necessary things for CMAK

[fridericus]

in my opinion mainly 2 things went wrong with cmbb compared with cmbo.

1. tanks need 1-2 min for rotating on one place. it might be on swamp, but on streets? i talked to german veterans, who talked, it took only seconds (10-20).

2. because of pathfindung-ai i have to give more orders to a tank, then in reality. when i want a tank moving around a corner without showing enemy his side, i need some more waypoints. but these orders i will punished with a delay upto 2 min? it isnt realistic. in 20 seconds you order a whole platoon how to move. although today a tank-platoon takes 1000m wide on battlefield and in ww2 only some hundrets m.

PLEASE BATTLEFRONT, dont repeat mistakes made in cmbb, but werent in cmbo.

THANKS ! :cool:

[rune]

And I disagree...

1. A veteran on this board stated he drove the M4, and the times feel right to him. It typically took 90 or more second to rotate in the same spot, requiring move forward/move backward, over and over again. A couple of the German tanks could neutral steer, but were the exception. Most tanks could NOT neutral steer.

2. Design decision. You aren't supposed to have that level of control, directing each tank meter by meter.

Rune

[fridericus]

hi

i cant agree.

1. maybe, that the m4 was such a lousy tank, but the german tanks didnt. their transmission allowed to move one track forward and the other backward, so it could rotate in seconds. also thatswhy the heavy german were much more mobile then the m4. so i change my wish: fast rotating only for german tanks. :cool:

2. if the ai could find the right way by itself, like every tankcommander in reality, i had never the problem of ordering the right way.

so you are not right. :eek:

[winkelried]

Originally posted by rune:

1. A veteran on ....

2. Design decision. ...

Can't answer to (1) I am not a tanker. But I agree with fridericus on (2). I think rune, you misunderstood. The penalty for additional waypoints is just too heavy. E.g. I order my tanks / jeeps etc to run along a street. For each bend the street makes I get punished. This is not how it works.

Our marching orders (I server 15 years with Recon) in combat were pretty simple (those before took usually longer):

1.) Target (Next Village, Border of the next Forest)

2.) How and in which formation to get to the target (On this Road in column, Across this field in line).

3.) What to do when arrived at the target (Hedgehog, secure, be ready to attack etc).

This usually too only seconds and not minutes just becaus the road had some bends.

[rune]

Actually most of the German tanks did NOT have neutral steer, only a few models did.

Your example is perfect, did the commander tell you to go to the curve, walk three meters from the house, turn to the left staying within 4 meters, then proceed to the next point? Or were you told to go to that point via the tree line? Too much control is outside the scope of the game. The reason the time element was added, so that conscripts and greens get the much added delay by giving complicated orders. Think WWII and a Russian conscript. Should you be able to control them to that degree?

I doubt you see any changes til CMX2.

Rune

[fridericus]

yes, rune

in reality the commander of the tank knows, how to go. but the problem in bb is, that the tankcommander dont know how to move. they are too stupid compared with reality, thatswhy the player has to help with more waypoints. for that helping for reaching reality you get punished

[rune]

Since we are on the topic, let's talk neutral steering.

Panzer I, II, III, IV did not have it.

Panzer VI did have it

Panzer V, anyone know?

This is my understanding, and can be corrected by those more knowledgable.

Waypoints, since we are talking the amount of waypoints, and the difficulty of directing a conscript or green troop, what has changed? Be it infantry or tank? What everyone forgets around here, is that until CMBO came out, there were NO games like this. Remember people telling Battlefront it cannot work, and they would fail? Yes, things aren't perfect, Steve, Charles, and the gang are the first to admit it. However, expecting everything in a simulation isn't going to work either. CMBO improved the model, and CMX2 will further improve the model. The bottom line, the decision was to add the time restriction is for waypoints to show the difference in Command and Control between conscripts vs elites.

Rune

[fridericus]

i think, cmbo is nearly perfect, but cmbb is a step back. so i am interested in a step forward in cmak.

[fridericus]

i forgot: sure, cmbb is far more better in graphics then cmbo, but this game lives from its realism. nice graphics are in most games

[JasonC]

To deal with the road with bends problem, what is really needed are some "follow terrain" movement commands, especially for roads, that are treated as single waypoints. Perhaps as a "shift" key option when selecting the speed? Then if you "shift-fast" to a road location endpoint on the same road you are on, the AI drives fast to that waypoint staying on the road instead of in a straight line. It counts as only one waypoint for delay purposes.

