Top dog? The F-35 put in a poor showing against an F-16 in a mock dogfight. (Lockheed Martin).
TIM ROBINSON puts virtual F-35s into perhaps the most accurate non-classified high-fidelity simulation of a future air combat clash. Who comes out on top?
Last week saw what seemed like a fairly damning flight test report leaked to the blog War is Boring - where a F-35A test pilot was unable to triumph over a 70s-era F-16D with two external tanks in a close-in mock dogfight.
In the report the unnamed pilot noted the F-35's lack of energy manoeuvrability and its restrictive flight control software led to it being unable to turn the tables against the Viper. He also criticised the aircraft's lack of rearward visibility.
The story - as might be guessed - sent ripples through amateur and professional air warfare experts – with critics claiming that it is yet more evidence the aircraft is an expensive disaster. Proponents, meanwhile, lined up to defend the fighter. The news story, picked up by other outlets, moved the F-35 JPO (Joint Program Office) to rebut the report saying: "The F-35's technology is designed to engage, shoot, and kill its enemy from long distances, not necessarily in visual 'dogfighting' situations".
The JPO also noted that the aircraft involved, AF-2, was a flight test prototype and thus lacked its stealth coating, mission, sensor and weapon systems that would have afforded off-boresight missile shots.
Crucially, however, it did not dispute the authenticity of the report nor the test pilot's comments about agility.
Simulating more representative combat
Harpoon on steroids - Command Modern Air Naval Operations is a high-fidelity real-time tactical simulator.
While the JPO rebuttal may have not addressed the elephant in the room as regarding the F-35's unimpressive kinematics – it is correct that a 1 vs 1 fight, with a non-production version F-35 stripped of its advanced sensors might well be an unrepresentative test. Even in WW2 most kills by the top aces were achieved without the victim being aware they were being bounced, rather than the aerial melee seen in 'Top Gun' and elsewhere.
With this in mind a curious, impartial mind might ask - how well might a F-35 do in a more operationally representative scenario? Beyond manufacturer's slick marketing videos and pithy soundbites about 'gamechanging' 5th generation technology, it is difficult for an outsider to evaluate the F-35s potential - especially in the air-to-air arena. Is it true revolution in air combat or an overpriced lemon?
Fortunately we now have a pubically available tool to take an informed look at least some of the claims made for the fighter.
To this, we turn to Command Modern Air Naval Operations - a hyper-realistic tactical PC simulation/wargame which models sensors, stealth and other factors in great detail. Having won plaudits from amateur and professionals alike in its detailed modelling – it recently is set to move into the professional military and defence world through a co-operation agreement signed with BAE Systems. The game features a Jane's style database of aircraft, ships, weapons, sensors and missiles from 1945 to 2020 with the whole world modeled and country-specific equipment lists.
Of course, simulations have been used before with the F-35 - with a famous RAND study concluding the jet would be 'clubbed like baby seals'. However CMANO is a substantial advance on Harpoon-era simulations - with a more detailed, higher fidelity air warfare model. For instance, it models the kinematic effects of aircraft losing energy dodging incoming missiles or SAMs - making salvos more important against highly-agile targets. The probability of a missile kill (Pk) is influenced by many factors - including seeker generations, range to target, agility of target aircraft, target aspect, countermeasures and pilot skill - making for a deep and complex simulation. The AI, too is clever enough to evade and try and 'beam' incoming missiles - making for a highly realistic BVR simulation. In the older Harpoon, for example, missiles could be fired rearwards from fighters, the air combat modelling being far more abstracted.
In this test we are role-playing the 'battlespace commander' with a top-down view of the air battle rather than the individual pilot - but in reality the F-35 is likely have this level of information provided to the pilots themselves in the cockpit - thanks to datalinks and sensor fusion.
How, then would it model the F-35 in an air combat scenario?
Western fighter pilots can expect to encounter the Su-35S in significant numbers. (Sukhoi)
To find a plausible air combat scenario these days it is not necessary to look too far away from the headlines.
The scenario imagines a Baltic crisis gone hot in 2020 with UK F-35Bs pressed into the air superiority/CAP mission. (In reality of course this would be more likely to be the Typhoons role).
For the purposes of this test we will imagine a flight of 4 x F-35Bs going up against 4 of Russia's latest production 4.5 generation fighter – the fearsome Sukhoi Su-35S. The F-35s are configured for long-range air superiority with 4 x MBDA Meteor BVRAAMs each - and thus relying on stealth. As support the F-35s have an E-3D AWACS & Rivet Joint. Meanwhile, the Su-35S have 10 AAMs each, comprising 6 x AA-12 Adder As and 4 x AA-11 Archers.
Both sides skill is set to equal & ROEs are such that only contacts positively ID'd as hostile can be engaged. Both flights start at high altitude (40,000ft).
Some caveats must be understood before we let loose the dogs of (virtual) war.
1) This is an unclassified consumer wargame with weapon ranges/sensor data drawn from multiple open sources (and very informed guesses). Real- world missile ranges and sensor performance therefore could well be better. While the simulation has undeniably accurate - it will still have some gaps and discrepancies. One oddity, for example, is the gunpod which was included as the standard F-35B LO air superiority load-out with 4 x internal Meteor missiles. Is this a major factor? Probably not, given that in around 15 tests only one saw the F-35 get into a guns fight.
2) Again, while the point of this test is to see how the F-35 performs in a more operationally representative test - it still omits a lot from a real-world air battle. Support assets would include SAMs, surface ships and friendly fighters as well as other assets to feed into the electronic order of battle. In particular, the presence of friendly Rafales/Typhoons/Gripens as non-LO assets could allow tactics to distract and feint the enemy while the LO F-35s set up ambushes.
3) This test is obviously being conducted by an amateur air power person. Those professionals who get paid to study, teach and train this subject full-time will no doubt be able to get much better results, more consistently.
4) Rules of engagement (RoE) may well be different for a real-world crisis – which short of WW3 may see civil air traffic in or near the battlespace. This could aid (in terms of providing a clearer air picture) or hinder (if political restrictions meant visual ID rules were imposed) an information-age, LO fighter like the F-35.
Lets then see what happens.
1. Instantly the F-35s EW suite classifies the bandits as Su-35s and tags them as hostile from around 350nm away – a huge advantage in deciding what to do next. This is particularly simple in this scenario since we have no other contacts to worry about, but can take time with other platforms to build the air picture and finally ID a contact.
2. With the contacts ID'd at this long range I can start coming up with a plan and can even 'sort' contacts and assign wingmen to targets. The Flankers PESA radar is something to keep in mind and I am aware the F-35 is not invisible – I therefore plan to send each pair of F-35s around to the side slightly to see if I can keep out of the radar and ambush them from the side. I am keeping the F-35s radar off and relying on passive sensors to sneak up on the Flankers.
3. Pincer move starts to take effect. Notice that the F-35s ESM suite is refining the position of the Su-35s – even though they have yet to enter the radar coverage of the AWACS.
4. The Flankers are now advancing into the jaws of the trap. How long I can remain undetected is unknown, however.
5. At this point – it looks like my gamble has paid off. The 2 southern F-35s have slipped past the Su-35's radar screen and are about in position to execute a turn and get into weapon range. Note also that thanks to datalinks from the AWACS the position information on the Flankers has firmed up.
6. With my southern pair ready to fire, I aim to fire 1 x Meteor each at max range to see if I can kill the Southern bandit. Note also I am aiming to 'crank' and turn to keep the threat at range to avoid closing to the merge.
7. A useful new feature introduced in a recent patch to CMANO is WRA (Weapons Release Authorization) – allowing the player to tailor missile salvos to the threat. With Meteor I can leave it at max range.
8. With my southern pair about to fire I get ready to arc my northern pair to hit the enemy from behind.
9. Weapon away. Meteor fired at max range.
10. Weapon guiding now – still the Flankers are unaware we are here.
11. First missile is decoyed by the Flanker's chaff – but a second one is now heading its way.
12. Second missile is also dodged. I risk going active with my radar to firm up a solution for 2 more shots.
13. Splash one Flanker. However my AI in automatic fire mode decides to loose its 4th and last missile – effectively putting it out of the battle - since I have no intention of closing in to the merge for a guns duel. I decide to RTB it.
14. I may have to take that back. The AI was smarter than I thought and was actually aiming at the second Flanker – hopefully setting up his wingman who still has missiles for an easy kill.
15. Southern F-35 no2 fires 2 x Meteors at the southern bandit.
16. Second Flanker splashed. However I now have 2 x Su35s bearing down on the south F-35 which is down to one remaining missile. I kick the northern pair into full power to get them to intercept the last two Su-35s.
