Two stoopid things: the problem with realism

[GreenAsJade]

I'll submit some save files ... maybe there's something I didn't think of that made this happen ... always possible

On the topic of the room:

lolz.

one word: abstract

This was the whole point being made in my original post. It's hard to have it both ways. If you are representing things "realistically" you generate visual expectations of realism, which cause a grating experience when they don't work properly.

It was good to hear the explanation of how the extra cover bonus works... I didn't know that, and it does explain how the tracers appear to go straight through the guy without killing him. Visually it's still a small bug IMHO. At the time that the game calculates that the shot that intersected the pixeltruppen is not actually a hit, the shot should at the very least terminate, not appear to go on through the other side... (actually, I better take a closer look and make sure my impression that the shots go straight through is correct... it's easy for the eye to be tricked, too!)

GaJ

[Blackcat]

"It was good to hear the explanation of how the extra cover bonus works"

Astonishing is the word I would use. All this talk about rounds hit where they hit, turns out to be rounds hit where they hit unless the pixeltruppen roll a 6 on a saving dice throw. I can see the need for this but it don't really fit in with all the hype (mind you it would be nice to have a manual where such things are explained).

The business in the building is not exactly new, and certainly existed in CMx1. It was mentioned in the manual for either CMBO or CMBB, and on more than one occassion I had troops occupy a buliding only to find later that it contained an enemy unit that they didn't ever see.

[ChrisND]

Hey GreenasJade, do you have a save of that infantry in the room ignoring each other? If so, PM me for my email address. I'd like to see it.

[dan/california]

"It was good to hear the explanation of how the extra cover bonus works"

Astonishing is the word I would use. All this talk about rounds hit where they hit, turns out to be rounds hit where they hit unless the pixeltruppen roll a 6 on a saving dice throw. I can see the need for this but it don't really fit in with all the hype (mind you it would be nice to have a manual where such things are explained).

The business in the building is not exactly new, and certainly existed in CMx1. It was mentioned in the manual for either CMBO or CMBB, and on more than one occassion I had troops occupy a buliding only to find later that it contained an enemy unit that they didn't ever see.

The level of abstraction has been reduced by an order of magnitude or more. An attempt to eliminate it completely would have your computer melting a hole in your desk. But Moore's law is our friend so it should keep getting better over time.

[Blackcat]

The level of abstraction has been reduced by an order of magnitude or more. An attempt to eliminate it completely would have your computer melting a hole in your desk. But Moore's law is our friend so it should keep getting better over time.

Fair go. As I said I can see the need for the stated level of abstraction, but would it have been better if this sort of stuff had been declared up front. If BF feel able to tell us about this stuff when people are wondering and complaining it clealry isn't trade secret material. I don't expect BF to publish their algorithms or deep detail but an overview of how comabt works would have been nice and, perhaps, more honest.

[akd]

Fair go. As I said I can see the need for the stated level of abstraction, but would it have been better if this sort of stuff had been declared up front. If BF feel able to tell us about this stuff when people are wondering and complaining it clealry isn't trade secret material. I don't expect BF to publish their algorithms or deep detail but an overview of how comabt works would have been nice and, perhaps, more honest.

Are you saying there needs to be a warning on the box? Pretty sure this is in the manual, but if not, it has been known and discussed openly for the last 5 years.

As far as multiple shots falling short despite clear LOS from the gunner to the target, I don't think this can automatically be assumed a bug. Imagine a gun elevated 1 foot above perfectly flat terrain firing at a target 1000m away that is also elevated 1 foot above the terrain. Let's also assume that the trajectory for this imaginary gun is nearly perfectly flat, but the dispersion at 1000m is 10 feet. Even if the gunner has the range pegged perfectly, some percentage of the shots fired are going to impact short of the target. And due to the random nature of dispersion, you could see every shot of a given number of shots impact short.

[Blackcat]

Are you saying there needs to be a warning on the box? Pretty sure this is in the manual, but if not, it has been known and discussed openly for the last 5 years.

A warning on the box, no. I can't find any reference to any of this sort of stuff in the manual, if you could point me to it I'd be grateful.

As for it being known about and discussed on the forums, you will forgive most CM players for not reading every post in every thread for the past five years.