As for rotating speed, I too find it slow. Part of the problem seems to be the way rotation in place comes up so often given the interface and command giving limits. In real tanks, they'd make most of the turns while in motion, not in place (with a few neutral steer vehicles only partial exceptions).

Rotations made on the move, as opposed to in place, seem slow to me in CMBB and in CMBO before it. I've ridden on M109 and M110 SP howitzers, hardly the nimblest tracked vehicles. They don't neutral steer, and take corners by stopping one track, resulting in a jolting skid that churns up ruts of mud. But they can still make a 30 degree turn in seconds, at speed.

In CMBB, I often find tanks getting thonked at the pivot points in their movement, as a fast mover protected somewhat by the difficulty of rotating to track him slows to pivot himself. It does not seem realistic. It leads to tactical gamey considerations like plotting even the smallest changes of direction behind houses or clumps of woods.

Perhaps a turning model that instead of a smooth rate allows the first 15 degrees or so to happen very rapidly, the next 15 somewhat slower, etc? Then very sharp turns or complete changes in direction would still take time. But minor course adjustments would just happen without even being noticable, while moderate ones would involve inconsiderable delay.

You don't have to worry about this being "gamed" by tons of waypoints, because the delay so many waypoints would cause would swamp any benefit from repeated "access" to the "first 15-30 degrees" "bonus".

Just some suggestions.

[fridericus]

@ jason:

nice ideas

with my leopard 2 i never rotate on one place except on streets. in open terrain i allways make the turns in motion.

[winkelried]

Originally posted by rune:

Your example is perfect, ...

I fully agree with you rune. Giving complicated orders is doomed to fail. We learned as rookies to follow the 3 point orders I mentioned in my last e-mail. But CMBB is quite realistic there: 31 seconds for a soviet conscript squad (1941) to receive orders seems ok to me. Could also be shorter (e.g. "Follow me!").

Now it comes to CMBB. The problem is, that you are punished when giving orders to a formation (especially better trained ones) to follow a simple order, but that you need a lot of waypoints to execute the order (e.g. following a road or going through a passage obligé or driving around an obstacle - a house).

This yields anomalies / punishments which are heavily unrealistic. E.g. take a conscript T-34 platoon (1941).

The platoon commander gives an order:

1) Target: To the ridge 500m ahead of us.

2) How: As fast as possible.

3) What to do there: Seek Hulldown.

Now this takes 39 seconds for the platoon leader to complete (waving flags etc). 31 for the fast, 8 for the Seek Hulldown. Seems ok with me.

The second tank also needs 39 seconds to give the orders. Ok too.

Now the third tank. It needs 62 seconds for two additional waypoints to drive around a house. Is this normal: 23 seconds to say what ?

[winkelried]

... continued:

I strongly believe that there should be some platoon level orders in CMAK. Like give the platoon leader an order and a formation (e.g. column, line, wedge etc.) and then the tanks follow the leader in the defined formation with the current distances and try to stay within the same formation. E.g. by slowing down or moving faster.

[Moon]

Guys, listen to yourself - 23 seconds this, 39 seconds that. TWENTYTHREE seconds is a very very SHORT period of time. Just raising somebody on the radio using the proper procedure can take longer than that in real life. And let's not talk about flag commands or other forms of signal.

Keep in mind - CM runs in REAL TIME. It might be, funny enough, the only *real* real time strategy game in this respect. There is NO abstraction in the passage of time in the game. Next time you watch an action sequence and see some of your platoons maneuver around the battlefield, try to remember this, and then put the 23 seconds you're talking about here in perspective.

Now, like Rune said, the command system is certainly not perfect, and we're the first ones to admit this. The TacAI can only do so much. And some situations show the few shortcomings of the system more than others - but overall the simulation holds.

What makes me... peeved... sometimes is when people make general statements like "it takes 2 minutes to do this or that". No, it doesn't. I am playing CM often, not as often as I'd like (and probably not as often as some of you), but more than any other game I have, and I have never ever seen command delays of 2 minutes.

Not unless you use dozens of waypoints - I tested this recently in fact in another thread - and certainly not when you use regular troops and not conscripts. And while there is nothing in the game to prevent the player from placing 100 waypoints on the map in an absurd way if he likes, it is simply not necessary under any circumstances. Frankly, if you play CM in a way that you feel you need to order your units on their exact location at the exact time, firing the exact type of shell at a precise target, in order to win - you have a totally faulty view of WWII combat, and we will not change CM to adhere to that view, in the old or new engine.

Don't like how long your tanks takes to pivot? Use a movement order instead of rotate. You will notice much faster turn rates. Don't like the time delay to plot a lengthy move down a road? Limit your orders to a couple of waypoints for each turn.

The game design, at its highest level, allows you to do pretty much anything you want, CM is very flexible in this way - but you have to consider the "built-in" pro's and con's. One of the con's for more micromanagement is that you give up reaction times - it takes longer to do stuff.

Many "complaints" I have seen in the past three years about various aspects of CM are, in my view, addressed wrongly, as they are not shortcomings of the game, but "shortcomings" that the real commanders of WWII faced as well.

This should not mean in any way that CM is perfect (if it was, we wouldn't need - and already be working on - a new engine), but that sometimes it might make sense for people to check what the real reason is for complaints before suggesting changes to the simulation.

Please note that I don't mean this thread specifically, it's just a general statement (though it sounds awefully lot like a venting, and I guess it is. Time for my medication I think )

Martin

[SFJaykey]

Jason- Agree with you totally on the need for a "Follow Road" or similar type of command. However one thing you've asked for is already in the game: vehicles underway do make small course corrections "at speed," without stopping to pivot. I'm not sure what the cutoff is in terms of turn angle but it seems like 10-15 degrees. Because of this it can sometimes be faster in terms of getting from A>B to plot an additional waypoint, in order to make smaller turns, rather than a more acute turn that requires a stop & pivot.

[Keke]

Originally posted by fridericus:

i think, cmbo is nearly perfect, but cmbb is a step back. so i am interested in a step forward in cmak.

Originally posted by fridericus:

i forgot: sure, cmbb is far more better in graphics then cmbo, but this game lives from its realism. nice graphics are in most games.

Sorry fridericus, but if you think that CMBO is more realistic than CMBB, then the problem is in your concept of realism.

[winkelried]

Originally posted by Moon:

Guys, listen to yourself - 23 seconds this, 39 seconds that. TWENTYTHREE seconds is a very very SHORT period of time. Just raising somebody on the radio using the proper procedure can take longer than that in real life. And let's not talk about flag commands or other forms of signal.

Hi Moon, exactly what I say in my post. The 29 secs are ok. Flag waving etc. What I don't understand is the ADDITIONAL penalty I get for my tank commander to solve a simple problem, like driving on a road, around a house. Or does it take you an additional 5 seconds before you leave home for each road bend you need to drive down to the office ?

[winkelried]

Originally posted by Keke:

I think that CMBB won compared to CMBO in the realism with the infantry model - suppression fire, move to contact.

Also the use of armor has won in realism with shoot/scoot, move to contact and seek hulldown. This is undisputed.

Where BB has lost realism, is on the driving around of the heavy gear. Especially when bringing in reserves (e.g. along a road).

Maybe reducing some features in AK would help?

[fridericus]

@ keke:

did you serve with tankforces?

[John D Salt]

Originally posted by Moon:

[snips]

Don't like how long your tanks takes to pivot?

I don't mind too much, but ISTM far too slow for vehicles that don't use the peculiar steering arrangements of the M4 (shared only with some early-war French tanks, I believe).

When we move on to CM:AK there will be substantial numbers of vehicles with Merrit-Brown or clutch-and-brake steering (there are many of the latter in CM:BB), and it would be nice if these could escape a penalty that seems to be justified only in the case of the M4.

However, I don't consider that half as important as fixing the rotation rates for ground-fired guns. These seem to me to be far low. I have seen no justification for the glacial traversing rates in CM:BB, and a ground-fired weapon doesn't have the option of speeding things up by using a movement order in the way suggested for vehicles.

All the best,

John.

[fridericus]

hi, john

i understand the low rotation-speed for heavy guns, but for i dont.

but worse is in my opinion, that a gun needs much time after moving for next shot. if you pull a light gun like pueppchen, atg37 or so by hand, you should have the next shot after only 5 seconds. (also in this case i talked with veterans from ww2)

[flamingknives]

I think that towed guns should move faster, i have to say. On a road, for sake of example, most guns can only crawl. By comparison, when moving laden trailers about by hand, and with a couple of hands working, I can get them up to running speed. Plus I can go the other way without turning round.

