homepage

AI use of "negative G" maneuvers.

Posted By: DukeIronHand

AI use of "negative G" maneuvers. - 11/28/16 09:17 PM

This is something I think needs a look at and I can't be the only one seeing it.
I have seen this fairly often in UE when pursuing an enemy but eh....things happen once in a while.

However was flying today with the "Hat in the Ring" squadron with N28's. Had a German formation below us and the entire squadron (both A and B Flights) performed a shallow dive for target. After about 20 seconds it was clear the intercept angle was wrong (I was not leading) so instead of doing something "logical" (like turning or circling for another pass) the AI squadron aircraft used the old "the shortest distance between two points is a straight line" thing and everyone (except me) pushed their nose way over and went into a negative G half loop/dive which resulted in a bit more then 1/3 of them stalling out or something but losing control and fluttering downward. Unknown if with stress/Over G damage or not. Looked a bit ridiculous for 8 or 10 planes.

Just FYI.
Posted By: Banjoman

Re: AI use of "negative G" maneuvers. - 11/28/16 09:48 PM

I see it quite frequently. Next time, let the AI control your plane in the initial stages of an encounter, especially when the enemy is below you. The AI will put your plane in a negative g maneuver.
Posted By: Winding Man

Re: AI use of "negative G" maneuvers. - 11/29/16 09:44 AM

Originally Posted By: DukeIronHand
This is something I think needs a look at and I can't be the only one seeing it.
I have seen this fairly often in UE when pursuing an enemy but eh....things happen once in a while.

However was flying today with the "Hat in the Ring" squadron with N28's. Had a German formation below us and the entire squadron (both A and B Flights) performed a shallow dive for target. After about 20 seconds it was clear the intercept angle was wrong (I was not leading) so instead of doing something "logical" (like turning or circling for another pass) the AI squadron aircraft used the old "the shortest distance between two points is a straight line" thing and everyone (except me) pushed their nose way over and went into a negative G half loop/dive which resulted in a bit more then 1/3 of them stalling out or something but losing control and fluttering downward. Unknown if with stress/Over G damage or not. Looked a bit ridiculous for 8 or 10 planes.

Just FYI.


This is tunable - we can look for an upcoming patch - maybe I'll make a workshops selection

WM
Posted By: DukeIronHand

Re: AI use of "negative G" maneuvers. - 11/29/16 09:50 AM

Thank you sir.
Posted By: Polovski

Re: AI use of "negative G" maneuvers. - 11/29/16 11:21 AM

Any FM mods or mods installed? Needs more details to help really. Is is always N28?

I just ran a few runs above the enemy with N28's and they all did natural dives, spirals etc generally.

These were 8 N28s v 8 DVas approx 3000 feet almost directly below. veterans.
Will try some with aces.

Do you have an idea of height distance apart etc?

Posted By: DukeIronHand

Re: AI use of "negative G" maneuvers. - 11/29/16 12:44 PM

This "negative G" thing happens with ALL (that I have flown) aircraft types on multiple occasions - mods or no as I have seen it both ways - and it's not just a N28 thing.
It does not always occur but certainly quite a bit.
Seems about a half to 1/3 when you are pursuing an aircraft. Unknown the AI "thinking" here. They are trying to get away - maybe on where they want to go?

With entire formations it certainly depends on, ultimately, where the AI are in the attack and the position of the enemy and the path the AI "thinks" is best to get there when the enemy is lower then you are.
Fly a few missions on AP and you will certainly see it in your aircraft when the enemy is lower.

For the above case the enemy formation was at our 10 o'clock and (guessing) 1500-2000' lower. Started with a shallow dive toward the target then, when above them, the AI goes into negative G dives/half-loop looking maneuvers to get at them.
Seen this in Se5's in the attack and in Albs when diving to evade.

This is the first time I have seen an entire formation do it (about 10 aircraft) but I don't always have a great view of everything around me like I did here.
Posted By: Hellshade

Re: AI use of "negative G" maneuvers. - 11/29/16 01:29 PM

Originally Posted By: DukeIronHand
This "negative G" thing happens with ALL (that I have flown) aircraft types on multiple occasions - mods or no as I have seen it both ways - and it's not just a N28 thing.
It does not always occur but certainly quite a bit.
Seems about a half to 1/3 when you are pursuing an aircraft. Unknown the AI "thinking" here. They are trying to get away - maybe on where they want to go?


I have never seen an entire formation do this, but I have seen individual aircraft do it on multiple occassions. Spads, Albs, Roland CIIs, Nieuports that I can recall but I feel certain that the behavior can be seen in nearly any type of craft. As Duke says, its not all the time to be certain but its not entirely unusual either. Most often I've seen it when they try to dive down to escape from me. If it could be fixed, that would be fantastic. I get how intertwined everything is though. Easy to fix one problem and unintentionally break 3 other things. Hopefully it's possible to find a balanced resolution to the issue.

Again, I have to give the highest praise to OBD for bug fixes after the sale.
Posted By: Winding Man

Re: AI use of "negative G" maneuvers. - 11/29/16 01:43 PM

Ran a lot of missions N28 Vs two seaters not seen it at all.

Pol has seen one aircraft do it in todays series of tests (similar to Hellshade I guess)- never a whole flight/most of the flight as you have seen.

If you are using the FM mods for N28 - take them off please.

What I am looking at doing is implementing a workshop switch that makes Vets/Novices even more less likely to do inverted maneuvers - but that is a separate issue from what you are seeing.

This switch is not a 'bug fix' at all but rather a 'preference switch' as to how gungho you think Novices and Vets should be, and we all have differing opinions as usual....
But no ETA on that yet - probably only next year


WM
Posted By: DukeIronHand

Re: AI use of "negative G" maneuvers. - 11/29/16 06:15 PM

If you have a moment what is meant by "...less inverted maneuvers?"
Except for this quirk the AI is doing quite well.
And, FWIW for your investigation, I don't recall this ever in WOFF 3 that I recall.

And, to be clear, this is the first time I have "seen" a formation do it (certainly the bulk of it) as I just happened to have a great view at the moment - I was flying on the left side of the formation and was looking over trying to find the enemy. Maybe the first or maybe a dozen times though logic would suggest if the flight leader does it the chance for the rest (in trying to follow) goes up mightily. I can't say though it is certainly NOT an uncommon AI solo move in my experience.

Thanks for you attention on this.
Posted By: Deacon211

Re: AI use of "negative G" maneuvers. - 11/30/16 01:41 AM

If it helps, I think Duke said something fairly significant.

The AI doesn't usually seem to choose negative G under normal circumstances. It never, for instance, chooses a belly in turn.

What it does occasionally seem to do is choose a negative G pushover as a guns defense (not a bad choice) but then "lose" its preference for positive G maneuvers. So, a negative G pushover might be followed by a negative G turn, even though a half roll and positive G would probably be more desirable.

Not sure it that helps any. I had it in WOFF3 as well, very rarely.
Posted By: JJJ65

Re: AI use of "negative G" maneuvers. - 11/30/16 06:29 AM

Yes, I have experinced this phenomenom only when AI is pursuing enemy aircraft too below (too low) and when enemy is closing. AI turns from steep dive (IMHO out of safe envelope either) to negative G dive turn. I have learned in bitter way to never take control of aircraft at this phase of flight wink.
Posted By: DukeIronHand

Re: AI use of "negative G" maneuvers. - 11/30/16 08:58 AM

Originally Posted By: Deacon211
If it helps, I think Duke said something fairly significant.

The AI doesn't usually seem to choose negative G under normal circumstances. It never, for instance, chooses a belly in turn.

What it does occasionally seem to do is choose a negative G pushover as a guns defense (not a bad choice) but then "lose" its preference for positive G maneuvers. So, a negative G pushover might be followed by a negative G turn, even though a half roll and positive G would probably be more desirable.

Not sure it that helps any. I had it in WOFF3 as well, very rarely.


Good point.
The continued use of the negative G maneuvers (for example an attempt at a negative G turn after the negative G dive) probably explains why some aircraft lose control (stall out?) while doing it. And I am fairly sure WOFF models fuel starvation in such maneuvers (which I avoid) so loss of engine power is probably another factor.

And it might be considered a great defensive move when pursued (I am reluctant to do it for a host of reasons when pursuing) but in...wait for it..."real life" I think negative G manuvers were rarely (never?) done for a lot of good airframe and engine reasons.
Posted By: Deacon211

Re: AI use of "negative G" maneuvers. - 11/30/16 01:21 PM

Negative G does have its uses, but they are mostly transitory, throwing off an opponent's aim, presenting a false aspect, etc. For displacing the aircraft however, there's a natural preference for positive G. But I've never gathered that the few occasions that you see the AI attempt to fly hanging in the straps that they were attempting something particularly tactical. It looks more like they would like to be upright, but can't decide the best way to go about it.

Purely an impression.
© 2024 SimHQ Forums