Resonance is supposed to be where the reactance is 0 and the impedance is minimum, anti-resonance is supposed to be where reactance is zero and impedance is maximum. Due to imperfections zero phase and minimum/maximum resistance might not exactly overlap.
N7PJ was suggesting driving it at anti-resonance.
Bizarrely my guesstimates at the equivalent circuit (100 mOhm Rs, 2 nF Cs, 10 nF Cp and whatever L is necessary for a given resonance frequency, 276 nH for 6.78MHz and 300 nH for 6.5 MHz) produce a near perfect 50 Ohm resistance at anti-resonance ... no matching needed whatsoever. Also dissipating the exact same power as the pi-match at resonance.