[YankeeDog]

In regards to the moving and rotation rates of small, light guns:

The Pueppchen is an awful case example as far as movement speed and setup time goes. It's more a rocket launcher than a gun, and therefore is substantially lighter and needs considerably less setup as far as anchoring it to the ground to keep it from moving on recoil. It is also a fairly short range weapon, so there is less need to carefully level the carriage before firing to ensure accurate shooting.

The 37mm ATG is a better example of a light, crew-movable gun, and an especially relevant gun to cinsider for AMAK - all the major combatants had a gun similar in size and weight (in fact, the American 37mm is largely a copy of the German one).

Even though I'm sure a reasonably well-trained 37mm ATG crew could drop the gun carriage, slap a round into the chamber and fire it off within about 5 seconds, they're going to be extremely lucky to hit anything beyond point blank range. In order to assure proper accuracy, the carriage has to be secured to the ground, the gun mechanism has to be leveled, and the crew also needs a few seconds to do things like take a look at the terrain in front of them and make some educated guesses as to range, etc. As such, CM's setup times for light guns seem quite reasonable to me.

Basically, there is no way even a 37mm gun crew is going to be able to go from pushing the gun across terrain to being ready to get off a shot with any accuracy in 5 seconds. A snap shot at a tank that suddenly rolls around a corner 75m away, maybe. It would be a nice thing if CM were able to model guns tanking close range snap shots at important targets before being completely set up, but the lack of such a feature is hardly ruining my enjoyment of the game for now. It strikes me as a pretty rare situation that would create the opportunity for such a shot.

At least one gun I know of, the 88mm PaK 43, was designed to be able to be fired without completely unlimbering, but firing the PaK 43 in this way placed severe restrictions on it's traverse, rate of fire and accuracy. Again, seeing this in the CM model falls into the category of "would be nice to see, but hardly a game-killer" for me.

As far as rotation rates go, the rates seem about right to me for most guns assuming the gun is deployed in an in-situ setup. The problem is, as near as I can tell, an ATG that is in a carefully prepared, leveled and camoflaged gun pit rotates it's carriage at the same speed as a gun that the crew has just hauled into position and setup as quickly as possible on whatever reasonably flat patch of ground could be found in the immediate vicinity. IRL, clearing and leveling the traverse area of a gun is an important element of building a proper gun position.

It seems to me like this should be relatively easy to model - basically, guns in their original setup positions should get a rotation 'bonus' to reflect a more prepared gun position. I actually think this bonus should apply to both the attacker and the defender - in this way, scenario designers can give the attacker overwatch gun positions in the setup zone that are somewhat more prepared if they so choose. Much like the current camoflage bonus, once a gun moves from its original setup location, it should lose any rotation bonus.

I also have a proposal for dealing with the whole "Follow Road" issue and command delays. For the record, I do believe that, in totem CMBB command delay system is vastly better - more complicated orders should take more time to relay. So long as you're working with radio-equipped vehicles, I don't even see the 'waypoint penalty' for following a road as much of an issue - with radio-equipped vehicles, you're looking at only 1-2 seconds per additional waypoint, which unless you're following an exceptionally curvy road is usually no big deal.

However, especially with non-radio equipped vehicles, those additional waypoints can really add up to a very long command delay just for simply following a road, even if the route doesn't go through any intersections. While not a game-killer for me, it would be nice to see this problem resolved so that players do not get unnecessarily penalized for the logical act of ordering reinforcements to move up a road rather then trundle in a straight line across field and stream.

Fortunately, I think this problem should be relatively easy to solve. There are at least two fixes I can think of that should work within the restrictions of the current game engine:

Option #1: Simply make the command delay for waypoints placed on roads only about half that of a normal waypoint. This doesn't totally fix the problem, and it is open to some gamey abuse, but I think on average it is an improvement. It makes sense, too - roads are obviously easily identifiable landmarks, and so relaying orders using roads as waypoints should be quicker and easier to communicate even if the unit isn't actually following the road. So while a player can 'take advantage' of a 'road waypoint bonus' by ordering his tank to turn as it crosses a road, this isn't necessarily unrealistic. Its a lot easier to communicate "turn due North when you hit the road" than it is "turn due North when you're 200m into the field (200m relative to what?, which field? etc.)

Option #2: CM obviously already incorporates a pathfinding algorithm - the StratAI needs it to plan its movements. Why not use this pathfinding logic to find road routes for a 'follow road' command as Jason has already described? The specifics of such a command might go something like this: If a vehicle is currently sitting on a road, and a player plots a waypoint that sits on another road, then a "Follow Road?" (similar to the current "Follow Vehicle" command) pops up. If the player says yes to "Follow Road?", the AI takes care of pathfinding the shortest route to the waypoint using only roads.

Pathfinding algorithms can be very complex and huge amounts of money and time are spent trying to create better ones for complex problems like routing data across networks, but in this case even a very simple algorithm should be able to handle the necessary calculations - no CM map I have ever seen contains an overly complicated network of roads, so the possible routes are very limited.

The beauty of this system is that the player has to 'take it or leave it' as the AI plots the movement order, which strikes me as a very realistic way of dealing with and order that is supposed to represent be equivalent to a real live command along the lines of "move up the road until you reach the bridge" or whatever. If the player is stupid enough to place a "Follow Road" command between two roads that don't actually connect, then he deserves to have the pathfinding AI go haywire and make is units run around like decapitated chickens.

It seems to me that something along these lines should be eminently possible, even within the restrictions of the current engine. Then again, my actual practical knowledge of programming is rudimentary at best, and my actual knowledge of exactly how the current game engine works is virtually zero, so perhaps I am wrong.

Whatever the case, IMHO, CMBO is damn good, CMBB is even better, and I'm sure CMAK will be better still.

Cheers,

YD

[fridericus]

small guns were never leveled in german army, but only artillery-pieces for indirect fire and 88flak for firing against airtargets. the 88flak could fire without unlimbering and leveling, but the hitchances were low at big distance without umlimbering.

leveling you need only for indirect fire.

[YankeeDog]

A good marksman with a rifle can't set himself up to hit a target 500m away in 5 seconds. I find it very hard to believe that a 37mm gun crew could. At the very least, they're still going to have to adjust the sight mechanism for the location and current conditions. While I've never looked through a German 37mm ATG sight, I have looked through other period sights, and in general any long-range sight mechanism I've ever seen takes a certain amount of setup time to adjust for conditions - you can't just look through the objective and line up the crosshairs like some video game fps. Most sights I have seen aslo have protective hardware designed to protect the lenses and components during travel that have to be removed before the sight can be used.

So setting up the sight mechanism, at least, is very important for direct fire at any reasonable range, and takes at least a few seconds. Ideally, you want to make some careful measurements of wind, range, etc, but a good shooter can certain do all this by 'seat of the pants' if necessary. Leveling is also important, especially for sights were you actually dial in the range to adjust the elevation (like most German WWII-era gunsights). If the gunsight and carriage aren't level, adjustments to range dialed into the sight will not work properly, and will actually throw off bearing as well as elevation.

Small ATG crews might not literally have taken out a bubble level and adjusted the carriage to be pool table flat, but I'm sure they took at least a few seconds to get the gun as level as they could before firing on targets at any range.

What would be helpful is actual documentation showing training standards for light ATG crews and how quickly they were expected to be able to go from moving to ready-to-fire. If I see a training manual saying that 37mm ATG crews were expected to be able to go from pushing the gun down a road to ready to fire at a tank 500m away in 5 seconds, I will believe it. Until then, I'm going to make the semi-educated guess that it takes the gun crew somewhat longer to get ready to fire than it does for me to get ready to hit a target at the same range with a good, high-power hunting rifle (something I have considerable practice with).

While I always find first hand accounts from Veterans very useful and educational, it's generally not a good idea to rely on them for empirical data - otherwise, you end up thinking that all German tanks are Tigers, all German guns were 88s, and Allied pilots could KO Tigers with .50 cal MGs.

It may also be an issue of what your Veteran meant when he said they could move the gun and then be ready to fire in 5 seconds. For example, was he talking about going from pushing the gun a long distance (when presumably ammo and tool would be stowed in some way to make them easier to carry), or just pushing the gun a few meters from a hiding position into a previously selected and prepared firing position?

If the latter, 5 seconds becomes much more believable to me. This kind of 'peek-a-boo' gun movement is currently modeled only very abstractly in CM in the sense that 'fighting crest' gun positions are very difficult to hit - a more detailed modeling of short gun movements, where the gun would not have to be limbered, and the ammo could remain in place, would certainly be welcome.

Cheers,

YD