I think to compare cut out by diving behind a 109 is not a good base for comparison as both types will have different dive behaviour, cut out or not.

I think: The g values used for cut out evaluation before patch are correct (all historic scources agree on this). They should be kept

I think the behaviour just needed a bit thweaking - the build-up of the effect should by SLIGHTLY made more inert. A small delay between time of g-factor for cut out and the realisation of cut out and a smooth transition between 0% cut out to Cough condition and between cough condition to 100% cut out would have done the trick to have a good representation of push over cut outs while still being insensitive to turbulences.