Tank Accuracy ....

[siffo998]

we are aware that firing on the move is not as inaccurate as it should be. Until we add a couple of Commands to the game and add the necessary TacAI, it is a problem that will remain for a while.

Steve

thats all i (and i think also alot of the people who recognized that issue) wanted to hear... thanks...!

[ClarkWGriswold]

Yes, I'm afraid to say AKD is correct. The test that siffo998 did has some problems with it. It is also a good idea to upload a test scenario somewhere so a tester can take a look at it.

As AKD states, we are aware that firing on the move is not as inaccurate as it should be. Until we add a couple of Commands to the game and add the necessary TacAI, it is a problem that will remain for a while.

I'm curious as to the reason why commands would have to be added? We already have four movement speeds available (Fast, Quick, Normal, Slow). Wouldn't it be possible to change the behavior for one or more while leaving it for the others? For instance, adjust the accuracy of firing on the move so that tanks can't hit so reliably and perhaps don't attempt shots as often if the target is far away. Leave Fast and Quick so that tanks fire on the move, but change Normal and Slow so that tanks stop, fire, then continue on their way.

I'm not sure how feasible this is, but it would seem to be a good solution if it could be implemented.

[Marc Anton]

I will also note this issue exists for all three CMx1 games.

Steve

That is not true. I made the test with both games, CM:BN and CM:BB with completely different results in terms of hitting and destroying enemy tanks on the move.

I mean, Steve, its not that we dont know the game and how it feels ...

For me i play it for 6 years now and the unrealistic accuracy/aiming speed of tanks was one of the FIRST things we (MP buddy and me) noticed while performing our first CM:BN MP test´s.

I does not need always a NASA test to proof some errors/issues ingame.

But anyway if you guys change the game from CM:SF with a WW2 mask into a realistic and worthy successor of CMx1 its ok for me.

Marc

[Battlefront.com]

I'm curious as to the reason why commands would have to be added? We already have four movement speeds available (Fast, Quick, Normal, Slow). Wouldn't it be possible to change the behavior for one or more while leaving it for the others?

Sure, but that doesn't fundamentally address the problem that tanks should be "smart" enough to halt to take shots and then keep moving. For that we need new Commands so when a player wants a unit to get from A to B without firing he can do that reliably, but in other situations have the vehicle either stop completely (like Hunt does now) or fire a round or two and then continue to move.

I'm not sure how feasible this is, but it would seem to be a good solution if it could be implemented.

Very feasible, actually. We are going to reexamine the speeds to see if one or more is too generous.

That is not true. I made the test with both games, CM:BN and CM:BB with completely different results in terms of hitting and destroying enemy tanks on the move.

That's not what I meant. I meant that in CMx1 tanks fire while on the move instead of coming to a halt. This is the fundamental problem the game, all CM games, have.

I mean, Steve, its not that we dont know the game and how it feels ...

For me i play it for 6 years now and the unrealistic accuracy/aiming speed of tanks was one of the FIRST things we (MP buddy and me) noticed while performing our first CM:BN MP test´s.

I does not need always a NASA test to proof some errors/issues ingame.

True, but without a "NASA" test to isolate the variables and quantify how much change is needed, then we might as well just change variables at random. That is never a good way to approach things.

But anyway if you guys change the game from CM:SF with a WW2 mask into a realistic and worthy successor of CMx1 its ok for me.

There's probably not a single variable in CM:BN that is the same as CM:SF. There's also a ton of things in CM:SF that are disabled in the code, a ton of new stuff added. There can be something wrong in CM:BN without having anything to do with CM:SF.

Steve

[LarsS]

...

Now something which really is "naturally" modelled in tank sim games (Red Orchestra/Darkest Hour is an adequate example) and NOT specifically modelled in any other games like CMBN AFAIK is....

-when attempting to fire from a moving tank, the following two factors ALL contribute to the increased difficulty of aiming at and hitting a target at any range:

a) the speed of the moving tank

and (this is a big one)

the terrain on which the tank is currently moving over

Most games/simulations like CM seem to just consider a) and completely disregard .

It may be possible that trying to aim and hit a target from a tank moving at 30km/h on a smooth flat road is easier than trying to hit the same target from the same tank moving at 2km/h on a very rough/rocky/undulating surface.

As far as I know CM (or any other game) has never considered the terrain a tank is currently moving on when determining it's ability to aim and hit targets. However I do wonder HOW the game deals with targetting while moving over shellholes for example....I might do a test. The result will be most telling.

...

Actually, the difficulty of the terrain the tank traverses has a great impact on accuracy when firing on the move. Since the modelling is "pixel to pixel" and "what you see is what you get", every bump the tank hits will change the gun's vertical alignment and therefore it's point of aim and accuracy if it fires before the gun has been raised/depressed to align with the target.

Having said that, there would seem to be no abstraction of "bumpiness" built into the terrain tiles, as is the case with abstracted cover for infantry. So I believe the accuracy would be the same when moving over rocky terrain, light forest or whatever, as it would be when going along a nice paved road. Unless the bumps are actually built into the 3D map, as with ditches, craters, and sudden changes of slope angles, the specific terrain type doesn't seem to be a factor for accuracy when on the move.

[JasonC]

Sounds like firing at "slow" is basically the same accuracy was firing stationary, "quick" is only a 50% accuracy reduction, and "move" is only about a 25% reduction. Not even a laser range finder stabilized targeting computer SEP Abrams can say that in real life. And for WW II it is just silly. I understand a desire to abstract short halt behavior in a turn based we-go game, but CMx2 has real time as an option, yes? So short halt behavior is something the player can create by directly intervening.

The CMx1 had shoot and scoot to allow it within the we-go turns. I can understand wanting to ditch that order, as it was complicated compared to the usual ones (associated with multiple waypoints etc). But I don't think it is a solution to allow "slow" move fire to be as accurate as stationary.

Of course, in the meantime it is highly relevant to CMx2 players to know what works right now. And it sounds like slow is great and move acceptable, and both superior to "hunt".

[stoex]

That's not what I meant. I meant that in CMx1 tanks fire while on the move instead of coming to a halt. This is the fundamental problem the game, all CM games, have.

Um....careful Steve, there was a "HUNT" command in CMx1 which caused tanks to move to contact, stop, fire until the contact was dead or out of LOS, and continue on with their move.

Actually, the return to this function of "HUNT" is what a lot of people have been advocating. I admit I don't understand the difficulties involved in coding such behavior, but "tanks fire on the move instead of coming to a halt" is by no means "a fundamental problem that all CM games have". This I know.

[YankeeDog]

Sounds like firing at "slow" is basically the same accuracy was firing stationary, "quick" is only a 50% accuracy reduction, and "move" is only about a 25% reduction. Not even a laser range finder stabilized targeting computer SEP Abrams can say that in real life. And for WW II it is just silly. I understand a desire to abstract short halt behavior in a turn based we-go game, but CMx2 has real time as an option, yes? So short halt behavior is something the player can create by directly intervening.

The CMx1 had shoot and scoot to allow it within the we-go turns. I can understand wanting to ditch that order, as it was complicated compared to the usual ones (associated with multiple waypoints etc). But I don't think it is a solution to allow "slow" move fire to be as accurate as stationary.

Of course, in the meantime it is highly relevant to CMx2 players to know what works right now. And it sounds like slow is great and move acceptable, and both superior to "hunt".

A human player can actually very create very sophisticated shoot & scoot/firing halt behavior in WEGO CMBN. All you need to do place an occasional waypoint along a movement path (wherever you want the AFV to stop and shoot), and put a pause (usually about 10 seconds is good) at this waypoint. The tank will stop at this waypoint and shoot at any enemy it sees. In fact, in many ways this is easier in WEGO than it is in RT because you have plenty of time to plot the precise orders for each unit; focusing too much attention to one unit and micromanaging it can really get you in trouble in RT.

Actually, the problem with the above technique is that if the tank sees an enemy while on the move before it gets to the halt point, it may shoot while still on the move just before it gets to the halt point, and then potentially spend most or all of its time at the halt point reloading.

So for human v. human, I see little need to make firing on the move accuracy higher than real life, as a compensation for anything.

I wonder if Steve's issue has to do with the fact that the Computer Player isn't sophisticated enough to pull of monkey business like this, so BFC is reluctant to make shooting on the move too hard, so that the Computer Player is a credible opponent. Purely speculation on my part here.

[stoex]

Interesting point, YankeeDog, and I personally think it is true to a certain degree that BF must limit the CM player in some ways to allow the AI to keep up. However, in the case at hand, the return of the CMx1 "HUNT" command would actually benefit the AI IMHO. If scenario designers had the option of scripting the AI units to HUNT to certain locations, it would make it easier to design attacking AI scripts for sure. Come to think of it, the implementation of HUNT would be the first real trigger in the AI scripts, allowing "if->then" behaviour of sorts.

Just think of the repercussions if you could tell the AI to interrupt its advance if it makes contact with the enemy as opposed to blindly moving forward to the objective area...

Would be an excellent addition to the game IMHO.

[Erik Springelkamp]

So we have to interpret shooting on the move as stopping, firing and moving again. And the accuracy has been more or less adapted to that abstracted behaviour?

[Lurker765]

That's not what I meant. I meant that in CMx1 tanks fire while on the move instead of coming to a halt. This is the fundamental problem the game, all CM games, have.

Steve

In CMx1 tanks will fire on the move if you tell them to. You can give them fast or move orders and they will proceed to their target while shooting (inaccurately). Or you can give them a hunt order and they will stop to engage a target and once that target is removed (knocked out or blocked due to visibility changes) the tank will continue moving on its way.

I don't think CMx1 has this fundamental problem as you state.

[stoex]

So we have to interpret shooting on the move as stopping, firing and moving again. And the accuracy has been more or less adapted to that abstracted behaviour?

I don't think it would be fair to call this abstracted. Speed of advance and possibility of getting hit yourself are very different for "firing on the move" vs. "stop to fire, then continue to move". Apples and oranges in my eyes. If you want to test how well this "abstraction" works, fire up a CMx1 test scenario with a couple tanks and see how your tanks perform with HUNT vs, say, FAST. (Don't forget to run the test 50 times to get a statistical spread)

[YankeeDog]

FWIW, I personally have very little need for a CMx1-style Hunt order in CMBN. I actually use Hunt for AFVs very rarely, and I don't think this would change if the Hunt order worked more like CMx1. I prefer to tell my AFVs exactly where and when they should stop and pause to look & shoot using the above method, not trust that they'll spot & stop to fight in an intelligent location along their movement path of their own accord. Stop next to he building that covers your flank; shoot from there if you see something, and then move on. Etc.

And in any event, given the amount of time it generally takes a tank to spot a target with WWII optics, engage, get a killing shot, and then loose a couple of MG bursts at the fleeing crew, the one-minute turn is usually over, IME. If I want it to start moving again, I can just order it to do so at the next orders phase.

I actually might have found more use for a CMx1 style Hunt order in CMSF, where Abrams really can run and gun, loosing shots at multiple targets seconds apart, with a good chance of multiple kills.

But to to each their own; I'm sure there are a lot of different play styles out there.

[Redwolf]

CMx1 most definitely had the kill probabilities between a standing shooter and a moving target then firing back screwed up, too. Much of the standing shooter failure had to do with the extremely limited/useless zeroing in going on in CMx1 but the bigger part was the deadly accuracy of moving tanks. Anyway...

But again, what does that have to do with having to insert stop commands? Let them fire on the move but make it less accurate. The player can halt them for now if he doesn't like it.

[Battlefront.com]

A human player can actually very create very sophisticated shoot & scoot/firing halt behavior in WEGO CMBN. All you need to do place an occasional waypoint along a movement path (wherever you want the AFV to stop and shoot), and put a pause (usually about 10 seconds is good) at this waypoint.

Sorta, but is is VERY "kludgey". And, as you pointed out...

Actually, the problem with the above technique is that if the tank sees an enemy while on the move before it gets to the halt point, it may shoot while still on the move just before it gets to the halt point, and then potentially spend most or all of its time at the halt point reloading.

And if you don't know the enemy is there, or pick the wrong place to put the Pause, or Pause for the wrong amount of time, or the enemy is temporarily out of LOS/LOF while Paused, or... cripes, there are so many things that could go wrong. It really isn't a good work around.

RealTime has it's own problems. You can perfectly execute firing from short halts, but only if you're micromanaging the unit and not doing anything else (including micro managing other tanks). So it's not all that great in RealTime either, though it is better than WeGo in that it can be far more adaptive.

So for human v. human, I see little need to make firing on the move accuracy higher than real life, as a compensation for anything.

I don't agree. From a philosophical standpoint REQUIRING kludgey, very difficult to execute work arounds is not good enough.

I wonder if Steve's issue has to do with the fact that the Computer Player isn't sophisticated enough to pull of monkey business like this, so BFC is reluctant to make shooting on the move too hard, so that the Computer Player is a credible opponent. Purely speculation on my part here.

That is a part of it. We are very leery of making behavioral changes that hobble the AI Player in some way.

In CMx1 tanks will fire on the move if you tell them to. You can give them fast or move orders and they will proceed to their target while shooting (inaccurately). Or you can give them a hunt order and they will stop to engage a target and once that target is removed (knocked out or blocked due to visibility changes) the tank will continue moving on its way.

