Previous Thread
Next Thread
Print Thread
Rate This Thread
Hop To
#4387085 - 10/28/17 03:57 PM Ripping the wings off the AI -- like a bug!  
Joined: May 2016
Posts: 303
BuckeyeBob Online content
Member
BuckeyeBob  Online Content
Member

Joined: May 2016
Posts: 303
Ohio, USA
I was reading the Developer's comments.pdf file found in the AI Rules of Engagement thread when I saw the following, which reminded me of something that I, and a few others, have noticed about the AI’s ability to avoid wing damage in a steep dive:

Quote
All AI craft now use the same flight models and engine models as the player craft - they have no advantages, so they are flying the same complexity of flight model as the player.


From the above quote, it seems that the AI is limited by the same aerodynamic factors (wing speed damage, wing stress damage, etc.) as the human player, but I have never witnessed an AI flown craft shed any of its wings in a dive, unless it has already been severely damaged by gunfire.

Is this because the AI automatically knows the exact instant to slow his descent in order to avoid wing damage, or is the AI truly limited by the same FM factors as the human player?

Assuming it is the former, is there a way to limit this AI super-human ability somewhat? For instance, might there be a way to limit this ability by skill level, so that rookies might occasionally “forget” to slow their rate of descent while an Ace pilot never does?

JJJ65 has discovered that wing overspeed damage results from staying above an FM controlled wing damage speed for a set number of seconds. Would it be possible to introduce a random delay of 0 to 10 seconds, for example (0 for Aces, 0 to 10 for rookies) before the AI “reacts” to wing overspeed damage?

Perhaps this post better belongs in the wishlist thread, but it seems to me this might be a more realistic way to model AI behavior in regard to this important aspect of WWI aviation. For the sake of argument, I am avoiding talking about how difficult this might be to code, but that is obviously a factor as well.

Sorry for the long post. I would be interested to hear other posters observations or comments on the matter, though.

#4387109 - 10/28/17 06:31 PM Re: Ripping the wings off the AI -- like a bug! [Re: BuckeyeBob]  
Joined: Jul 2010
Posts: 1,134
JJJ65 Online cool
Member
JJJ65  Online Cool
Member

Joined: Jul 2010
Posts: 1,134
Czech Rep.
+1
I have made several tests of overspeed damage. For these tests I have set an overspeed damage for AI planes to 60 miles per hour and overspeed damage time to 0 seconds. Then I have performed the test dives. While player controlled plane disintegrated after a short time at speed 180 mph (the preselected overspeed damge value for player plane was set to 180 mph), AI planes following in close formation player craft in dive did not sustained any damage even their dive speed was far more than 60 mph. But, according to AI FMs overspeed damage value set to 60 mph they should have received an overspeed damage at 60 mph and should be destroyed earlier than player controlled plane. confused

#4387119 - 10/28/17 08:18 PM Re: Ripping the wings off the AI -- like a bug! [Re: BuckeyeBob]  
Joined: Nov 2014
Posts: 1,413
Fullofit Offline
Member
Fullofit  Offline
Member

Joined: Nov 2014
Posts: 1,413
Ajax, ON
+ 1000 Haven’t seen AI rip a wing off from overstress. Ever. Clear evidence of damage not being applied to AI is in a dive. Unless you are in a SPAD you cannot follow AI in a dive without damage, even when AI is flying the same plane. Further proof of that can be had if you leave your own plane under AI control. The AI controlled plane will be able to reach incredible speeds in a dive without any damage. As soon as you turn the AI control of your plane off you will incur speed damage. I would really like to see novice AI planes rip their wings off or fold their wings occasionally (I know, that damage is not modelled). In fact, they should sometimes disintegrate while falling out of control after being shot down. Typically they fall in one piece. Thanks for mentioning it BuckeyeBob and Triple-J.
Pol, is this something that one day can be “fixed”, or at least point the modders in the right direction how to tweak it? salute


"Take the cylinder out of my kidneys,
The connecting rod out of my brain, my brain,
From out of my arse take the camshaft,
And assemble the engine again."
#4387129 - 10/28/17 09:30 PM Re: Ripping the wings off the AI -- like a bug! [Re: BuckeyeBob]  
Joined: Jun 2014
Posts: 2,391
Banjoman Offline
Member
Banjoman  Offline
Member

Joined: Jun 2014
Posts: 2,391
Antigua, Guatemala
+1


Member and provider of banjo music for the Illustrious BOC
#4387136 - 10/28/17 10:17 PM Re: Ripping the wings off the AI -- like a bug! [Re: BuckeyeBob]  
Joined: Mar 2011
Posts: 20
VonS Offline
WWI Flight Sims on a Mac
VonS  Offline
WWI Flight Sims on a Mac
Junior Member

Joined: Mar 2011
Posts: 20
A stimulating conversation gents'. While I wouldn't know anything about FM tweaking for WOFFue, from what I'm reading in the posts here, I am assuming that the AI can tolerate higher Maximum G limits across lift surfaces than the human-flown plane can - this is something that I discovered for FE2. Implementation of the same MaxG values across lift surfaces (for left/right elevator, also wing panels), for both the player aircraft and AI-flown craft solves this problem.

On occasion wings/tail sections now fall off in FE2 for the AI as well - especially if the AI has been hit and enters a dive, or reckless AI behavior now has a greater chance of breaking things off the plane. I'm sure there's a way to implement similar tweaks in WOFFue. By the way, WOFFue is a beautiful flight sim - hope to spend more time in it in the future (FM tweaking for FE2 takes up a lot of my time).

Happy flying,
Von S smile

Last edited by VonS; 10/28/17 10:18 PM. Reason: Fixed some typos.
#4387152 - 10/29/17 01:33 AM Re: Ripping the wings off the AI -- like a bug! [Re: BuckeyeBob]  
Joined: May 2016
Posts: 303
BuckeyeBob Online content
Member
BuckeyeBob  Online Content
Member

Joined: May 2016
Posts: 303
Ohio, USA
Thanks for the replies, gentlemen. And welcome to WOFF, Von S! I know what you mean by spending all of your time tweaking. wink

JJJ, from your testing, it seems the AI's ability is working as designed, perhaps to give the AI a fighting chance (especially the aces) against the human player. Or, it could be a bug.

Fullofit, perhaps when WOFF was under development, the AI simply crashed too often if they were subject to the same wing speed damage setting as the human? Or, maybe it has something to do with the conversion of a WWII simulator (CFS3) into a WWI simulator (WOFF)? Due to engineering advances, pilots in WWII did not have to deal with the problem of wings coming off in a dive anymore -- except perhaps in catastrophic situations. So, getting the WOFF AI to deal with this may be a bit of a problem?

#4387167 - 10/29/17 07:04 AM Re: Ripping the wings off the AI -- like a bug! [Re: BuckeyeBob]  
Joined: Jul 2010
Posts: 1,134
JJJ65 Online cool
Member
JJJ65  Online Cool
Member

Joined: Jul 2010
Posts: 1,134
Czech Rep.
I think that AI pilots do not read flight manuals and do not keep their speeds and overG limits in safe margin. That is why the DEVs decided to release the AI pilots from overspeed and overG damage testing.
Just my 2 cents.

#4387186 - 10/29/17 01:04 PM Re: Ripping the wings off the AI -- like a bug! [Re: JJJ65]  
Joined: May 2016
Posts: 303
BuckeyeBob Online content
Member
BuckeyeBob  Online Content
Member

Joined: May 2016
Posts: 303
Ohio, USA
The devs may not have had a choice, given CFS3 was originally a WWII simulator. By 1940, damage from over-speed and over-G was likely no longer a concern, except maybe in extreme circumstances. So, perhaps in CFS3 (and now WOFF), the AI never "sees" this as a problem.

This is another reason why I wish we could modify induced and parasitic drag, which was possible before WOFF 3. Some planes in WOFF, IMHO, can still reach faster dive speeds than I think possible in real-life, given the poor aerodynamics of the time, including large wing surface areas due to multiple wings, plus all those wires and struts, etc. Others, of course, like the Albatros series, could dive comparatively fast, but suffered from inherent wing weakness which also effectively reduced their dive speed. WWII pilots largely no longer had to deal with these issues, so it doesn't surprise me that our WOFF AI pilots don't either!

At least if we could control the drag variables, we could "slow" the maximum speed that could be obtained in a dive to more realistic (IMO) levels.

#4387257 - 10/29/17 09:52 PM Re: Ripping the wings off the AI -- like a bug! [Re: BuckeyeBob]  
Joined: Mar 2011
Posts: 20
VonS Offline
WWI Flight Sims on a Mac
VonS  Offline
WWI Flight Sims on a Mac
Junior Member

Joined: Mar 2011
Posts: 20
I seem to remember an artificial and/or induced drag "factor" that was implemented back in the days of RB3D, either to control the max. ceiling of the aircraft in RB3D...or for some other reason. Perhaps introduction of an extra, artificial variable in WOFFue would achieve a similar effect, of acting as a brake on max. dive speeds.

The SPADs and Se5a, among others, could come close to 430-450 kph in a near vertical dive (about 300 mph) without losing wings, providing that dives were exited gently. Albatroses, for example, while slippery and aerodynamic, should avoid speeds beyond 250-260 kph in a dive, and should especially avoid rough exits from dives exceeding such speeds. Have never tested AI dive speeds in WOFFue, but I tend to avoid diving when flying in the Alby. and Nieuport types, unless I'm flying the Alb. D.I and D.II, or the Pfalz.

Anyway, keep the thoughts coming gents' - very stimulating reading.

Happy flying,
Von S


Moderated by  Polovski, Sandbagger 

Quick Search
Recent Articles
Support SimHQ

If you shop on Amazon use this Amazon link to support SimHQ
.
Social


Recent Topics
How Marines Recover a Lost Cell Phone
by Arthonon. 11/21/17 04:34 PM
Finally dead... Charles Manson
by Raw Kryptonite. 11/20/17 11:47 PM
Geostorm
by DM. 11/20/17 10:52 PM
Radioactive pet
by Chucky. 11/19/17 05:54 PM
Another Icon of My Childhood Gone: Mel Tillis
by vocatx. 11/19/17 04:59 PM
Big John W
by Haggart. 11/18/17 08:02 PM
Malcolm Young RIP
by JimK. 11/18/17 02:16 PM
Obscene Pilots
by Timothy. 11/18/17 05:23 AM
Powered by UBB.threads™ PHP Forum Software 7.6.0