This is a long-standing debate and I don't thing that it makes sense to dive too deep into that in this thread. I'll just say that I don't agree with that and point you to Bruce Eckel's essay Strong Typing vs. Strong Testing[1]. Not to mention that there are numerous large (not huge) systems written in dynamic languages that prove otherwise.

https://docs.google.com/View?id=dcsvntt2_25wpjvbbhk