I've used both time-based and value-based and my answer is an unequivocal "It depends on the project." I'm a pretty creative person and I hate to box myself in by saying "Never do it this way", especially when some very smart people have advocated both/multiple ways. I suggest taking a risk-analysis approach towards estimating as well whereupon you figure out what the mean time to complete, best, and worst times. This can give you confidence (after you've done it awhile) in your guesstimates and that confidence allows you to move more into the value-based pricing. I definitely wouldn't suggest value-based pricing for someone who had not had loads of time-based pricing experience.