Jump to content

Wow AI ace level 109 is Godlike


Pandacat

Recommended Posts

Surprised?

Dogs of War Squadron

Call sign "HeadHunter" P-51D /Spitfire Jockey

Gigabyte EP45T-UD3LR /Q9650 3.6Ghz | 16GB DDR3 1600 RipJaws | EVGA GTX-1060 ACX3 FTW | ThrustMaster 16000m & G13 GamePad w/analog rudder stick | TurtleBeach EarForce PX22 | Track IR5 | Vizio 40" 4K TV monitor (stuck temporarily with an Acer 22" :( )

Link to comment
Share on other sites

I could hardly get the AI into my front quadrants. The only time I was successful was after I did a few rolling scissors and got it to overshoot me. Then we did a headon pass when he turned back around. I got a few shots on him but after that, I got blasted out of sky.

Link to comment
Share on other sites

The AI also uses an SFM, which gives them an advantage from the word go.

ASUS ROG Maximus VIII Hero, i7-6700K, Noctua NH-D14 Cooler, Crucial 32GB DDR4 2133, Samsung 950 Pro NVMe 256GB, Samsung EVO 250GB & 500GB SSD, 2TB Caviar Black, Zotac GTX 1080 AMP! Extreme 8GB, Corsair HX1000i, Phillips BDM4065UC 40" 4k monitor, VX2258 TouchScreen, TIR 5 w/ProClip, TM Warthog, VKB Gladiator Pro, Saitek X56, et. al., MFG Crosswind Pedals #1199, VolairSim Pit, Rift CV1 :thumbup:

Link to comment
Share on other sites

Ya, but what is amazing is AI Ace Fw190 is good but not unbeatable. 109, though, is a totally different story. I mean both of them have SFM. How come the performance of the 2 are so different. This gives me a feeling that 109k4 significantly outperforms D9, but in real life that's not true.

Link to comment
Share on other sites

Has anyone test using D9 against AI Ace lvl k4? I mean if you argue k4 outperforms p51d because its higher power/weight ratio, wings more efficient for lower speed, higher rate of climbe, I got all that. But D9's power/weight ratio is pretty high too. It also has faster roll rate, powerful engine. At least in vertical, D9 shouldn't lose that badly.

Link to comment
Share on other sites

Has anyone test using D9 against AI Ace lvl k4? I mean if you argue k4 outperforms p51d because its higher power/weight ratio, wings more efficient for lower speed, higher rate of climbe, I got all that. But D9's power/weight ratio is pretty high too. It also has faster roll rate, powerful engine. At least in vertical, D9 shouldn't lose that badly.

 

I'd be willing to bet that in a neutral start position, the 109 would destroy the 190.

[sIGPIC][/sIGPIC]

29 October 2009

CGNR-1705 / USMC Cobra

* * * * * * * * *

Semper Memoria, Semper Vigilens

Link to comment
Share on other sites

So in a low-speed turning contest at low to moderate altitude between a P-51D and a 109K4, the 109 has a lot of advantages. It turns betters, it climbs better, accelerates better and is slightly faster.

 

The P-51 of course has its own set of advantages. It dives and zooms better, it handles much better at high speeds (rate of turn and roll), and it should start to out perform the 109 at high altitudes (25,000 ft plus). If you meet a K4 at medium alt and co-energy probably the best bet is to sucker him into a 400mph+ game of scissors, the P-51 should have more advantages in that situation. If you don't like the way the fight is going, and still have altitude, then dive away; the 109 would be foolish to follow as the P-51 dives and zooms much better, and is much happier at high speed than the 109.

 

Not that this is unusual for the P-51, it wasn't known for being a great turner or climber. It was known for its amazing dive and zoom, its high speed, good high-speed handling, its range and its high-altitude performance. Until now the only play partner for the P-51 has been the FW-190D9, and that aircraft is atypical because it's even worse at turn fighting than the P-51, so you could get into a turn fight with one and win.


Edited by Tomsk
Link to comment
Share on other sites

There's also something to be said for the ease of handling in the mustang.

It responds sooooooooo well and smoothly to input that's it's just a dream to fly.

I prefer the 109, always have always will, but it is such a constant pain in my 4th point of contact.

The mustang feels like it responds to your very thoughts, the 109 fights you.

It seems like the two German fighters are at polar ends of the spectrum. One flys and fights exceptionally well at low alt and speed, where the other only excelled above 400kph. The mustang seems it lives between the two extremes.

Link to comment
Share on other sites

WW II birds ranking

 

There must be hundreds of posts about the unrealistic AI performance, and I think we have to accept the fact that the AI pilots on one hand have only limited smarts; while at the other hand they have maximum SA, their fire is very accurate and they never stall, always going to the limit of their airplane. It's also true that the AI fighters can take a huge number of hits before they are destroyed. I would like that to be fixed so a P-51D for example cannot take 100 or more hits before if crashes. The best antidote to that is to try and hit the pilot, or to cut a piece of a wing off.

The pecking order between the AI birds (FW-190, P-51D and Bf-109) seems realistic though. I do a lot of dogfighting between these 3 and I find that the Bf-109 gives me the best chance and the FW-190 the worst, with the P-51 in the middle. I have also fought the FW-190 from a Bf-109 and I can kill it almost every time. It's always myself against against an AI just to be clear.

Where I struggle is the poor performance of the FW-190 compared to the Bf-109, because this does not seem to match the actual reputation of the Focke-Wulf. I still hear Erich Brunotte saying that the FW-190D was the Non-Plus-Ultra, and in this sim it certainly isn't!

LeCuvier

Windows 10 Pro 64Bit | i7-4790 CPU |16 GB RAM|SSD System Disk|SSD Gaming Disk| MSI GTX-1080 Gaming 8 GB| Acer XB270HU | TM Warthog HOTAS | VKB Gladiator Pro | MongoosT-50 | MFG Crosswind Pedals | TrackIR 5

Link to comment
Share on other sites

There definitely are issues with the AI. It uses a simplified flight model and damage model. As a result it almost never loses energy and can take an insane amount of damage that would leave a human pilot struggling to keep the bird even flying, never mind combat effective.

 

It's however not surprising the FW 190 is the worst at dogfighting, that's what you'd expect. It is, however a boom-and-zoom monster. Dives and zooms really well, amazing high speed handling especially the roll which no other plane can even get close to. Also has great weapons and great cockpit visibility. But yeah it turns like crap and definitely doesn't climb like a 109 (after all what does?).

Link to comment
Share on other sites

109 beat all currently because of SFM, in previous version the P-51 AI was the same ED tone down the AI flight model of the P-51. Only 109 left.

Once you have tasted Flight, you will forever walk the Earth with your Eyes turned Skyward.

 

[sIGPIC][/sIGPIC]

9./JG27

Link to comment
Share on other sites

Btw, I heard a lot about use zoom technique. I even remember Johnson's book about how he beats spitfire with zoom in 47. So in game, how do I do that? Do I quickly yank stick back or smoothly pull it back? I am always pretty bad at timing when to start my zoom and how quickly to do it.

Link to comment
Share on other sites

After P-51 is quiet heavy you have to climb only slightly in a Zoom, and do your manoeuvres always gentle, because quickly yank the stick cost allot of energy. And 109 is lighter airframe in a hard climb the Mustang lose faster his speed because of gravity. After I am not the top of the Notch Mustangs possible wait for more answers :thumbup:

Once you have tasted Flight, you will forever walk the Earth with your Eyes turned Skyward.

 

[sIGPIC][/sIGPIC]

9./JG27

Link to comment
Share on other sites

@Pandacat:

 

So the P-47 has much better zoom abilities than a Spitfire. This is because the P-47 is a lot heavier than a Spitfire, but not hugely dissimilar in aerodynamic efficiency.

 

In practice what this is means is that if a Spitfire is doing 400 mph in level flight, then if it zooms straight up it might gain 2000ft. In contrast if the P-47 is doing 400 mph in level flight then if it zooms straight up then it might gain 3000ft. It zooms much higher for the same energy. So if a Spitfire is following a P-47 and they both zoom up, then P-47 will end up 1000 ft higher than a Spitfire. That's quite a big energy advantage, and the P-47 might be able to use that extra energy to defeat the Spitfire, even though they started at the same altitude and the same speed.

 

Similarly with diving. If a Spitfire dives 1000 ft he might gain 100 mph, but if a P-47 dives 1000 ft he might gain 150 mph. Again because a P-47 is much heavier, but has similar aerodynamics. Although they started at the same speed / altitude, the P-47 is much better able to convert that altitude to speed (or vice versa).

 

A really nice example of this I saw is from a Warthunder Youtuber called ramjb who explained the concept really well.

 

 

So caveats, it's a WT video and it's not even the most realistic mode of WT. But War Thunder's modelling of energy is accurate enough that the correct principles apply - even with all the mouse aiming shenanigans. The execution in a sim like DCS is a little different, but the idea is basically the same.

 

In the first half of the video he defeat a Typhoon of similar energy by using the better dive & zoom abilities of the FW-190 (which is an amazing BnZ plane) to build a significant energy advantage over it in what looks like a simple chase. Once he's gained enough separation he can convert this energy advantage (in terms of speed) to an altitude advantage over the Typhoon and use that to defeat it.

 

In the second half he fights a Spitfire of superior energy, and uses the zoom ability of the FW-190 to zoom much higher than the Spitfire would expect he could. His FW-190 zooms much better so it hides its energy well, even though it looks like he's at an altitude disadvantage that can easily be made up for by the excellent ability of the FW-190 to convert speed into altitude.


Edited by Tomsk
Link to comment
Share on other sites

:huh: What?! Boom-n-Zoom™ is diving on the enemy, firing and then climbing again, without engaging in a turning fight.

 

A heavy aircraft like the P-47 and P-51 do have more energy than much lighter aircraft, like the Spitfire or Bf-109, but to climb it must also exert much more energy to have its mass overcome gravity: you trade speed (kinetic-energy) for altitude (potential-energy).

 

But the full advantage of the P-47 and P-51 is not their ability to carry higher potential-energy through their weight, but their better handling at high speeds. Thus diving to gaining high speed, puts them into a flight mode in which they excel compared to the Spit and the 109.

 

That video only demonstrates the failed physics engine in WT.


Edited by Captain Orso

When you hit the wrong button on take-off

hwl7xqL.gif

System Specs.

Spoiler
System board: MSI X670E ACE Memory: 64GB DDR5-6000 G.Skill Ripjaw System disk: Crucial P5 M.2 2TB
CPU: AMD Ryzen 7 7800X3D PSU: Corsair HX1200 PSU Monitor: ASUS MG279Q, 27"
CPU cooling: Noctua NH-D15S Graphics card: MSI RTX 3090Ti SuprimX VR: Oculus Rift CV1
 
Link to comment
Share on other sites

I thought zoom climb was a quick high angle climb where you quickly convert speed into altitude. Because of your heavier mass, you can potentially gain more altitude by inertia. But after the initial climb, once you enter a sustain climb, lighter high rate of climb plane will win. That how I understand what Johnson was doing.

Link to comment
Share on other sites

A heavy aircraft like the P-47 and P-51 do have more energy than much lighter aircraft, like the Spitfire or Bf-109, but to climb it must also exert much more energy to have its mass overcome gravity: you trade speed (kinetic-energy) for altitude (potential-energy).

 

So heavier planes with good aerodynamics really do zoom higher (and dive better) than lighter ones with similar aerodynamics. Seems counter intuitive, since as you say purely looking at it in terms of energy you'd think it'd cancel out. After all

 

Kinetic energy (KE) = 1/2 * mass * velocity^2

Gravitational potential energy (GPE) = mass * g * height, where g is approx 9.8 m/s^2

 

If we have a plane flying along level, and then we point the plane straight up and convert all the kinetic energy into gravitational potential energy then all the kinetic energy becomes gravitational potential energy and:

 

GPE = KE

mass * g * height = 1/2 * mass * velocity^2

g * height = 1/2 * velocity^2

height = (1/2 * velocity^2) / g

 

Here mass is on both sides of the equation so it cancels out, this equation is independent of mass. A faster plane has more energy, but takes more energy to climb to a given height. Exactly as you say.

 

If planes travelled in a vacuum this would be the end of the story, QED. However, planes don't travel in a vacuum and this changes things. The key difference is air resistance. So let's look at the forces acting on a plane in an engine-off straight up vertical zoom.

 

The first force is the force of gravity pulling straight down on the aircraft (nose-to-tail). The second force is air resistance, resisting the plane's motion through the air, again straight down in this case since the plane is zooming straight up. We can calculate the total forces acting on a zooming plane as:

 

Total force = Gravitational force + Air resistance

 

We can calculate the Gravitational force using simple F = ma.

 

Gravitational force = mass * g, where g is again approx 9.8 m/s^2

 

It's more complicated to accurately calculate the air resistance but fortunately we don't need to. The key thing is that air resistance only depends on velocity as well as shape (cross sectional area etc.) it does not depend on mass. We'll say that the air resistance at velocity 'v' is R(v), i.e. some complex function of velocity based on shape and drag coefficients and such.

 

Air resistance = R(v)

 

So now we can express the total force acting on the plane as:

 

Total force = mass * g + R(v)

 

We can now use this total force to calculate the acceleration that applies to the plane, again using F = ma.

 

mass * a = mass * g + R(v)

 

Rearrange to get the acceleration (in this case 'deceleration' since the plane is zooming up):

 

a = (mass * g + R(v)) / mass

a = g + (R(v) / mass)

 

Here we can see that as the mass gets bigger the deceleration due to gravity doesn't change. However, the deceleration due to air resistance gets less. R(v) doesn't vary with mass, and as the mass gets bigger R(v) / mass gets smaller. The result is a heavier plane decelerates more slowly when zooming up than a lighter one, assuming the planes are both the same shape. Because it decelerates more slowly it will reach a greater height.

 

You can use the same maths to show a heavier plane also accelerates more quickly in a dive (and will reach a higher terminal velocity) than lighter planes of the same shape.

 

That video only demonstrates the failed physics engine in WT.

 

The WT physics engine definitely has its problems, but it does model both gravity and air resistance well enough to get approximately correct behaviour here, and this effect is well known. Heavier planes zoom higher and dive faster than lighter planes (if they have the same shape and thus same air resistance). You can do the same experiment in DCS and you'll get the same result.


Edited by Tomsk
Link to comment
Share on other sites

You are forgetting the extra energy necessary to accelerate the extra mass.

 

Think about this. You have two identical rockets, except Rocket-A weights 10,000lbs, while Rocket-B weights 15,000 lbs, because its payload is full of bricks :D. Both are moving straight up away from Earth at 10,000 mph.

 

If you cut the engines on both at the exact same time Rocket-B will continue to climb longer than Rocket-A, because Rocket-B's greater mass (inertia) works for it against the gravitational and air resistances.

 

Now, start both on the ground, motionless, with the exact same rocket engines, with fuel for both to burn the exact same amount of time.

 

Which flies higher?

 

Carry that over to aircraft, because the principle is the same.

 

Also, when an aircraft changes direction to go into a climb, it is changing its velocity. The greater the mass the greater the energy to change velocity.

 

So, given two identical aircraft, except one being heavier, both flying at the same altitude and speed, if they both begin a climb through identical maneuvers, the heavier aircraft will already lose more speed than lighter aircraft through the maneuver.

 

So the question is really, how much the additional energy to change velocity compares to the additional inertia to overcome air resistance provided by the additional mass. That on top of getting both aircraft to the same altitude and velocity.

When you hit the wrong button on take-off

hwl7xqL.gif

System Specs.

Spoiler
System board: MSI X670E ACE Memory: 64GB DDR5-6000 G.Skill Ripjaw System disk: Crucial P5 M.2 2TB
CPU: AMD Ryzen 7 7800X3D PSU: Corsair HX1200 PSU Monitor: ASUS MG279Q, 27"
CPU cooling: Noctua NH-D15S Graphics card: MSI RTX 3090Ti SuprimX VR: Oculus Rift CV1
 
Link to comment
Share on other sites

Now, start both on the ground, motionless, with the exact same rocket engines, with fuel for both to burn the exact same amount of time.

 

So that's a completely different scenario. I'm only talking about zoom climbing (which is primarily based on climbing using inertia), not powered climbs which are mostly based on power to weight ratios. Yes, for a powered climb the lighter plane will climb quicker. But engine power is not the dominant factor in a zoom climb, that's what makes it a zoom climb :-)

 

So the question is really, how much the additional energy to change velocity compares to the additional inertia to overcome air resistance provided by the additional mass.

 

So based on simple energy conservation, no not really. It definitely does take more energy to accelerate a heavier aircraft, however it also has more energy. Much as in the GPE vs KE question the difference cancels out.

 

However, it is the case that heavier planes tend to have higher wing loadings, and planes with higher wing loadings tend to lose more speed in turns. So if the question were "If you fly along level and then zoom up by making a 15G turn" this might be important. However, assuming that the turn is reasonably gentle then the difference in speed lost in the turn is small.

 

That on top of getting both aircraft to the same altitude and velocity.

 

It absolutely does take more energy to get a heavier plane to the same altitude. Of course it also has more energy than a lighter plane at the same altitude. But the question is not which can get higher for the same fuel load, the question is given they are at the same altitude and speed already if they zoom up which will go higher. The answer is the heavier one will go higher. The fact it burnt a lot more fuel to get to that initial altitude isn't really relevant.

 

The question is definitely more complex than my simple illustration. Most zoom climbs are not power-off climbs, so engine power relative to weight does play a role. A lighter plane gains some advantage here, although typically heavier planes also have heavier and more powerful engines to compensate for their additional weight. More importantly, for a zoom climb power-to-weight is not a dominant factor, planes can use zoom climbs to temporarily climb much faster than their power-to-weight ratios would allow.

 

Another factor, as you point out, is that if you make a very hard turn into the vertical then that will bleed a lot of energy and the heavier aircraft will likely bleed more. However, if the turn is reasonably gentle then that isn't going to be a big factor.

 

I think we can both agree the issue is actually quite complex. However, what has been found in practice is that when all things are considered, heavier planes with good aerodynamics such as the P-51 (about 3500kgs empty, 4300 kgs loaded) tend to zoom much better than lighter ones such as the 109 K4 (about 2700kgs empty, 3100Kgs loaded). This effect can be used in combat to gain an advantage.


Edited by Tomsk
Link to comment
Share on other sites

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...