Rocket effectivness

[maturin]

Physics that no one bothers to simulate.

 

Actual militaries make decisions with probability.

 

Danger Close radius is calculated by the distance at which a friendly casualty is a chance within 100 or 1000. The accuracy of a piece of artillery is its probability-rating of CEP. And Casualty radius for grenades. And hit rates for ATGMs and on and on.

[GGTharos]

His point is that for ED to reasonably show differences between weapons, they need to know the physics that result in wounding or destruction mechanics.

 

At minimum you would like to know average fragment amount, size, mass and velocity after detonation, the rest can be estimated.

 

For bonus points, fragment standard deviation for mass and size would be nice as well as the frag pattern, and for extra bonus points, physics of explosive landscaping would be fantastic :)

 

I don't want to go too far into 'perfect modeling', you can do it even with probabilities and you can probably derive various mechanics for dealing damage to different targets - eg. heavy mavericks use a blast-frag warhead which has a different mechanic from a blast, or frag, or shaped warhead.

 

Why do we want more of this instead of the simple stuff you mentioned? Because it's also the very same physics that AAM/SAM warheads use, and I know that on some level, ED wishes to simulate them.

[Fri13]

Pardon me if I sound impolite now, but are you a programmer? If yes: Have you worked on such a thing already? It doesn't sound like it.

Doing such things halfway realistically for multiple objects at the same time in real-time is hard on any CPU.

 

No, not a game programmer but enough to know what is possible as what has already been done long time ago and knowledge of what kind things can be done.

 

And you need to ask yourself what is your definition of "real-time"? Like is it realtime like in RTOS (Real Time Operating System) (talking about in microseconds) or like in common FPS games (talking about 100ms latencies).

 

I would not require a so deep modeling and simulation for shrapnels flying from explosion that it requires a 2 hour calculations when there are 1000 shrapnels flying so they are got correctly modeled.

But I would not either accept a typical online FPS game style where shrapnels are modeled but they are only a visual ones.

 

Every tree, building, vehicle etc object in DCS needs to be calculated, tracked and calculated again. Using different manners what makes it possible to have a games like DCS.

There is no reason to draw objects what you don't see, until you look at that direction. But it is still required to calculate objects behavior even when you don't see something, but when it doesn't "matter" you, it can be done in much smaller impact to CPU and quickly.

 

Now a normal case is that player fly a plane and drops a bomb at target 2km below him.

The required code to calculate shrapnels effect is still same as if player would be following bomb drop or being in one of the targets. But the impact would be smaller when you look it far away as there is no need to render individual shrapnels but just the "big explosion".

If player is looking closer the situation, then everything else around player can be just calculated without drawing it and focus more about the explosion drawing itself.

 

A todays typical computer can do so many calculations in second that a rocket explosion with 512 shrapnel is easy task to do in few milliseconds. It isn't a hard thing to do even if there would be a 32 rockets.

 

Why it has low impact? Because we are dealing about ways to actually "blend time" to do calculations, it isn't required to be done realtime (every microsecond track and draw every shrapnel velocity, direction, rotation etc). The calculations can be started already a second before impact. The effect can be pushed even to a 1-2 seconds after explosion (fire and smoke etc) or even further as if player is not close by or watching at that direction, the calculations can be done in longer batches.

 

There are own dedicated algorithms for collinsions, flying objects and other physic modeling.

It doesn't require to take in use a complete physics engine (like PhysX or Havoc etc) to be able a have thousands of flying objects with physics modeling.

 

The main problem would be what level of modeling players would accept for shrapnel modeling?

A real real-time or delayed one? Does it need to be a visual one (that adds lots of requirements) or can it be a calculation only? How well every shrapnel needs to be modeled, like is every explosion really required to be a full physics simulation or can it be just calculated with information what weapons manufacturers give like a warhead has pre-fragmented shell what includes 512 small shrapnels and 128 bigger shrapnels? Is there need to calculate shrapnels what are always flying or just shrapnels what are flying toward game objects (cars, airplanes, infantry etc)? As there can be done easily a randomization (like it actually would be so easy to do a real randomization (this by security point of view)) to spread shrapnels from point of explosion, collect only the ones what are going toward important game objects, valuate objects priority (example a running infantry should be calculated before a shutdown MBT without crew) and then start doing shrapnel calculations individually like penetration, flight distance, time etc. And for that it is enough to use existing tables from penetrations.

