It all depends on how good the baulk rings are and how much tolerance either side of the matched speed they allow. I imagine that cars that are very fussy about clutchless gearchange have very effective baulk rings which prevent the synchromesh cones coming into contact until the speeds are very closely matched, and make the characteristic graunching noise if the speeds are not closely matched. That's good for the cones because there will be very little rubbing at contact point. If the baulk rings are less effective, they will allow contact when there is a greater disparity of speed, which potentially puts a greater strain on the rings if you happen not to match perfectly. OK, so you'd feel it as a sudden retardation or acceleration of the car.
On my car I can usually change down OK, because you just increase the throttle gradually until the gear slips in. I find changing up more difficult for some reason, even though in theory it's just the same throttle adjustment in the opposite direction.
Clutchless gearchanges aside, I always try to match my engine and road speeds reasonably well - I've got to know roughly how much the rev counter needle needs to move from its speed for the old gear when changing to a new gear. At the very least, I keep the engine revs constant during the gearchange, and preferably I actually change the engine revs the right amount in the right direction. What I don't do is what some people do: let the engine revs fall to idling, let the clutch up on the idling engine and then reapply power (*). That causes horrendous lurches and must do horrible things to the clutch which has to take the strain of the mismatch. I got a lift with a woman who had been driving a few years longer than me but had never learned about rev-matching. After she'd apologised for the n-th time about her jerky gearchange, I rather diffidently suggested that there might be a "different" (ie "better") way of doing it. She let me demonstrate and I talked her through the process (which surprisingly difficult to analyse when you do it subconsciouly). She was gobsmacked. Goodness knows whether her instructor taught her badly or whether she'd slipped into bad habits afterwards. I suspect the latter, because the gearchanges she was doing would not have got her to pass the test.
It's a skill. No-one is born knowing how to do it. It takes a lot of practice. The miracle is that having acquired the skill, it is transferrable from one car to another and doesn't have to be re-learned to take into account different clutch bite point, different responsiveness of throttle, different spacing of gear ratios etc - it just requires a bit of mental and muscle-memory tweaking of the parameters in the mental algorithm.
(*) The technique that was required for drivers of 1st generation DMU trains, to allow time for several different gearboxes on engines along the train all to change gear by remote (drive-by-wire?) control. I have "fond" memories of the DMUs on the Aylesbury-Marylebone service in the 1970s and
80s which would accelerate hard in first gear after leaving a station, then disengage and let the engine idle for what *seemed* like almost a minute during which time the train not only stops accelerating but actually starts to slow a bit, and then there was a sudden surge of power in the new gear.