[WAD] The new dogfight engagments are very awkward (beta 1243.5)

[Mickeys91]

I am a huge DCS and BMS player and have been playing for years with real life retired fighter pilots in few sim groups. This post is my issue of the way the planes in game avoid incoming air to air missiles compared the older (pre-1243 beta) was much more realistic. I'm not an expert in A-A engagements but I have talked about this issue to a retired F18 pilot the other day and even he said that this new coding of how the AI avoids missile is a little awkward. Air to Air engagements are one if not the biggest scenarios for this game so I would love to see it simulated more realistic.

Not sure how to explain this better so here is a Example: I put 2 F16s on one side and 2 Su27s or Migs on the other (with similar skill level and altitude or mix it up). I make them spot and engage each other either through BVR or within close range (With or without AWACS). Sometimes one side gets off missile first sometime they both fire at same time. Now instead of diving and using the notching tactic (going 90 degrees and diving) like the previous versions of the game (which was done pretty well) one or if not BOTH sides will go into a defensive mode and turn 180 degrees and try to bleed the missile out of its energy (which is fine since that is also a real life tactic). The issue now is the ai will lose sight of the missile (usually because the missile has slowed down and they cannot visually see it behind them) so the AI will than assume the missile is dead and turn around, but than it will detect the missile AGAIN and go into a defensive mode AGAIN. Realistically the pilot in the plane should simply just assume the missile has lost too much energy and fly up, left or right of the missile. Instead the AI just continues the opposite way again into defensive stance even though there is no way the missile has the energy to catch up. Meanwhile the enemy AI side is simply taking advantage of this situation and just keeps a lock and firing off its missiles because the AI that is being chased has no idea how to break out of this defensive stance. As I said previously most of the time both sides are stuck in this endless spam battle of missiles where they're shooting and turning, than once clear they shoot again. Sometimes I got to points where both sides are exhausted of missiles. Now I understand these situations could happen in real life but it is very rare. The AI simply cannot comprehend using its energy to just go around the dead slow missile. Maybe its a game limitation? but in my opinion the old notching method was much better. Pilots usually got kills after 1 or 2 or 3 missiles and still had a great Situational awareness of the incoming missiles which it avoided.

So my question is there a way to get the AI to use both this new altitude hold 180 degree turn and burn method AND the older notching method? This will mix the engagements up because right now its way too predictable with what your or the enemy AI will do. Everything else in the new beta is amazing and keep up the great work dev team!

[Dimitris]

This was previously discussed on a now-deleted Reddit thread, so copying my responses from there to provide context.

This is an "admiral / AF general vs individual pilot" perspective difference.

You, on your theater plot, see the missile [bleeding speed] and think it's no longer a threat. The pilot, in his plane, sees (at best) a dot that is still chasing him. Has it slowed down enough to not be threat? Maybe, maybe not. Is it worth the risk to stop running away and turn into any direction other than away from it? Remember, you get it wrong and you may DIE.

See also this video of BVR evasive maneuvers in F4 BMS: https://www.youtube.com/watch?v=T7y-NUzhaKk . Notice how the pilots re-join the engagement only after the incoming shots have died out. Survival instincts prevail.

Command does not model perfectly-obedient robots (mostly). It aims to model human crews who try to execute their task at the best of their abilities, but at the end of the day still want to come back home in one piece. Sometimes this goes against "the plan", and that's OK. This is how RL combat frequently unfolds.

In fact the "beam and dive" evasion is still there; the difference is that now it is used as a last resort, when the pilot deems he has been unable to outrun the missile (or the missile is close enough that outrunning it is not an option), rather than being the default behavior.

We too have talked with people who do this for real, and they concur the new behavior is a better representation of _most_ evasion cases.

It's interesting you should mention [entire airgroups scattering because of incoming missiles], because in Vietnam it was a common tactic to break up aircraft formations by firing missiles at them, sometimes even unguided.

Same thing happened sometimes in Desert Storm; read up on "Package Q" as an example.

The pilots weren't stupid; they wanted to live.

[thewood1]

Using DCS as a model is losing proposition. Living room pilots have tried to use it a lot. But even Eagle Dynamics admits its AI reactions to a2a combat leave something to be desired. There was a big thread here not too long ago that brought quotes in from ED and players about the need to focus on AI reactions to missile shots.

I'll track it down. Once you read it, you'll never use DCS as a benchmark again for AI reactions in a2a combat.

[thewood1]

https://www.matrixgames.com/forums/view ... S#p4974669

[Mickeys91]

This was previously discussed on a now-deleted Reddit thread, so copying my responses from there to provide context.

That is one of the members I played with. We had the same discussion and was told he would put it on reddit. I never checked it out myself but I assume it was pretty similar to what I wrote. Not exactly sure why it was deleted.

As far as what you said it being used by real pilots (as many of the pilots I play with use this tactic). yes you are correct but the issue is the CMO AI does not use it as it is suppose. Why not add both 180 and the older notching? The AI is too confused on understanding difference between a dead slow missile and one that is a actual threat. I can fire one missile at a group of 15+ planes and that ENTIRE group will go into defensive mode and turn around. Same as if I have a top flight group and a low flight. The Top is suppose to provide over watch type station in case the low flight gets engage. But the issue is the top will also flee even if the missile is being fired 25,000 feet below it at a complete different group.I guess the only way is for me to take over and micro manage firing missiles. But that is kind of a paint if I have maybe huge amounts of groups scattered around.

Using DCS as a model is losing proposition. Living room pilots have tried to use it a lot. But even Eagle Dynamics admits its AI reactions to a2a combat leave something to be desired. There was a big thread here not too long ago that brought quotes in from ED and players about the need to focus on AI reactions to missile shots.

I'll track it down. Once you read it, you'll never use DCS as a benchmark again for AI reactions in a2a combat.

Yes but I don't use DCS AI. The AI is terrible. I avoid it at all costs. I play only online against human players and a decent amount of the player base on the servers we use are ex AF and Navy pilots. And the number one tactic to avoid incoming missiles is Dive notching 90 degrees which was used pretty well by the CMO AI prior to this new Beta update and yes the new current tactic of 180 turn and burn as we call it is also used but only depending on the situation.

And I will check that link out when I have time later tonight appreciate the response!

[DWReese]

I'm not sure if this is real tactics or not, but from what I see from CMO is this:

Two groups of fighter planes approach each other. Each group fires at the other. Each group turns and runs. All of the missiles peeter out, and no one gets hit. This seems to happen every time. All of the missiles, regardless of maneuverability and speed, are treated the same. The planes with the most missiles could repeat the process.

As an alternative, I tried setting it up to fire at 75 percent distance, and then at 50% distance, and the outcome was the same, although the missiles almost caught some of the respective planes at 50%.

I assume that the tactic to take he is to have the pair of fighters split up, and attack from two different directions. That seems like a lot of effort, however.

[thewood1]

"I assume that the tactic to take he is to have the pair of fighters split up, and attack from two different directions. That seems like a lot of effort, however."

This. This especially valid with the new formation capability. You can use formation presets and load them in based on the tactical situation if you want to manage them that way. This isn't dependent on a scenario designer. You can build all kinds of templates and execute then as you see fit.

[Mickeys91]

I'm not sure if this is real tactics or not, but from what I see from CMO is this:

Two groups of fighter planes approach each other. Each group fires at the other. Each group turns and runs. All of the missiles peeter out, and no one gets hit. This seems to happen every time. All of the missiles, regardless of maneuverability and speed, are treated the same. The planes with the most missiles could repeat the process.

As an alternative, I tried setting it up to fire at 75 percent distance, and then at 50% distance, and the outcome was the same, although the missiles almost caught some of the respective planes at 50%.

I assume that the tactic to take he is to have the pair of fighters split up, and attack from two different directions. That seems like a lot of effort, however.

Exactly on your first half. Real BVR engagements are not rinse and repeat. That is exactly why there are certain tactics in real life to counter this situation of just spamming $500k-$1M+ missiles at each other only for there to be no kills and everyone returns to base. Pilots are taught to be aggressive and not just run away when ever there's a missile in the air especially one that is not a danger to them. The one that is currently used, the previous notching are only 2 of many. The issue that I see is the AI only using one basic maneuver when I think at least another 2 or more could be coded in (if the game engine allows it).

What I am trying to say is the previous version prior to this update (IMO) was much more realistic and enjoyable with actual kills compared to this new AI behavior. And I never had to touch any Settings since the AI seemed to stay very aggressive. I I also assume the new missile physics of engine burning out plays a big role into it as well. So I guess for now my best bet is to work with the settings to get the AI to behave how I would like them too. I do hope the devs can take a look into flight AI dynamics in BVR and WVR in future (If the CMO engine allows it) would make for some epic air engagements.

[DWReese]

"What I am trying to say is the previous version prior to this update (IMO) was much more realistic and enjoyable with actual kills compared to this new AI behavior. And I never had to touch any Settings since the AI seemed to stay very aggressive"

I totally agree. This version might be more realistic, but it's definitely less fun.

[DWReese]

This especially valid with the new formation capability. You can use formation presets and load them in based on the tactical situation if you want to manage them that way. This isn't dependent on a scenario designer.

Speaking od scenario design, what exactly does this do to all of the old scenarios that have been created? None of those are set up to utilize formations, so when you play those scenarios you will be back to the shooting-and-running tactics that we are seeing right now without the use of formations. Does that make them somewhat obsolete if they aren't updated to use the new formation capability?

[boogabooga]

I know that I'm a month late on this one, but I wanted to weigh in as I'm getting ready to release a new 12XX-friendly version of my A2A BVR Duel Set for the Quick Battle Generator, and I have been looking at this problem a lot.

Let's look at what DCS actually says:

2022-06-19.png (517.14 KiB) Viewed 816 times

This is from the DCS F-15C manual. It is showing that Rtr, the effective range for an AIM-120C against a maneuvering target is just 10 nm (!). That is about 40% of the Rpi range, which is the maximum range against non-maneuvering targets. And yes, that is for a closing target.

I've tried this many times in CMO, and if both sides adjust their WRA range to 40% (whoever said 50% was close- but no cigar), then indeed you can get fairly spectacular (and I would say very DCS-like) results. At that range, missiles can not be outrun and it forces notching or combined radial + notching tactics as the OP expects. The Pk is about what one would have expected from CMO 'classic.' But 12XX opens up a whole new tactic (for both sides) of taking shots outside of 40% for suppression, rather than for ensured kill- which can perhaps be exploited.

I do believe that older scenarios are going to be 'broken' in that they very rarely account for the drop-off in effective range in their WRA, since this was not really well modeled before. Some remedies might be

-Making the new missile fly-out model an optional scenario feature
-Giving a new doctrine setting for practical vs. kinetic missile range, the same as is done for guided torpedoes now
-Some sort of script to adjust the WRA Max range to 40% (or whatever)

[Gizzmoe]

But 12XX opens up a whole new tactic (for both sides) of taking shots outside of 40% for suppression, rather than for ensured kill- which can perhaps be exploited.

Yes, suppression shots can very effective. If we would set the WRA to a realistic 40 or whatever % for both sides then the AI must still somehow be able to launch the occasional suppression shot, otherwise we players could totally exploit that in certain situations.