I don't have a beef with what the game does and how it does it, I would just prefer it if that information was made clear, up-front, so we could all have a chance of understanding how it works.

[Blackcat]

As far as multiple shots falling short despite clear LOS from the gunner to the target, I don't think this can automatically be assumed a bug. Imagine a gun elevated 1 foot above perfectly flat terrain firing at a target 1000m away that is also elevated 1 foot above the terrain. Let's also assume that the trajectory for this imaginary gun is nearly perfectly flat, but the dispersion at 1000m is 10 feet. Even if the gunner has the range pegged perfectly, some percentage of the shots fired are going to impact short of the target. And due to the random nature of dispersion, you could see every shot of a given number of shots impact short.

Lots of assumptions there that may or may not apply to the situation that GAJ was talking about. Furthermore I don't think anyone has suggested that the circumstances he described can be assumed to be caused by a bug.

[Vinnart]

I hear ya. I had a guy run up to an enemy, kneel down right next to him, doesn't shoot him mind you, then gets up and runs away from the enemy only to get pluged in the back. Unfortunatly AI technoligies, or whtever needs to develop more. It's an imperfect world, but the AI is usually isn't that bad most of the time.

[easytarget]

This forum is an enabler for wasting a lot of time about small inconsequential things.

It's evil, step away before it's too late.

/too late

[Holman]

A warning on the box, no. I can't find any reference to any of this sort of stuff in the manual, if you could point me to it I'd be grateful.

As for it being known about and discussed on the forums, you will forgive most CM players for not reading every post in every thread for the past five years.

I don't have a beef with what the game does and how it does it, I would just prefer it if that information was made clear, up-front, so we could all have a chance of understanding how it works.

I think the fact that every building in the game is a big empty box is the tip-off that something is being abstracted there.

[Blackcat]

I think the fact that every building in the game is a big empty box is the tip-off that something is being abstracted there.

Martyr, I know about abstractions in buildings, they were mentioned in the manual of CMBO or CMBB - shame BF didn't carry on the tradition because then GaJ might not have felt as he did when he started this thread.

P.S. My post which you quoted was about the other issue raised by GaJ, not the buildings thing.

[Holman]

I just looked at my post above, and I see I struck the wrong tone. I sounded snarky when I meant to be droll. My apologies for that.

[Blackcat]

I just looked at my post above, and I see I struck the wrong tone. I sounded snarky when I meant to be droll. My apologies for that.

No sweat, mate. I did wonder if I was reading you aright.

[Macisle]

-Saw the vid. I have no problems with what happened. As others have said, building interiors are abstracted. In the confusion inside a building with different rooms, furniture, darkness and gunsmoke, the German got the drop on the Ami. Simple as that. Frustrating for the player, but shouldn't be hard concept to take in and accept.

Given the above--nothing unrealistic here.

[guachi]

The level of abstraction has been reduced by an order of magnitude or more. An attempt to eliminate it completely would have your computer melting a hole in your desk. But Moore's law is our friend so it should keep getting better over time.

CM:BN uses one core of my four core processor. Soon, I'll be able to own an eight core processor, and then CM:BN can use 12.5% of my maximum processing power.

Turns in CM:BN take, in the scenarios I've played so far, no more than 30 seconds. In CM:BO, I had turns for battalion sized battles take over 15 minutes to calculate.

Let's just say I think we are already there in terms of having a computer that can handle greater fidelity in modeling.

[gunnergoz]

I don't think the challenge lies in fidelity of modeling, more in getting AI to act anywhere near as smart as the dumbest human without investing a small country's GDP in the programming development process...otherwise, we'd be playing against unholy smart computer games by now. Instead the most common complaint people have about games is their lackluster AI.

That said, the AI in this game is quite good compared to most and seems open to even more development and improvement.

[dan/california]

CM:BN uses one core of my four core processor. Soon, I'll be able to own an eight core processor, and then CM:BN can use 12.5% of my maximum processing power.

Turns in CM:BN take, in the scenarios I've played so far, no more than 30 seconds. In CM:BO, I had turns for battalion sized battles take over 15 minutes to calculate.

Let's just say I think we are already there in terms of having a computer that can handle greater fidelity in modeling.

I am under the impression that BFC really doesn't want to cut off a the trailing edge of its customer base. They also don't want to rewrite the engine until the performance gain is truly eye popping. High end video cards can already be used to run things like LOS checks with a LOT of cores. But this is nowhere near common yet.

