Goodbye DCS?

[Aquorys]

1 hour ago, Dragon1-1 said:

Quite the other way around. Missiles in DCS never suffer from random failures, unlike IRL, where missiles are real world machines that sometimes break.

Did you read my initial post? Are you suggesting that 3 AMRAAMs, that were all fired well within parameters, will realistically all miss the target, assuming that they have NOT randomly failed?

1 hour ago, Dragon1-1 said:

If your missiles miss, it is typically because you fired them out of parameters.

They were not fired outside of their parameters

1 hour ago, Dragon1-1 said:

It can also be because AI took proper steps and defeated your missile.

Are you suggesting that slowing down and trying to pull a Cobra maneuver in an F-15 at ~25 AoA while facing towards a missile and doing a couple aileron rolls is a "proper step to defeat a missile"? Asking, because every single piece of information I have ever seen or heard from what I would call reliable sources suggests quite the opposite.

1 hour ago, Dragon1-1 said:

IRL, modern missiles are not a magic death wand, though a lot of people have that misconception.

That is true, but you might want to ask some former pilots from Serbia who were smoked by AMRAAM Bs while trying to notch them around 1995 (which seems to be players' #1 favorite successful AMRAAM evasion maneuver in DCS). Unless AMRAAM Cs have somehow de-evolved since then, I am pretty sure that an AMRAAM C can, most of the time, guide towards and successfully intercept an almost head-on target that is doing Mach 0.8 and is flying in a straight line.

1 hour ago, Dragon1-1 said:

Flares and chaff in DCS do need some work, they're very simplified currently, but it's also wrong to suggest that AIM-9X can't go for a flare if the conditions are right.

I did not say it absolutely can't. I said it apparently commonly does in DCS, because it seems a little bit unlikely that 3 AMRAAMs and an AIM-9X in a row, all fired well within parameters, would all fail to intercept the target.
How many missiles would you suggest to fire at a single F-15 in real life to have a reasonable chance of intercepting it? Like, 6 AMRAAMs and 4 Sidewinders?

There is not a lot of air-to-air combat going on these days. It would be quite interesting to hear how many missiles hit or failed to hit each of the targets that were intercepting recently - e.g., during the Iranian attack on Israel. Some of the drones and cruise missiles were targeted by fighter aircraft, presumably using AMRAAMs. The ballistic missiles were shot down using SM-2s or SM-6s. The SM-6's sensors were developed based on research and technology from the AMRAAM.

[Dragon1-1]

22 minutes ago, Aquorys said:

How many missiles would you suggest to fire at a single F-15 in real life to have a reasonable chance of intercepting it? Like, 6 AMRAAMs and 4 Sidewinders?

One at a time, following a proper BVR timeline. Remember, your aim is not necessarily to destroy the F-15. If he goes bingo or winchester and has to abandon his mission, you won. In some cases, if you delay it enough that your strikers get through or backup gets there, you've won. You don't have to merge with the F-15 or shoot Sidewinders at it. It's a very capable fighter with very powerful engines, lots of gas and plenty of weapons. It's not uncommon for weapons in a BVR duel to be defeated kinematically. Serbian pilots didn't have anything that could threaten an AMRAAM carrier, so NATO pilots could get close and fire those high PK shots with little risk of retaliation. 

Once again, it's not that the AMRAAM's sensors are crap, or that there's something wrong with missile dynamics. The AI FM is not the best, but it's a known issue. Was your opponent set to "Ace" level? On this level, it will use every trick in the book against you, augmented by its unrealistic FM. 

[Aquorys]

17 hours ago, Dragon1-1 said:

One at a time, following a proper BVR timeline.

Sounds like what I did

17 hours ago, Dragon1-1 said:

Remember, your aim is not necessarily to destroy the F-15.

Well, it is in DCS, because the AI chased me to my home airport, despite multiple friendlies in the area, which is another problem with realism. There are very few pilots in the real world who would go on a suicide mission to shoot a couple more missiles at a random enemy aircraft. As it is now, AI aircraft in DCS are all Kamikaze pilots.

To be fair, getting shot at while landing at my own home base is also a very common problem with mission design (this was on a public server). A couple years ago, I published several training missions in various levels of difficulty for the F-16. If you look around in these missions, you will see that the airspace near the home airbase is covered by an air defense system.

17 hours ago, Dragon1-1 said:

It's not uncommon for weapons in a BVR duel to be defeated kinematically.

Which is why I survived that long. A kinematic defeat means that a properly guiding missile is unable to reach you, that was not the case with the AI aircraft. Especially in the second case, the missiles did reach it with energy to spare, they just flew by instead of guiding properly.

17 hours ago, Dragon1-1 said:

Was your opponent set to "Ace" level? On this level, it will use every trick in the book against you, augmented by its unrealistic FM. 

I guess that it was (as I mentioned, this happened on a public server). If the "Ace" level behaves irrationally and causes the game to glitch by doing so, it should be disabled. And I am not talking about super-human skill, for example, with regards to reaction time, precision, G resistance, etc. - none of those are a problem, a setup like that can be useful for training purposes. However, training against an opponent that gives you irrational and/or random results instead of an indication of how effective your tactics are against less capable opponents doesn't do anything for you.

[Dragon1-1]

Well, there you have it. The public server sounds like it's got serious mission design problems. Using "ace" AI is one of them, because it does things like that (it'd be hard to just disable it because it'd possibly break a lot of things). What the server owners can do is set it one step lower. There should also be a trigger making the AI abort and return back to their station once you're no longer a threat and inside friendly air defense ring.

[Aquorys]

An easy temporary fix would be to link the Ace AI level to the Veteran AI level, so that selecting Ace and selecting Veteran effectively do the same thing, until the Ace AI is fixed to behave properly, and the missile guidance is fixed not to fail because an AI aircraft pulls off maneuvers that only work in Hollywood movies, but would be ineffective in the real world.

[Dragon1-1]

14 minutes ago, Aquorys said:

and the missile guidance is fixed not to fail because an AI aircraft pulls off maneuvers that only work in Hollywood movies, but would be ineffective in the real world.

I'm pretty sure that's not the case. Missile physics an guidance are reasonably accurate. It's not that the AI aircraft pulls off maneuvers that'd be ineffective IRL. It's that it pulls off ones that would be impossible. As a matter of fact, if they were possible, they'd probably be pretty good at trashing missiles, because fast, erratic turns would force the missile to make rapid guidance adjustments and bleed its energy. I'd be worried if humans were capable of dodging missiles this way, because human-flown fighters have a realistic FM.

Again, missiles are not magic death wands, and their guidance is not as infallible as the popular hype expects them to be. They are subject to laws of physics, and the fact that they can only guide to a position they expect the target to be, based on its current equations of motion. A mid-2000s missile can't observe the target for a while and go "oh, he's doing that that corkscrew thing, I'd better fly straight", it checks where the target is, how fast it's moving, and its acceleration at the given moment, then figures out where those parameters will get it. Missiles in DCS follow those constraints and limitations, AI aircraft... don't. Especially when set to "Ace", which is one reason to avoid it. Other settings suffer from this, but to a lesser extent.

GFM should resolve those issues and force AI to behave realistically. This will likely require adjusting AI dogfight logic and especially formation flying behavior.

[Aquorys]

41 minutes ago, Dragon1-1 said:

I'm pretty sure that's not the case. Missile physics an guidance are reasonably accurate.

Explain this, especially considering that you just said

49 minutes ago, Dragon1-1 said:

the fact that they can only guide to a position they expect the target to be, based on its current equations of motion

 

[Dragon1-1]

That one looks like it lost guidance entirely. Did you lose lock at any point? The bandit began to maneuver, and the missile continued to where he'd be if he haven't. This could be a bug, or a lost radar track before pitbull. Either way, I'm not having any of those problems. In fact, most people seem to be able to hit with those missiles just fine.

[Aquorys]

2 hours ago, Dragon1-1 said:

the missile continued to where he'd be if he haven't

Nope. Different view angle, about 3 seconds later.

This missile is not headed anywhere where this F-15 could ever have ended up, with or without maneuvering.

Meanwhile, I had time to test in singleplayer missions.

