Well I sorted it anyway.
For anyone else wondering, it is possible to have a great idle at ~760 RPM (and part throttle driveability) with aftermarket cams (Badger5), 1000cc injectors (running at 6BAR FPR, equivalent to approx 1500cc), 4" MAF housing with RS4 MAF sensor, I.E. Intake manifold AND 80mm hemi throttle.
Attached is the log with o2 correction turned off (for tuning injectors), super stable and the log itself is the same for over 10 minute period.
I did have to really tune down warm up enrichment by factor of 4 (not sure if much higher fuel pressure puddles less?) and same for deceleration enleanment plus the injector dead times have to be perfect.

In the end I studied the FR in more detail and only changed the PID values for the Air path (which are DMLLRLMNN. DMLLRLMXN, PVLDNEZV, PVLDNEZ, also there is a D air path but on my file they were all set to 0 anyway), I divided those by 6 as on low throttle the throttle angle vs airflow values were about a 6th of the stock values.
The issue above was caused by the KFZWOP map, as the ECU was decreasing the load via the throttle, the KWFZOP map had decreasing angle values. This caused the spark angle to decrease, which made the RPM drop, so the ECU would re-open the throttle to compensate.
To fix it I increased the values in KFZWOP map at really low load/RPM values. This meant the the ECU could keep the spark angle the same as load decreased, not causing a drop in RPM.
Most likely KFZWOP  should be re-calibrated fully with the new cams but as I dont have access to a dyno I will just have to keep it stock in the higher load areas.

Next issue to solve is the drop in RPM when releasing the throttle, it drops way below the idle RPM before stabilizing...

Ok thanks for the feedback.
I have played further with KFMSNWDK/KFWDKMSN at below 1.5 degree angle and now my KFMSNWDK/KFWDKMSN are spot on (although they very pretty good before anyway).
I also increased torque reserve as you suggested and the ignition angle has decreased to 10* on average on idle. As expected this resulted in throttle plate angle of 1.57* (before it was 1.17 with the higher ignition angle). Although this may not be strictly necessary in my case, as I removed A/C and replaced the power steering pump with an electro hydraulic one (so I guess it still does have to compensate for the extra alternator load when turning steering wheel on idle).
Also I think my original issue was partly caused by KFWZOP map, I increased KFZW to compensate for the aftermarket cams but I did not increase KFWZOP. While this was OK on warm idle, on cold the the timing retardation map was interfering and causing some strange oscillations loops. Once I upped KFWZOP, the issues went away.

Anyway, I am still fighting the issue where the revs undershoot on throttle release (even with the increased torque reserve). The controller is closing the throttle to 0.8* but at this angle there is no airflow (because I do not have PCV valves EVAP etc). I have set all the additional airflow values to 0 in the map (e.g. MSNDKO) and I have set KFMSNWDK/KFWDKMSN to 0 for all opening below 1*. So I do not understand why it still closes it below 1.6*. How can I stop this from happening?? I have tried to play with KFMIOP/IRL but this causes some other unwanted behaviour when pulling away.

I am in the exact same problem. I have the RS4 MAF on my S3 1.8t, scaled it, but got the same oscillations. I already whooped around with KFWDKMSN which already made it better already. I found out in another thread that KFMSNWDK also needs some modifications so I already did that but I need to test it right now.
My car is also pretty modified with performance cams, dahlback intake etc. etc.

I also increased (advanced) ignition timing at idle because I was seeing -12 atdc most of the time, but I would never pull it more than 12-14 for idle. (also I dont care much if its running a bit rough at idle because of the cam overlap timing)

Thanks for pointing Prj. I just checked my log again and as you are saying rl_w > rlsol_w.

FWIW here 2 logs from where I begun today and where I already am, I noticed when modifying KFWDKMSN it already stabilized a bit more but still has the drop to 600 rpm and the bump to 1600 rpm once in a 2-3 secs instead of going up and down from 600 to 1500 every second, I will calibrate like Prj just adviced and will give an update here when I am on it again.

You are changing things without having any clue how they work.
KFWDKMSN and the inverse are not "tuneable" they are the flow characteristics of the throttle. You set them to what the manufacturer specifies and you're done.
They also have absolutely nothing to do with the MAF. It is a totally different flow model, in no way connected.

I had the same cyclical oscillations but only on warmup. To do a quick check, I swapped injectors to the previous ones I had. My problem went away which Nyet proposed was wall film model related....There are a lot of components to ESUK and I've been reading it for a while.

Perhaps you are dealing with issues of wall wetting as well.

FWIW here my log of the last adjustment.