I'm inclined to agree with adamors, any time a framework says you can plugin any templating engine, it's almost always more of a hassle than not. If you choose a templating engine that is not a first class citizen of the framework, you're often blocked from key features or functionality offered by the framework. Even if a third party library bridges the feature gap, that library often trails development of the framework or ends up unmaintained and outdated.
I also agree that it's disappointing Twig wasn't selected. It's syntax is incredibly common among templating engines so ramp up time for new developers is usually short - and once you get into the more complex features it's pretty dang cool.