Breakpoint for P&G in N vs. DFCO

[8$PG]

I've been trying to figure out when to use N or just let the electronics cut of the fuel in gear. I realize this will depend on a number of things varying between cars and conditions. I just want to know how you think about this.
Up until now I have always used the dfco out of lazyness I guess or fear of damaging the transmisson (which I dont think it will done right since the transmission is already in the right gear).

One thing about coasting in N is that you maybe need to match the rpm on the engine to the speed. For a short second you are injecting gas in the engine without any real gain in driving the vehicle forward, except saving the transmission. This means there is no point in doing this for short periods of time. A longer slope is needed. And the slope should not be too steep so that you are gaining speed down the hill, possibly getting more wind resistance or speed that makes the planning ahead more difficult.

I believe it also depends on the revs of the enginge at that speed you are traveling at before you let your foot of the throttle. At 100km/h the revs im at is about 2000rpm. The engine is a 2.0L s? the engine brake is not that strong. I do alot of city dfco-coasting at low speeds. Sure coasting in N makes a substantial difference but if the slope is steep enough for the speed to be fairly constant the opportunity to give a bit more throttle using the engine more effectively reducing pump-losses when going up the next hill might be worth it.

At low speeds <56km/h the transmission will downshift to 4:th at about 900rpm. If I'm going just slightly downhill at the same time the speed would be constant and I'm using no fuel at all. Then when going up the consumption is maybe 8-12l/100km. I think this would be better in average than consuming 7L/100km at a constant rate, varying the speed instead or coasting i N on the way down consuming 1.8-3L/km.

Same thing when driving on a winding road with lots of turns and straights. Giving some extra throttle accelerating on the first half of the straights and then decelerating into the turns might give som extra efficiency compared to just going in a straight line at constant speed.

How would you go about testing this? I have a scangauge connected to the car now. I also do alot of flat highway commuting at constant speed but with not so much traffic (mostly lorrys), I'm a bit curious about how much I could gain by skipping the light trottle going more pulse-or-nothing.

Sorry for the long text.