1: \begin{abstract}

2: We give a unified treatment of the convergence of random series and the

3: rate of convergence of strong law of large numbers in the framework of

4: game-theoretic probability of Shafer and Vovk \cite{ShaVov01}.  We consider games

5: with the quadratic hedge as well as more general weaker hedges.  The latter

6: corresponds to existence of an absolute moment of order smaller than two 

7: in the measure-theoretic framework. 

8: We prove some precise relations between the convergence of centered random

9: series and the convergence of the series of prices of the hedges.

10: When interpreted in measure-theoretic framework, these results characterize

11: convergence of a martingale in terms of convergence of the series of

12: conditional absolute moments.

13: In order to prove these results we 

14: derive some fundamental results on deterministic strategies of Reality,

15: who is a player in a protocol of game-theoretic probability.  

16: It is of particular interest, since  Reality's strategies

17: do not have any counterparts in measure-theoretic framework, ant yet

18: they can be used to prove results which can be interpreted in measure-theoretic

19: framework.

20: \end{abstract}

21: