[seqfan] Finitude of A161620

[Hans Havermann]

Let f (A060797) be the floor of the square root of primorial-k (k#,  
A002110). The sequence f*(f+1)-k# for k=1,2,3,.. begins  
{0,0,0,0,42,72,0,420,6162,15560 
,-272370,1743902,14074002,77960070 
,-571197768 
,-569848020,32849821160,-238745336670,247830905532,-9203096199960,..}.  
The zeros place primorial numbers in A161620 which, as far as I am  
aware, has not been proven finite. Notwithstanding the start of this  
sequence and the subsequent sign-changes, the growth of the absolute  
values of its terms seems to me quite regular and seemingly  
inexorable. Couldn't this be shaped into a proof of a finite number of  
zeros?