1: \begin{abstract}

2:   Sequence transformations are valuable numerical tools that have been

3:   used with considerable success for the acceleration of convergence and

4:   the summation of diverging series.  However, our understanding of their

5:   theoretical properties is far from satisfactory. The Euler series

6:   $\mathcal{E}(z) \sim \sum_{n=0}^{\infty} (-1)^{n} n! z^{n}$ is a very

7:   important model for the ubiquitous factorially divergent perturbation

8:   expansions in physics. In this article, we analyze the summation of the

9:   Euler series by Pad\'{e} approximants and the delta transformation

10:   [E. J. Weniger, Comput. Phys. Rep. \textbf{10}, 189 (1989), Eq.\

11:   (8.4-4)] which is a powerful nonlinear Levin-type transformation that

12:   works very well in the case of strictly alternating convergent or

13:   divergent series.  Our analysis is based on a new factorial series

14:   representation of the truncation error of the Euler series~[R. Borghi,

15:   Appl. Num. Math. \textbf{60}, 1242 (2010)]. We derive explicit

16:   expressions for the transformation errors of Pad\'{e} approximants and

17:   of the delta transformation.  A subsequent asymptotic analysis proves

18:   \emph{rigorously} the convergence of both Pad\'e and delta. Our

19:   asymptotic estimates clearly show the superiority of the delta

20:   transformation over Pad\'{e}. This is in agreement with previous

21:   numerical results.

22: \end{abstract}

23: