A counterexample of the entropy of a skew product (Q1279727)

The classical formula of \textit{L. M. Abramov} and \textit{V. A. Rokhlin} [Vestn. Leningr. Univ. 17, No. 7 (Ser. Mat. Mekh. Astron. No. 2), 5 -13 (1962; Zbl 0156.06102)] expresses the entropy of a skew-product transformation as the sum of the base entropy and a conditional fibre entropy. This was extended to amenable group actions by \textit{T. Ward} and \textit{Q. Zhang} [Monatsh. Math. 114, No. 3-4, 317-329 (1992; Zbl 0764.28014)]. Here a slightly different question is addressed. If \(T\) is a single transformation on \(X\), \(\phi:X\to G\) is a measurable map into a countable group, and \(F\) a \(G\)-action on \(Y\), then \(\widehat{T}(x,y)=(Tx,F_{\phi(x)}y)\) is a single transformation on \(X\times Y\). What is the entropy of \(\widehat{T}\)? For the case \(G={\mathbb Z}^2\), there is a notion of ``directional'' entropy, and when the fibre action is of a certain form, the entropy of \(\widehat{T}\) is the sum of the base entropy and this directional fibre entropy. In this paper it is shown that this cannot hold in general, by building a fibre action of \({\mathbb Z}^2\) with prescribed directional entropy behaviour.