Bushmanni

I think pure dogfighting with modern fighters in any sim is mostly done as having fun or training with friends. People who really want to just dogfight online usually fly WW2 stuff.

The basic problem with both fights was that the Mig wants to have a one circle fight after merge and you want to have a two circle fight and the Mig gets what he wants as all it takes is a little reversion of turn direction for the Mig to do it. In the first track the AI doesn't manage to track you after the merge as he loses speed and control of the aircraft in the vertical while you fly a wide circle around him. Generally if you go for a vertical fight at the merge you want to be pointing upwards to get use of the gravity assist in turning. You drop your speed but keep it still a bit higher than the Mig at all times. You try to turn with the Mig with in-plane turns while having higher speed which doesn't work out as the Mig has shorter turn radius due to slower speed and hence stays out of your gun sights. You would have had to either slow down to be slower than the Mig or use out of plane maneuvers (rolling scissors or displacement roll) to gain turning room. At 1:35 you make a high yo-yo that finally slows you down and gives you more turning room and tighter turn radius to get in to his six. When you overshoot you happily fly in front of him to be an easy target for revenge. If you are going to overshoot you should pull up into vertical (or to the side) early enough to not fly into his gun sights and make a lag roll to get back to his six. In the second fight it seems like you are purposefully trying to initiate a one circle fight but still going for in-plane sustained turn rate turn instead of small turn radius. The Mig gets you in his sights at 0:43 with very little effort merely 17 seconds after the merge. While it's a difficult shot for the Mig if you know how to defend against it, it would still be a dangerous gamble in your part against human opponent. After this you use your turn rate advantage and simply fly almost to his six. If you had slowed down at 1:12 you would have actually got there. This time when you overshoot you have so much more speed than the Mig that the the window for gunshot is practically non-existent. At 1:52 Mig gets chance for a snap shot. At 1:57 you reverse your turning direction and throw away the possible two circle fight you might have got and wanted. At some point after 2:08 the Mig might have got another single chance to shoot if flow by a human but this time you use your turn rate advantage like you should and then slow down for the final turn for the gun kill like you should have done at 1:12. AI is good for practicing visual tracking of the bandit and making maneuvers in relation to the other plane but tactically AI gives you a lot of slack in that he shoots you only in super easy situations and he doesn't turn very effectively and maintains a relatively steady speed over the course of the fight.

One basic thing is to keep track of the bandit visually. If you don't know where he is you can't maneuver for better position against him. This requires you to fly while looking mostly at the bandit instead of the HUD which takes practice. Against human opponent it's especially important as you need to know immediately when he's changing his plans. You should know how to get the maximum performance out of your aircraft which means digging through performance charts. Some basic points of interest are optimum speeds for sustained turns at different altitudes and optimal climb speed schedule. Regarding optimal fighting speed it depends. You should read some theory about one circle and two circle fights to get the basic idea. Basically in one circle fight maximizing your turn rate is irrelevant and minimizing turn radius is everything while in two circle it's just the opposite. You maximize turn rate by turning at the optimum sustained turn speed because if you bleed your speed right in the beginning you might gain angles fast initially but would be stuck with poor turn rate for the majority of the turn. Turn radius is minimized simply by minimizing flying speed ie. flying near stall speed. Most often you end up with one circle fight ie. a slow fight as the slower flying bandit usually manages to keep itself inside the faster ones turn radius forcing him to slow down also. When you add third dimension it gets more complicated as you can use corkscrew maneuvers like rolling scissor and displacement rolls to keep up your speed and still gain a shooting position on a slower and tighter turning bandit. One of the distinctive skills of scary skilled online aces I have seen in action is gunnery skills. They will shoot you down from pretty much any aspect and situation they get you in their sights (this with WW2 planes). This kind of skill simplifies tactics a lot as you are no longer concerned about getting into enemys six but just turning your guns towards him. With modern fighters gunnery is easy with rapid rate of fire, high muzzle velocity, real time lead calculating sights and relatively easy handling characteristics of the planes but still with practice you can make harder shots that you couldn't make otherwise.

In your latest video you were constantly losing sight of the bandits and you weren't very effective scanning the sky to find them again. If you don't know where the enemy is you can't effectively maneuver your aircraft to better your position but end up running around the sky while the enemy is maneuvering into your your six. You should practice flying while looking at the bandit and only taking intermittent quick glances to HUD to check your speed and nose position. You will also sometimes inevitably lose sight of the bandit when he gets masked by our aircraft during maneuvers so you should learn to predict where he's going to reappear again.

Apache would be my choice because of Hellfires and two man crew. Russian ATGMs are just not adequate in a two way shooting range because of long exposure time during guidance.

You said in the video that air combat is a perishable skill but from performance psychology standpoint it just means that you don't know what you are doing. You did something that worked last time but because you really didn't understand why it worked or what you exactly did you can't replicate it after some time. Put more effort into analyzing your fights and reading some theory and you will improve faster and retain it over time. Besides target fixation you entered one circle fight way overspeed and got quickly wasted as expected. Su-27 and Su-30 like to keep their speed up in a dogfight which makes them easy targets in a one circle fight if you simply pull hard and bleed your speed like every rookie does. Some other AI fighters like F-15 and Mig-29 like to turn with low speed so if you want to do a one circle fight with them you need to start your turn with equal or less speed than the bandit. A better option though would be to go for two circle fight and do a max sustained turn rate turn. F-15 manual explains how to make sustained max performance turns in different conditions. Typically you want to keep your speed at 0.9M and pull just enough that the speed stays the same. You should be able to pull about 7g (for as long as you have fuel) at the altitude you had in the video. Check the circle flow subtopic to understand what I'm talking about one and two circle fights. It's not a bad idea to read the rest of it too. Understanding some of it might need some experimenting and analyzing of those experiments. http://en.wikipedia.org/wiki/Basic_fighter_maneuvers

It seems like there's no actual penetration modeled in the game. Modern APFSDS round would go through several BMPs but in DCS it will be stopped by the first one. It also seems like this holds for aircrafts damage sections. If you happen to hit the nose with an AP round you make only superficial damage as anything behind the nose is saved even though in real life the round would have went through the aircraft and exited through the rear.

Manual explains pretty well the basics, you can/should start with that.

I was able to replicate it and I didn't need to pull more than about 20 degrees pitch on average to make it work. I tried it with full fuel, missile and tank load. I didn't check if it's accurate to the feet and second but it seemed more than close enough to the real numbers. You cant climb vertically at high speed due to drag even if your T/W is little above 1.

I don't know if you know what you are saying but assuming your knowledge is gained from videos (not so in depth) or books (better) you are going to have a nasty surprise. In my opinion you can't have in depth understanding of the theory and principles of dogfighting unless you have at least average skills to execute those theories and principles in practice. First of all dogfighting is completely different mental exercise from anything else that normal person won't know what's really important and what's not unless having experienced it (dogfighting in Battlefield or ARMA doesn't count). The other thing is that dogfighting is a dynamic process that can't be predicted analytically. While it's governed by science as in what can be done it's up to the opposition to choose what he will do within the boundaries set by rules of nature. Unless you have experienced the process you can't understand it. Don't expect too much from yourself at the beginning and ask questions from more experienced players. I don't know how long it takes until you can't learn anything new about dogfighting, it probably never happens. It takes few years to learn just the basics when doing it professionally so you shouldn't expect more of yourself. Most other players on the servers have years of head start on you and the best have been playing since the first somewhat realistic sims came out so you can imagine how it is for newbies. You don't have a chance in multiplayer before you can face off and kill multiple AIs without breaking a sweat. You are still probably just a target after this but you have enough skill that when you get the opportunity to kill someone you might actually be able to do it. If you can manage to find someone to practice with it will help immensely with basic 1vs1 skills as you can get back to the fight right away after getting shot down and not waste time flying around looking for a fight. Tactical judgement is crucial skill on public servers but that you can learn only through the hard way by playing on them.

I have been doing quite a lot of fighting with Hinds lately and they seem be acting pretty smart but don't have a spotting ability of a psychic so you actually have a change of seeing them first, especially if they are already busy engaging their targets. For once it seems chopper AI is working really nice. I hope that after EDGE ED would update the simulation engine so that AI wouldn't break after every update in some way.

The thing with CFD is that it also needs to be validated with real data. You can't just put together a 3D model and expect realistic results. If you could build a model with nearly infinite amount of detail and really dense mesh (the volume where the flow flows) you might be able to do that but it would be computationally too ineffective. The 3d model and mesh will be built with detail only where it really is needed and for that you need experience or data how the flow behaves in real world with the shapes, scale and parameters you are dealing with. You might also not calculate all the possible physical effects like temperature changes to simplify the model for better computability. I don't have enough experience myself about CFD to know how much error you would get with certain kind of shortcuts but this is how the pros explain it.

I think the control assignment UI is pretty good, at least I don't know how to make it better. But it's really stupid and frustrating to re-do all the controls of a plane from scratch when a change is made to the available commands. There should be some kind of automatic system that will transfer all the previous bindings to the new configuration.

You need to use string concatenation ie. string1 = '[blue]' string2 = '[plane]' string3 = string1 .. string2 Now string3 is '[blue][plane]'

Did you get any reply when you used the "Get to the chopper" radio command in F10 menu? Anyway the mission shouldn't work in the current version as there has been some changes made to the UH-1 module that change the behavior of some critical SSE commands and essentially break the troop carry scripts. I already have an updated working version of the scripts though but need some time to update the mission.

Well it seems to me they don't. I don't have experience with most recent versions but for example fuselage lift wasn't calculated very accurately. It seems like a plane with a rectangular fuselage cross-section behaves like one with a round one for example. At least body cross section shape didn't seem to make much difference in behavior in my experiments. But as is stated in their own website, they don't use CFD and there's no precalculation using CFD in the Plane Maker either. X-Plane gives probably very good results for planes with somewhat circular cross sections ( ie. cylindrical body) and traditional wings sticking from it regardless how the wings itself are positioned or shaped. For the rest you need to improvise the geometry to some degree to get good results like in case with planes with blended wings. Blade Element Theory is very effective with airfoils but you can't use it for fuselage or how the flow behaves after the airfoil. I'm speculating here but X-Plane probably has something like SFM for the fuselage that's tuned with the general dimensions of the fuselage to give plausible figures regarding its drag, lift and torque. If you throw in forces from airfoils modified with some wake effects simulation you will get a pretty close match as long as your fuselage doesn't have too funky shape.

X-Plane doesn't use simplified CFD but Blade Element Theory just like DCS. http://www.x-plane.com/desktop/how-x-plane-works/ DCS uses apparently some more advanced methods for determining propwash, downwash and wingtip effects (and most likely some other effects). X-Plane uses some generic look-up tables for all aircrafts while DCS can have customized tables for each airframe or even some kind of simulation for these effects. Ka-50 is a good example that shows simulation of rotor interaction with each other and itself that is just unbelievable when you start to examine it more thoroughly.

It still seemed to be the best plane available for the AFAC job at that time. Kosovo is more of an example of how hard it's to find and destroy ground forces hiding in a forest than how bad the A-10 is as a CAS plane. A-10 didn't do CAS that it was designed for as there wasn't any ground forces that needed air support.

I'd set the curves only as a last resort if nothing else helps. CH has enough accuracy to use linear curve if there's no mechanical problems like loose play in attachment of potentiometers. Curves makes flying by the muscle memory very hard so you have to make all maneuvers very carefully ie. slow to avoid problems. I tried curves once but didn't feel it was worth the cost.

With the airfoil nature I meant the typical purpose of directing flow. This is kind of hair splitting but it's not the flow direction or the shape of one airfoil that creates the diffusion effect but that they form a diverging duct as they are placed close together that is the essential feature for diffusion. I just wanted to point out that it's not the flow redirection but the diverging duct feature why stators slow the the flow and increase pressure. It's true that the stators thickness variation is like in typical airfoil (or at least something like it).

I have been pondering the mystery of how jet engines work in other thread and while mechanically simple they are nothing like that. I have huge respect for people who figured this tech out for the first time. Then I stumbled upon this. How to build a TURBOJET engine How to build a TURBOJET engine -The Maths JET ENGINED BBQ.....The TURBAQUE Granted it doesn't do much else than sustain itself but still impressive.

Ok, found the secret of the stators from the Jeppesen book. They actually form a diverging duct as they are much thicker at the front. The airfoil nature has nothing to with the diffusion but only changes the flow direction to be optimal for the next compressor stage.

How are the stator wanes arranged in relation to the compressor blades ie. in what kind of angle they are put in relation to the flow and the compressor blades? The airduct in compressor is converging which means the duct would speed up and drop the pressure of the flow. It seems a bit odd that airfoils increase the pressure of the flow and slow it down (or prevent acceleration in this case) so there's definitely something interesting going on.

After reading some theory from the Jeppesen's book I think I figured out what I was after. First of all compressor forces the exhaust out from the back of the engine instead of both ends. Pressure inside the engine is highest just after the last compressor stage before combustion chamber. The ducting and other flow constrictors are shaped so that the pressure (total pressure = static + ram) after the combustor declines as the flow moves towards exhaust. As the pressure difference always creates an acceleration towards low pressure the flow keeps going from forward to aft. If the compressor would fail to provide enough pressure or there was enough extra constriction added after combustor the exhaust would start going out from both ends and the engine would stop functioning. The engine is pushed forward by air molecules mostly from combustion chamber, secondly from compressor blades and a little bit by turbine exhaust cone. The thrust created by compressor is similar to any fan or propeller, ie. airfoils moving fast through air. The combustion chamber is basically a pressure vessel that has larger outlet than inlet. Compressor forces air through the inlet creating pressure inside the combustion chamber and the expansion of the gas inside the chamber makes up for the greater flow through the outlet so that the pressure stays up. Turbine exhaust cone is bacwards facing cone which is pushed forward by the pressure inside engine exhaust (or afterburner) section. Turbine blades push the engine backwards due to drag of the turbine blades and exhaust nozzle due to pressure pushing the inside surface of the nozzle backwards. When the afterburner is in operation the nozzle can be opened so that the pressure pushing the nozzle from inside is directed more to the side and less backwards than during mil power. You need to speed up the gas to produce the same backpressure in the open nozzle than when the nozzle is closed. The pressure inside afterburner stays the same but the force pushing the nozzle is directed more to the side so the force component facing rearward is smaller. I was puzzled by the thrust equation where thrust created by pressure and that of flow acceleration is combined in the same equation. It contains the other half of my confusion that I was having with jet engines. Basic physics dictate that there are four basic forces ie. electric force, gravity and two nuclear forces that are relevant mostly to nuclear physics. Electrical force is responsible for atoms sticking together, colliding and bouncing off from each other. There's no other mechanism for fluid and solid to affect each others movement but by molecule collisions at the surface of the solid. The expanding and accelerating gas and the jet engine don't just agree to go separate ways and magically move in different directions. They have to interact through collisions which is to say through pressure. How this exchange of momentum is done in practice is what I am interested to know as by looking at the jet engine it's quite hard to see. Also having force from pressure and from change of momentum in the same equation for thrust is puzzling at first. Simple answer is that pressure gradient accelerates flow and applies a force to the walls that contain the flow. If you integrate the force applied to the wall and the force applied to the flow by the pressure gradient (for which you can then calculate acceleration or change in speed) you get the same force (but they are to the opposite directions) for both as should. If you know how much the flow accelerated inside the engine you know how much force the pressure gradient applied to the gas and hence how much it applied to the engine. You don't need to know where and how hard the gas molecules are pounding the engine to calculate the thrust. All you need is the pressure and area at the exhaust nozzle of the engine and the force that is applied inside the engine to the flow through velocity change of the flow inside the engine. The part for change of momentum has the "hard to calculate with"-pressure gradient hidden inside of it. Talking and writing about this stuff with others is educational and reduces the chances of making some stupid mistakes or assumptions. And it's of course more fun too.

Thanks. I'll probably build one of these for myself one day if nothing better comes along.