So the 109 DID change, Aha!

Why does the new high poly Bf109E-4 seem different to the older one ?  Mainly because the new paint job is baked into a new higher poly model, you can see this the minute you compare the two visually, like the E-1 the E-4 is a higher poly model not just a new paint job.

At low speed in the take off roll the new E-4 model does require more rudder, but it requires this with exactly the same flight data, so it's speed, climb, roll and turn rates, and it's stalling point, are not incorrect once you are airborne.

We first noticed this with the E-1 (which handled differently on take off to the low poly E-4 and F-4 and G6) and now that the E-4 uses the higher poly model it inherited the same take off rudder requirement.

As a result of this discovery, we went over the flight model very carefully to make sure the speed, climb rate, turn and roll rates were what they should be, because change always makes one go "what happened here ?" especially when the data files are not changed but feel is. So, yeah, it does need that extra rudder in the take off roll like the E-1 did when it became the first 109 to use the higher poly physical 3D model, but it flies the same.

When we give the F-4 and G-6 the higher poly model I will not be surprised if they inherit the same take off characteristic, which BTW, is much more realistic because Bf109s were renowned for requiring rudder on the take off roll and when they rotated.

What did not happen is the E-4 did not inherit the wrong data (like from the E-1) ... this would have required copying the hundreds of performance flight profiles that comprise the flight model from the 109E-1 directory to the entirely different 109E-4 directory, which, I can assure you, did not occur. The two are totally different file structures and only a deliberate and very extensive file copy action can make one replace the other. These were checked anyway and this did not happen, and it would have beenvirtually impossible for it to have happened by accident in the first place.  We would never have done this on purpose either.

The reason the handling in the takeoff roll is more realistic and requires that extra rudder is rooted in the changes to a higher poly model and the shape being more complex now than the older 109 models are. When Granik first built the E-1 two things showed up:

1) He struggled with the geometry of the landing gear and wheels to make them look right down and locked and to seat properly in the wings at the same time
2) Because the shape is more detailed the landing gear and wheels are not the same geometry or precise location (if you overlay 1 model on the other)

The flight model data didn't change and this is reflected in the performance numbers when we wind tunnel the aircraft. But it's interaction with the ground during the take off rollout did change because the geometry and detail of the landing gear and wheels are different. It took a while to figure this out because it isn't the first thing you would think of, but in hindsight I remember the bugger of a time Granik had to get the landing gear and wheels to satisfy both their down and locked appearance, and the up and seated in the wings flush fit. Because of the greater detail in the shape (more curves and angles) in the new fuselage, wings, wheels and even the landing gear itself, he had to alter the geometry over and over again to get it to sit right in both up and down cases. This included the angle of the wheels to the landing gear as well as the angle of the landing gear to the aircraft. Basically the interaction of the landing gear, and even the rotation of the wheels/tires with the ground is noticeably (in detail at least) different.

So it feels different to the experienced pilots of 109s because suddenly the only aircraft in the game, that never required rudder in the take off roll for years and years, needs rudder in the take off roll! That's going to be a bit of a shock after 1000's of hours of 109 flight experience with slam-throttle-no-rudder-off-ya-go under your belt. I didn't think of that consequence until much later, I guess I was too preoccupied with other work.

Just for the record I never did figure out why the old 109 models were the only ones that required no rudder input when taking off. It never really made sense to me but I didn't build those aircraft so I'll probably never know now. Their performance is correct so it was always something I never found time to really dig into. In real life Bf109s ALWAYS required rudder in the take off roll. We did go over the flight performance and retested the E-4 after this higher poly model update because, even though the flight data was not changed, you can never be too sure; and the whole ground handling thing had been sort of forgotten since the higher poly model creation experience with the landing gear goes back to the E-1 creation and not the E-4 update. The flight data  checked out as expected so what we ended up with is a better looking Bf109E-4 (and a new paint job) that takes off more like a real 109 would, and much more realistically than the old lower poly model did.