Air Superiority Fighter Competition Continued

[ZLM-Master]

Yopyop !

This is my 1st time with kerbal Konstruct & i try this .

KSC 1 to KCS 2 , start directly over the sea with 29.1 KM far away .

Both runaway face on . Just finish the building & not test dogfight yet .

Minimal Building for performance .

EDIT : Not really active this day but we have beautiful weather  so i most outdoor than in my house  

EDIT 2 : Great formation in both team with all crafts spawning with vessel mover .

Edited April 19, 2018 by ZLM-Master

[Box of Stardust]

Progress on combat locations continue. Got 3 locations set up today, about maybe 12 to go. Progress will go faster now that I'm getting used to the process. Not sure if there's a correlation between latitude on Kerbin with 'stable' distance. The farther away I got from the equator, the shorter 'stable' distances it felt like. I was able to get a 26km separation between spawns at a location near the equator, but couldn't get past 20km safely once I got farther from the equator.

There's also now a new 'standard' spawn arrangement, placing #1 on the left center line, #2 on the right edge line, and #3 on the left edge line. Seems to work well.

Anyways, here's the next battle, showcasing another one of the new locations. Something different for visual variety! It also seems like there wasn't much problem with 'first takeoff glitch' this time, with both teams settling into identically tight formation. Speed is a factor though, and formations tend to be tighter with slower aircraft. Nonetheless, this method is proving to work well to keep formation consistency.

Gunbrick vs @dundun93's TFD:

 

 

Yeah, results are more of the same. Not much to comment on.

The next one though is Gunbrick vs Vampire Squirrel, which is where things get interesting. Honestly not sure how this one will work out, given the new 'certain, consistent' spawns.

[Box of Stardust]

One more battle for tonight, showcasing an 'old but new' location.

We're back fighting over KSC, but this time with spawns facing directly at each other. This location was redone and kept as a combat zone because, quite frankly, sometimes it's nice to be lazy and be able to quickly recover aircraft after a quicksave load right from the space center screen, instead of having to go to the tracking station to do it.

Also, I'm planning on adding an optional 'easter egg' to it later.

So here we go, the original battle of debate which prompted creating easier spawn points in the first place: Gunbrick vs @goduranus's Vampire Squirrel:

 

 

I decided to try something different for the third battle, under the long-ago-discussed effects of a stationary, landed observer, as opposed to following around aircraft.

 

 

Spoiler

Analysis:

Much more interesting.

It's not as decisive as I thought it would be, nor was it as decisive as the tests I used to run using the original KSC-Island spawns. Those used to be far more in favor in the VS and its MLRS, or what I'll call from this point onward, a pre-routing device (PRD). What that means... I'm not quite sure, since these still aren't perfectly identical spawns, nor is the terrain they fight over clear of obstructions, which may have an effect.

So I redid the battles at Eaglin Straits, a 100% equal battle environment, with equal height spawns facing each other, and the combat is over water (this was the first location shown, in Gunbrick vs Viper), and the spectating was done from the central observer post to remove game-physics-bias towards either side. Quite frankly, this is probably the only location we'll really need to test aircraft, if we want to be completely, uncompromisingly fair. I could even put up the save file and KK instance files right now if anyone wants the preliminary version with Eaglin Straits in it, since it's really the only location we'll ever need if no one minds the boring landscape.

...So yeah, so apparently VS actually loses the majority of the time under completely equal conditions, at least the way I did it. Maybe someone else will have different results. But that's not what was recorded, so... I guess they don't count?

But what happens is that, while VS's PRD does eliminate the merge, if the enemy reacts quickly enough or if the VS is not able to capitalize on the routed opposition (and sometimes even then), it turns into a standard dogfight. In most tests at E.Straits, the Gunbrick team all turned to evade in the same direction. Even if a VS managed to down one Gunbrick, if they couldn't down a second, the ensuing dogfight turns in favor of Gunbrick... weirdly enough.

And with Gunbrick, it's simply the ultimate theory that excessive gun firepower is a valid solution.

So pretty much, in the recorded battles... VS got lucky. Which is another reason why I feel like we need to move to a format that is more than just best of 3. Doesn't feel... right.

As a final note, I feel like previously highly aggressive aircraft are now notably less aggressive when I put them into fights, and I don't know why... but I will also note that I feel like in the old KSC-Island format, the KSC spawn was generally more aggressive. Hm. Food for thought.

 

So, unless anyone wants to re-test themselves in an ultra-fair environment, the recorded results of Gunbrick v. VS will be that Gunbrick stays at #3.

 

 

[Box of Stardust]

Daily update so far:

I've decided that since the KSC combat zone is arguably the most important due to ease of access from the space center screen (able to quickly access spawn beacons and recover aircraft), I re-did the spawn locations to make it an unbiased location. For some strange reason, it also only kind of occurred to me that, with the ability to place runways literally anywhere, I could even simulate in-air interceptions by placing runways in the sky.

What this does it extend the altitude floor when aircraft battle. And what that means is that there's now arguably a much more equal footing between aircraft, as the minimum altitude setting comes into play less, as aircraft start at a higher altitude by default.

This realization came about when I re-examined my private testing between Gunbrick and Vampire Squirrel at Eaglin Straits, realizing that the runways, while for the most part 'on the ground', start at 300m height. And note that the minimum altitude setting uses terrain height, not sea level height. So, what this means is that, when aircraft take off, first they reach their minimum altitude over their current terrain, then in-combat over Eaglin Straits, there's actually a 300m additional buffer to the usual combat altitudes from the original KSC-Island format, because water is at sea level; 0m. So instead of having only, say, 200-300m to maneuver and evade missiles, there's more like 500-600m.

One of the problems with the original format was that by always launching from the ground, there was (and still kind of is) a tendency for the lower-altitude aircraft to have an advantage, simply because they'd pull the opposing team down to their altitude. This effect is quite pronounced with the KSC-Island spawns, as aircraft will turn-in towards the opposing team once they have taken off and reached minimum altitude. And when the minimum altitude is so low, the AI always has to juggle maintaining minimum altitude and evasive maneuvering, which also partially gives an advantage to the aircraft with a lower minimum altitude. Which is kind of dumb, but the planes are also risking flying into the ground/ocean, which is... kind of even dumber. This is an air superiority competition; avoiding the ground shouldn't be a concern if we only want to focus on the 'air combat' portion.

Interestingly, this was less-pronounced if the starting method used is to start both teams off in a common location and make them fly away from each other to the competition start distance, then turn back, because in this method, aircraft will keep climbing until they have either reached the default altitude setting or the competition start distance is reached. So this generally meant that combat altitudes were at least around 1000m, unlike the ~500m we usually see with KSC-Island.

Progress on setting up the new locations is going smoothly, and I now place runways at equal heights if it makes sense to, and in the air to add that buffer zone. There's still a few that have the combat arena close to the terrain, but now it's mostly just for 'flavor'. The KSC 'arena' after I reworked it can easily be the only location used for aircraft testing due to minimized bias and ease of operations. Its runways are placed at 750m height, which translates to, after takeoff, roughly 900-1000m altitude. That's enough of a buffer to ensure that the minimum altitude setting infrequently comes in to play.

Lastly, there is, again, a new spawn arrangement. Now freed from the default spawn location of aircraft on the KSC runway, since all aircraft are now spawned with Vessel Mover, I figured we should do something that looks more like an actual aicraft formation. Lead aircraft left center line, #2 on right edge line, and #3 on left edge line. This, I think, optimally arranges aircraft so that they minimally interfere with each other in-flight to target, since there's no trailing aircraft that's going to have both aircraft in its way crossing in front of it when the BD AI does its little 'position flip' thing after takeoff.

 

No new battles for tonight. I might re-test Gunbrick vs Vampire Squirrel once more with the finalized KSC arena, in a best-of-5 for better statistical significance.

[ZLM-Master]

@Box of Stardust Just an idea ; If you create 3 runaways in a perfect formation onto the floor ( or in air ) with grouped different name ( Runaway 1 , Runaway 2 ... / V Formation Spot 1 ... / Invert-V Formation Spot 1 ...) Craft spawn perfectly on each runaway . But the issue is that i must back to SPH every time , but the result is perfect !!!

Need a way for spawning 3 crafts in formation directly on the runaway ..... Should be great ^^

( Maybe having better interact between Vessel Mover & KK )

EDIT , Working on a way for spawning different formation on a single runaway , with tiny trigger on the floor .

Nice day on Kerbin .

Edited April 21, 2018 by ZLM-Master

[ZLM-Master]

Ok , i do this , a perfect spawning V-Formation made with Trigger just under the Runaway .

Just have to duplicate this for other formation

( Someone know how to have a screenshot on the Runaway in the SPH ? )

[Box of Stardust]

2 hours ago, ZLM-Master said:

Ok , i do this , a perfect spawning V-Formation made with Trigger just under the Runaway .

Just have to duplicate this for other formation

( Someone know how to have a screenshot on the Runaway in the SPH ? )

I admire the consistency, but having to go back to the SPH every time is too much.  I don't think I can handle spending 1-2 minutes to spawn each plane.

The idea I had was to make a tool in the SPH, with I-beams, that has the right placement spots for planes on the runways. Or you could place KK markers/lights on the runways without colliders turned on, for where you put planes. Not sure if you can save structures as combined structures though.

[ZLM-Master]

Maybe a visual helper for the Spawn Spot !

Simple enough and accurate enough too .

EDIT : Start with V formation or Invert V formation with the same tool , just rotate it with vessel mover !

Quite Fun

EDIT 2 : Multi Spawn Tool , Rotate to the desired formation .

Tool Formation Helper on KerbalX .

Edited April 21, 2018 by ZLM-Master

[Box of Stardust]

2 hours ago, ZLM-Master said:

Maybe a visual helper for the Spawn Spot !

Simple enough and accurate enough too .

EDIT : Start with V formation or Invert V formation with the same tool , just rotate it with vessel mover !

Quite Fun

EDIT 2 : Multi Spawn Tool , Rotate to the desired formation .

Tool Formation Helper on KerbalX .

That's pretty much what I was thinking, except that those spawn points are too close together. They'll cause the AI to do 'avoid collision' things.

[ZLM-Master]

Easier is better . 13 Parts Tweakscaled .

Edited April 21, 2018 by ZLM-Master

[ZLM-Master]

Yopyop , I finally use a Grid Tool , so i can use most formation in the way i want . I should think about that earlier ^^ Just Move it after placing craft .

Yopyop .

[Box of Stardust]

1 hour ago, ZLM-Master said:

Yopyop , I finally use a Grid Tool , so i can use most formation in the way i want . I should think about that earlier ^^ Just Move it after placing craft .

Yopyop .

Haha okay that's pretty creative.

[SuicidalInsanity]

On 4/20/2018 at 10:03 PM, Box of Stardust said:

Progress on setting up the new locations is going smoothly, and I now place runways at equal heights if it makes sense to, and in the air to add that buffer zone. There's still a few that have the combat arena close to the terrain, but now it's mostly just for 'flavor'. The KSC 'arena' after I reworked it can easily be the only location used for aircraft testing due to minimized bias and ease of operations. Its runways are placed at 750m height, which translates to, after takeoff, roughly 900-1000m altitude. That's enough of a buffer to ensure that the minimum altitude setting infrequently comes in to play.

Heh, reminds me of when I started experimenting around AI limitations back in the BAD-T III. Admittedly, my solution was to remove the ground...

As for spawn point markers, you totally can clip KK statics into each other, so you could mark the center of the grid squares so you only need to spawn the tool once, rather than once per fight; would also save having to go back to the Space Center to remove it. Just remember to leave colliders on the marker statics off.

[Box of Stardust]

8 minutes ago, SuicidalInsanity said:

Heh, reminds me of when I started experimenting around AI limitations back in the BAD-T III. Admittedly, my solution was to remove the ground...

As for spawn point markers, you totally can clip KK statics into each other, so you could mark the center of the grid squares so you only need to spawn the tool once, rather than once per fight; would also save having to go back to the Space Center to remove it. Just remember to leave colliders on the marker statics off.

I was considering it only if I could save KK creations as 'assemblies' of multiple statics, which I haven't found a way (though admittedly I have not searched too thoroughly).

It already took me a few hours to place down... 15? pairs of runways (though admittedly a lot of the difficulty was measuring stable spawn distances, which I've discovered to be correlated with altitude/LoS to each vessel; i.e., world curvature not blocking it; weird). To place down markers for each runway... no thanks.

I just eyeball the spawn points now; I figure a few meters off of difference won't be significant variance that the AI pilots won't induce in the air anyways.

Edited April 23, 2018 by Box of Stardust

[SuicidalInsanity]

I feel your pain - imagine building whole airbases out of statics instead of runways (the BAD-T arenas took awhile) - but you can save assemblies, at least for packaging and export; though if you mean saving a KK creation so it can be pasted somewhere else, not so sure that can be done - I generally resorted to custom statics at that point.

Hmm... are you using the same runway static for all arenas? If so, I could come up with a EZ place (copy/paste parent runway runway coords and it will snap into correct position) runway marker for them.

Edited April 23, 2018 by SuicidalInsanity

[Box of Stardust]

43 minutes ago, SuicidalInsanity said:

I feel your pain - imagine building whole airbases out of statics instead of runways (the BAD-T arenas took awhile) - but you can save assemblies, at least for packaging and export; though if you mean saving a KK creation so it can be pasted somewhere else, not so sure that can be done - I generally resorted to custom statics at that point.

Hmm... are you using the same runway static for all arenas? If so, I could come up with a EZ place (copy/paste parent runway runway coords and it will snap into correct position) runway marker for them.

Yeah, they're all Runway Strips.

For my intents and purposes, the runways are simply tools to get aircraft to the arena, which is just some nice unique landscapes. Some more for exhibition combat, like a river ravine.

[SuicidalInsanity]

Something like this work?

KKRunway Strip Markings
Unzip into Gamedata/KerbalKonstructs/Statics. Copy/paste Runway coordinates to the Runway Markings static to snap to alignment with the runway strip.

[Raptor kerman]

Hey are we allowed to enter as a constructor (airframe providor) where people download our craft clean before adding their own weapons. Just like how in F1 most teams buy the engine and gearbox components from constructors. 

[Box of Stardust]

1 hour ago, Raptor kerman said:

Hey are we allowed to enter as a constructor (airframe providor) where people download our craft clean before adding their own weapons. Just like how in F1 most teams buy the engine and gearbox components from constructors. 

Er, well, under my authority, I guess I can work out a way for this to work, but I don't really see the point.

Weapons are part of the equation, but how an aircraft flies is a big part of this competition, as well as AI tuning. With that  much work already being put in by competitors here in the aircraft itself, I'm not sure anyone would really be interested in just taking an airframe and adding weapons, if they're just going to have to modify it to make it fly their way as well.

 

Basically, I don't see the current competition environment having a place for that kind of thing to be able to exist. While there's no rule against it, there's nothing encouraging it either.

 

However, since this thread is ending and will soon enter a re-continuation, I might be able to put in some rules in place for it to work out. Maybe. 

[Raptor kerman]

25 minutes ago, Box of Stardust said:

Er, well, under my authority, I guess I can work out a way for this to work, but I don't really see the point.

Weapons are part of the equation, but how an aircraft flies is a big part of this competition, as well as AI tuning. With that  much work already being put in by competitors here in the aircraft itself, I'm not sure anyone would really be interested in just taking an airframe and adding weapons, if they're just going to have to modify it to make it fly their way as well.

 

Basically, I don't see the current competition environment having a place for that kind of thing to be able to exist. While there's no rule against it, there's nothing encouraging it either.

 

However, since this thread is ending and will soon enter a re-continuation, I might be able to put in some rules in place for it to work out. Maybe. 

Awesome, but I'm thinking of something akin to F1 rules, where only the basic airframe, aerodynamic surfaces and powerplant is provided, leaving the entire weapons and control system to be developed and integrated separately while the aerodynamic properties are already figured out, it would be more complicated than 'just adding missiles'. I think it would be a good idea for players who for whatever reason can't run mods on their machine but want to participate. It could be a 'quick and dirty' way for new players to get acquainted to the challenge and while this challenge is EXTREMELY popular i think this would give a welcome boost. If this does work, why not have a spoiler with the list of 'approved constructors' and their kerbalX link. To become approved the constructor would have to submit at least 3 independent designs that exceed expectations when integrated with the weapons system and is met with some success. Independent where the designs are not based of each other or a common design.

Edited April 23, 2018 by Raptor kerman

[Box of Stardust]

21 minutes ago, Raptor kerman said:

Awesome, but I'm thinking of something akin to F1 rules, where only the basic airframe, aerodynamic surfaces and powerplant is provided, leaving the entire weapons and control system to be developed and integrated separately while the aerodynamic properties are already figured out, it would be more complicated than 'just adding missiles'. I think it would be a good idea for players who for whatever reason can't run mods on their machine but want to participate. It could be a 'quick and dirty' way for new players to get acquainted to the challenge and while this challenge is EXTREMELY popular i think this would give a welcome boost.

That's what I figure you were getting at but I'm not sure what you mean by 'can't run mods' hindering participation. Wouldn't players already be at the 'running mods' part if they're entering something in the competition?

The only barrier that 'not being able to run mods' I can see is that they can't test their own aircraft, but honestly, adding weapons and working out the AI for them really isn't that hard. 

As well, if a player is unable to run mods effectively to test their aircraft, then they kind of lose out on arguably the hardest part of the 'adding weapons' part of the process, which is seeing what works and what doesn't in terms of weapons management and flight AI. While this part is actually difficult, again, I don't see how providing an airframe makes this any easier for new players.

And lastly, our mod list, while open-ended, isn't very extensive; it's mostly just AirplanePlus and Tweakscale, and upcoming in BDA 1.1, DCK.

You do know that competitors aren't obligated to run the matches, right? That's all left to those who volunteer as judges, so there's never any worry about straining machines trying to get potentially stressing 3v3 matches recorded. 

 

If you want to know my solution for your proposal, it would be to have a separate leaderboard where veterans cannot place in, to give new designers a chance to learn.

Edited April 23, 2018 by Box of Stardust

[Raptor kerman]

6 minutes ago, Box of Stardust said:

That's what I figure you were getting at but I'm not sure what you mean by 'can't run mods' hindering participation. Wouldn't players already be at the 'running mods' part if they're entering something in the competition?

The only barrier that 'not being able to run mods' I can see is that they can't test their own aircraft, but honestly, adding weapons and working out the AI for them really isn't that hard. 

As well, if a player is unable to run mods effectively to test their aircraft, then they kind of lose out on arguably the hardest part of the 'adding weapons' part of the process, which is seeing what works and what doesn't in terms of weapons management and flight AI. While this part is actually difficult, again, I don't see how providing an airframe makes this any easier for new players.

And lastly, our mod list, while open-ended, isn't very extensive; it's mostly just AirplanePlus and Tweakscale, and upcoming in BDA 1.1, DCK.

You do know that competitors aren't obligated to run the matches, right? That's all left to those who volunteer as judges, so there's never any worry about straining machines trying to get potentially stressing 3v3 matches recorded. 

 

If you want to know my solution for your proposal, it would be to have a separate leaderboard where veterans cannot place in, to give new designers a chance to learn.

I'm speaking from the designer's perspective as the ones being unable to run mods (like me, i can just about run KSP but even installing KER or HyperEdit makes the computer grind to less than 5fps). And yes, i do know the rules  . Thanks for the answer, its appreciated!

[Box of Stardust]

10 minutes ago, Raptor kerman said:

I'm speaking from the designer's perspective as the ones being unable to run mods (like me, i can just about run KSP but even installing KER or HyperEdit makes the computer grind to less than 5fps). And yes, i do know the rules  . Thanks for the answer, its appreciated!

Ah, I suppose that makes sense from that perspective then. If it's the airframe designers that can't run mods.

Unfortunately, it then comes down to people that are willing to take designs from airframe constructors and do something with them. Not that it doesn't make it impossible, it's just that I personally don't see it being viable.

The closest thing I know of happening similar to this is... me, where I toyed around with someone else's plane (twice) because it interested me to rework it a fair amount and mess around with it. That's the kind of person you'd need to sort of partner up with- people willing to toy around with other's designs and make it work, but these kinds of people I would think would be the kind to be motivated to make their own craft from scratch.

But again, there will be a new thread up, which may present new opportunities and new interests, so who knows.

[DoctorDavinci]

Just a PSA:

BD Armory Continued v1.2 haveen released for KSP v1.4.x

 

[Box of Stardust]

30 minutes ago, DoctorDavinci said:

Just a PSA:

BD Armory Continued v1.2 haveen released for KSP v1.4.x

 

Aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa