If you (re)match for the same applied power, the gain by having lower
loop resistance is 20log(R1/R2).  Your data indicates a gain of +0.7 dB
from resistance ratio.

I think it should be 10log(R1/R2) :

Assume you put 100W into 100 Ohm, this gives a current of 1A. If you reduce
the loss to 50 Ohm (and still put 100W into it) the current rises to 1.42A,
not 2A !

So reducing R from 25 to 22 Ohm is (theoretically) only 0.56dB gain, but I
am willing to believe Christer if he noticed a higher gain.
I had a similar experience when I added and 'elevated' loading coil to my
LF antenna a few months ago. Calculations (and simulations on PC) predicted
a gain of about 0.3dB, but reports and field measurements gave about 5dB.

So calculations are OK (and often very helpfull), but never rely blindly on
them. Or as they say in German : 'probieren geht über studieren'

73, Rik  ON7YD