“…The authors show there that (similarly to the extremal case studied in [4]), there exists a polynomial P (z) of degree ≤ m + 2 which is positive on the interval (−1, 1) if and only if (M, J) admits a K-extremal Kähler metric in ζ, and in this case, the extremal Kähler metric is given by the Calabi ansatz. On the other hand, the computations from [6,40] also show that for each z 0 ∈ (−1, 1), P (z 0 ) computes (up to a positive multiple) the relative weighted Donaldson-Futaki invariant of a smooth Kähler test configuration (M , A z 0 ) associated to (M, J, ζ); furthermore, (M , A z 0 ) is a polarized test configuration precisely when z 0 ∈ (−1, 1) ∩ Q. Thus, the weighted K-stability established in Theorem 1 only yields that P (z) must be positive on (−1, 1) ∩ Q should a K-extremal metric exist, whereas Conjecture 5.8 mentioned above would imply the positivity of P (z) everywhere on (−1, 1), i.e., the existence of a K-extremal metric of Calabi type.…”