 

Like how many simmer would be happy that a rocket is calculated with a 1024 shrapnels from what 128 are "big" and rest of them "896" are "small". All shrapnels are first distributed randomily different directions, checked the possible target(s) (trucks, infantry, MBT etc) plane direction and only shrapnels heading at those planes are taken for calculations and rest simply forgotten. Calculate a damage caused by shrapnel "size", based the distance of point of explosion and target "armor"?

 

Example if the situation is that a S-8 rocket hits 10 meters rights from un-armored truck.

From 1024 shrapnels if it is distributed to whole sphere (of course there should be own directional modeling but lets just think now about sphere so it is easy) only a 6 shrapnels is flying toward the target and from those only one is "big" and rest (5) are "small".

Would it be enough for simmers to consider the fragmentation modeling over current (non-existing)?

As after all, most vehicles have basic level of impact modeling. They have healtbars and calculations seem to be "a live", "mobility damage" and "destroyed"?

 

Calculations from 1024 shrapnels, based distance, direction and size is very very low impact for CPU. Even when firing 32 rockets it would be small impact as most of the shrapnels and their calculations can be forgotten before the explosion is even visually happened so from 32768 shrapnel the calculation can be done only for example 768. Some of them based target the result is simple like any shrapnel hitting a infantry is a kill so if infantry is on target list, they are dead. If there are heavily armored targets like MBT, the "light" shrapnel calculations toward them can be forgotten and only calculate impact of "heavy" shrapnel. And because we know that a shrapnel would not penetrate the front or side armor the vehicle is operational and because we don't have complex modeling in them, we can just check does the targets tracks get cut, optics braken etc (if there even exist such modeling).

 

And there maybe doesn't even need to be any kind shrapnel modeling like that. Lets just randomize the effect as well. We know amount and "size" of shrapnel flying toward target from specific distance, lets throw a dice and see what target suffers based "HP".

[Dejjvid]

We wouldn't even notice if there was a fragmentation zone outside the blast zone.

In the fragmentation zone you can do a really quick randomization based on IRL specs.

 

A piece of cake for a programmer to do. And the only boring work is the .lua for each weapon that need new fields for frag damage, spread and velocity.

 

Modeling each fragment is an absurd way of trying to tell us fragmentation is fine.. We're not asking for an over kill..

[Fri13]

Found the PDF about artillery test against MBT

http://sill-www.army.mil/firesbulletin/archives/index.html

(I don't directly link to the PDF as I am not sure would domain owner like it.)

 

The article is "2002 November-December" and it is pages 8- 11 with topic "Who Says Dumb Artillery Rounds Can't Kill Armor?" by Major (Retired) George A. Durham.

 

And I did remember wrong, tests were not from WW2 era but from late '80s.

 

This article reminds the Army of the lethality of area fire

Field Artillery by examining the background of modeling and

simulations data used to portray artillery effects in a study

beginning in 1988 and reviewing the results of this four-year

test of the effects of the basic Soviet 152-mm and US 155-mm

rounds.

 

A literature search conducted

before the test found a Soviet report titled “Firing for

Effects Against Strongpoints.” This report became the “base

case” for comparing US and Soviet effectiveness estimates.

The Soviets claimed that 122-mm and 120-mm mortars and

152-mm artillery rounds achieved very high levels of damage

against tanks and armored personnel carriers. The explanation

of the disparity ranged from the Soviets simply overestimating

their effectiveness to the difference in the way damage

criteria were established. Other possibilities offered were that

the database for US models was incorrect or not used properly.

Further research revealed the Soviets had a large database

developed by an extensive live-fire test program.

