WWII BAD-T V: The AI Strikes Back - BD AI Dogfight Tournament

[Box of Stardust]

So far, the true winner of this competition seems to be GLOC.

I suddenly don't have high hopes for my own plane.

[SuicidalInsanity]

At the risk of spoilers, you needn't worry about the Sea Fairy. The next few fights are free of GLOC incidents.
As for those craft that have knocked themselves out, the second tier ladder will be fought with GLOC off, A) so we can see what these craft are actually capable of, and B) to avoid matches that consist of both teams uncontrollably cartwheeling across the sky until one runs out of fuel/into the ground.

[Box of Stardust]

Yeah, I suspect the single engine fighters, given that they have much less power, are less prone to GLOC-ing themselves in most cases outside of a high(er) speed dive.

 

[Pds314]

9 hours ago, SuicidalInsanity said:

I can always throw up a bonus video with it at some point.

But back to the main event. Match 4, between @dundun92's Tallyhawk MkXXVIa and @splatn't™'s TH-C.

 

That was... not what I expected. Congrats to @splatn't™ and the TH-C for surviving against the Tallyhawk though.

2 hours ago, Box of Stardust said:

Yeah, I suspect the single engine fighters, given that they have much less power, are less prone to GLOC-ing themselves in most cases outside of a high(er) speed dive.

 

That's true but they also have a better potential power/weight. And don't have the drag of extra wing area and engines that most 2-engine designs have. In practice, this can often be offset by the ammo weight, which will pretty much always be proportionally greater in single engine designs, but I'm not sure how big the difference really is.

Edited October 27, 2019 by Pds314

[Jeb-head-mug kerman]

2 hours ago, Pds314 said:

That was... not what I expected. Congrats to @splatn't™ and the TH-C for surviving against the Tallyhawk though.

That's true but they also have a better potential power/weight. And don't have the drag of extra wing area and engines that most 2-engine designs have. In practice, this can often be offset by the ammo weight, which will pretty much always be proportionally greater in single engine designs, but I'm not sure how big the difference really is.

Wow.... I definitely expected the Tallyhawk to win!

[splatn't™]

9 hours ago, Jeb-head-mug kerman said:

Wow.... I definitely expected the Tallyhawk to win!

The secret? Basically no AI tuning.

[GDJ]

12 hours ago, Pds314 said:

That's true but they also have a better potential power/weight. And don't have the drag of extra wing area and engines that most 2-engine designs have. In practice, this can often be offset by the ammo weight, which will pretty much always be proportionally greater in single engine designs, but I'm not sure how big the difference really is.

It can be significant. Twin engine planes typically have 3 times the drag (You're pushing more stuff thru the air) and are heavier. Careful planing and adjusting will help with the drag issue, but it takes a lot of tweaking (I tweaked the AVRO Defender to death and it has the same transonic drag as a single engine plane, but it took forever to adjust, and some of the adjustments were literally one pixel at a time, and that one pixel could mean a 5% increase or decrease).

[Box of Stardust]

43 minutes ago, GDJ said:

and some of the adjustments were literally one pixel at a time, and that one pixel could mean a 5% increase or decrease).

Yeah, thanks FAR, because moving my rudder down and making my vertical stabilizer shorter should worsen my L/D and Cl.

FAR is great, but it still ain't perfect...

[SuicidalInsanity]

Could be something to do with area ruling, if the shortened stabilizer resulted in a change in wave drag?
But, yes, the voxelization method FAR uses for determining aerodynamics is ultimately an abstraction that does leave room for errors.
Might be something to ask in the FAR thread.

....

Match 5, between @Box of Stardust's FB-97B-5 Sea Fairy and my IA-42XB Skaris:

 

[Pds314]

1 hour ago, Box of Stardust said:

Yeah, thanks FAR, because moving my rudder down and making my vertical stabilizer shorter should worsen my L/D and Cl.

