I kind of think that the discussion about whether or not the 3% should be added and how you would look like a dufus if you needed the 3% but didn't allow for it is a discussion where people are not seeing the wood for the trees.
There are so many variables that go into operating off limiting strips, runway length and TODR are such a small component.
You could allow the 3% and then lose all of that 3% and much more as you slowly advance the power levers in order to stop the prop governor from hitting the stops.
You could be taking off from a country runway with patches of water on the runway and as you run through the puddles you compromise the acceleration of the machine - you are effectively operating off a contanimated runway.
You could be using standard weights to determine the weight of the aeroplane and be well over the weight you think you are. Not a biggie in a jet but in terms of scale a much bigger issue in an aeroplane that weighs in the 5700 kg mark.
You could have a cargo pod on the aircraft which hasn't been properly accounted for in the performance calculations or operate it into icing conditions with the pod fitted. Many operators in days gone by had cargo pods fitted which if you read the STC said do not operate in icing conditions, which everyone to a man conveniently disregarded.
And none of that even starts to consider the fact that in those types of operations there is generally little to no consideration given to terrain clearance procedures in the event of an engine failure i.e. Pre planned escape procedures nor do the performance figures derived from the book consider obstacles so while the aeroplane might be able to get off the runway if anything goes wrong you are in no mans land.
I appreciate that there is a difference between transport category and part 23 certified aeroplanes, but in my view focusing on whether or not a 3% reduction in TODR is required due to a particle separator is not looking at all the components of the equation.
Having trained a lot of pilots on a number of turboprop and Jet types, I often saw people totally invalidating the performance calculations with line up and power application technique.
The biggest problem is that they very rarely, if ever, saw the aeroplane operate at performance limited weights, and when they did it was with both engines operating so it wasn't a true indication of how the machine would perform in the failure case.
In my experience people are horrified when they sit down and calculate how many track miles it would take to get to a 3 or 4000 ft LSA when the aircraft was performing at its minimum certified performance.