plan. That seems extremely basic to me -- at least this many years in.

But yeah, it would be good to have some fresh threads for both Scenario and QB AI Plan design. Can't get into that yet, though. No time right now.

Think the main issue remains the "speed vs. cover" dependency as evade or rout mostly uses fast move mode. Any slow going or bottleneck terrain (doors!) remains neglected then. The other thing is the internal definition of "cover". Most efficient cover is terrain placed between concerned friendly unit and any enemy weaponry. Highest order "cover" terrain would be terrain mesh (ditches, folds. crests, depressions, craters etc), or "slopes" beeing the BFC term for it. Secondary then is walls, buildings (behind cover) etc. Considering the thin skin nature of buildings, particularly vs smallarms at 300m or less, it´s no wonder the TacAI considers them as of little cover value, when the enemy is near so they preferably do not go in and rather around. Also each pixeltrooper to his own judgement which poses quite a serious problem for all pathfinding in the game. Any cohesion of units is sacrificed for the sake of.... I don´t know. For same reason it´s extremely hard to get an AIP maneuvering plan to work in effective ways. There´s workarounds though (thoughtful group assignments, timings and triggers) and some the V4 new AIP orders (facing, retreat) help some, but it´s still all limited.

And yes, best to wait for what next patches bring on fixes and improvements, before starting a (new) solo mission design thread.

Link to comment
Share on other sites

7 hours ago, RockinHarry said:

...IIRC map edges weren´t always working as advertised as I can´t otherwise explain my previous conviction they do not! It´s years back when I went through all of that testing and I just can guess it either got fixed or implemented with some the patches released between ~2015 and present. Odd. :unsure:...

It's quite possible that something changed/got fixed with the recent patches. For example, prior to the patches, AI infantry would generally not set up in buildings, unless the painted action spots were actually limited to building tiles. That's why SP QB infantry was always found outside hugging walls and hedges. Post-patch, you can paint a general area on an urban map and have most/all of the AI infantry set up in buildings.

My experience of the current AI setup on my urban map is that it tends to leave perimeter defense gaps that a human player would not, but is actually quite good at finding juicy ambush spots. Pretty frequently, it finds devilishly clever spots that are hard to gain fire superiority on.

As I mentioned before, the AI is still crap at positioning tanks in a mixed group (again, assuming an urban block environment), but if the designer sets up good tank positions with the tanks in a separate group, he can paint whole blocks with an infantry+ATG group and they seem to support the tank positions pretty well, and sometimes very well.

So, even though it didn't get much attention, there was a pretty massive fix for AI infantry set up in the patches. Maybe something happened with map edges, too.

Link to comment
Share on other sites

'Nother suicide rout example. This time, multiple units. First up, the highlighted team on the second of two levels panics, due to fire from an infantry team, marked "Threat." There is another infantry team out-of-pic in the same NE diagonal direction about 70-100 meters away. Both enemy infantry teams are firing rifle grenades, along with small arms fire (and a grenade from the marked team). As with the previous example, which was this area a few turns ago, the picture top edge is N and bottom S. N is enemy edge and S friendly.

CMRT_suicide_example2_1.thumb.jpg.1682cec8b86013d29af51f21c6935fc8.jpg

 

Rather than simply dropping back an action spot S into safety, the panicking team runs downstairs and exits NE via the side door, running directly towards the Threat unit. One man is lost exiting and the other is gunned down in the street by the Threat unit.

CMRT_suicide_example2_2.thumb.jpg.5385e70c2b57aad96cd4691a4e90be2d.jpg

 

Next up, we have the point man, who has drawn fire and is under orders to stay put and hide.

CMRT_suicide_example2_7.thumb.jpg.187afb4796f6276310d11c528b56b3cb.jpg

 

He decides to boogie and moves SW. Better, but closest safety is directly S into the factory two action spots away.

CMRT_suicide_example2_9.thumb.jpg.fef0488471a8832eca7e5ce0fbe9fbf1.jpg

 

He turns rounds the corner and moves N a few meters. This is unscouted territory. Luckily, he does not draw fire.

CMRT_suicide_example2_10.thumb.jpg.96cd8234e90ce2e7f42bf37d50940dd2.jpg

 

During the same minute, another team has freaked out.

CMRT_suicide_example2_3.thumb.jpg.b613924e477ddcf88d126eb1fa197c83.jpg

 

Rather than move to one of the safe adjacent action spots, they exit the building, some men going N and some S. One of the men exiting N is gunned down.

CMRT_suicide_example2_4.thumb.jpg.64f7c93508f1e4f3bd8b4297f815d21d.jpg

 

Rather than entering the safe S block where the company and battalion HQ are, they follow the scout, round the corner and head N unto unscouted country.

CMRT_suicide_example2_5.thumb.jpg.e4b4eb49fc292033f8997abe63f50660.jpg

 

Whereupon, the unit size draws enemy attention and a previously unknown MG42 covering the street opens up, promptly dropping the team leader. Now, at the end of the turn, they are in quite a bad spot.

CMRT_suicide_example2_6.thumb.jpg.a0961f3c8a42f5fe7c2c314329c4518c.jpg

Edited by Macisle
Link to comment
Share on other sites

5 hours ago, Macisle said:

'Nother suicide rout example. This time, multiple units. First up, the highlighted team on the second of two levels panics, due to fire from an infantry team, marked "Threat." There is another infantry team out-of-pic in the same NE diagonal direction about 70-100 meters away. Both enemy infantry teams are firing rifle grenades, along with small arms fire (and a grenade from the marked team). As with the previous example, which was this area a few turns ago, the picture top edge is N and bottom S. N is enemy edge and S friendly.

CMRT_suicide_example2_1.thumb.jpg.1682cec8b86013d29af51f21c6935fc8.jpg

 

Rather than simply dropping back an action spot S into safety, the panicking team runs downstairs and exits NE via the side door, running directly towards the Threat unit. One man is lost exiting and the other is gunned down in the street by the Threat unit.

CMRT_suicide_example2_2.thumb.jpg.5385e70c2b57aad96cd4691a4e90be2d.jpg

 

Next up, we have the point man, who has drawn fire and is under orders to stay put and hide.

CMRT_suicide_example2_7.thumb.jpg.187afb4796f6276310d11c528b56b3cb.jpg

 

He decides to boogie and moves SW. Better, but closest safety is directly S into the factory two action spots away.

CMRT_suicide_example2_9.thumb.jpg.fef0488471a8832eca7e5ce0fbe9fbf1.jpg

 

He turns rounds the corner and moves N a few meters. This is unscouted territory. Luckily, he does not draw fire.

CMRT_suicide_example2_10.thumb.jpg.96cd8234e90ce2e7f42bf37d50940dd2.jpg

 

During the same minute, another team has freaked out.

CMRT_suicide_example2_3.thumb.jpg.b613924e477ddcf88d126eb1fa197c83.jpg

 

Rather than move to one of the safe adjacent action spots, they exit the building, some men going N and some S. One of the men exiting N is gunned down.

CMRT_suicide_example2_4.thumb.jpg.64f7c93508f1e4f3bd8b4297f815d21d.jpg

 

Rather than entering the safe S block where the company and battalion HQ are, they follow the scout, round the corner and head N unto unscouted country.

CMRT_suicide_example2_5.thumb.jpg.e4b4eb49fc292033f8997abe63f50660.jpg

 

Whereupon, the unit size draws enemy attention and a previously unknown MG42 covering the street opens up, promptly dropping the team leader. Now, at the end of the turn, they are in quite a bad spot.

CMRT_suicide_example2_6.thumb.jpg.a0961f3c8a42f5fe7c2c314329c4518c.jpg

That pretty much confirms my basic assumption on how things appear to work with evade and retreating units, particularly in urban environments.

1. Buildings are bad cover at close ranges, below 300m. Walls are penetrable by most small arms, particularly ones with higher V0 like rifles and MG´s. To make matters worse, ceilings as well are paper thin and the more abstracted openings (windows, doors, any sort of invisible internal damage) allow hitting and killing troops from odd locations. Like i.e through windows at ground level, up through ceilings (or maybe staircases as well) at next level up. An apparent height advantage is more or less neglected this way.