17. For my last Meteor from my southern F-35, I decide to switch to manual targeting and allow the Flanker to get closer to 50nm to put it at the heart of the Meteor's engagement envelope. A risky move, but it could pay off.
18. My last Meteor from the southern F-35 is about to go active.
19. Splash Flanker no3! Waiting to launch the Meteor until it was closer paid off. I now have two F-35s Winchester, two F-35s still with 4 x Meteors and one bandit left.
20. With the Flanker now vectoring south – there is about 25nm before it comes into Meteor range of my northern pair.
21. Coming in range – I risk going active with radar to firm up the firing solution.
22. Switching the radar on has the effect of getting the Su-35s attention and it turns to north to meet the new threat. Four Meteors streak away to the last Su35.
23. 4 x Meteors inbound to the bandit.
24. First Meteor is spoofed, the second connects. The last Flanker is down. In the whole scenario - Red Air only got 2 x missiles off.
SIDE: Red Air
4x Su-35S Flanker
2x AA-12 Adder A [R-77, RVV-AE]
9x Generic Chaff Salvo [4x Cartridges]
From the Red Air perspective
1. A chance to look at the scenario from the enemy point of view.
One of the very cool aspects of a sandbox simulation like CMANO is its ability to jump into a Gods-eye view showing all sides, or even to switch sides. Let's take a look at this scenario from the Flankers point of view.
2. With my four Flankers spaced roughly 40nm apart I am hoping that I can pick up an F-35 in our wide sweep. Intel has said there are F-35s in the Baltic – but where are they?
3. Getting closer – but my radars are still not picking anything up.
4. Meanwhile it is a very uncomfortable experience to be flying into a zone where stealth fighters have been reported. I have PESA radar and 40 missiles between us – but we still haven't picked up a sniff of the F-35s.
5. Still no contacts. Could they be further in?
6. At time of firing – still no contact.
7. @~&^! my southern Su-35S spots a missile contrail arcing their way – at only 13nm away.
8. After decoying two missiles a third Meteor obliterates my southern Flanker. Luckily I now have an ESM spike which shows a radar to the south. Time to turn to meet the threat.
9. At about 76nm away I finally am able to classify the threat as an F-35 using my IRST.
10. A second Flanker goes down – again from a F-35 I haven't seen. However I still have a contact on a F-35 to the south, although frustratingly there is no speed or altitude information.
11. Another missile warning – this time from a different bearing to the F-35 we are chasing!
12. A third Flanker from my flight is blasted out of the sky. At this point, with three aircraft down any sane fighter pilot would be thinking about egressing and how to escape – but since only 1s and 0s are losing their digital lives here, I'm going to press on. Perhaps I can kill the pesky F-35 we did have a track on.
13. Whaaaat? A FCR (fire control radar) spike pops up behind me? Turn to face this or keep after the original contact?
14. I turn north to meet the new threat. Three wingmen down, but at least I now seem to have a radar contact on the edge of my WEZ.
15. I fire two AA-12 at the contact – but predictably there are two missile contrails heading down my throat too. The last Flanker goes down in a ball of flame.
Game over for Red Air.
The challenge for any 'Red Pilot' to solve is that if F-35s keep their radar off it is extremely difficult for the Flankers – even with AESA radar to detect them. The Meteor BVRAAM meanwhile with its high agility in the end-game mean that even the highly manoeuvrable Su-35S can lose energy dodging these shots which appear out of nowhere. Even armed with 10 missiles each, the Flankers need a reliable target before they can engage, which the F-35s simply do not provide.
Even with the Flankers EMCON set to passive - AWACS was able to direct the F-35s to classify the hostiles as Su-35s at a range of around 96nm using their passive ETOS.
As a final test - I decided to hand complete control of the F-35s to the AI, assigning them a CAP zone to defend and switched to Red Air with the intention of finally beating them. I also loosen up the RoE for both sides allowing the fighters to fire on anything not friendly, rather than hostile contacts. But, even using sneaky tactics (1 x Flanker with radar on as bait, the rest silent relying on passive sensors) the result was much the same - with missiles appearing out of nowhere and from unexpected directions. I finally managed to down a F-35 when the AI made the mistake of switching its radar on deep inside my WEZ - a mistake that a well-trained human F-35 pilot probably wouldn't make. However by that time I had lost 3 x Flankers - a Pyrrhic victory at best.
Trying to even the balance
Adding the A-50 as AEW support to Red proved pointless as the F-35s still were invisible to its radar.
