Comparing Bennequin-type inequalities (Q2223065)

Suppose \(K\) is a topological knot type in \(S^3\). Let \(SL(K)\) be the maximal value of the self-linking numbers of transverse knot representatives and call it the maximal self-linking number of \(K\). \textit{L. Rudolph} [Bull. Am. Math. Soc., New Ser. 29, No. 1, 51--59 (1993; Zbl 0789.57004)] proved the slice-Bennequin inequality: \(SL(K)\leq 2g_{4}(K)-1\), where \(g_{4}(K)\) is the four-ball genus of \(K\). \textit{O. Plamenevskaya} [Math. Res. Lett. 13, No. 4, 571--586 (2006; Zbl 1143.57006)], \textit{A. N. Shumakovitch} [J. Knot Theory Ramifications 16, No. 10, 1403--1412 (2007; Zbl 1148.57011)], and \textit{T. Kawamura} [Topology 46, No. 1, 29--38 (2007; Zbl 1114.57010)] proved the \(s\)-Bennequin inequality: \(SL(K)\leq s(K)-1\), where \(s(K)\) is the Rasmussen \(s\) invariant and \(s(K)\leq 2g_{4}(K)\). \textit{O. Plamenevskaya} [Algebr. Geom. Topol. 4, 399--406 (2004; Zbl 1070.57014)] proved the \(\tau\)-Bennequin inequality: \(SL(K)\leq 2\tau(K)-1\), where \(\tau(K)\) is a concordance invariant introduced by Ozsváth-Szabó and \(\tau(K)\leq g_{4}(K)\). In the paper under review, the authors exhibit a family of knots \(K_n\), where \(n=1,2,\dots\), such that the defect of the slice-Bennequin inequality \(\frac{1}{2}(2g_{4}(K)-1-SL(K))\) is \(2n\), whereas both the defect of the \(s\)-Bennequin inequality \(\frac{1}{2}(s(K)-1-SL(K))\) and the defect of the \(\tau\)-Bennequin inequality \(\frac{1}{2}(2\tau(K)-1-SL(K))\) are \(0\). On the other hand, the authors prove that all these three defects are \(0\) for a quasi-positive knot.