While many stealth proponents point to Red Flag exercises as proof of superiority of stealth aircraft, reality is, as usual, very different from official accounts. It is simple truth that exercises never can replicate reality perfectly; limitations must be placed to ensure safety of pilots and aircraft; other limitations are also in place which would not be in war, such as operational g limit being placed at 9 g for most modern fighter aircraft, despite many of them being able to pull 11-12 g turns (albeit at expense of airframe life). Further, all shots are simulated; and as many factors cannot be taken into account, or at least not completely, results are never as they would be in real world.
But there is more to that. Often, higher branches of military fall into reality denial for various reasons, including pushing for new projects. They have good incentive for that too, as successfully propagating expensive weapons secures them lucrative posts in industry after retirement.
To offer an example: while no missile in the world achieved Pk above 70% against literally defenseless opponents (70% is value for visual-range IR missiles in Falklands against opponent that did not have a countermeasures and did not attempt to evade the missile, trying instead to break missile lock by no longer useful tactics; BVR missiles never achieved more than 50% Pk in same conditions), in Red Flag Alaska Typhoons managed 16 “kills” from 18 simulated missile shots, a Pk of 90%, while 2005/2006 tests between the F-22 and the F-15 used Pk of 65%. However, against capable aircraft, BVR missile Pk can be expected to be less than 10%; with Pk in actual BVR combat being no more than 3-7%. Even if F-22 is carrying 8 BVR missiles, real world Pk for all 8 missiles combined will be 24-56% at beyond visual range – less, if fired in salvos. Aircraft usually used by agressor squadrons is the T-38 trainer, which can pull a maximum of 7,2 g and does not have a good EW suite (if it does have one at all).
Tactics used by agressors are nothing to write home about either, as they use outdated Soviet tactics (“damn the AMRAAM, full speed ahead”), and heavy limits are put by Air Force. Agressors also “simulate” older Soviet aircraft, newest being Su-27 and MiG-29, latter of which does not have shining turning performance thanks to high wing loading, or even older aircraft like MiG-23.
Further, no foreign aircraft ever played the agressor, while US fighters – excepting F-35 – have no IRST, and most have went without avionics upgrade in very long time; as such, most are incapable of jamming, or even detecting, AESA radars. As they cannot detect AESA radars, it is impossible – all “shots” being computerized and not physical missiles – to see where the shots came from, or even realize when they are under the attack. In real combat, however, missile trajectory is easily reconstructed by using missile warners and possibly IRST, and pilots (if competent) start maneuvering as soon as they come under the attack, lowering missile Pk significantly.
Actual force ratios are not represented either. Taking into account cost and sortie rate, 12 F-22 would be expected to face 102 F-16Cs, which means that F-22s in stealth configuration would still be outnumbered by time of the merge even if BVR missiles had Pk of 90%. Yet in all exercises where F-22 dominated, there was near-parity in terms of numbers, or relatively small Red Force advantage.
There is no representation of the fog of war, and fallibility of the IFF systems, both of which require the visual identification before attack can be made – and the F-22 has no IRST or other optical system, which coupled with its large size means that the advantage in engagement range is reversed in favor of the new European fighters (Rafale, Typhoon, Gripen E/F) or the Russian fighters equipped with the IRST (Su-30, Su-35).
Heavily-scripted Red Flag exercises are quite similar to the computer simulations which predicted and predict huge exchange ratio advantage for newer fighters. F-15A was predicted to have a 955 to 1 exchange ratio against the Russian MiG-21. F-14s missiles were predicted to have 100% Pk.
2005/2006 tests between the F-22 and the F-15 assumed:
a) a force ratio of 2 F-15s for each F-22 (a realistic ratio would be 2,2x advantage in procurement numbers * 2,19x advantage in sortie rate = 5 F-15s for each F-22)
b) a radar guided BVR missile Pk of 0,65 (it was 0,34 in Desert Storm against non-maneuvering (cooperative) targets with no countermeasures; incidentally, Pk of 0,65 gives the F-22 two salvos with Pk of 95,7% each, meaning that only one in 22 F-15s might come to the visual range, whereas Pk of 0,34 would give two salvos with Pk of 64,1% each, allowing one in 3 F-15s to the visual range – and that is assuming that they fired salvos instead of simple additive calculation of individual missiles, as in 0,65+0,65 = 100% Pk, or 3-4 targets shot down by each F-22, in which case 2 F-15s to 1 F-22 would result in no F-15s getting to the visual range)
c) perfect BVR IFF (in reality, only visual IFF – either through Mk I eyeball, binoculars, telescope or IRST – is reliable, and the F-22 does not have last two)
So it is quite obvious how seemingly not so large differences in Pk assumptions and force ratios can change results (in 5 F-15s vs 1 F-22 and radar missile Pk of 0,65, one in 5 F-15s would be getting to the visual range, with missile Pk of 0,34 one in 2 F-15s, and with missile Pk of 0,08, seven in 8 F-15s would be getting to the visual range).