I tried graphing the data. A COP of 1.22 seems a bit too far-fetched to me.
I don't see why it is far fetched. Look at the power level of 75 W.
The temperature difference is 25 °C (K) thus by the stefan-boltzmann-radiation law.in the case of the fueled reactor
we have(450°C-25°C)^4 =3.2625e+010
In case of the calibration run we have(425°C-25°C)^4 = 2.5600e+010
thus we have a ratio of 3.2625e+010 / 2.5600e+010 = 1.27 corresponding to a COP that is even a little bit higher.
Of course certain effects have to be taken into account, c. f. the post above by GlowFish,
but I do not agree with him that these effects local temperature change etc.
can be enough to explain a difference of 25°C.
IF your TC is really unbiased.
Thus I think probably we have a real COP between 1.15 and 1.2 or so.