On the average congruence class bias for cyclicity and divisibility of the groups of \(\mathbb{F}_p\)-points of elliptic curves (Q6950331)

The problem of studying the behaviour of an elliptic curve defined over the rational field when reduced modulo an integer prime has a long and fascinating history. It features very deep and important results in the area, but also a great many of open problems and conjectures still waiting to be solved.\N\NIn 2009 \textit{W.D. Banks} and \textit{I.E. Shparlinski} [Isr. J. Math. 173, 253--277 (2009; Zbl 1250.11064)] studied two of these problems: the cyclicity (that is, when the reduced curve is a cyclic group) and the \(m\)-divisibility (that is, when the number of points of the reduced curve is divisible by a previously fixed integer \(m\)). Their study was focused on prime density: they fixed an upper bound for the coefficients of the curve and estimated the number of primes \(\leq x\) for which the property in question held. The results involved some constants and the \(\operatorname{Li}(x)\) function (used a a proxy for the prime-counting function).\N\NIn the paper under review, the author adds a congruence condition on the primes (essentially forcing them to lie in a certain arithmetic progression), which enables him to add new insights. For example, it renders explicit formulas for the relevant constants and it allows the interesting study of the so-called average congruence class bias, which measures how the said congruence condition affects the constants as it varies.\N\NThe author goes to some lengths in order to make the paper essentially self-contained and easily readable and the article benefits greatly from that.