Infantry Don't Bennfit From Low Walls

[ChrisND]

Big thumbs down to some of the "arm chair generals" here who are criticizing the validity and design of the tests trying to examine a SINGLE variable without running any of their own.

Part of vetting test results is questioning the test itself. If they stand up to scrutiny, they are that much stronger.

[Vanir Ausf B]

Interesting stuff, guys. I'm going to run a bunch more tests today. They will all be of the head-to-head variety since I believe the invincible-unit-shooting-at-infantry type I ran yesterday is worthless for anything except showing walls are a good place to cower or hide behind.

[ChrisND]

The results have interested me enough that I'm running my own as well.

[Vanir Ausf B]

Just be aware that if anyone sees you doing this they will not think you're "normal" anymore

[sburke]

Just be aware that if anyone sees you doing this they will not think you're "normal" anymore

LOL yeah really "what did you do yesterday?"

"I tested to see if I was better off behind a low wall if people were shooting at me"

So here is what I came up with in changing the proximity of the wall.

This is with the range still split with 10 lanes having no wall behind them. Numbers are units in open wnd/dead, units behind wall wnd/dead.

1) 21/22 62/75 69 with wall 8 squares back

2) 15/15 65/82 71 with wall 6 squares back

3) 30/27 58/68 75 with wall 4 squares back

4) 31/31 49/53 33 with wall 3 squares back

5) 38/43 38/48 30 with wall 2 squares back

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Map range altered so all 20 lanes have rear wall

numbers are units in open wnd/dead units behind wall wnd/dead

with wall 3 squares back

1) 33/21 29/44

2) 47/40 24/28

with wall 2 squares back

1) 48/24 28/27

2) 38/27 25/33

with wall 1 square back

1) 45/37 11/17

2) 47/43 13/17

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

And now back to the original question of whether a low wall is better or worse in and of itself. All 20 lanes have no rear wall. All units are regular/normal/no leadership bonus/fit run as veteran 2 player hot seat. Numbers are units in open wnd/dead, units behind wall wnd/dead.

1) 28/20 67/76

2) 20/26 77/87

I'd have to say I am not likely to place my units behind a wall unless there are other circumstances to contribute to their security. At least not a rough stone wall. Am I really gonna go verify against other type walls....sadly yeah

[Vanir Ausf B]

That is positively weird.

[sburke]

So on the other 2 low wall types I found generally the same behavior.

All 20 lanes have no rear wall. All units are regular/normal/no leadership bonus/fit run as veteran 2 player hot seat. Numbers are units in open wnd/dead, units behind wall wnd/dead. Distance is approx 230 meters.

15/20 63/98 stone wall

15/23 70/82 brick wall

[Battlefront.com]

Big thumbs down to some of the "arm chair generals" here who are criticizing the validity and design of the tests trying to examine a SINGLE variable without running any of their own.

As Normal Dude says, it's called "vetting" or (as I like to call it) "peer review". I've seen literally hundreds of customers run thousands of tests over the years and we've learned to never take a test seriously until it's been kicked first. We've wasted too much time on flawed tests.

Besides tests which aren't really all that well constructed, the bigger reason to be mindful of the test conditions is that for us to focus in on what is wrong we need to rule out as many things as possible. There was a sticky on this Forum for quite some time about a Beta Tester testing round that eventually lead to finding a very bizarre bug (TC was off center), but we only found it after many different tests. There were too many variables and therefore the results were too generic to figure out what was going wrong.

We're looking into this ourselves using tests similar to the ones now being conducted.

Remember, most bugs get discovered by someone saying "this doesn't seem right". We know that too. We also know that there's a lot of crying wolf. It's important to spend our limited resources on real issues, in as specific a way possible, instead of non-issues.

Steve

[sburke]

We're looking into this ourselves using tests similar to the ones now being conducted.

Remember, most bugs get discovered by someone saying "this doesn't seem right". We know that too. We also know that there's a lot of crying wolf. It's important to spend our limited resources on real issues, in as specific a way possible, instead of non-issues.

Steve

I hear that. Honestly I can't even say I know what the results SHOULD be for these tests. I can see an argument going either way that a prone unit would be harder to hit versus a unit with head and shoulders above a wall or a unit behind a wall could hunker down a bit and have the protection of the wall. I just want to make sure I am not sending my GIs up to that wall for cover as they advance in the mistaken assumption that it is the right thing to do given a certain set of circumstances.

The behavior exhibited when they have a wall to their rear within 8 meters is something else, but I have no idea what the results might be showing or not. I don't come across many small walls that close to a tall wall so I don't expect it is going to make much difference in the long run.

[Offshoot]

I don't come across many small walls that close to a tall wall so I don't expect it is going to make much difference in the long run.

I see you've tried with other terrain features behind the troops, but out of interest have you tried with a building behind them? I imagine this is a relatively common instance of finding a wall and terrain feature 'close' together. If you do check, probably a building without a door on the troops side would be best

Thanks for the testing everyone, After seeing bullets bouncing off walls my automatic assumption would have been that they were better than no cover as well.

[sburke]

I see you've tried with other terrain features behind the troops, but out of interest have you tried with a building behind them? I imagine this is a relatively common instance of finding a wall and terrain feature 'close' together. If you do check, probably a building without a door on the troops side would be best

Thanks for the testing everyone, After seeing bullets bouncing off walls my automatic assumption would have been that they were better than no cover as well.

Not yet, but I have been wanting to do some general testing of buildings anyway so will likely be hitting that this week.

[Vanir Ausf B]

Latest (and hopefully last) test.

US armored infantry behind low rural wall vs. US armored infantry in open on pavement (the reason I keep using US armored infantry is because they don't have any automatic weapons so the guys lying prone in the open don't gain an advantage from fully deploying).

14 firing lanes separated by bocage. Each lane has 1 squad from each side facing each other at 200m. I made sure the 9-man squads were only matched up against other 9-man squads. The HQs are isolated so no C2. 154 men per side total including HQs.

Each test was run for 3 turns/minutes on Warrior difficulty. Test was run 40 times.

[U]BEHIND WALL[/U]


KILLED:  1344

WOUNDED: 1572

TOTAL:   2916

AVERAGE: 72.9 / 47.3%


[u]IN OPEN[/u]


KILLED:  1422

WOUNDED: 1131

TOTAL:   2553

AVERAGE: 63.8 / 41.4%

Save game

http://www.2shared.com/file/HOHRGk3N/Isolation_H2H_no_C_001.html

Scenario

http://www.2shared.com/file/4W_WcEyG/Isolation_H2H_no_C2.html

[BigJ62]

10 USA rifle squads versus 10 USA rifle squads. Each lane is 100m wide separated by a high wall and distance is 210m. Each squad is regular, fanatic and 0 leadership. Hotseat, fow veteran.

1st Test Allied no buildings, Axis with back buildings, 6 runs of 11 turns each

Allied wia 270, kia 418, total 688

Axis wia 81, kia 67, total 148

2nd Test Allied with buildings Axis no back buildings, 6 runs of 11 turns each

Allied wia 91, kia 67, total 158

Axis wia 276, kia 398, total 674

3rd Test both sides with back building, 3 runs of 11 turns each

Allied wia 98, kia 90, total 188

Axis wia 98, kia 106, total 204

And just for fun I gave the axis side 1 double wide one story building in each lane and the allies nothing and it was a blow out so buildings do help greatly.

So for a total of 12 runs of 11 turns each we get 1362 casualties for the side that has no building behind them and 306 casualties for side that has a building behind them with an empty grid between it and the building. In every single run I did not have any outliers that might indicate a gross flaw in the test.

[dieseltaylor]

SO, the big question in my mind is - was this true of CMSF??

Which I never bought.

[PhantomApple]

So is it confirmed that troops benefit from nearby terrain/cover? I just got done building my shooting range, so will need to make adjustments if this is the case. If it was answered apologies, the amounts of numbers in this thread started to blur together

[dieseltaylor]

Well, the story so far.

1. Infantry firing from behind walls lose to prone infantry. However if they hide behind the wall they do not take any casualties[ ignoring rifle grenades etc]

2. Infantry placed in front of walls and buildings do not suffer as badly as infantry in the open. So essentially if you want a fire fight put your infantry in front of walls not behind them. ANd depending how far in front affects the number of losses.

I look forward to your report Placing troops in front of fences! Barbed wire? Do advancing troops fire more accurately than stationary troops?

[BigJ62]

I wouldn't mind running a test using dead cows but unfortunately we don't have any dead cows.

[YankeeDog]

Hmmm... I do wonder if all of this isn't due to some sort of abstraction of the effects of skylining in the game. That is, the game considers the soldiers that have a dark background behind them (i.e., a wall or a building) harder to see than the soldiers that have nothing but open sky behind them. So the units with a wall or a building behind them are gaining a concealment advantage, not a cover advantage.

Granted, if this is what is going on, based on the test results posted here so far, the effects seem much stronger than they should be for a relatively close-range, infantry firefight. But it's the only thing I can think of that would explain these results.

[BigJ62]

I also tested with the wall 1 extra grid further back and it didn't seem to help that unit anymore so no I dont believe skylining is the effect I am seeing.

[YankeeDog]

I also tested with the wall 1 extra grid further back and it didn't seem to help that unit anymore so no I dont believe skylining is the effect I am seeing.

Actually, if properly modeled, the wall could be quite a distance behind a prone soldier and still provide a "dark background" bonus. It would have to do with the shooters LOS -- if the line from the shooter's eye that goes through the soldier he is trying to spot/shoot passes also through the wall further downrange, then the target is against a dark background, and might be harder to see.

Once again, I have NO IDEA if this is what is actually going on. And in fact it would surprise me if the game actually had this kind of modeling in it, because it's the first I've heard of it. Just pointing out that distance between the target and the background, in and of itself, is not important to the skylining effect (or lack thereof).

Edit: Whups, I think I misunderstood your post; I interpreted as "didn't seem to help that unit any more [than in the previous tests]", when I think you actually meant that the wall no longer provided any benefit when moved an additional action spot away.

Gotcha. I Dunno. It is possible the skylining effect (if any) is abstracted, so it only takes into account nearby backgrounds. It's hard to imagine the game could actually take into account the concealment bonus provided by a dark, tree covered hill 500m behind a unit...

It's also possible, I suppose, that there is some sort of abstract "cover bonus" for being close to a wall (similar to the abstract cover bonus the game provides to units in heavy woods terrain or broken ground, for example), and that this bonus is currently being misapplied to incoming fire that does not cross the plane of the wall.

Just speculating here... they are very strange results and I'm wondering what exactly in the game engine could be going wrong to cause all this.

[BigJ62]

My theory is the cbt code thinks the wall is in front of the unit when it actually is behind the unit and behind the unit when it actually is in front of the unit. If it is a bug and they fix it do the casualties seem right for units firing from behind a wall?

[Sgt Joch]

SO, the big question in my mind is - was this true of CMSF??

No, I never saw that behavior in that now classic wargame.

Just to be sure, I ran a very quick test. A U.S. infantry squad vs a Syrian Republican Guard infantry squad, both regular experience/motivation. One behind a low stone wall, the other out in the open 100 meters away. In both tests, the squad behind the wall (U.S. or Syrian) wipes out the squad in the open inside of 3 mins, while suffering no casualties or only a few lightly wounded.

One critical factor in CMSF appears to be that the units out in the open quickly take casualties and suffer suppression/morale hit, cower and stop firing while the units behind the wall almost all keep up a high ROF.

So it works as expected in CMSF(although a very limted test). Not sure what is going on in CMBN, but we will get to the bottom of it.

Which I never bought.

You don't know what you are missing.

[Sgt Joch]

I'll have the results of the normal motivation test up soon. However there is an odd TacAI behavior that is taking place that is affecting the results. Units in the open that panic and run are invariably running towards the enemy instead of away. This is obviously adding to their casualties. I don't know if this is because both sides are US. I did make setup zones on both side of the map. Or maybe it's because the wall is the only cover on the map?

when you setup the scenario, you have to specify to the AI (in Mission/Data) which is the friendly map edge, so it knows which way to run. The AI will always run towards the edge you specify as the friendly edge.

[Steiner14]

No, I never saw that behavior in that now classic wargame.

Just to be sure, I ran a quick test. A U.S. infantry squad vs a Syrian Republican Guard infantry squad, both regulars. One behind a low stone wall, the other out in the open 100 meters away. In both tests, the squad behind the wall (U.S. or Syrian) wipes out the squad in the open inside of 3 mins, while suffering no casualties or only a few lightly wounded.

So it works as expected in CMSF. Not sure what is going on in CMBN, but we will get to the bottom of it.

But if someone says, that the results he saw in CMBN were extremely unrealistical to a degree, that makes it unplayable for him, he is rudely attacked...

I have played only one single battle with a low wall and immediately recognized, that it either doesn't work, or in the case of intention, CMBN being unplayable for me. So if i, with my very low experience, could recognize that immediately, what are the testers looking for, if they haven't recognized that? We are not talking here about fine tuning of weapon effects in the low percentages. We are talking about inverse effects.

Although CMx2 is not based on the outcome, but the outcome is the result of an engineered model, i think testers should not worry too much about the depth of the engine. That should stay part of the developers. I think testers should recognize, if the results, that are presented, are logical or not - and not search for explanations, why the results could nevertheless be plausible because of the depths of the model.

If i construct a bridge and a bus driver mentions, that "it swings too much", i don't expect from him, that he understands why it does so. Much worse would be, if the bus-driver would ignore this thought, because he trusts in the engineer, who built the bridge.

Therefore it is important, that testers stay away from the inner game mechanics. Their task is to look at the object from another persepective and judge the outcome but not find explanations for the outcome.

It's enlightening to read this thread to see, how the CM-community has turned into a fanboys dominated one. I don't have the impression it was that way in CMx1 days. Now a storm of criticism, often enough subletely covered, unleashes against critics and they are not taken serious.

Therefore my big thanks to all the wall-testers, that were able to prove, that something is wrong in CMBN right now.

That makes hope, that they will find the problem and solve it. I'd be happy to finally start playing CMBN and see the potential of the engine unfold!

[Blackcat]

Steiner14,

Do I take it from that last post that you have stopped playing CMBN?

All the best