I was able to dodge an AMRAAM by departing controlled flight. I did not try this often, it worked on the second attempt, no idea how easy it would be to reproduce this. Needs more testing if anyone feels inclined to try, I'm not too motivated to try perfecting my skills in doing completely unrealistic things. Also not sure if I broke the record with my 84.7 degrees AoA in an F-16 at roughly 160 knots. Didn't know you could flat-spin it... kind of...

What's possibly more surprising is that I was able to recover from this departure.

I also believe that I have a reproducer for the AIM-9X failure that happened on the multiplayer server. Turns out that the AIM-9X in DCS can not track targets head-on. Tried it 22 times, with and without radar lock, slight variations in position and angle, slightly below and slightly above, distance within the no-escape zone, mostly between 3.5nm and 2nm, got 19 misses and 3 hits (one of which was from 1 nm, which seems to work better), for a probability-of-kill of approximately 14%. This also reproduces with Veteran AI, so this has nothing to do with Ace AI.

I flew almost all of those engagements until I splashed the F-15, and almost all of the missiles fired later at different angles hit, so I am pretty sure it wasn't me doing something wrong every time when firing the first missile, and then somehow always getting it right the second time - extremely unlikely.

I also tried reproducing the same situation, but with AMRAAMs (some 10 times or so, 5 of them within the same distance, some of them even without a radar lock) for comparison, and all of them hit, so this is also not a case of being impossible to guide or maneuver for a missile in general.

 

Edited February 15 by Aquorys

[Dragon1-1]

8 hours ago, Aquorys said:

I also believe that I have a reproducer for the AIM-9X failure that happened on the multiplayer server. Turns out that the AIM-9X in DCS can not track targets head-on. Tried it 22 times, with and without radar lock, slight variations in position and angle, slightly below and slightly above, distance within the no-escape zone, mostly between 3.5nm and 2nm, got 19 misses and 3 hits (one of which was from 1 nm, which seems to work better), for a probability-of-kill of approximately 14%. This also reproduces with Veteran AI, so this has nothing to do with Ace AI.

This is definitely something worth analyzing further. However, do note that tracking a head on target which maneuvers away is not an easy job for the missile, because a small change in target aspects generates a large movement of the intercept point, necessitating energy-depleting maneuvers. NEZ (especially as indicated on the HUD, based on current closure rate) is less important than kinematics here, which is why the shot at 1nm works - the bandit has no time to turn away. High aspect shots should be expected to have low Pk against maneuvering targets, even with the AIM-9X.

I've actually had experience with this in the other Viper sim, which provides a really nice test scenario with its HMCS training mission. Hitting a high aspect shot with AIM-9X wasn't a sure thing there, either. Both MiG-23 and -29 usually thrashed the face shot, the -21 usually didn't. A sub-1nm shot directly to the face really seems to be the way to go in this case if you want the missile to hit (if you don't care about that, forcing the bandit to defend before the merge is a great for allowing you to slot in behind and shove another heater straight up its tailpipe, or even gun him).

8 hours ago, Aquorys said:

I was able to dodge an AMRAAM by departing controlled flight. I did not try this often, it worked on the second attempt, no idea how easy it would be to reproduce this. Needs more testing if anyone feels inclined to try, I'm not too motivated to try perfecting my skills in doing completely unrealistic things. Also not sure if I broke the record with my 84.7 degrees AoA in an F-16 at roughly 160 knots. Didn't know you could flat-spin it... kind of...

What makes you think it shouldn't work? Nobody tested the AIM-120 against a target in a flat spin, to my knowledge. In fact, I don't know of any instance of a missile being fired at a target that just spun out, much less one spinning out when defending against a missile. It's hard to believe it'd work, and it's not something anyone would even think to try IRL. It is, however, a very interesting question what the missile would do in this situation.

We should ask Musk to get a general to order such a test. Deliberately depart a target drone defending against an AMRAAM launch, and watch what happens. 

Edited February 15 by Dragon1-1

[Aquorys]

8 hours ago, Dragon1-1 said:

However, do note that tracking a head on target which maneuvers away is not an easy job for the missile, because a small change in target aspects generates a large movement of the intercept point, necessitating energy-depleting maneuvers.

As I said, the AMRAAM could do it reliably, so the AIM-9X being a dedicated dogfight missile should be able to compete.

8 hours ago, Dragon1-1 said:

What makes you think it shouldn't work? Nobody tested the AIM-120 against a target in a flat spin, to my knowledge.

The fact that there seems to be a complete absence of air combat instructors across the entire world who teach any tactics that involve low-speed, low-energy, low-maneuverability (or rather, no-maneuverability in this case) as a viable option to defend against a missile. Missiles that are sufficiently able to intercept high-speed maneuvering targets are usually able to intercept low-speed targets on a ballistic trajectory.

In fact, at least western missiles have over-performed rather than under-performed compared to their specifications. An AIM-9X shot down a hot air balloon, and it was not exactly designed for that type of target. An SM-3 shot down an out of control satellite reentering the earth's atmosphere, and it was not designed for that type of target either, but performed flawlessly (that target was destroyed by the kinetic energy of the collision, there was no explosive warhead on that weapon, that is how accurate modern guidance systems can be).

[Ala13_ManOWar]

On 1/1/2025 at 3:27 AM, kitten40 said:

It is with much regret that I decide to part ways with DCS. 

The thought has been brewing in the back of my mind for some time now, but I kept putting it off, making excuses for this or that, I have other things to worry about, etc, etc.

Lately i have been absent from actually flying, and concentrating on fulfilling my agreements with custom liveries ..

This is where things really hit home for me.. Not only is the DCS core full of bugs and problems (which we all know about and have come to love), especially with each update, but even the simplest of tasks (and i mean that figuratively) like painting custom skins, has become such a headache.

Yes, I'm ecstatic that ED provides paintkits .. But .. they are incomplete "IN MY MIND" .. What should have taken approx no more than 10 minutes, took more than an hour to try and figure out the Board numbers for the Flanker.. Why? Because those files weren't included in the 'kit'.

It's just one more thing on my growing list of incomplete things in DCS that I seem to be trying to overcome lately..

On the topic of paint, where is the Tomcat kit? How long ago was it promised? Is it a Heatblur thing or ED thing? I'm not hating on either, I love them both to death (check my purchases lol) 

How do we paint the new F5.. the descriptions are absolutely horrendous when trying to figure out F5 vs FC2024, etc.

Overall I'm not disappointed in the general state of DCS, but more like I feel it's very disorganized and haphazard, and almost seems to be trending towards the worst.

Just my 2 cents, that really in Canadian currency means nothing, but to me, it's a concerning observation.

Ok, bye bye drama queen. We won't miss you whatsoever…

[Rudel_chw]

9 minutes ago, Ala13_ManOWar said:

Ok, bye bye drama queen. We won't miss you whatsoever…

 

You should have read the whole thread:

 

 

[Ala13_ManOWar]

23 hours ago, Rudel_chw said:

You should have read the whole thread:

I know Rudel. BTW, I did read that back when he wrote it. But it's deeply annoying how every now and again people write their essays on "why I leave DCS because devs don't do what I like how I like…", and then again those threads go on for months without an end with people really concerned and thinking about "leave or stay" like that was important.

And I mean, yeah, we all do go through whatever it is and go and come back. I don't fly DCS right now since I can't remember when due to personal reasons, life and those silly things happening all the time, you know. Ok, so I don't fly, I can't fly, I whatever you wanna call it because I'm unable to do so or whatever. But who cares!!! It's a damn game, and entertainment, a pastime. Sometimes we do, sometimes we don't, but, giving that much of an importance to a hobby is stupid on itself!!

Edited February 17 by Ala13_ManOWar

[SOLIDKREATE]

Even with the bugs, I still love this SIM.

[Beirut]

42 minutes ago, SOLIDKREATE said:

Even with the bugs, I still love this SIM.

 

Damn straight!

 

Where else can we get hyper-detailed military jets to fly around? The civvy sim military jets have FMs that feel like melted ice cream and don't look nearly as good. And they have no "Kaboom!"

 

We get the tight FMs, the best aircraft eye candy, and the "Kaboom!"  

[Dauntless]

"Kaboom!" 

[MAXsenna]

18 minutes ago, Beirut said:

We get the tight FMs, the best aircraft eye candy, and the "Kaboom!"  

Bang! Pow! Boom!

https://open.spotify.com/track/4JfcEkOP08hSwP7cSSz6AW?si=tUgjaFeHT460Q6DT-VQ_yA&pi=akSsDPhaQEWHn

Edited February 17 by MAXsenna

[Aquorys]

8 hours ago, Beirut said:

Where else can we get hyper-detailed military jets to fly around?

I was going to say, that might be part of the problem, because since there are virtually no other options, there is also no pressure to fix problems. BUT the harsh reality is that there is no shortage software products that are buggy, unreliable and unstable in completely different areas of information technology that are offered in market segments where there is a lot of competition, so even having competing products available on the market does not necessarily mean that you will get a better product from anyone.

As I said before, I guess, it is what it is, but to be fair, not only in DCS.

[Dragon1-1]

On 2/15/2025 at 8:35 PM, Aquorys said:

As I said, the AMRAAM could do it reliably, so the AIM-9X being a dedicated dogfight missile should be able to compete.

At a comparable range? AMRAAM has a lot more energy, so if you fire it at a distance you would the AIM-9X, the same maneuver will not save the bandit. AIM-9X is more maneuverable than the AMRAAM, but it has a smaller, shorter burning motor. Case in a point, being able to hit a sub-1nm shot on a rapidly closing target is actually quite remarkable. That's what AIM-9X excels at, however keep in mind that making large turns scrubs off a lot of energy.

[Beirut]

2 hours ago, Aquorys said:

I was going to say, that might be part of the problem, because since there are virtually no other options, there is also no pressure to fix problems.

 

But it also means that without DCS, all we'd have would be the civvy sims with mushy flight models and no "Kaboom!"

 

I think we're lucky to have that one option.

[Aquorys]

1 hour ago, Dragon1-1 said:

At a comparable range?

Yes, same range. The range where the AIM-9X would miss very consistently was about 2 nm to 3.5 nm. I tried the same distance range with the AMRAAM, and it hit its target every time.

[SOLIDKREATE]

17 hours ago, Beirut said:

 

Damn straight!

 

Where else can we get hyper-detailed military jets to fly around? The civvy sim military jets have FMs that feel like melted ice cream and don't look nearly as good. And they have no "Kaboom!"

 

We get the tight FMs, the best aircraft eye candy, and the "Kaboom!"  

 

 

[Dragon1-1]

9 hours ago, Aquorys said:

Yes, same range. The range where the AIM-9X would miss very consistently was about 2 nm to 3.5 nm. I tried the same distance range with the AMRAAM, and it hit its target every time.

The enemy evasive maneuvers are causing the AIM-9X to run out of energy. Remember, you're both closing in rapidly, and the AIM-9X accelerates to the predicted intercept point. The enemy then turns around, still going fast, and suddenly, the intercept point is way off to the side. The missile has to turn and accelerate towards this new point, only, most of its motor burn time has been spent accelerating to the old one. The AMRAAM, having a much longer burning motor, will still be gaining speed at this point, and will easily make the intercept.

To fix this issue, forget the DLZ and launch the AIM-9X at very short ranges in a head-on merge. This is pure physics, and not unrealistic in the slightest, just surprising, especially with how modern missiles are hyped. The "common wisdom" about what those missiles can do is based on ranges and Pk against nonmaneuvering targets (likely because advertising brochures use that). Against maneuvering targets, you need to adjust your launch parameters. 

Now, I'm not quite sure if the AI should be able to spot a head on launch at all. I personally find it difficult to see a tiny heater at 3nm, and the AIM-9X is smokeless, so unless it's leaving a contrail, perhaps the AI shouldn't be able to react. Of course, against missiles with a smoky motor, or with MWS, it's a different story.

[Aquorys]

4 hours ago, Dragon1-1 said:

The enemy evasive maneuvers are causing the AIM-9X to run out of energy.

Not the case in the scenarios I tested, because in most cases, it hardly maneuvered at all, and then missed by flying essentially straight and low. I'll see if I can make a video and/or track of it, so y'all can see what exactly I mean. It should be easy to get it on camera, since it reproduces so consistently.