It's not that IP aren't relevant, it's that you seem to apply them as a leading indicator but then throw out things like "Well Perez is actually a #2 starter he just has had some bad games" or "Pitcher X has changed his delivery and that is why suddenly he's pitching 2 runs better on his ERA".
Learning about small sample sizes and metrics like swing and miss rates and hard hit ball % are going to tell you a lot more about what a pitcher is going to do in the future than IP.
Learning about small sample sizes and metrics like swing and miss rates and hard hit ball % are going to tell you a lot more about what a pitcher is going to do in the future than IP.