Page 1 of 2
WVR dogfights.. no leading?
Posted: Tue Mar 08, 2022 9:30 pm
by wyskass
Hi,
Just finishing all the air tutorials and am mostly comfortable with the air game.
But noticed some issue with WVR fights.
1. Attackers don't seem to lead to win a circle chase, even when it's painfully obvious and in good position to cutoff the defender.
2. Trying to manually control fight with plotting course often makes it worse. For example trying to go for a 2 circle, they seem to slow down in following course. Or if plotted too close, they will actually turn the opposite way, ruining any tactical position and plane.
3. Generally trying to control maneuvers with plotting, seems to be frustrating, so easier to let them fight automated and frustratingly watch the circle chases where they don't take opportunities for advantage.
What's some advice on WVR fights, and/or manual control vs automated.
(This was most evident on the tutorial with WW2 fighters and dogfights)
Thanks.
Re: WVR dogfights.. no leading?
Posted: Wed Mar 09, 2022 5:52 am
by boogabooga
There are even more problems than you mentioned (energy retention, no out of plane maneuvers, etc.)
I think you just need to accept that WVR isn't CMO's strong point right now.
Re: WVR dogfights.. no leading?
Posted: Wed Mar 09, 2022 6:12 am
by Dimitris
Replied to this on Discord but reposting here.
wyskass wrote: Tue Mar 08, 2022 9:30 pm
Or if plotted too close, they will actually turn the opposite way, ruining any tactical position and plane.
This often happens, deliberately so, when the objective of the turn (a target or a waypoint, or the base when landing etc.) is "inside" their turn radius. They will then turn to the opposite direction in order to extend, give themselves more space, and try again. This is consistent with RL manouvers (on the horizontal plane). This is how energy fighters frequently deal with more manouverable opponents, for example (exit the turning fight, turn around and engage).
Generally if you are down to micromanaging fighters in a dogfight you are not really "commanding". The most you should be involved should be the level of "you two interceptors take those 4 bandits" and sit back and let them do their thing. This is not DCS or Falcon-4. You are not the pilot at the cockpit.
Some background reading on modern air-combat trends:
https://csbaonline.org/research/publica ... uperiority
Re: WVR dogfights.. no leading?
Posted: Wed Mar 09, 2022 9:41 pm
by wyskass
boogabooga wrote: Wed Mar 09, 2022 5:52 am
There are even more problems than you mentioned (energy retention, no out of plane maneuvers, etc.)
I think you just need to accept that WVR isn't CMO's strong point right now.
Yea, I will have to. I already had TacView from use in XPlane, so wanted to see some of the air maneuvers in CMO. Kept shouting, "What are you doing, turn turn.."
But ya, fair enough, there are many other aspects of CMO to enjoy, and am now learning on ships.
It would help if AI could lead an intercept a bit better though.
Re: WVR dogfights.. no leading?
Posted: Wed Mar 09, 2022 9:47 pm
by wyskass
Dimitris wrote: Wed Mar 09, 2022 6:12 am
Replied to this on Discord but reposting here.
wyskass wrote: Tue Mar 08, 2022 9:30 pm
Or if plotted too close, they will actually turn the opposite way, ruining any tactical position and plane.
Generally if you are down to micromanaging fighters in a dogfight you are not really "commanding". The most you should be involved should be the level of "you two interceptors take those 4 bandits" and sit back and let them do their thing. This is not DCS or Falcon-4. You are not the pilot at the cockpit.
Some background reading on modern air-combat trends:
https://csbaonline.org/research/publica ... uperiority
Sure, that's a good way to think of it. No admiral or general is telling pilots where to turn.
I just noticed the lack of leading also from small boat returning to ship, so they end up in a chase. The software can calculate a proper intercept as with torpedoes, so odd that it's not performed for air or boats. Less important with large speed difference. But oh well.
Thanks for the useful info, will give it a read, to see if common knowledge is still valid for future.
BTW, while nobody wants another world war, it seems likely that actual air superiority fights will return at some point over decade, and all those new fighters and tactics will be tested. I wonder if the F22 is what it's cracked up to be, though of course pilot ability is still a bigger factor among peers. Not really getting much intel on Russian air to air performance in Ukraine
Re: WVR dogfights.. no leading?
Posted: Wed Mar 09, 2022 9:58 pm
by Dimitris
wyskass wrote: Wed Mar 09, 2022 9:47 pm
I just noticed the lack of leading also from small boat returning to ship, so they end up in a chase. A simple algorithm as in the early Sidewinders to maintain bearing rather then direct would be nice. But oh well.
This was actually tried at one point, with a classic intercept algorithm. It works OK as long as the headings and speeds of the parent and the child differ significantly. But if they happen to be similar (e.g. parent is at heading 0 and the child is at heading 350, with similar speeds) then you can easily end up with a VEEEEERY long-duration intercept, with the child frequently being stranded in the sea.
(Fun fact: AAW missiles in DCS World suffered from a similar issue a while back, for the same reason, which was why they were very susceptible to "snaking" evasion maneuvers - far more so than IRL. Pretty sure there are still YT videos out there that demonstrate this. I _think_ they've fixed this more recently but haven't checked either way).
A pure-pursuit algorithm is indeed less efficient most of the time, but has the underappreciated advantage of reliability under a very wide range of conditions.
Re: WVR dogfights.. no leading?
Posted: Thu Mar 10, 2022 2:03 am
by wyskass
Dimitris wrote: Wed Mar 09, 2022 9:58 pm
A pure-pursuit algorithm is indeed less efficient most of the time, but has the underappreciated advantage of reliability under a very wide range of conditions.
True, good point
Re: WVR dogfights.. no leading?
Posted: Thu Mar 10, 2022 4:13 am
by boogabooga
Dimitris wrote: Wed Mar 09, 2022 6:12 am
Generally if you are down to micromanaging fighters in a dogfight you are not really "commanding". The most you should be involved should be the level of "you two interceptors take those 4 bandits" and sit back and let them do their thing.
This is not DCS or Falcon-4. You are not the pilot at the cockpit.
Right, not DCS, but it's not Eagle Day to Bombing the Reich, either. (That has a purely abstract/probability based WVR combat model.) CMO is something in the middle.
I get that this is
YOUR software, but another way to think about CMO play is not "commanding" per say but putting human intelligence into the model to augment the AI- and that is sort of occurring simultaneously over many 'command levels', including the pilot. As it is, letting the AI "sit back and...do their thing " in CMO often gets all of your units killed. Getting a better result than that yourself is why CMO is "fun".
Re: WVR dogfights.. no leading?
Posted: Thu Mar 10, 2022 4:43 am
by stww2
My general impression has been over the years has been that, as annoying as the AI's simplistic WVR tactics can be, it is something that is simply not a factor in the vast majority of air engagements in most CMO scenarios. Granted, part of that is probably because the community, for whatever reason, by and large seems to eschew any scenarios that occur before 1980, but nonetheless.
Re: WVR dogfights.. no leading?
Posted: Thu Mar 10, 2022 11:52 am
by fulcrum28
Regarding to this, I have some cases in which a bomber is being attacked by a fighter at relative close range, and I assigned an order return to base or directly assigned a direction to move opposite to the attacking fighter. After few seconds, the bombers turns around instead of following my order. Is that an AI evasive response to the fighter?
Re: WVR dogfights.. no leading?
Posted: Thu Mar 10, 2022 12:57 pm
by boogabooga
stww2 wrote: Thu Mar 10, 2022 4:43 am
My general impression has been over the years has been that, as annoying as the AI's simplistic WVR tactics can be, it is something that is simply not a factor in the vast majority of air engagements in most CMO scenarios. Granted, part of that is probably because the community, for whatever reason, by and large seems to eschew any scenarios that occur before 1980, but nonetheless.
It's a bit of a factor when playing in CWDB (which has some very good scenarios out there, IMHO). It might go the other way too, that there would be more interest in earlier stuff if the WVR model was a bit better.
Also, probably most CMO scenarios are naval engagements anyway.
Re: WVR dogfights.. no leading?
Posted: Thu Mar 10, 2022 3:07 pm
by stww2
I definitely agree the CWDB is underutilized.
Re: WVR dogfights.. no leading?
Posted: Thu Mar 10, 2022 3:45 pm
by BrianinMinnie
boogabooga wrote: Thu Mar 10, 2022 12:57 pm
stww2 wrote: Thu Mar 10, 2022 4:43 am
My general impression has been over the years has been that, as annoying as the AI's simplistic WVR tactics can be, it is something that is simply not a factor in the vast majority of air engagements in most CMO scenarios. Granted, part of that is probably because the community, for whatever reason, by and large seems to eschew any scenarios that occur before 1980, but nonetheless.
It's a bit of a factor when playing in CWDB (which has some very good scenarios out there, IMHO). It might go the other way too, that there would be more interest in earlier stuff if the WVR model was a bit better.
Also, probably most CMO scenarios are naval engagements anyway.
Booga, what do mean when you say "when playing in CWDB", Thanks
Re: WVR dogfights.. no leading?
Posted: Thu Mar 10, 2022 9:08 pm
by Tcao
Cold war data base?
Re: WVR dogfights.. no leading?
Posted: Thu Mar 10, 2022 11:14 pm
by Figeac
boogabooga wrote: Wed Mar 09, 2022 5:52 am
There are even more problems than you mentioned (energy retention, no out of plane maneuvers, etc.)
I think you just need to accept that WVR isn't CMO's strong point right now.
I would be ok to accept that. But it has been 10 years since the game's release now...
Re: WVR dogfights.. no leading?
Posted: Thu Mar 10, 2022 11:43 pm
by thewood1
Yes. They've done nothing to improve anything in years, let alone air combat. If you are that disappointed in the development of the game, I'll buy your copy from you and you can move on with a clear conscience.
Re: WVR dogfights.. no leading?
Posted: Fri Mar 11, 2022 1:50 am
by stww2
I remember the Dark Ages (i.e CMANO) back before an aircraft immediately astern could trigger auto evasion...definitely prefer the system in CMO!
Re: WVR dogfights.. no leading?
Posted: Fri Mar 11, 2022 6:35 am
by Figeac
thewood1 wrote: Thu Mar 10, 2022 11:43 pm
Yes. They've done nothing to improve anything in years, let alone air combat. If you are that disappointed in the development of the game, I'll buy your copy from you and you can move on with a clear conscience.
Deal.
Re: WVR dogfights.. no leading?
Posted: Fri Mar 11, 2022 11:27 am
by thewood1
PM me your venmo handle. Once I send the money, send me all the serial numbers.
Also, once the TX is completed, post here that its done. That should be the last time you need to post on the CMO forum.
Re: WVR dogfights.. no leading?
Posted: Sat Mar 12, 2022 3:18 am
by wyskass
thewood1 wrote: Fri Mar 11, 2022 11:27 am
PM me your venmo handle. Once I send the money, send me all the serial numbers.
Also, once the TX is completed, post here that its done. That should be the last time you need to post on the CMO forum.
Complaining is one of the main enjoyments of forums. At least there is more response here than a wall, or blank stare from your spouse.