Adding to problems is needless bunching up of soldiers at single windows, or better said, abstracted regions around them.  Soldiers fight among themselves to get to any window in order to snap a shot at visible enemies. Working with small teams (lowering unit density) doesn´t really help much as they´d bunch up at windows at same opportunities. Here´s also what makes enemy rifle grenades such a great "window cleaner". One hit at or near the window = multiple guys down. One can use it as tactic or cheat, however one likes to take.

The TacAI is aware of the building cover situation but likely rather through quick accumulation of suppression points, not from knowing the building providing actual little cover.

2. Selected move mode for evade and retreat requires to avoid mentioned bottlenecks and slow going terrain (gapped bocage beeing the bugged exception). So obviously internal movements within buildings are treated similarly, like moving through doors from in- or outside. When the TacAI decides for the destination waypoint it can just consider quickest possible pathing towards it, not considering any possible threats on the way. While map edges seem considered for general direction of evade waypoint, it´s still the maximum speed pathing that is weighted the most.

These are the main issues, but I see indications for other variables like loss of C2 to friendlies and HQ beeing weighted in as well. Whether game mode (warrior, iron...) has an influence I can´t judge yet. I tend to believe it does not, particularly for the AIP which appears to be given certain advantages anyway.

In my own test mission I deleted the troublesome gapped bocage and replaced with straight hedge. The infantry team (firing team from a split german squad) that formerly liked to evade move beyond into enemy LOS/LOF now stops in front of it. This team had just verbal contact to both its Plt HQ and squad leader split team in a building two AS to its left. The altered evade move WP puts the team now in both visual and verbal C2 (like it did before with gapped bocage) so I assume that seeking (better) C2 to superior is another weighting factor here. However the evade moving team could´ve achieved the same by moving rearwards into a building just 1 AS away. Instead it moves parallel to frontline 3 AS through rather open terrain with all following consequences.

If I´d had opportunity to interrupt the evade move while it starts I had changed both WP and move mode (slow) back to the building, then "hide" and recover. Usually in next orders phase it´s mostly too late to interrupt and anyway the bad state of the unit makes it plot a new (or the same) evade move instantly. So better let go and just change move mode to slow maybe. There´s some tactics to help that unit recovering and remaining low, but they´re not so often applicable (own forces deployment, providing covering fires from places away from the evading unit). Extremely difficult in MOUT, particularly with regard to little cover buildings and mentioned bad deployment of soldiers within buildings (bunching up, wrong guys at the desired windows etc).

So far, so (not so) good.

Link to comment
Share on other sites

15 hours ago, Macisle said:

It's quite possible that something changed/got fixed with the recent patches. For example, prior to the patches, AI infantry would generally not set up in buildings, unless the painted action spots were actually limited to building tiles. That's why SP QB infantry was always found outside hugging walls and hedges. Post-patch, you can paint a general area on an urban map and have most/all of the AI infantry set up in buildings.

My experience of the current AI setup on my urban map is that it tends to leave perimeter defense gaps that a human player would not, but is actually quite good at finding juicy ambush spots. Pretty frequently, it finds devilishly clever spots that are hard to gain fire superiority on.

As I mentioned before, the AI is still crap at positioning tanks in a mixed group (again, assuming an urban block environment), but if the designer sets up good tank positions with the tanks in a separate group, he can paint whole blocks with an infantry+ATG group and they seem to support the tank positions pretty well, and sometimes very well.

So, even though it didn't get much attention, there was a pretty massive fix for AI infantry set up in the patches. Maybe something happened with map edges, too.

I never play nor deal with QB maps in any way so any changes might´ve surely escaped me (incl reading patch notes). Just like in regular mission designs I avoid leaving anything setup related to what the game finds beeing appropiate. To me it´s all FUBAR, so I either lock AIP units in place (no setup zones) or restrict and fine tune them. Any larger generic zones lead to messed up unit deployments and further loss of cohesion once the movements start. Counts both for attacking and defending AIP alike. Off course different for human players forces if part of the mission includes free deployment in a given map area. Might test larger AIP deployment zones again (in case something has changed unnoticed by me), but I don´t expect much for the better to be honest. Dealing with AIP move zones and orders is another topic, better be left out here since it gets lost and scattered.

Edited by RockinHarry
Link to comment
Share on other sites

10 hours ago, RockinHarry said:

...That pretty much confirms my basic assumption on how things appear to work with evade and retreating units, particularly in urban environments.

1. Buildings are bad cover at close ranges, below 300m. Walls are penetrable by most small arms, particularly ones with higher V0 like rifles and MG´s. To make matters worse, ceilings as well are paper thin and the more abstracted openings (windows, doors, any sort of invisible internal damage) allow hitting and killing troops from odd locations. Like i.e through windows at ground level, up through ceilings (or maybe staircases as well) at next level up. An apparent height advantage is more or less neglected this way...

For my part, using my modular building constructions, I'm not finding the protection level of buildings to be a problem. The problem is that when my units rout, the often don't use the buildings in the protective way that real soldiers would.

I agree that soldiers bunching up can lead to annoying, undesirable outcomes. However, the only areas where I'm seeing a problem that definitely needs looking at is with group spacing inside some building types and Soviet MG teams (maybe other Allied as well, but I haven't played them lately). As I'm sure you know, some building types tend to cause a soldier or two to stay outside when the squad takes position on the ground level. This, of course, gives away the sqad's location and can draw fire. I often put units on the second level as to avoid this.

The issue with the Soviet MG teams is that a crew member often has part of his body sticking outside walls of the building they are in. This gives away their position and draws fire. The protruding body also seem to be treated as being in the open with regards to received attacks. Facing can help, but the guy is often the first casualty for the team. I don't notice this with German HMG teams.

10 hours ago, RockinHarry said:

I never play nor deal with QB maps in any way so any changes might´ve surely escaped me (incl reading patch notes). Just like in regular mission designs I avoid leaving anything setup related to what the game finds beeing appropiate...

Understood on not leaving setup to the Engine. I do the same thing...at least I did until the new patches. Now, I think a good way to go would be to see what the Engine does on its own first and then use that as a starting point for micro-managing setup as the designer. That's what I plan to do for my SP work for my map slices. Like I said, the AI has a knack for finding some excellent ambush points. So, analyzing what it comes up with, keeping the wheat and manually fixing the chaff might be a good working method.

But, honestly, I'm having a lot of fun with my semi-blind quick testing for newly-completed map areas. To simulate SS German motorized units, I take a standard Grenadier battalion with excellent equipment, up the stats,  reinforce each platoon with 1HMG, 3 LMG, 1 scout, 1 TH, 1PS and one PaK40. Then I divide it into 9 groups (HQs above Plt are usually stripped) and divide the German area into 9 solidly painted areas, which are usually large block areas of the urban map. To this, might add a dedicated HMG group and/or a FO group with more possible locations than teams, so I don't know exactly where they are. Those are single painted tiles spread all over the map. To that, I add say, two platoons of tanks (usually Panthers) and choose the locations for those. Then, maybe I have one or two combined arms counterattack groups that are terrain triggered. Plus lots of arty (Oh, PROBLEM, in this dense urban environment, I'm seeing very little enemy arty use, unless I place plentiful TRPs).

Then, I take a typical Soviet 44 infantry battalion (44 has best splitting options) with a customized "combat engineer" platoon, plentiful AFV (I economically pull platoons piecemeal from the pool to avoid having to set up reinforcement waves) and arty support up against it. Having the AI be 1:1 or more in infantry is intended to simulate a human player being dynamic with his forces.

It's not perfect, but it is quite fun and gives me pretty good traction for testing my map areas. I really like not know where the enemy infantry is. Another thing I try to do is set up when I'm sleepy so I forget things. Like with the current playtest, I actually forgot where the counterattack triggers are and exactly what paths they take. Just remember it's Tigers and infantry. 🙂

Edited by Macisle
Link to comment
Share on other sites

9 hours ago, Macisle said:

For my part, using my modular building constructions, I'm not finding the protection level of buildings to be a problem. The problem is that when my units rout, the often don't use the buildings in the protective way that real soldiers would.

I agree that soldiers bunching up can lead to annoying, undesirable outcomes. However, the only areas where I'm seeing a problem that definitely needs looking at is with group spacing inside some building types and Soviet MG teams (maybe other Allied as well, but I haven't played them lately). As I'm sure you know, some building types tend to cause a soldier or two to stay outside when the squad takes position on the ground level. This, of course, gives away the sqad's location and can draw fire. I often put units on the second level as to avoid this.

The issue with the Soviet MG teams is that a crew member often has part of his body sticking outside walls of the building they are in. This gives away their position and draws fire. The protruding body also seem to be treated as being in the open with regards to received attacks. Facing can help, but the guy is often the first casualty for the team. I don't notice this with German HMG teams.

Group spacing is and remains a problem for sure unless certain SOP´s would be introduced to the game. When looking closely at teams or full squads deploying it always seems like a "who comes first" pattern is applied here. That´s why it always looks like the pixeltroopers can´t agree among themselves on what place to take in an AS and within buildings quite in particular. Not much that one can do about it. CA´s and face command sometimes help but most the time rather not. Regarding MG teams I noticed serious problems with mentioned assistant gunner in german squads. Less or none of it with seperate 2 men lMG teams. While the assistant guy is tied to its gunner internally he nevertheless always got to fight with other squad team members for his final place beside the lMG gunner. That´s when I see that guy left out in the open when the remaing squad deployed in the basement of a building. I´m bits of working around in my own missions by using "Ersatz" (replacement) type infantry which usually have no lMG´s. Cutting down on headcount and then (re-) assigning 2 men lMG teams to the Plt HQ gives me better options for Plt and squad tactics, particularly beyond rifle fire ranges (>= 300 - 350m ). Single lMG teams also have the benefit that just the gunner engages targets (assistant remains idle unless at very close range), making them a rather hard to spot and targetable unit at ranges beyond 350m (preferably 400-600m). Area fire prefered as lMG´s only shoot straight when the gunner´s lying prone. Otherwise ammo is wasted 90% as shooting from kneel stance almost always overshoots and support from windows and walls is non existent obviously. Makes german MG34/42 quite ineffective in particular.

I don´t have CMRT so can´t tell if lMG gunners in soviet squads have similar ties to a gunner assistant (if there is any).

Regarding HMG teams I usually set headcount for germans to 50%, as otherwise it would just be a huge enemy target spotted all to quickly. The (full strength) 6 men team is simply too large and tightly spaced for a single AS. In real life just the gunner and an assistant would operate the weapon in a combat situation while the ammo carrier guys remain somewhat off in full cover. In game that would be mostly 1 AS away. Wondering that gun and ammo carriers are seperated in other nations TOE as it was german SOP to have them seperated (2 teams) by default.

Re soviet MG teams, do you refer to the lMG DP28 or Maxim HMG ones?

9 hours ago, Macisle said:

I'm seeing very little enemy arty use, unless I place plentiful TRPs).

TRP´s and a security range to forward friendlies of at least 300m does the trick for AIP usually. Less than 300m when just small to medium mortars concerned. And off course a FO placed in a spot with LOS to many those TRP´s. Hard to find in urban environments but if it comes to helping the AIP one can tweak the map to make it happen more oftenly (AIP Arty use).

In a mission of mine parts of the AIP attack plan is to clear some AS´s with good observation into the player defense area (rear slope). Once the area is clear from the player units, an AIP FO moves in (trigger) and gives the player a heavy beating with couple of medium Arty batteries. So if the player does not defend actively and aggressively for key terrain he´s doomed to loose from (US) AIP artillery alone. Remaining in close contact with AIP units (within 300m security range) can prevent the heavy beating as is stubborn defense of that key terrain. To make matters worse, I don´t mark nor tell the player of that important key terrain (it´s in terrain objectives, but hidden to player). So he got to figure out by himself what terrain is key (to his defense and survival) and what is not. But this on a fairly open map so there´s enough chances for the 300m security range to apply or deny for both players forces.

In MOUT environments that 300m range is of equal importance, but more for the rather automatic spotting of units at normal rifle range (< ~350m) and small arms building wall penetration capability. A sniper or HMG team in MOUT is dead meat quickly if closer to the enemy than said 300-350m. For an AIP defense one might think about move to switch positions, triggered when player units coming into said rifle and spotting range. But that requires single unit groups tied to terrain triggers and can eat up groups maybe needed at another place. But one can create interesting "active defense" situations for the AIP this way and help it survive longer. Just like in RL a well placed crack sniper (single, not team) beyond 350m can wreck havoc on the enemy if he can be helped to survive long enough. A headcount 50%, high experience,, low morale and moving to alternate positions oftenly gives me good results most the time.

Link to comment
Share on other sites

18 hours ago, Macisle said:

 However, the only areas where I'm seeing a problem that definitely needs looking at is with group spacing inside some building types and Soviet MG teams (maybe other Allied as well, but I haven't played them lately). As I'm sure you know, some building types tend to cause a soldier or two to stay outside when the squad takes position on the ground level. This, of course, gives away the sqad's location and can draw fire.

I've noticed the behaviour you describe across both allied and axis forces in Final Blitzkrieg, so it is not confined to the Russians  I certainly didn't notice the problem of a team, or individuals staying outside of a building until the introduction of Engine 4. Some people had in the past complained about "squads" bunching, or moving in single file. So rather than splitting them into "teams" and avoiding the bunching themselves, they wanted the behaviour changed by BFC, or at least that is how I read it at the time, I maybe wrong on that score.  I don't know if the hanging around outside of a building has been addressed in either patch 4.01 or 4.02 as I now confine myself to playing on engine 3. 

Link to comment
Share on other sites

2 hours ago, Warts 'n' all said:

I've noticed the behaviour you describe across both allied and axis forces in Final Blitzkrieg, so it is not confined to the Russians  I certainly didn't notice the problem of a team, or individuals staying outside of a building until the introduction of Engine 4. Some people had in the past complained about "squads" bunching, or moving in single file. So rather than splitting them into "teams" and avoiding the bunching themselves, they wanted the behaviour changed by BFC, or at least that is how I read it at the time, I maybe wrong on that score.  I don't know if the hanging around outside of a building has been addressed in either patch 4.01 or 4.02 as I now confine myself to playing on engine 3. 

That's my recollection as well - that the spacing change was in response to community request. The tendency of one or two guys from a squad to drift outside when on the ground level seems to happen most often with long narrow building types. It can sometimes be fixed with a face command to point the squad at one of the long walls. However, it is often faster and easier just to have them go to the second level.

I always stay with the most current Engine version. For the last couple of years, I've spent most of my game time working on my dense urban map for CMRT. Aside from suicide rout, which did seem to get slightly worse in the .02 patch, Engine 4 has worked rather well overall for that combat environment. Part of that could be the specific dynamics of typical Soviet attacker vs high-quality German defender, though. In terms of current CM, that fits the environment like a glove.

Link to comment
Share on other sites

I see the stray german remain outside building guy mostly beeing the mentioned german infantry squads assistant lMG gunner. In full strength squad it´s mostly one of the last 2 riflemen (8th or 9th) in team No 2. Then mostly when pushing a full size squad into a rather small building (8x8m). But I´ve also seen it with split teams or otherwise reduced size squads, so lack of space is not all that matters. There´s some inherent issue with that particular assistant gunner tied to the lMG guy. To me it looks like there´s a limited time chunk available for deploying troops in a building. If that time chunk (or limit) is spent and the guy didn´t found its place yet, he remains outside. Renewed setting of waypoint into same building oftenly gets the guy moving again , but not always. Sometimes the TacAI sorts out itself without further player actions. But as long as the guy remains outside he makes all of his squad mates a target to any enemy in LOS (fire aimed at him inevitably hits the building and occupants with all its consequences).

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...