FAR is great, but it still ain't perfect...

Complex and counterintuitive wing interactions are definitely realistic though. Not sure if what FAR does with them is but especially transonic aero doesn't always match expectations. TBH though I don't design prop planes for a bad AI to fly in a transonic windtunnel anymore. My first entries to older BADTs were designed that way but TBH even my jets entries were far too focused on transonic performance, considering that most combat tasks place below those speeds and they still weren't supersonic at sealevel the way some of my 2 engine designs were.

Edited October 27, 2019 by Pds314

[Box of Stardust]

I have to say, I'm a bit confused at the sequence of events and how the IA-42s always start off more split apart.

That said, I think my primary flaw was assuming that altitude is a good thing when dealing with low power aircraft, but seemed to forget my knowledge of BDA dictates that the AI really defines how battles play out, and that trying to gain altitude is a really, really dumb idea with BDA.

[SuicidalInsanity]

The starting far apart, at least, is something I've seen with other flyingwing type designs as well. The flightlead flies fine, but the AI for the wingman drunkenly wobbles behind the leader for some reason, bleeding energy; when the match starts, the lead turns towards the opposing team, while the wingman more slowly turns, and then immediately dives to try and regain speed.
Which events are you confused about?

 

[Box of Stardust]

The split off part, really. That seems... in some ways, advantageous.

[Pds314]

1 hour ago, SuicidalInsanity said:

The starting far apart, at least, is something I've seen with other flyingwing type designs as well. The flightlead flies fine, but the AI for the wingman drunkenly wobbles behind the leader for some reason, bleeding energy; when the match starts, the lead turns towards the opposing team, while the wingman more slowly turns, and then immediately dives to try and regain speed.
Which events are you confused about?

 

This also happens about 30% of the time for my plane since it has a pretty small rudder. I've seen the wingman end up at 35 m/s wobbling back and forth whilst the lead is at 200. Usually this means the battle will never start and you need to stop and restart the competition.

Edited October 27, 2019 by Pds314

[Pds314]

1 hour ago, Box of Stardust said:

The split off part, really. That seems... in some ways, advantageous.

In some ways, although it does make it more likely that both of the side that didn't split go for the same target and shoot it down before the other one can rescue it.

[SuicidalInsanity]

The AI quirk I'm wondering about in that match is the one that occurs starting about 7:34 or so, when the focused IA-42 starts an attack run against the crashed Sea Fairy, instead of heading to assist the other IA-42; all the guns have been set to only target air, so the AI shouldn't have been able to see the crashed plane as a target in the first place.
...

Sunday Double Feature! Match 6, between @Dwerto's PD-2 Honeybee and @The Dunatian's S 13 Brighton:

 

[Dwerto]

 

18 minutes ago, SuicidalInsanity said:

The AI quirk I'm wondering about in that match is the one that occurs starting about 7:34 or so, when the focused IA-42 starts an attack run against the crashed Sea Fairy, instead of heading to assist the other IA-42; all the guns have been set to only target air, so the AI shouldn't have been able to see the crashed plane as a target in the first place.
...

Sunday Double Feature! Match 6, between @Dwerto's PD-2 Honeybee and @The Dunatian's S 13 Brighton:

 

 

Good to see that the PD-2 didn't end up in permanent GLOC or any such thing. I particularly like the maneuver at 2:55, it seems I got an ace pilot for this fight.

Edited October 28, 2019 by Dwerto

[Pds314]

45 minutes ago, SuicidalInsanity said:

The AI quirk I'm wondering about in that match is the one that occurs starting about 7:34 or so, when the focused IA-42 starts an attack run against the crashed Sea Fairy, instead of heading to assist the other IA-42; all the guns have been set to only target air, so the AI shouldn't have been able to see the crashed plane as a target in the first place.
...

Sunday Double Feature! Match 6, between @Dwerto's PD-2 Honeybee and @The Dunatian's S 13 Brighton:

 

@Dwerto's PD-2 Honeybee is a scary plane. Accurate, fast, and maneuverable on all axes, with good evasion. Surprised it's so good with a large polar bear of extra weight. I could see that one being a winner.

@The Dunatian's S 13 is somehow even more overweight. Deciding to cram an American Bison if empty weight in there just in case I guess. Somehow ends up at over 4.6 tonnes loaded weight. So at launch it's got two Bisons strapped to it. Despite this, I feel like it suffers from very limited AOA performance and therefore low CL, compounding the moderate-speed turning issues and making it very crash-happy. I will say however that it does look pretty trying to dodge bullets like that.

Edited October 28, 2019 by Pds314

[Mr. Peabody]

10 hours ago, Pds314 said:

@The Dunatian's S 13 is somehow even more overweight. Deciding to cram an American Bison if empty weight in there just in case I guess. Somehow ends up at over 4.6 tonnes loaded weight. So at launch it's got two Bisons strapped to it. Despite this, I feel like it suffers from very limited AOA performance and therefore low CL, compounding the moderate-speed turning issues and making it very crash-happy. I will say however that it does look pretty trying to dodge bullets like that.

Eh bien. I tried. 

At least we will know for next time. Vivre et apprendre...

 

[SuicidalInsanity]

Match 7, between two familiar aircraft - @sturmhauke's SAI K9 Krakenhound, returning from last year, and @OmegaForce's ACF-88 Thalo, drawing inspiration from BAD-T II's Tytonid. Will the Thalo continue its precursor's legacy, or will it slip up  before the hounds?

 

 

[sturmhauke]

Cry havok!

[Pds314]

4 hours ago, SuicidalInsanity said:

Match 7, between two familiar aircraft - @sturmhauke's SAI K9 Krakenhound, returning from last year, and @OmegaForce's ACF-88 Thalo, drawing inspiration from BAD-T II's Tytonid. Will the Thalo continue its precursor's legacy, or will it slip up  before the hounds?

 

 

That is probably the fairest fight we've seen so far, especially if you discount the Thalos GLOCing out on the first round.

[Pds314]

My predictions for match 8:

At least two AL-25s suffer an RUD via controlled flight into terrain, without being unconscious.

Any match where the initial separation is over 11 km due to wingman wobble or something will be a victory for the AL-25s due to an La-9 entering GLOC at 220+ m/s.

At least one La-9 will GLOC during one of the matches. This will not result in a crash.

All matches will be pretty short unless someone goes after ground targets.

At least one AL-25 will suffer loss of an engine while remaining controllable.

No AL-25s will GLOC.

At least one match will lose one combatant on both sides from the initial merge or before they can reconverge, resulting in a 1v1.

Edited October 29, 2019 by Pds314

[OmegaForce]

back engineering

On 10/28/2019 at 7:30 PM, SuicidalInsanity said:

Match 7, between two familiar aircraft - @sturmhauke's SAI K9 Krakenhound, returning from last year, and @OmegaForce's ACF-88 Thalo, drawing inspiration from BAD-T II's Tytonid. Will the Thalo continue its precursor's legacy, or will it slip up  before the hounds?

 

 

 

The Thalo came about from recreating/back engineering Ferram's Tytonid somewhat accurately . Their was many issues encountered with an asymmetrical design and more of a trial and error approach was taken to get a flyable aircraft. The GLOC problem though was probably on account of not knowing what the G limits of the pilots would be. I would also say after seeing the match accuracy could be improved upon as the 30mm's were more of a late design change.

If anyone would like to take a look at this design here's the craft file:

https://www.dropbox.com/s/ijw62pnd1r8te62/AFC-88 Thalo.craft?dl=0

Edited October 30, 2019 by OmegaForce

[SuicidalInsanity]

Match 8, between @Pds314's La-9 DPRK and @Alioth81's AL-25 Vulture: