Top Gun AI - The Official Tournament Thread

[Alphasus]

What was the steer limiter?

[RuBisCO]

Well I set it to 0.3

[CreativeCombat]

What was the steer limiter?

Well I set it to 0.3

Having your AI set cleverly matters a lot more than having a better plane.

[GDJ]

Well I set it to 0.3

Wow. That usually isn't enough elevator to keep a plane flying unless it's balanced to the pixel (CL inside the CG sphere. A touchy balance in the best of planes).

The problem with box planes like the video portrayed or tubular fins (if you're into rocketry) is the tubes only allow so much air to pass thru them (like Alphasus was hinting at) and at a certain speed, the tubes cannot pass anymore air. So air piles in front of the tubes and you essentially have a brick wall or enough air pressure at the front of the tube to overcome the air trying to rush in, and it spills over. That's where the drag really hits tubular fins or box planes (they're essentially the same thing).

There was some supersonic tube fin testing done by the British in the late 1950's to early 1960's and it basically ended up that if one ran a cone that had a 3 degree taper, smaller end at the front and bigger end at the back, there would be sufficient vacuum at the rear of the tube that it would suck any air that would build up at the front, hence allowing more air to enter the tube fin assembly. Unfortunately they also concluded that the tube fin is not suitable for supersonic flight due to shock waves bouncing inside the tube, creating more pressure fronts, blocking more air, and eventually the tube fin self destructed from over pressurization.

That's where Grid Fins come in. They are really REALLY bad at drag in the subsonic to transonic region, but when they are exposed to supersonic air velocities, the shock waves don't build up because the thickness of the Grid Fin is so small. They are very good at low drag for supersonic flight (some missiles have grid fins, as well as bombs that are deployed at supersonic velocities).

Edited September 12, 2015 by GDJ

[ferram4]

Some points worth noting on the box design compared to the other designs:

1) it had much more lifting area, which would have made it much more maneuverable.

2) it had much more in the way of control surfaces, which would have made it more maneuverable.

3) it appears to have the CoL closer to the CoM, which would have certainly made it more maneuverable.

4) it had 3x as much engine, which would have made all the extra drag from the extra wing area irrelevant.

5) wing area in virtually every direction ensures that it has some extra protection for deflection shots.

6) using multiple parts for the wings ensures that damage at one location doesn't wipe out a large wing part, ensuring that damage is minor.

7) keeping everything close together ensures that for minor damage that doesn't destroy a part, heat can quickly dissipate throughout the vehicle, effectively nullifying all of the damage.

It is a well designed vehicle that is clearly optimized wrt stock's aero model and BDArmory's heat-damage model, as well as being able to get guns on target to force the enemy AI to go into evasive maneuvers and become less of a threat. Yes, it's ugly, but when you have that much engine, it's certainly realistic to be that ugly and still fly.

[CreativeCombat]

The second match, MarsRocks vs SasquatchM is uploaded

Round 1:

SasquatchM takes the victory.

Round 2:

The recording cut off for some unknown reason, but it eventually ended in a draw.

Round 3:

MarsRocks takes the victory.

Round 4:

SasquatchM takes the victory.

Round 5:

Everything dies in an explosion so massive and cool that it crashes the game. Technically one of MarsRock's ships was still flying, but only as a cockpit with a wing strapped to it. Draw.

That's all for tonight. I'll continue this tomorrow.

- - - Updated - - -

I did about an hour of independent testing on that exact subject. The elevon 4s offered better performance in all aspects.

Pull up with the Elevons (press S) before the plane has enough speed to take off. If you have enough control surfaces strapped on, it'll decelerate.

Edited September 12, 2015 by CreativeCombat

[RuBisCO]

The second match, MarsRocks vs SasquatchM is uploaded

Sooo SasquatchM wins?

[Alphasus]

Some points worth noting on the box design compared to the other designs:

1) it had much more lifting area, which would have made it much more maneuverable.

2) it had much more in the way of control surfaces, which would have made it more maneuverable.

3) it appears to have the CoL closer to the CoM, which would have certainly made it more maneuverable.

4) it had 3x as much engine, which would have made all the extra drag from the extra wing area irrelevant.

5) wing area in virtually every direction ensures that it has some extra protection for deflection shots.

6) using multiple parts for the wings ensures that damage at one location doesn't wipe out a large wing part, ensuring that damage is minor.

7) keeping everything close together ensures that for minor damage that doesn't destroy a part, heat can quickly dissipate throughout the vehicle, effectively nullifying all of the damage.

It is a well designed vehicle that is clearly optimized wrt stock's aero model and BDArmory's heat-damage model, as well as being able to get guns on target to force the enemy AI to go into evasive maneuvers and become less of a threat. Yes, it's ugly, but when you have that much engine, it's certainly realistic to be that ugly and still fly.

I perfectly understand the realism with that many engines. But it goes against the spirit of the contest by doing the same thing as wing clipping but better. Wing clipping added more wing area with less surface area. Boxes do the same thing. This happens with 1 engine box planes too. That is why you have to allow both or neither in a tournament, because they do the same thing.

Edited September 12, 2015 by Alphasus

[ferram4]

I gather by "surface area" you mean "projected area"... but under that understanding, wouldn't that ban all multi-planes, including X-wing designs and multiple vertical tails? Both provide the same advantage, that of lower projected area for specific directions for the same lifting area as their monoplane / single vertical tail counterparts. If not, then... what's the crucial, objective difference?

[Alphasus]

Easy enough, they have a larger surface area than boxes. X planes have a larger side surface area and multiple vertical tail counterparts cant fly while missing the tails unless they are spammed to insanity. With a box, it can lose 1/2 or more of its control surfaces and fly. The X plane can do that as well, but there is a noticeable loss in performance.

[RuBisCO]

So your saying there can't be redundant lift and control surfaces? Ok what is the spirit of the competition: beauty, realism. or to win within the rules provided?

MarsRocks's Bullet was the first box I encountered and it inspired me to make boxes, but try as a might mine were not as survivable as his. I had to figure out how to survive attacks from missiles and thus came to the epiphany that the real secret is speed to dodge missiles altogether. This comes from dropping the steer limiter and having enough engines to accelerate a flying brick to mach 2. Everyone was trying to make their planes really maneuverable and small but this makes them really slow, they lack engine power and they loss all their speed in the turning and bobbing, from a missile's perspective they are barely moving, easy to hit. A plane flying just straight but really fast is much harder to hit. The XIPHOS V is a normal plane design, but I decided to go back to the XIPHOS IV because with the steer limiter set so low and the stupidity of the AI, the plane is grossly unmaneuverable at the low speed needed for cannon combat. The AI can't adjust its steering limit dependent on speed or if it is attacking or dodging. Once out of missiles or close in the AI tries to do dog-fighting cannon combat regardless if the plane even has a cannon at all! So the XIPHOS V was woefully incapable of fighting close in dogfights, the XIPHOS IV on the other hand could take a lot of hits from cannons easy, and if both of the team survived the missiling stage of combat as they often did, then they could gang up on a more maneuverable opponent even after it had stung back.

Edited September 12, 2015 by RuBisCO

[dangerouspeople1]

Actually is there any high thrust but efficient engines in ksp? If I use aero spike engines for a fighter, will it run out of fuel too early?

[Alphasus]

So your saying there can't be redundant lift and control surfaces? Ok what is the spirit of the competition: beauty, realism. or to win within the rules provided?

MarsRocks's Bullet was the first box I encountered and it inspired me to make boxes, but try as a might mine were not as survivable as his. I had to figure out how to survive attacks from missiles and thus came to the epiphany that the real secret is speed to dodge missiles altogether. This comes from dropping the steer limiter and having enough engines to accelerate a flying brick to mach 2. Everyone was trying to make their planes really maneuverable and small but this makes them really slow, they lack engine power and they loss all their speed in the turning and bobbing, from a missile's perspective they are barely moving, easy to hit. A plane flying just straight but really fast is much harder to hit. The XIPHOS V is a normal plane design, but I decided to go back to the XIPHOS IV because with the steer limiter set so low and the stupidity of the AI, the plane is grossly unmaneuverable at the low speed needed for cannon combat. The AI can't adjust its steering limit dependent on speed or if it is attacking or dodging. Once out of missiles or close in the AI tries to do dog-fighting cannon combat regardless if the plane even has a cannon at all! So the XIPHOS V was woefully incapable of fighting close in dogfights, the XIPHOS IV on the other hand could take a lot of hits from cannons easy, and if both of the team survived the missiling stage of combat as they often did, then they could gang up on a more maneuverable opponent even after it had stung back.

If wing clipping within the rules is taking a giant .... on the rules of the contest, than making things that beat out wing clipping in every possible way is also. And so is the unrealism, hence why the Double down was banned on page 11.

Edited September 12, 2015 by Alphasus

[RuBisCO]

Oh come on have you seen some of the other design? Redundent Wings directly on top of other wings! The whole idea of wing clipping was to hid the wings and make the craft a tiny target to hit. My wings are not hidden, i'm a large target to hit. At least my wings have some space between them!

[Alphasus]

Here, fight the goldenhawk then. Your design was fair in area. The other boxes weren't. goldenhawk download here, this won last tournament and got banned for wing clipping.http://www./download/6ljokp7gmnuul99/FA-33_GoldenHawk.craft

[RuBisCO]

Well I would but i'm out of town until Wednesday so if you could post pics because my laptop is for work not play (no VC).

[Alphasus]

&index=32
A video of it in combat. Its final fight before it was retired/relegated to second line use.

[ferram4]

Easy enough, they have a larger surface area than boxes.

...I have no idea what you mean by this. If you mean projected area, then if we compare your planes to the box-planes that you fought against, the boxes have a larger projected area from the front/back, larger projected area from above/below, and about the same projected area from the side. All of which points to the box-planes being easier to hit than yours, which doesn't seem to be the argument you're making.

So basically, you're pushing for a ban on biplanes of all types then? You need to be a lot clearer and straightforward with the rule you want.

X planes have a larger side surface area and multiple vertical tail counterparts cant fly while missing the tails unless they are spammed to insanity.

Except your X-plane doesn't have greater side-surface area because the wing tips aren't appreciably higher than the fuselage, so no, that's not guaranteed, by your own example. And I gather from the vertical tail argument that the sides of the box plane aren't part of the problem, after all, it's necessary for them to be there.

With a box, it can lose 1/2 or more of its control surfaces and fly. The X plane can do that as well, but there is a noticeable loss in performance.

If a plane is well designed, it can lose half of its control surfaces and still fly, regardless of its shape. There was a fighter submitted to the WWII BDArmory challenge that looked very much like a P-38 that could lose half of its tail, half of the boom for that side, and continue to fly and still take down other fighters as if it didn't lose anything. It's not something inherent to boxes, it's something inherent to robust, damage-resistant design, which this particular box plane obviously is, while your X-wing obviously isn't.

[CreativeCombat]

...I have no idea what you mean by this. If you mean projected area, then if we compare your planes to the box-planes that you fought against, the boxes have a larger projected area from the front/back, larger projected area from above/below, and about the same projected area from the side. All of which points to the box-planes being easier to hit than yours, which doesn't seem to be the argument you're making.

Allow me to explain our concept of 'surface area'. If the plane has a larger wingspan, then it's a bigger target because the massive wings mean it has more 'surface area'. The Albatross, for example, has a HUGE surface area compared to the Aeris 3A. The smaller your surface area is, the harder your plane will be to hit because it's a smaller target. On the other hand, the smaller your surface area is on a traditional two-wing plane, the less lift you will have because your wings are smaller. Box planes achieve minimal surface area, while maximizing lift, by wrapping their wings above and below the fuselage and keeping them very short so that their surface area is almost the same as if they had no wings at all.

Allowing wing clipping allows a plane to have an arbitrarily low surface area by placing its wings inside the fuselage, and inside each other.

- - - Updated - - -

Actually is there any high thrust but efficient engines in ksp?

The Turbojet.

If I use aero spike engines for a fighter, will it run out of fuel too early?

Using a standard rocket engine gives you the advantage that your plane will take off much faster because it won't have to spool up, but it will run out of fuel very quickly.

Edited September 12, 2015 by CreativeCombat

[SasquatchM]

Personally I understand the concerns about the flying boxes, but there was no rule posted that said 'Must look classically plane-like". This being the case, personal preferences aside, the box type planes would seem to be perfectly valid entries as long as they do not violate any of the other rules.

[CreativeCombat]

Personally I understand the concerns about the flying boxes, but there was no rule posted that said 'Must look classically plane-like". This being the case, personal preferences aside, the box type planes would seem to be perfectly valid entries as long as they do not violate any of the other rules.

The question is what to do in future tournaments.

[Alphasus]

I would say must look classically plane like. Or, allow boxes and wing clipping. Both have absurdly low surface area and turn very well.

[SasquatchM]

Personally I like the classic plane idea. More challenge to it.

[Alphasus]

More challenge... Which do you prefer? I can remove that rule and say wing clipping is allowed, or I can keep that rule.

[GDJ]

Well, you guys figure this out. As for me, I'll build standard planes, no to little wing clipping, and they will look and feel like jet planes.....

Just like it should be!