I don't think CMx1 has this fundamental problem as you state.

It's there and I've not overstated it. CMx1's implementation of Hunt definitely gave players a less kludgey workarounds than CMx2 currently has, but fundamentally there was absolutely no support for firing at short halts.

Hunt, for example, would stop the tank and it would fire from a stopped position. However, if any viable target, including infantry, present itself then the tank won't continue on. This is not what a tank should be doing. It should be stopping, firing, maybe firing again, then leaving even if the target is not destroyed or other targets appear. Well, unless there's no real threat that is.

Move to Contact just stops the tank dead in its tracks like it does in CMx2. So on that count CMx1 and CMx2 are the same.

But again, what does that have to do with having to insert stop commands? Let them fire on the move but make it less accurate. The player can halt them for now if he doesn't like it.

Absolutely the degree of accuracy is something separate from the abstraction of firing at short halts. We can tweak the accuracy and yet keep the same basic behavior. We're already doing our own tests to see how much and under what situations things should change.

I will say we internal types have always advocated a lower accuracy while firing on the move. But with hundreds of equally important things (in their own way) to get fixed up we couldn't remain fixated on this one thing too long.

Steve

[Lurker765]

It's there and I've not overstated it. CMx1's implementation of Hunt definitely gave players a less kludgey workarounds than CMx2 currently has, but fundamentally there was absolutely no support for firing at short halts.

Hunt, for example, would stop the tank and it would fire from a stopped position. However, if any viable target, including infantry, present itself then the tank won't continue on. This is not what a tank should be doing. It should be stopping, firing, maybe firing again, then leaving even if the target is not destroyed or other targets appear. Well, unless there's no real threat that is.

I used hunt with covered armor arcs quite well in CMx1. My tanks would hunt for armored targets and engage them at a stop and then continue on while ignoring infantry. I had no complaints in this area in CMx1

[Battlefront.com]

I used hunt with covered armor arcs quite well in CMx1. My tanks would hunt for armored targets and engage them at a stop and then continue on while ignoring infantry. I had no complaints in this area in CMx1

Which is fine for you, but that doesn't change the fundamental fact -> there is no support for firing from short halts in CMx1. The Cover Arc, if used, won't keep your tank moving if there's more than one non-infantry target. It also isn't flexible to have your tank stop, firing one shot at an AT Gun if you had it set for Armor Arc. Etc. So while you may be generally pleased with the kludge work arounds in CMx1, they are still kludges. And that means the basic behavior is being worked around with tools which weren't intended to be used in that particular way, which in turn means the outcomes are not optimized to compensate for the missing behavior.

Put another way, one can pound screw into a wooden board using a hammer. But that doesn't mean the hammer is a the right tool for screws. A sledgehammer is

Again, both systems have automatic firing on the move behavior. Both have kludges to get a unit to stop and fire. Some players have more or less problems with either the automatic behavior or the kludge work arounds. But in all cases, no matter how you slice it, neither game system fundamentally supports firing from short halts.

Steve

[Lurker765]

Which is fine for you, but that doesn't change the fundamental fact -> there is no support for firing from short halts in CMx1. The Cover Arc, if used, won't keep your tank moving if there's more than one non-infantry target. It also isn't flexible to have your tank stop, firing one shot at an AT Gun if you had it set for Armor Arc. Etc. So while you may be generally pleased with the kludge work arounds in CMx1, they are still kludges. And that means the basic behavior is being worked around with tools which weren't intended to be used in that particular way, which in turn means the outcomes are not optimized to compensate for the missing behavior.

Put another way, one can pound screw into a wooden board using a hammer. But that doesn't mean the hammer is a the right tool for screws. A sledgehammer is

Again, both systems have automatic firing on the move behavior. Both have kludges to get a unit to stop and fire. Some players have more or less problems with either the automatic behavior or the kludge work arounds. But in all cases, no matter how you slice it, neither game system fundamentally supports firing from short halts.

Steve

It is all about how much time you can spend on developing these things. Sometimes kludges that achieve acceptable results are fine. The time required to code something like if your tank should stop based on X targets being around is unacceptable.

Getting a CMx1 hunt command with armor cover arcs solves the vast majority of cases.

[Redwolf]

It is all about how much time you can spend on developing these things. Sometimes kludges that achieve acceptable results are fine. The time required to code something like if your tank should stop based on X targets being around is unacceptable.

Getting a CMx1 hunt command with armor cover arcs solves the vast majority of cases.

I'm not sure. It's actually worse than Steve says. It is not only multiple armoured targets that will bring your tank to a stop. The same target, as long as it isn't eliminated and known to be so, will stop your tank for good. Where what is needed is stop-fire-continue. Remember there are more tanks behind that lead one, you can't stop the whole show to fire at a halftrack.

I really wish that CM would go with explicit SOP settings. "halt to fire?" yes/no. "continue after halting-for-fire" yes/no. "stop if vehicle in front stops" yes/no (aka the traffic jam switch).

[YankeeDog]

I'm not sure. It's actually worse than Steve says. It is not only multiple armoured targets that will bring your tank to a stop. The same target, as long as it isn't eliminated and known to be so, will stop your tank for good. Where what is needed is stop-fire-continue. Remember there are more tanks behind that lead one, you can't stop the whole show to fire at a halftrack.

This. I'd definitely make use of some sort of AFV movement order that told the tank to occasionally make firing halts along the path (if targets presented themselves), than the old CMx1 Hunt -- "stop as soon as you see something, keep shooting as long as you can see a living enemy, and then move on".

Actually, functionally, the kind of order we're talking about here isn't that different from the way infantry behaves when moving under a Quick order right now.

[Redwolf]

Actually, functionally, the kind of order we're talking about here isn't that different from the way infantry behaves when moving under a Quick order right now.

Yeah. Except for the traffic jams and that's the big elephant in the room. The CMx2 code still occasionally turns vehicles > 90 around, presenting not only side but back armour to the enemy, on traffic conditions. That would mix very unwell with code that randomly stops vehicles.

IMHO a solution to the traffic bumping needs to come first or at the same time (namely, instead of inventing a new path to go around a temporarily stopped vehicle, just wait if the code knows the stop is temporary).

[YankeeDog]

Yeah. Except for the traffic jams and that's the big elephant in the room. The CMx2 code still occasionally turns vehicles > 90 around, presenting not only side but back armour to the enemy, on traffic conditions. That would mix very unwell with code that randomly stops vehicles.

IMHO a solution to the traffic bumping needs to come first or at the same time (namely, instead of inventing a new path to go around a temporarily stopped vehicle, just wait if the code knows the stop is temporary).

Good point. IME, vehicles in CMBN *usually* pause and wait for a vehicle ahead of them to clear path, rather than attempting some wonky detour, but not 100% of the time. It's markedly better than CMSF, but every once in a while, you still get the crazy AFV dance.

It's not such a big deal to me because I've gotten pretty used to the limitations of the vehicle pathfinding code, so I keep a pretty good margin of safety between vehicles. And in any event, CMBN's damage and ricochet modeling gives a pretty strong incentive to keep a good distance between vehicles. On one of the few occasions I have put two tanks really close together in CMBN, I actually had a shreck rocket aimed at one tank sail long, hit a tree next to the tank behind it, and kill the TC. Oops.

BUT... I can see how things could get really "interesting" really quickly if AFVs were frequently taking firing halts, so the traffic flow issues definitely are something to consider here.

[dieseltaylor]

I was justing reading today how in North Africa the US Shermans advanced halted and fired, and then advanced again. Panther crews were only assessed - when graduating - on halting and taking out their four targets on the range.

SO apart from the British early war practice with the tiny 2 pdr it seems hard to give much credence to firing on the move - when aiming at another tank or ATG. Of course there are exceptions but does the game need to make the rare occurrence the common method? It does suck against the realism.

Shoot and scoot worked very well in CMAK provided one understood that you could actually advance with the order. There was a higher hit chance than blindly firing during a fast move to the same spot. Unfortunately most players thought shoot & scoot was a move up and reverse order.

[Battlefront.com]

It is all about how much time you can spend on developing these things. Sometimes kludges that achieve acceptable results are fine.

This has been at the core of our development philosophy since the very beginning. So obvious I agree 9 times out of 10. When I don't agree it is because a) the problem is significant enough to warrant a real solution and the kludges are either too annoying or too narrow. In this case...

Getting a CMx1 hunt command with armor cover arcs solves the vast majority of cases.

The work arounds are both too annoying and too narrow We really need a better solution than what was in CMx1. The posts above this are good examples of why.

Steve

[MikeyD]

I can't recall the last time BFC added a new command to this game engine. Perhaps it was the introduction of the 'Fire Smoke' command for Challenger 2 tanks in the CMSF Britsh module. Adding more and more commands is a slippery slope. There's already (non-CMx1) newbies complaining about the complexity of the game. And BFC does hope to haul in more happy purchasers than just the CMx1 old fogies.