'Replication' and 'estimation' are totally different things, in Statistics. Taking the average of first two, just to get third one, as a 'replication' is not correct. Each 'replication' should be independent 'observation'/ experiment.

thank you for your answer

Check out this: http://en.wikipedia.org/wiki/Bootstrapping_%28statistics%29

First commandment of statistics: Thou shalt not make up data.

To act as though you had three replications when, in fact, you only had two would be wrong. A cheat. Bad.

thanks for answers