Tricks like http://developer.nvidia.com/content/gpu-ai-technology-preview

[GreenAsJade]

I don't think the challenge lies in fidelity of modeling, more in getting AI to act anywhere near as smart as the dumbest human without investing a small country's GDP in the programming development process.

The challenge that I was observing that racking up the level of realism brings these stoopid hard to deal with behaviours that you don't have to worry about at the higher level of abstraction.

Of course, everyone knows this, so I was just motivated to post to share the experience and compare notes on how close or not the game currently us to matching it's modelling (I mean the AI behaviour modelling human behaviour) to the level of realism of the artwork.

GaJ

[WillLight]

The level of abstraction has been reduced by an order of magnitude or more. An attempt to eliminate it completely would have your computer melting a hole in your desk. But Moore's law is our friend so it should keep getting better over time.

Moore's law hasn't been true for quite a while now, at least not in the classic sense. Rather than faster CPUs we are getting more of them (aka cores). If a program doesn't take advantage of the parallelism then it isn't gaining the benefits of having multiple cores. And a game (as opposed to say a web server) is not a problem that lends itself very well to parallelism. Typical games follow an endless loop of render-input-decision-render-input-decision-etc. Without knowing a single thing about how CM is implemented, but being a know-it-all IT consultant I would venture that changing the design to parallelise some parts of the loop isn't as easy as some might think it is. As usual I'm probably talking out of my arse, but then what else is new?

[Other Means]

I was wondering if it’s possible to virtualise a home PC so it seems single core, then running BN on it...I really doubt it, as there's a lot of low-level calls going out but maybe, just maybe...

[Sarjen]

The level of abstraction has been reduced by an order of magnitude or more. An attempt to eliminate it completely would have your computer melting a hole in your desk. But Moore's law is our friend so it should keep getting better over time.

Not if the game would support all cores of the CPU

[YankeeDog]

In re: multiple cores

Steve made some comments on this a while back. IIRC, they went something like this:

The potential benefits of rejiggering to code to take advantage of 2 cores was only on the order of 10-15% or so, but there were more potential benefits if the code were able to run on 4 cores or more.

So, given that a large proportion of the CM customer base is probably still playing the game on dual core machines (or less), BFC doesn't think it's worth it to put in the considerable time and effort required to allow the game to run on multiple cores... for now. Once the majority of the customer base has 4 cores or more, this equation will probably change, and BFC will put a higher priority on multi-core coding.

In any event, such a major shift is something that will certainly come in with a major new title release (Bulge, Ost Front, CM Modern 2, etc.), not as an update to an already released game.

Also, I agree that levels of abstraction you see in the game don't necessarily have anything to do with available processing power. While I'm sure more powerful computers will make modeling some things easier, there's also the fact that BFC is a two coder shop, and things like AI behaviors are EXTREMELY complex to code and debug. I don't think the abstractness of building interiors, for example, is so much due to lack of computing resources as it is the fact that having realistic representations of building interiors in the game, and getting the soldier AI to interact with said interiors in a realistic way, would be a HUGE drain on BFC's programming resources. So getting realistic building interiors in the game would require postponing or delaying other significant game features.

[sburke]

In re: multiple cores

I don't think the abstractness of building interiors, for example, is so much due to lack of computing resources as it is the fact that having realistic representations of building interiors in the game, and getting the soldier AI to interact with said interiors in a realistic way, would be a HUGE drain on BFC's programming resources. So getting realistic building interiors in the game would require postponing or delaying other significant game features.

That and they'd have to listen to us b***h about how the kitchen table doesn't look like Norman furniture, or the couch needs a nicer slip cover and don't get me started on the interior to Cafe Grammont!

[Blackcat]

That and they'd have to listen to us b***h about how the kitchen table doesn't look like Norman furniture, or the couch needs a nicer slip cover and don't get me started on the interior to Cafe Grammont!

Can you imagine the threads that would go on about whether a .30 would penetrate an oak dinning table and, in the cafe, whether a German officer enjoying the pleasures of room 6 (with the flying helmet and wet celery) could realistcally be expected to "spot" a US squad in the corrdor outside? Some level of abstraction is best left in.