Jump to content

The M1A1 was destroyed by a near miss.


F.T

Recommended Posts

  • Members

FT's arguments are valid. Where we diverge are the conclusions with respect to the damage model in Steel Beasts.

 

All models are wrong. Some models are useful.

 

Steel Beasts is not designed as a simulation tool for quantitative stochastic analysis of battle damage results. It's supposed to educate crews about "this is likely bad for you/for the enemy" and "here you can somewhat trust your armor to protect you". I will concede that our model is massively overengineered for this purpose, and that the higher-than-necessary level of detail might help to create the misleading impression that our model will deliver much more detailed and at the same time highly accurate predictions. We have a pretty good model for APFSDS and HEAT jet interactions. With full-bore AP rounds things get iffy, but fortunately they play only a minor role these days. the HESH model is, as of yet, still incomplete. The HE/frag model works very well in combination with infantry in the open, but not as detailed as I wish it would when it comes to tanks and fragments, and most certainly we do not model the propagation of shockwaves across different mediums (such as air-armor-air transitions, or inside non-homogeneous armor arrays). We have squeezed from the current model what we can . A future model might be able to improve the overall predictive quality a little bit in selected areas.

 

 

 

But let's not forget the whole forest while studying the bark of individual trees. A model is only useful if it can be parameterized properly, and for a lot of things these parameters simply aren't available; either, because they haven't been studied, or because the study results are classified, or because there's no decent standardization. We need to use a model that we can apply to our whole fleet of virtual vehicles, ranging from main battle tanks to motorcycles and flimsy quadcopters, or helicopters, where the protection concept is based on massive redundancy rather than an outer hard shell. We need a model where the research effort can be kept at a reasonable level, a model that suits the limitations of a real-time simulation, a model that is fit for the purpose of educating crews and junior officers about the lethality of the contemporary battlefield.

Link to comment
Share on other sites

On 10/9/2022 at 12:50 AM, Assassin 7 said:

Yeah like my Tank in Iraq hit by 3 buried 155 rounds. All Crewmembers survived this hit btw. As stated above I think the overpressure is an overkill. Its not as bad as many make it out to be. I agree with Ssnake comment "Don't read too much into Steel Beasts battle damage."

 

IED in Iraq.jpg

Where was the detonated IED placed relative to the tank?

Link to comment
Share on other sites

On 10/9/2022 at 5:06 AM, F.T said:

This article talks about the effectiveness of artillery against armoured targets. Including but not limited to fragments increases the kill range. But it's still not directly related to what I'm talking about.

You didn't seem to grasp the topic you started, now you switch to the next thing?

Odd...

Link to comment
Share on other sites

3 hours ago, F.T said:

Where was the detonated IED placed relative to the tank?

Buried next to the road around 4 feet. When the IED exploded the Tank left track lifted up about a foot off the ground. I was in the Gunner’s seat and was given a concussion for a few minutes. My Driver didn’t get hurt and was able to drive the Tank slowly back to the FOB as the Power Pack switched into Protected mode. My Loader and Tc suffered a Concussion too but was okay after the fact. The Tank behind us said they seen our Tank lift from the ground and disappear from the dust and smoke. 

Link to comment
Share on other sites

7 hours ago, Assassin 7 said:

Buried next to the road around 4 feet. When the IED exploded the Tank left track lifted up about a foot off the ground. I was in the Gunner’s seat and was given a concussion for a few minutes. My Driver didn’t get hurt and was able to drive the Tank slowly back to the FOB as the Power Pack switched into Protected mode. My Loader and Tc suffered a Concussion too but was okay after the fact. The Tank behind us said they seen our Tank lift from the ground and disappear from the dust and smoke. 

You are so lucky. If the explosion had happened under the chassis, it would have been very bad.

Link to comment
Share on other sites

  • Moderators

It may have been covered already (I haven't read every post), but the overpressure kills from artillery, airstrikes, IEDs, mines etc are not strictly representing the crew being harmed inside the vehicle. Did some of the crew get wounded? Maybe, maybe not at all. But like the vehicle hitting a tree at speed causing a crew member to be randomly "disabled" in SB (with the probability being dependent on the tree's mass and the vehicle's mass and speed), this represents wounds that take them out of the fight from say, an injury like a broken bone or whatever. People often interpret this as the crew getting killed, but that isn't the case.

 

Overpressure is a similar abstraction to structural damage to the vehicle (not the crew), but is far more detailed. The point is though, like the tree collisions, it is representing something that we cannot directly model within the limitations of SB. I will try to explain...

