What I mean is, despite it being only an approximation and perhaps not entirely appropriate to the situation, if its within an order of magnitude correct its still acceptable and the "relative effectiveness assessment" is still good.
No, it would not be good.
An order of magnitude is a factor of ten, just like 'db', and in radar detection, a loss of 10db or a reduction of RCS of 10db will allow the aircraft to cut the effective distance in half before it is marked as a valid target.
Here are APA's own words...
A Preliminary Assessment of Specular Radar Cross Section Performance in the Chengdu J-20 Prototype
The PO computational algorithm performs most accurately at broadside or near normal angles of incidence, with decreasing accuracy at increasingly shallow angles of incidence, reflecting the limitions of PO modelling.
Remember, in radar detection
EVERYTHING works off
REFLECTED signals and their intensity. So if you want to know as much as possible the power intensity of a reflected signal, you want to see it at its fullest and what better than perpendicular?
What that sentence mean is that Physical Optics algorithms works best when the return is straight on as in 'normal angle of incidence'. The moment the viewer depart from perpendicular, the less accurate the algorithms can give us regarding the power intensity of the reflected signal. The more the viewing angle approaches horizontal the higher the percentage of that power intensity is away from the viewer. So oddly enough, the only time when PO can give us an accurate RCS value of a complex body is the frontal view where there are few surfaces that can give us 'normal angles of incidence'. In other words, in the frontal view, most deflected signals will be so far 'downstream' of the radar view those signals will become irrelevant. This is why the frontal RCS of every aircraft is the lowest.
But when we (the radar) are looking at the sides, which includes 'top' and 'bottom', we
KNOW that there would be times when we would not be capturing all or a significant percentage of a reflected signal because we would not be in such a favorable 'normal' incidence angle all the time and that mean any RCS estimation would be suspicious. This uncertainty is compounded by the presence of as complex a body as an aircraft with its many reflecting surfaces forcing uncaptured percentages of reflected signals to interact with each other. This uncertainty was partially cleared up by Ufimtsev with his edge diffraction equations, with surface wave behavior algorithms, with null fields method, and many other.
Go back to what I said: That a 10db loss of reflected power intensity equals to half the effective detected distance, meaning the aircraft can come closer by half before it is qualified as a valid target.
I did not specified if that loss is through legitimate means such as shaping or absorber. That loss could be through sloppy math from a flawed analysis using an inadequate tool. In other words, if you know that PO is inappropriate
BY ITSELF but you proceed anyway and if you record a 10db difference as you measure the aircraft from various viewing angles, you would be fooling yourself that you have just cut in half the effective detection distance of the J-20.
The problem here is that instead of a complex body like an aircraft, we have recorded as great as 30db differences between viewing angles
FROM THE SIMPLER AUTOMOBILE USING MORE ADVANCED TOOLS THAN PHYSICAL OPTICS. So what make you think that we should accept PO alone?
Garbage in. Garbage out. Why do you think we have other tools to compensate for PO's known shortcomings?
What APA did was no different than a training exercise for the new junior engineer on how well he knows his software. In my days, instead of software we took the newbie down to the range to see if anyone has a full size model of any aircraft and see how well he set up his measurement regiment. If there are no aircraft then we use a car we got from the junk yard. Sometimes the person like that aspect of the job so much that he transferred out of the design dept. and into the range facility and become a range test designer.
Look at what your man said...
When it comes to stealth, you can't use "looks" as a standard.
There is a huge difference between "looks" and SHAPING. I am discussing SHAPING and you can't tell the difference.
What is the difference between some aerodynamics values such as drag coefficient or aerodynamic efficiency and radar cross section? Keep in mind that both depends on the interaction of a body and a medium.
In aerodynamics, we measure some things and come up with some values from the target's perspective, meaning we measure the response of the body when aerodynamic forces are present, to put it simply.
But that does not mean if I shoot a 100 km/h air blast at a car I am going to get a return blast of air that also travels at 100 km/h when that blast of air make contact with the car.
But that is how radar detection works: From a viewer's perspective. Not from a target. That mean if I shoot an EM beam at the car, I will get an EM return and both signals will travel at the speed of light. It is from this return signal that I will perform many analysis and come up with a conclusion about that signal.
Your man violated his own rule, did not know it, and worse yet did not know why.
He assumed that just because the F-22 and the J-20 have similar shaping therefore they must have similar frontal RCS value and he made that assumption based upon looks, not from a viewer's analysis of reflected signals, of which no one has access to those data.
Do you remember those 'bow tie' RCS graphs?
The Radar Game
A Tale of Three Shapes
Combat aircraft in today's inventory employ a number of different techniques for reducing their Radar Cross Sections, which are of three different shapes. The Fuzzball, Pacman, and Bowtie shapes are highly simplified symbols for basic signature patterns. Actual signatures are considerably more complex, of course, and information about them is restricted. The three shapes are used to depict how general patterns of RCS reduction give attackers a revolutionary edge.
...the theoretical Bowtie shape has a 15 dB reduction in RCS in its front and rear aspects.
The 'bowtie' shape cannot be measured by sitting in the aircraft's cockpit. It must be allowed to form from reflected signals from the viewer's perspective. See that 15db reduction figure? That will cut effective detection distance by slightly more than half. We cannot calculate that by sitting in the aircraft's cockpit.
A Ferrari and a Lamborghini can have very similar aerodynamics and does have similar shaping to achieve those aerodynamic values. But because unlike aerodynamic forces where there are no returns we have feedback with EM signals so that despite similar shaping there can be great enough differences in db that will mean life or death.
Again...Your man violated his own rule, did not know it, and worse yet did not know why.
He based his assumption from looks, totally devoid of any credible data, and in favor of the J-20, just because the shaping of the J-20 is visually similar to the F-22 in shape. Then he pompously crowed about how many 'Thanked' him for his 'useful' (more like borderline useless) posts, how many viewed his silly video, and he called others names just because they challenged his baseless assumptions.
You guys can take this crap back to your two playgrounds where every gullible conscript reject will trip over each other and 'Thank' you for it and where the admin staff will protect you from criticisms. But you will not be so protected here.