I think the difficulty comes from modeling a system of factors that influence a set of values, instead of modeling the set of values directly. If we were to simply have X top speed and Y service ceiling, things would be very easy. What we have however is the set of equations that govern these values.
Incidentally, that's more or less how sims in the early 90s worked. You had only a set of values that described each aircraft and they were constant.
The upside is that it always works. The downside is in the real world things don't remain as constant as that: a lot depends on atmospheric conditions, other parameters of flight, etc.
For example, in an older sim a developer could get away with modeling all fighters at their maximum turn rate. However, we know today that turn rates are not constant, they depend on things like available thrust and airspeed at the time of turn, etc etc.
It's just a case of how much you model. The more you do, the better the representation once you get it right, but also the harder it is to get it right because things are interconnected: you make a small change here and it influences 3-4 other parameters down the road, which forces you to go back and reevaluate the model you use itself.
On the matter of tests now, they reflect top performance most of the times. Top performance is rarely continuous in the real world though. It's what we had in the previous IL2 series, where all aircraft could be pushed to the limit and stay there all day long, which was equally inaccurate.
The current system is a better reflection, because it takes more parameters into account and makes for highly situational combat (like the real thing), with features like CEM and so on. But like i said before, it's harder to get right.
When all is said and done and the sim's FMs are finalised at some point in the (hopefully, near) future, i don't really expect our aircraft to be pushing their top speed 24/7 either.
One day you'll get jumped by a 109 that's been in another engagement previously and you'll survive in your Hurricane, because your engine was cool and his was hot, allowing you to to use emergency boost for bit longer than the enemy. The next day you'll have the superior aircraft and lose because you will be the one who's coming from a previous engagement and get bounced by a 110 that has been lazily cruising along 2km above you. Which is pretty much the kind of thing we've all been reading in pilot memoirs and historical documents
P.S. It's not just the RAF models. The 109s are also slower. Plus, you'd be surprised but probably the fastest thing in the sim (at least around a specific altitude band) should be the 100 octane versions of the 110 that are currently missing. It would still accelerate like a brick and bleed speed like a pig (so it's dead meat anyway if it turns with you), but it's still missing some speed and the correct cannon shells
By the way, i was the guy in the 110 that you chased back to France today on ATAG, after i attacked the ships