 

In the case of a vehicle getting killed by overpressure, it primarily represents damage to the vehicle's structural integrity that occurs that knocks the vehicle out of action. What we have is a system where there is a % of a kill calculated from overpressure that depends on the overpressure force of the explosion based on the distance from said explosion, depending on the overpressure protection level of the target type. MRAPs (at three different protection levels in SB) being the best to survive such explosions for obvious reasons, followed by "mine protected" vehicles, followed by ordinary tanks, then light armored vehicles, then unarmored vehicles, and so forth all the way down to micro drones.  All of these different protection levels are represented in SB.

 

In SB this translates to a "100% threshold", that is the radius distance from an explosion that there is a 100% "kill" (destroyed) chance from the point of impact. From that 100% radius, it then rapidly decreases in probability to 0% chance over distance, the further away from that radius you get. Meaning, its not strictly a yes or no situation here, its a 100% "yes" at a certain distance, then there is distance it is possible before becoming "no", all depending on the target's protection and explosive power of the round (which in itself is calculated from many variables that even factors in the type of explosive - TNT, RDX, HMX, C4 etc.).  So, the distance on what would be considered 100% is very detailed.  This isn't some arbitrary number happening here, its something that is carefully handled while allowing some "chaos" to exist. 

 

The very same explosion in the image is likely a low % chance of an overpressure kill happening at that distance on that target type, of that round type, and it was by pure chance that it happened, otherwise it would have likely been a long list of non-kill damages as the result. Bad luck, if you will, but of course had it not happened then likely there would have been crippling mobility damages. So in this case, a kill occurred and this was allowed to exist in the values because we want there to be cases where it represents the "gray area" where fundamental structural damages can occur (that don't exactly include in the relatively simplistic damage types that are explicitly modeled ("engine", "suspension" and so on -- we are very limited in that regard).

 

Actually, out of curiosity and the desire to test if everything looked OK here, I took some time to do a manual calculation to test the situation. Based off an estimate of distance in the image (based of how wide I know the M1 tank is), the 3OF-83 BB-HE at that distance comes out to something like and overpressure kill probability of 31% at ~4m radius, dropping off to 8% at ~5m radius, dropping off to 0% at ~6m for an ordinary (non-mineprotected) tank, with only a direct hit from a single 155mm HE being something like a ~70% kill chance. This means that the explosion has just barely occurred inside the non-zero radius, which is not unreasonable. And to have overpressure, which is a very complicated thing, be represented in a dynamic way that allows some randomization "die rolls", this is about as best as can be expected without it being a completely binary approach (strictly a 100% kill or 100% no-kill result). Of course someone might question the exact probability over distance to 0%, but these levels are not unrealistic here and we do not entertain subjective opinions - we instead base it off known cases to establish the 100% situation, then let the math do the rest without putting a hand on the scales. 

 

Hopefully that simplified explanation helps explain how overpressure works in SB. It is quite detailed (probably way more detailed than it really needs to be, but hey -- I think its one of the best features in SB), and it allows for all sorts of overpressure protection levels to be represented, 100% preventing an MRAP from being destroyed by the same explosion - saving its infantry inside to allow them to dismount and defend in an ambush, for example. But it allows for normal tanks to have a chance of being destroyed by near 155mm artillery explosions from IEDs (the more artillery rounds in the IED explosion the higher [but not exponentially] the destroyed probability radius). It also allows for actual mine protected vehicles to have a better chance of avoiding destruction, by decreasing the radius distance at which it drops to 0% kill, if that makes any sense, which is quite nice. 

Link to comment
Share on other sites

On 10/8/2022 at 10:33 PM, Mirzayev said:

@F.T I'd give this a read: 

 

 

In short, a tank being destroyed by a near-miss from a 152mm HE shell isn't outside the realm of possibility. Especially when said unit is hit with 152 fragments. 

 

image.thumb.png.d821c5fe8970f18455aa44617979bc80.png

 

Shown here are the results of a test with a Bradley. Both the main gun and the TOW launcher are definitely inoperable. I'd imagine the track is pretty torn up as well.

 

image.png.210cdaf74cae36dce2260d230ade8c6c.png

 

Such as with this T-55, which was hit by fragmentation from a near-miss of a 155mm HE round impacting 30 meters away. 

 

 

Such as with this T-55, ya, I need new glasses :)

Link to comment
Share on other sites

  • Members

Maybe my answer was needlessly terse; the point is, we're trying to use a model that is sufficiently easy to work with, gives adequate results over a wide range of situations, based on hard parameters that are documented (such as the amount and type of explosive filling that a munition may have).

