1: \begin{abstract}

2: We present strongly convergent explicit and semi-implicit adaptive numerical schemes for systems of stiff stochastic differential equations (SDEs) 

3: where both the drift and diffusion are non-globally

4: Lipschitz continuous. This stiffness may originate either from a linear

5: operator in the drift, or from a perturbation of the nonlinear

6: structures under discretisation, or both.

7: Typical applications arise from the space discretisation of

8: an SPDE, stochastic volatility models in finance, or certain ecological models.

9: We prove that a timetepping strategy that adapts the stepsize based on the drift alone is sufficient to control growth and to obtain strong convergence with polynomial order. %An initial exploration of this class of timestepping strategies was presented in \cite{KeLo2016}, and the current article presents a comprehensive rigorous treatment in a general setting. 

10: %By contrast with recently developed fixed step methods, we do not need to explicitly prove a moment bound on the numerical solution, instead we can combine existing results with the time stepping strategy.

11: The order of strong convergence of our scheme is $(1-\varepsilon)/2$, 

12: for $\varepsilon\in(0,1)$, where $\varepsilon$ becomes arbitrarily small as the number of available finite moments for solutions of the SDE increases. 

13: %\GJL{As well as considering more general SDEs than in \cite{KeLo2016}, requiring new analysis, we also give a rigorous proof, incorporating the random number of time-steps required.} 

14: Numerically, we compare the adaptive semi-implicit method to a fully drift implicit method, three tamed type methods and a truncated method. Our numerical results show that the adaptive semi-implicit method is well suited as a general purpose solver, is more robust than the explicit time stepping methods and more efficient than the drift implicit method.

15: \end{abstract}

16: