I don't quite believe that a slick tyre grips better in every possible on-road scenario. Point number one, if I quote Brandt:
"Tread patterns have no effect on surfaces in which they leave no impression."
Okay, I'll go with that. Now lets look at a picture. Try to see the relevance of the above statement here:
That's a bog standard rack and pinion arrangement. The rack itself is obviously a hard solid surface and we can think of the pinion as a wheel/tyre. I think we could all agree that the specially shaped pinion will provide maximum 'grip' on the rack whereas a smooth disc would be useless. But, the pinion is as hard if not harder than the rack, and it leaves no impression on it...clearly in this case the 'tread' of the pinion is far superior to a slick gear.
So what's the relevance to bike tyres? Well, when you get down to it, no surface is perfectly smooth so you might consider rough asphalt to be a bit like that rack on the very small scale. In which case it's possible that a very fine tread might just help out a bit. However irrelevant you might think this example is it proves the 'no effect on harder surfaces' rule isn't a universal law.
Out of theory-land and into reality and I'll quote from one of my favourite articles on tyres:
"The importance of tread pattern is no surprise to the off-road world but common wisdom says it's a non-factor on the road, where slick treads are assumed to deliver the greatest surface contact with the ground and thus, the best grip. However, asphalt is far from a perfect – or even consistent – material. Certain tread designs can provide a measureable mechanical adhesion to the ground."
So here's a source saying basically the opposite to what Brandt and Sheldon assert. I don't doubt that on very, very smooth surfaces like a racetrack a slick is best. But, now think about the variety of surfaces that all come under 'on-road', I'm especially picturing rough-but-hard asphalt that even to the eye is quite uneven. Again 'common wisdom' has it that narrow tyres roll faster, which makes sense, but then when you get round to measuring it properly the opposite is true.
To be clear I'm not about to swap out the slicks on my road bike for any form of knobbies but this is the sort of thing that interests me. I'd believe that in 99% of everyday riding scenarios a slick is best but I won't take it as gospel.