 

A basic difference existed in the way damage criteria were

used. The Soviets had two criteria: “damaged,” which means

unable to continue to fight, and “destroyed.” The US used

“estimated time of repair” as the yardstick for damage in

models. If damage could be repaired in less than 30 minutes,

it was not counted; 30 minutes to one hour was assigned as a

value; one hour or more had a value; and so on.

 

The targets were manikins

placed in fighting positions, US trucks, M113 and M557

armored vehicles, and M-48 tanks. Several different computer

models were used to predict results. The test was fired

three times using 56 HE rounds with point-detonating (PD)

and variable-time (VT) fuzes.

The resulting effects on the trucks and personnel were close

to model predictions. However, the effects on the armored

vehicles and tanks were significantly higher than model

predictions.

The model predicted 30 percent damage to armored vehicles

and tanks; however, 67 percent damage was achieved. Fragmentation

from the HE rounds penetrated the armored vehicles,

destroying critical components and injuring the manikin

crews.

 

The SAE live-fire tests proved that HE

fragmentation rounds are very effective against the most

modern US and Soviet stationary armored systems.

[Neon67]

so that means if a closeby arty shell can damage a tank, a mk82 or so should do it as well ?

[Fri13]

so that means if a closeby arty shell can damage a tank, a mk82 or so should do it as well ?

 

Almost any explosion with fragments parts should do it when 30-100m distance from target.

 

Of course soft targets gets destroyed but tanks like IFV and MBT gets at least disabled, IFV more likely destroyed. Lightly armored vehicles like APC gets destroyed.

 

Crew of modern MBT of course survives from fragments a live.

[Dejjvid]

Modern tanks are packed with sensitive sensors, if you destroy them the tank is essentially useless. I can promise you that tank crews will haul ass if they are fired upon with 12cm+ shells

[PFunk1606688187]

What I have questioned is whether a hack today is better than a proper implementation tomorrow.

 

Well it would be better for me because then I wouldn't groan audibly everytime I thought about having to carry rockets for anti-infantry duty.

[Dejjvid]

If we get a quick fix now we can wait a year or two for a proper fix. It's not like the proper fix is coming tomorrow anyways.. ;)

[Dombl]

I don't see what makes you guys think the fix is in anyway improper.

It is sufficient, and any more seems like overkill to me.

 

Make a fragmentation zone, find targets in the zone, ignore any that would not take fragmentation damage, use their distance from the center of the zone (and maybe some input from height difference or the like) to determine the chance that the object is hit.

Optionally, if it is hit, draw a trace from the center of the zone to the object, if nothing else is inbetween, it is hit, otherwise the encountered object is hit.

If you want extra detail for aircraft, run a hit trace for individual parts (wings, tail, nose).

 

Simple, keeps traces to a minimum, and good enough to simulate something invisible.

Anything more would be like simulating what doors are open or closed inside buildings. Sure, it's more realistic, but completely useless and a waste of resources.

[shagrat]

If we get a quick fix now we can wait a year or two for a proper fix. It's not like the proper fix is coming tomorrow anyways.. ;)

totally agree. I don't care about precision modelling,a simple fragments zone and rolling virtual dice for damage points is better than what we have now. remember the vehicles use hitpoints anyway.

Currently infantry is virtually immune against any anti infantry frag weapon! So yes anything is an improvement no matter how gamelike it is.:music_whistling:

[Fri13]

totally agree. I don't care about precision modelling,a simple fragments zone and rolling virtual dice for damage points is better than what we have now. remember the vehicles use hitpoints anyway.

Currently infantry is virtually immune against any anti infantry frag weapon! So yes anything is an improvement no matter how gamelike it is.:music_whistling:

 

Neither does fragments from A-A missile are usable against aircrafts!

 

Like that is one reason why Vikhr missile is nearly useless against helicopters as the fragmented shell doesn't count when firing missile in A-A mode with a proximity fuse engaged. You need direct hits and you want to fire two missiles at the time to cause some kind damage to enemy helicopter.

[GGTharos]

Don't expect a quick fix.

 

If we get a quick fix now we can wait a year or two for a proper fix. It's not like the proper fix is coming tomorrow anyways.. ;)

[ShuRugal]

But what would you prefer if you got to choose between doing it that way now (and letting it be like that for the foreseeable future), or waiting however long it might be (I have no clue) and then getting fragmentation done RIGHT where the listed issues would not occur.

 

It's the whole "do it right or not at all" things, I guess, and we might have different preferences there. I've never been bothered by this issue, so for myself I'd rather wait. YMMV. :)

 

Honestly, I'd rather have the hack fix than true fragmentation computation. We have enough trouble with framerate loss and server dumps because the engine craps itself when a single cluster bomb triggers. Now we want to have it try and calculate fragments for each bomblet, when it barely handles the bomblets?

 

Thanks, but no. I'll take a simple "blast zone + frag zone + LOS check + randomizer" solution. I want to kill soft vehicles and infantry, not game clients and servers.

[Dejjvid]

Don't expect a quick fix.

 

Yea, you really had to answer that one didn't you? ;)

 

I fully understand ED have to milk the cow and focus on the lo-fi segment of the market..

[GGTharos]

You fully assume, you mean. ED has plenty of other features and bugs to deal with, and they're not often interested in a 'quick fix' for anything ... they prefer to try and do things right when they get around to dealing with a given feature.

 

Yea, you really had to answer that one didn't you? ;)

 

I fully understand ED have to milk the cow and focus on the lo-fi segment of the market..

[Teknetinium]

You fully assume, you mean. ED has plenty of other features and bugs to deal with, and they're not often interested in a 'quick fix' for anything ... they prefer to try and do things right when they get around to dealing with a given feature.

 

 

I wish I could share you opinion GG. EDs actions for this two years that have passed should tell us a lot of how they approach the product they are working on.

Edited October 27, 2013 by Teknetinium

[shagrat]

I do not blame them for their wish to do it right, but honestly: The ground vehicles use hitpoints varying for sides... on that level, it would be at least fair to give the area effect weapon some effect on areas / hitpoints? Don't you think? We lived with the hitpoints for quite a while, and it is not bad.

What is annoying, especially with more and more choppers (Huey, Mi-8) is the inbalanced weapons you carry against infantry! A huey against Tanks? BMPs? No chance, but against Infantry???

At least you should have a "chance" to use the FFARs as they are intended.

 

If at least someone, may point us to blast radius, I could even live with enhancing that for rockets!

[GGTharos]

IMHO ED looks at it like this:

 

Spend a bunch of time to band-aid it now and 'do it right' later, or just spend time to do it right only.

 

It's just a matter of efficiency - but again, IMHO.

[Irregular programming]

I'd rather they spend time on fixing bugs than changing the damage done by rockets.

 

They are much more effective in game as is than what some of you people are saying they are.

[shagrat]

try to kill some infantry with them! Then tell us if you think it is not a bug! Rocket fragmentation is what kills infantry - rockets are helicopters primary weapon against larger groups of infantry - so for the Huey, Mi-8, A-10C, A-10A, SU-25, SU-25T and to a lesser degree for all the fighters that can load rockets it is a BUG! It makes the rockets useless in their primary role.

I don't tell ED to stop fixing radar and Air to air missiles because there are "more important " bugs... like the rockets for example, either. Or whatever you consider important.

[Irregular programming]

Of course it's not a bug, it's by design and how rockets work and have always worked in game. Fixing them in anyway other than fixing all explosives would probably be a huge time sink and something I'd rather not see them doing if that is the case.

 

Rockets are hardly "useless" even against infantry.

[ShuRugal]

I'd rather they spend time on fixing bugs than changing the damage done by rockets.

 

They are much more effective in game as is than what some of you people are saying they are.

 

are they? I don't remember the last time i killed anything with an S-8, though i once let off 80 in a single pass... completely saturated the enemy troops. Not one fell down though.

 

I will grant that S-13s will get the job done much more reliably, but i can't carry enough of those to be useful against area targets....

[SESova]

I wish I could share you opinion GG. EDs actions for this two years that have passed should tell us a lot of how they approach the product they are working on.

And we are going to remember that.