As ever, all that B'macc says I agree with.
It's maybe worth just adding that with wear on the heel of the opening point and on the camring (or pushrod and camplate on a face-cam thing like a magdyno) you'll get some internal timing 'error' over time (compared to the maker's setting) so tweaking to get a fraction more internal advance isn't necessarily a bad thing sometimes.
Fitting new points or a new or different camring changes things again though. You can only really see the differences on a dynamic rig.
The difference in low speed spark performance between full advance and retard of the ecccentric can be more than you'd think although as B'macc says there aren't many degrees of movement here.
It is often possible, where a K2F won't spark the Lucas spec test gaps at the rpm specified by the makers, to gain significant improvement. Typically, there is an upside of maybe 30 or 40 rpm on the mag between 'late' and 'spot-on' on this screw, which makes a useful difference when trying to start, say, a Velocette with its low-geared kickstart, not to mention the problem with the late spark on the second cylinder of a V twin. That problem is compounded by the use of a manual camring on a V, where you can be asking the second spark to occur at 45° of retard compared to optimum (and it still works, albeit not below maybe 250+ rpm). Only the use of twin mags, or conversion to coil igniton with a distributor, will get round the unsuitability of using a device not designed to work asymmetrically.
The best I have ever seen a K2F do is 'all sparks across 3 point 5.5mm gaps' at 90rpm of the mag. Standard spec is almost all sparks at130rpm, excellent (to me anyway) is about 110rpm, a bit sad is 150rpm (double these numbers for engine speed). In all cases the beast will start, and the mag will make good sparks at the flick of a wrist. (When you think that some mags for commercial vehicles in the vintage period would produce their sparks at THIRTY rpm, it shows how effective these things can be at low speed!)
In the context of minor departures from performance specs, other things are just as likely to be relevant too, however. The magnetism, the state of the cb points, whether there is wiggle on the opening point's pivot, whether there is wear on the opening ramps of the camring (which denies the points a snappy opening) . . . and maybe above all the air gap between armature ad magnet on a several-times rebuilt instrument. All of which can have a worse effect than a minor maladjustment of the eccentric pin.
So yes, a modest tweak of the eccentric screw won't probably have much visible effect . . . but all the same, best to try to set it right and leave it there, I'd say.
