I wonder whether anyone can formulate a convincing answer to the following
question.

Let us assume you wish to design a 3-element Yagi antenna for good performances,
but with NO REGARD AT ALL TO ITS IMPEDANCE (i.e. assume that you anyway intend
to have a matchbox at the antenna terminals, so that the antenna impedance is of
no concern at all for you).

At this point you run a good antenna design program, giving it no constrain in
terms of antenna impedance. The program will give you the "optimal" solution in
terms of element length and spacing, typically also showing the antenna
impedance at the feedpoint, that we here assume to be e.g. 30 ohm.

At this point you change your mind and, instead of using a 30-to-50 ohm
matchbox, you decide to match the antenna to the cable using the well known
hairpin technique, by which the radiator length is reduced somewhat (to create
some capacitive reactance in series with the antenna resistance) and an inductor
is put in parallel to the antenna (to resonate the residual capacitance after
resistance transformation).

Doing so the radiator length will result to be shorter that the length
calculated by the computer program, and one could then presume that the actual
antenna radiation characteristics (gain & front-to-back performance, radiation
pattern) do not correspond to those calculated by the program any longer.

Even though the changes may be small, I would anyway be interested to understand
whether, in principle, the hairpin technique results in a change of the antenna
radiation characteristics.

If the change is confirmed, we would inevitably also come to the conclusion that
the antenna radiation characteristics depend on the technique which is adopted
to match the antenna to the cable (e.g. matchbox or hairpin). This would sound
real odd!

Thanks and 73

Tony I0JX