Apart from the argument of Wicket\'s simplicity (that is, Wicket is a simpler system IMHO) and GWT\'s responsiveness in the client (GWT\'s client side state and JavaScript -
It's somewhat not fair to compare GWT with wicket (or likewise) since they are really coming from 2 different camps. The former is a framework for building JavaScript front-end applications while the latter is a classic Java web application framework.
So the points below are not as much as GWT vs. wicket but rather general list that was compiled when we decided to use GWT for advanced JavaScript/AJAX web application:
I can think of several reasons why GWT is a better choice than Wicket, for typical business web apps:
I've been involved in a GWT based project for the past few months. I was, having been part of the Wicket development team for years, looking forward to a change, and expected a lot from GWT (which I've always touted as another great Java framework).
Honestly, I am disappointed when it comes to working with GWT. I feel - my whole team in fact - that productivity took a big hit. Theoretically GWT is great. But when you factor in the quirks and limitations of the framework, mediocre error reporting (particularly when it comes to serialization errors), the long compile times (anywhere between 3 - 10 minutes, and our project isn't even that big yet), the fact that when all is said and done, you still need to test for all browsers and find tweaks and workarounds, the fact that it produces a huge initial download (almost an MB, which we're cutting back gradually, but with a lot of effort), etc, etc, I feel Wicket is much easier and quicker to work with.
I don't hate working with GWT. It's still a lot better than most Java frameworks. It's just that I expected a lot more from working with it; I even expected it to be possibly nicer than Wicket. But in the end, it is just not imho. Hopefully GWT 2.0 will improve a lot of things, and hopefully some of the quirks of the Eclipse plugin will be straightened out soon too.