While a 152mm artillery shell with 8.3kg RDX exploding about two meters away may not guarantee disabling a tank, I find nothing wrong with the possibility that it does. We've been studying that very same article from the US Artillery magazine (and other sources). Our model does not simulate wave front propagation, reflections, or impulse transfer. It must be simple enough to perform adequately in a real-time application. I explained all this already.

 

What we won't do is a complete disclosure of all our methods and models so that lazy competitors can simply copy the stuff we're discussing here rather than inventing their own solutions.

Link to comment
Share on other sites

49 minutes ago, Ssnake said:

What we won't do is a complete disclosure of all our methods and models so that lazy competitors can simply copy the stuff we're discussing here rather than inventing their own solutions.

 

Considering this was the reason for the answer, I think it being terse was quite appropriate. 

Link to comment
Share on other sites

3 hours ago, Ssnake said:

What we won't do is a complete disclosure of all our methods and models so that lazy competitors can simply copy the stuff we're discussing here rather than inventing their own solutions.

This decision is reasonable. Although I'm just curious about the manual calculation.

Link to comment
Share on other sites

9 hours ago, F.T said:

This decision is reasonable. Although I'm just curious about the manual calculation.

 

I think this answers your question

 

12 hours ago, Ssnake said:

What we won't do is a complete disclosure of all our methods and models so that lazy competitors can simply copy the stuff we're discussing here rather than inventing their own solutions.

 

You might have to derive your own formula.

 

Sorry.

Edited by Hedgehog
Link to comment
Share on other sites

seems like there's a lot of fuss caused by some misunderstandings by perhaps some poor wording in steel beasts.

perhaps the word "killed" should be changed to "incapacitated" or "out of action" to reflect the state of the vehicle more accurately. 

especially since it's a "catch-all" for multiple simulated states, from "everyone inside are dead but vehicle can be repaired"  to "vehicle is dead and on fire"

Link to comment
Share on other sites

1 hour ago, dejawolf said:

seems like there's a lot of fuss caused by some misunderstandings by perhaps some poor wording in steel beasts.

perhaps the word "killed" should be changed to "incapacitated" or "out of action" to reflect the state of the vehicle more accurately. 

especially since it's a "catch-all" for multiple simulated states, from "everyone inside are dead but vehicle can be repaired"  to "vehicle is dead and on fire"

Used by most armies at unit level sent to repair units.

 

Operational condition of an vehi indicating it can perform at least one and potentially all of its designated missions. Mission-capable is further defined as the sum of full mission-capable and partial mission-capable. Also called MC.

 

MC-1 fully MC

MC-2 user repairable 

MC-3 1st line repairs

MC-4 Repair depo required

MC-5 NON repairable (KIA)

Link to comment
Share on other sites

22 minutes ago, 12Alfa said:

Used by most armies at unit level sent to repair units.

 

Operational condition of an vehi indicating it can perform at least one and potentially all of its designated missions. Mission-capable is further defined as the sum of full mission-capable and partial mission-capable. Also called MC.

 

MC-1 fully MC

MC-2 user repairable 

MC-3 1st line repairs

MC-4 Repair depo required

MC-5 NON repairable (KIA)

 

right.  SB basically takes 3 of these, and merge it into the single word "killed". 

so changing "killed" to "mission incapable" would signal that this doesn't mean MC-5 (KIA) but that it is merely unable to perform it's intended tasks. 

Edited by dejawolf
Link to comment
Share on other sites

  • Members
29 minutes ago, dejawolf said:

seems like there's a lot of fuss caused by some misunderstandings by perhaps some poor wording in steel beasts.

perhaps the word "killed" should be changed to "incapacitated" or "out of action" to reflect the state of the vehicle more accurately. 

especially since it's a "catch-all" for multiple simulated states, from "everyone inside are dead but vehicle can be repaired"  to "vehicle is dead and on fire"

I got to say, the only customers with deep inquiries about our damage model are PE users. Not once has an army customer, in all these years, raised the topic (be it that they inquired about the model itself, or the wording to characterize the damage). It seems to me that these finer distinctions about how much a unit is out of action is not very relevant for the intended purpose of SB Pro as a tool for training and education. It may raise points of discussion similar to this in AARs, and that's then where some of the learning about the finer points of "kill" definition may come into play.

Steel Beasts is no tool to predict battle outcomes in great detail, and we do not aspire it to become one. We try to become better in all that we do because we take pride in our work. But inventing an electronic crystal ball to divine the future is explicitly excluded from our list of development goals because it's a completely elusive goal. You know the decisive factors of a real battle only after a real battle was fought, and analyzed. You can identify many contributing factors in advance, but at the end of the day there's always the chance that some seemingly irrelevant detail turned out to be pivotal, and be it just the infamous "human factor" (morale, battlefield psychology), or a random mechanical fault that makes a tank lose track at a most inopportune moment. Ancient people knew that "the gods" or "fate" had a big swing in all they did, especially in battle. It's a folly of modern times that some believe that everything is predictable. Some things are, others might be one day, but there's always that little extra surprise that you identify only afterwards.

Link to comment
Share on other sites

7 minutes ago, Ssnake said:

I got to say, the only customers with deep inquiries about our damage model are PE users. Not once has an army customer, in all these years, raised the topic (be it that they inquired about the model itself, or the wording to characterize the damage). It seems to me that these finer distinctions about how much a unit is out of action is not very relevant for the intended purpose of SB Pro as a tool for training and education. It may raise points of discussion similar to this in AARs, and that's then where some of the learning about the finer points of "kill" definition may come into play.

Steel Beasts is no tool to predict battle outcomes in great detail, and we do not aspire it to become one. We try to become better in all that we do because we take pride in our work. But inventing an electronic crystal ball to divine the future is explicitly excluded from our list of development goals because it's a completely elusive goal. You know the decisive factors of a real battle only after a real battle was fought, and analyzed. You can identify many contributing factors in advance, but at the end of the day there's always the chance that some seemingly irrelevant detail turned out to be pivotal, and be it just the infamous "human factor" (morale, battlefield psychology), or a random mechanical fault that makes a tank lose track at a most inopportune moment. Ancient people knew that "the gods" or "fate" had a big swing in all they did, especially in battle. It's a folly of modern times that some believe that everything is predictable. Some things are, others might be one day, but there's always that little extra surprise that you identify only afterwards.

 

i agree. Steel beasts is not a tool to predict battle outcomes in great detail. and even if we work on it for another 20 years, it still won't be a magic eight ball. 

i think changing the wording will reflect this more accurately. 

i would say the PE customers are important too. while PE doesn't bring in the big bucks, it does give Steel beasts valuable PR, since many PE players are army veterans, Ex-tankers, and sometimes even 

active duty tankers in militaries who are not customers of Steel beasts pro. and if we treat them respectfully, they might put in a good word or two, and bring in a completely new army contract. 

Link to comment
Share on other sites

We often hear during online sessions when the Co asks for a status update (Sit rep) ..I have no tracks, FCS is gone, driver KIA, fuel leak...and I'm still firing into the enemy postions at their tanks. We use a long version, we Could use the MC.

 

It depends on the commander. On asking for a status, one may get "Just look at the map",  failure to communcate is why  missions fail for one side, and if  I may add, who to pick for your side, one wants the best transmitters of battle info, even though there is Emotional Damage to do so.

 

I have no issues with the current status.

Link to comment
Share on other sites

55 minutes ago, Ssnake said:

It's a folly of modern times that some believe that everything is predictable. Some things are, others might be one day, but there's always that little extra surprise that you identify only afterwards.

 

i am not sure who espouses that particular view, but somewhat distinct or perhaps to put a finer point on it is not that everything is predictable but that everything is deterministic; even probability is deterministic- that is to say, there is no state of things at any given moment where something that happened could have happened a different way than the way they occurred- everything that happens occurs the way they did given the only the situation that could have happened. i have personally seen what they are doing with machine and AI learning in non-gaming applications, primary in financial transaction modelling because of the regulations in the finance industry to have a comprehensive anti-money laundering detection practices in place, and the models are getting more accurate all the time- flagging patterns of behavior against known human behaviors and you get fairly consistent results; one bad actor thinks he's being clever, not understanding against him are all the other times a bad actor used a similar technique to evade detection. multiply that sort of behavior over time, hundreds, thousands, or millions of times, and it all becomes a system- far from what people believe is essentialy unique about human beings or free-will, as if they were unique or distinct from the environment around them- and it also works quite well in the insurance industry because of their actuarial models- if those things weren't at least accurate much of the time, the insurance industry would have failed a long time ago. contrary to that we're seeing the opposite happen, and the industry of assigning risk to each and every one of us all has become an important area of interest for people to make money off of our consumption habits and life maps from beginning to conclusion.

 

back to the point- and i think you allude here as well, your PE customers have additional benefit over your professional military customers in providing feedback of where there might be strange or undesired behaviors, bugs etc, and have the time necessary to report those perceived anomalies and submit them for analysis

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
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...