Poll: More Survivable Wingman AI?

[D-Scythe]

I'm not speaking for ED or anything, this is simply something I'm interested in. Would anyone like the wingman AI to be made a bit more survivable? Through realistic means, of course ;)

 

I know technically it isn't "realistic" to have your wingman be superior to normal AI, but this is more a gameplay issue I think. The AI is already pretty good at sacrificing themselves, and I just feel that wingman should at least be made a little bit harder to kill. I'm not really a huge fan of watching the other three guys in my four-ship be wiped out in a single volley.

 

Don't get me wrong, I don't mean to increase it's lethality or anything, to make them like uber-Darth Vader type pilots that kill everything in sight, but more like they'll last a bit longer. Increase their missile reaction range, or the rate of chaff/flares they drop, whatever. You know, so you don't have to micromanage them all the time.

 

Anyway, if you feel the same, please vote and make this issue known :)

[SuperKungFu]

yea it would be nice to increase the AI. So the trip back home wont be so lonely :D

[Weta43]

Agreed

[S77th-GOYA]

Missle evasion is not their problem. It runs deeper than that. They're morons.

[SuperKungFu]

lol well said...all of you.

[D-Scythe]

Missle evasion is not their problem. It runs deeper than that. They're morons.

 

I dunno about that. With the right kind of managing, they can be pretty good at helping you kill things, although they have some serious survivability issues.

 

Just some ways you can effectively manage your wingman:

 

• Tell them to split low about 5 miles out of missile range then tell them to engage. This helps them kill things (look up).
  
• When they are trying to reform, slow down to about 600-700 kmph (~385 kts) and wait for them to rejoin. They'll save a LOT more fuel that way.
  
• Telling them to split left or right can help you with any targeting problems you may encounter. For example, in a two vs. two, splitting up your own flight will also cause the enemy flight to split in two. Thus, now you know exactly where your enemy is.
  
• Wingmen are also pretty good with A/G missiles.
  

 

You're right in that missile evasion won't solve the basic problem, but it would keep them alive longer and thus make the whole wingmen managing ordeal more forgiving and gameplay friendly. IMO, of course. There are obviously other ways at looking at the problem - a complete overhaul of the AI system for example would be welcome. But that unfortunately isn't possible.

[S77th-GOYA]

Earlier tonight I fired 120s at a head-on 25T and an escort 27 (both AI) from 7nm. Both missed. Someone else got the T and I came back at the 27. We both fired from 7nm again as he was turning into me. The 120 missed again. This is with a missle slider at 100%. You want the AI to be better at missle evasion?

[Harven]

Did the missles miss as in it just flew by.. or did it go after some chaff?

[D-Scythe]

Earlier tonight I fired 120s at a head-on 25T and an escort 27 (both AI) from 7nm. Both missed. Someone else got the T and I came back at the 27. We both fired from 7nm again as he was turning into me. The 120 missed again. This is with a missle slider at 100%. You want the AI to be better at missle evasion?

 

Shoot from BELOW your target Goya. Set up your AMRAAM so that at the point of impact, the missile is mostly pointing up at the target and mostly blue sky. You'll double your PK easy this way.

 

And no, I don't want the AI in general to be better at missile evasion. It's just that I feel the whole "wingman control" system not to be as extensive as it should be, and that player wingman AI is more prone to flying straight into SAM sites, burn off too much fuel using AB or some other suicidal manuever (e.g. while trying to rejoin with you, trying to attack your enemy), thus IMO they seem a bit more vulnerable than normal AI.

[Konovalov]

From what I have read here you have to micro-manage the AI wingman to avoid them getting killed. I voted yes but not in the sense that the wingman should be better at dodging inbound missiles but rather in their overall AI behaviour and not getting themselves into crazy suicide situations.

[ARM505]

I get tired of watching them loiter over SAM sites, oblivious to the beeping of the RWR, then reacting with brilliant evasion tactics the millisecond the missile is actually launched. The problem isn't their evasion skills (which are uncannily good sometimes) it's the skill of not getting into that situation in the first place. They should have some sort of situational awareness model, that says 'ok, I got a spike from there, and now it's there. There should be a SAM roughly there then, so best I give that general zone the appropriate respect, and maybe I should fly a bit lower or something...' Something like that. Easier said than done I suppose.

[D-Scythe]

Yeah, I know exactly what you mean. My point is, however, that since wingmen have the tendency to put themselves in a dangerous situation, more so than computer AI, that they do in fact end up in the end being more easily shot down since missile evasion logic is the same for all AI.

 

So maybe making their missile evasion logic better than the AI in a small way would be a good way to keep them alive longer and make the whole frustration that comes with trying to manage them a bit lower. You know, to compensate for their increased tendencies to put themselves in danger. Just my opinion.

[SwingKid]

For anyone who answered "yes":

 

Show us how.

 

Let's see a trk of YOU evading a launched missile, from an equal starting point, "better" than an Excellent AI.

 

-SK

[D-Scythe]

For anyone who answered "yes":

 

Show us how.

 

Let's see a trk of YOU evading a launched missile, from an equal starting point, "better" than an Excellent AI.

 

-SK

 

Um, it's not that hard for a player to evade missiles...just check the stats on-line. Often times, just pumping out decoys two or three pairs at a time would easily fool any missile, even at 100% setting.

 

Plus there are manuevers that can defeat missiles even without decoys. A simple beam plus a max G pull up seconds before impact, a relatively simple manuever (beam then pull up), can almost gaurantee you're survival, at any portion of the missile's employment zone.

 

The main problem for most players isn't evading missiles, but rather being able to see the missile to react to it. Once you know the intricacies of how missile works (i.e. decoys most effective at low aspects, poor look-down performance) in LOMAC, it's actually not that big of a problem.

 

I think this is the right track, I'm not sure. Missile slider at 100%.

4vs.4.zip

[SwingKid]

Plus there are manuevers that can defeat missiles even without decoys. A simple beam plus a max G pull up seconds before impact, a relatively simple manuever (beam then pull up), can almost gaurantee you're survival, at any portion of the missile's employment zone...

I think this is the right track, I'm not sure. Missile slider at 100%.

 

Ok you showed me, that was pretty impressive by any standard, nicely done!

 

All the AI however are of "Random" skill, and this appears to make a difference, since some of the AI are foolishly trying to climb away from the missiles (thus critically slowing down), instead of correctly dive and beam.

Beaming and pulling up seems to also require being at supersonic speed to make it work.

 

After the initial encounter, most of the AI are also shooting downwards, while the player is then shooting upwards - good offensive tactics by the player, but debatable whether it could be called "equal starting point" for evaluating AI missile evasion.

Even so, the AI manages to evade every third missile (in the "Joust" trk).

 

So, while dazzling fun to watch, I'm not sure what concrete tips it's pointing out to future AIs, besides "ask D-Scythe to make you Excellent". ;)

Does an AIM-120 Pk against Random AI of ~70% seem unreasonably high?

 

Thanks!

 

-SK

[D-Scythe]

Ok you showed me, that was pretty impressive by any standard, nicely done!

 

All the AI however are of "Random" skill, and this appears to make a difference, since some of the AI are foolishly trying to climb away from the missiles (thus critically slowing down), instead of correctly dive and beam.

Beaming and pulling up seems to also require being at supersonic speed to make it work.

 

After the initial encounter, most of the AI are also shooting downwards, while the player is then shooting upwards - good offensive tactics by the player, but debatable whether it could be called "equal starting point" for evaluating AI missile evasion.

Even so, the AI manages to evade every third missile (in the "Joust" trk).

 

So, while dazzling fun to watch, I'm not sure what concrete tips it's pointing out to future AIs, besides "ask D-Scythe to make you Excellent". ;)

Does an AIM-120 Pk against Random AI of ~70% seem unreasonably high?

 

Thanks!

 

-SK

 

Yeah, the tracks were recorded when I "discovered" that a look up situation is much better for radar missiles than a look down situation. Thus, even though everyone starts at high altitude (my equal starting point I guess), I did stack the odds with a 50-60 degree dive in my first move BVR.

 

Accordingly, PK for my missiles were pretty high, much higher than what it would be in any other situation in Lock On. I think that is a LOMAC thing though, as the PK drops drastically in a look-down situation, cause all AI, Average or Excellent, can execute a perfect beam thus lose itself in ground clutter in a look down situation.

 

In any case, the AMRAAM has a 60% PK combat record, so 70% against Random AI isn't too high, especially if the player employs his weapons in its most effective zone.

 

I think the kill ratio between my wingmen and the enemy fighters was pretty bad, like 2 to 1 or something. Thus it brings me back to my point that even with micromanaging, the player's wingmen put themselves in danger more often then normal AI, thus seems more vulnerable, to me at least.