After about 15 runthroughs with the Su35s being shot out of the sky I decided to add more Red support assets. The first being a A-50 'Mainstay' to provide AEW coverage and the second a Su-24MP Fencer F EW variant to provide jamming capability. Surely this would jam the missiles and allow the Su-35s to get to the merge?
1. As you can see - despite jamming from my stand-off ECM aircraft (top right) on my missiles, my northernmost F-35 has dispatched the northern Su-35 with its first missile. Other missiles from F-35s - (staying radar silent) are on the way.
2.Despite ECM two more Meteors find their mark. It may be that the jammer needs to be closer to have a better effect - however that risks turning it into an easy target for my F-35s.
3. Last Su-35 is about to be shot down. At this point I have dispatched 4 x Flankers and still have 3 x Meteors left with no F-35s lost. If I was feeling really aggresive I could now go after the defenceless AWACS and Fencer to finish them off.
Another runthrough – this time with the F-35s heading more directly to the merge. In this instance 'Gods Eye' view is turned on, but the Flankers are still blind to the F-35s presence despite being in the AESA radar coverage..
This of course, was a quick and dirty look at a possible future air combat scenario using the F-35 rather than exhaustive simulation and testing that goes on in military or defence industry labs. However it does throw up some interesting observations. In more than 15 runthroughs the kill ratio was 3-0 or 4-0 to the F-35s, with a couple of instances of 3-1. So what does this tell us?
First. The F-35 certainly does not suck at air combat, providing it keeps within its own realm. As this testing demonstrated, the challenge for any future 'Red Air' pilot will be detecting the F-35 and then getting close enough to nullify its LO features in the merge. Though CMANO simulation is extremely powerful in modelling kinematics, sensors and is a huge leap from the earlier Harpoon, it does not model a 3D ACM encounter in high-fidelity like say Falcon 4 or DCS. Post merge, like real-life, it then becomes more matter of chance. However a third playthrough, ironically, did see a F-35 close to guns range and destroy a Su-35 leaving the score at 3 Flankers to nil F-35Bs lost. In around 15 runthroughs, the F-35 came out ahead each time, with the worst result being 3 x Su35s lost to one F-35 shot down. As noted above, professional air warfare tactics experts would undoubtedly be able to do better. In only one of these runthroughs did the fight enter the merge - long and medium range shots being the norm.
Two. A LO fighter, with high-end sensors to detect (and importantly classify) targets at range when paired with the Meteor BVRAAM is extremely potent. While the F-35 was able to classify the Flankers at extreme range, the Su35's sensors were still only able to classify the F-35 as a 'multi-role' – even when it was nearly within weapon range. This may be fine in a simulation without other hostiles, friendlies and civilians air contacts to sort and track, but undoubtedly would be more complex in real life.
Note also that the game is conservative about the Meteor's true range – giving it an effective range of 75nm. The real range is likely to be more than this, (think AIM-54s 100nm) and the Meteor is expressly designed to be lethal all the way out to maximum range, unlike other rocket-powered missiles which 'coast' and thus lose energy in the end-game - and thus are easier to evade at long ranges. Thus the real-life Meteor BVRAAM may be far more lethal than this simulation models - as it seems to apply a reduced Pk missile effectiveness across the board for all missiles.
[UPDATE the developers inform me that the Meteor's ramjet giving better Pk at range IS modelled- another example of the incredible detail in this simulation]
Also, while the Meteor certainly can be spoofed by chaff in a last-ditch defence, as the West's newest generation air-to-air missile it seems extremely resistant to ECM/jamming - despite icons clearly showing my Fencer F was having some effect.
Finally, a couple of runthroughs with the Meteors exchanged for internally-carried 4 x AIM-120C AMRAAMs also produced similar results – with 4 Flankers shot down in short order.
Three - it's the human, not the machine. Smart tactics & cunning outmatch technology each time. One observation is that I could have made the 'jaws' of the the trap tighter and still avoid being detected by the Flankers radar enough to put the enemy even closer within the Meteors WEZ. A subsequent playthrough saw me head all F-35s directly into the Flankers path and resulting in all 4 Su35s being shot down within about a minute and a half of the first Meteor shot – an even better result than the test described in detail here.
Four – it is extremely frustrating to play as Red Air and somewhat unnerving to have missiles appear out of thin air. While a previous simulated look at the F-35 (the infamous 'clubbing baby seals' study) concluded that sheer numbers of J-11s would prevail against F-22/F-35s facing masses of Chinese fighter pilots all happily flying into certain death, here the psychological factors were more apparent. If one, two or three of your flight vanished suddenly in explosions and you still couldn't get a reliable track/lock on the enemy - at what point do you decide to withdraw and escape?
Fifth – A fifth (surprising) observation is that the support enablers I added, the E-3D and Rivet Joint appeared to contribute little to the air battle when the Su35s were emitting. This may be due to the F-35's impressive ESM suite – or potentially my non-optimum placement of these assets behind my fighters. Where the AWACS did make a difference was in a couple of runthroughs with the Su35s staying 'radar-silent' - and allowing the F-35s to close to ETOS range to passively ID their targets for missile shots.
Does agility still matter?
Another runthrough – this time with the F-35s heading directly into the merge saw 2 x Su35Ss shot down within seconds of each other.
So does agility even matter? It certainly seems to - even for BVR combat - because the game takes manoverability of the target into account in calculating missile effectivness.
Interestingly, the game database merges fighter 'generations' and 'agility' to give one overall number - rather than as seperate values. So the F-35 in CMANO has a fighter generation/agility in the '5' class – on a par with the F-22/Typhoon/Rafale – while the the Su35 is 4.5 – which may account for the guns kill at close range by the F-35. Statistical anomaly, sheer luck or a necessary simplication that while the ultra-agile Su-35 should theorectically be able to have the F-35 for breakfast in a visual dogfight, there may be other factors at work (eg sensor fusion, HMD, 360degree electro-optical distributed aperture
Indeed, while the leaked dogfight report lays bare some of the F-35s deficiencies in close-in ACM - the recommendations from the test pilot shows that at least some of them can be addressed by adjusting the FBW system. However, although software tweaks to the FBW systems will be able to improve the pitch rate, AoA blending & remove some of the flight control restrictions - the overall energy sluggishness – a heavy fighter with a high wing loading will still remain and pilots will need to adjust their tactics and skills accordingly. A close reading of the report also suggests that while the test focused on the F-35 pilot using high AoA and nose-pointing (like an F/A-18) to engage the F-16, it might be more effective fighting at a higher speed/lower AoA - like the F-16 itself. As ever, as the aircraft reaches operational squadrons and we see more DACT encounters, pilots will evolve specific tactics for the fighter. Don't, however, expect that this will be the last mock dogfight that the F-35 loses - even F-22s have appeared in the HUDs of Typhoons, Growlers and T-38s.
An analogy of the F-16 being replaced by the F-35 might be the US WW2 Eagle squadrons in the UK trading their nimble Spitfires for the heavy lumbering Jug - the P-47. While there was initial reluctance from the pilots - it was found that with the right tactics a P-47 could beat a Spitfire in 1 vs 1 combat - and pilots soon got to appreciate its other advantages in the ground attack role.
The key question then is - will the manoeuvrability of a production F-35 with only comparable agility to legacy fighters be outweighed entirely by the situational awareness advantages it provides its pilots? Is 'information the new 9G'?
In 1982 - a previous STOVL fighter fielded by the UK FAA ended up surprising the world when it shot down supersonic delta-winged jets that on paper were superior in performance. Could F-35B do the same? (Lockheed Martin)
As noted above, while these simulated tests give an interesting insight into air combat using LO fighters, they do come with a number of caveats and should not be taken (as is so often the case, firm evidence to support conclusion X). Your mileage may indeed vary. However, they do highlight the extreme difficulty for an adversary of getting to the merge with assailants, who, if playing 'unfair', maximises their LO and sensor advantages. That is not to say that WVR air combat cannot happen. Leakers, decoys and pop-up threats mean the enemy always gets a vote - and thus F-35 pilots will still need to train how to fight in the visual arena, and learn the strengths and weaknesses of their aircraft vs any threat aircraft.
For those nations, air forces looking to draw conclusions from this single F-16 vs F-35 leaked 'dogfight' report (in reality a dynamic flight test around the stability of the fighter at high AoA and fine-tuning the FBW) - it would seem to be unwise to underestimate the F-35. Get close-in with a highly agile fighter in a 1 vs 1 and you may be able to beat it, but as these tests seem to indicate, the real challenge will be getting that close without getting turned into burning wreckage.
If you are fighting LO fighters in CMANO - get used to this message.