Bad places for the approximation property for finite groups (Q2155603)

Let \(k\) be a number field and \(G\) be a finite group with a continuous action of the profinite group \(\mathrm{Gal}(k):=\mathrm{Gal}(k/k)\) which is compatible with the group structure of \(G\). Such a group is called a finite \(k\)-group. A finite \(k\)-group \(G\) has approximation in a finite set of places \(\Sigma\subset\Omega_{k}\) if the restriction map \[ H^1(k,G)\to\prod_{v\in\Sigma}H^1(k_{v},G) \] is surjective. So, a finite \(k\)-group has the approximation property if it has approximation in every finite set of places of \(k\). This property is in general too strong and it is not satisfied, even for abelian groups. The author of the paper under review considers a weaker property. Let \(\mathrm{Bad}_{G}\) be the finite set of places that either divides the order of \(G\) or ramifies in the minimal extension splitting \(G\). The Tame Approximation Problem for \(G\) asks whether the restriction map \(H^1(k,G)\to\prod_{v\in\Sigma}H^1(k_{v},G)\) is surjective for every finite set of places \(\Sigma\subset\Omega_{k}\) disjoint from \(\mathrm{Bad}_{G}\). The author of the paper under review proves that there are finite abelian \(k\)-groups \(A\) where the map \(H^1(k,A)\to\prod_{v\in\Sigma_0}H^1(k_{v},A)\) is not surjective in a set \(\Sigma_0\subset\mathrm{Bad}_{A}\), which is the set of places that do not divide the order of \(A\) and ramify in the minimal extension splitting \(A\).