It's hard to argue with that. I have a feeling that if we only use a stationary dyno (like the only kind we have!), we'll miss something (e.g., the dyno shows negligible gains and we conclude it's not worth it). Consider what the wide variety of racing genres do here. You'll see all kinds of CAI setups in almost every form of racing. That suggests they value this setup, maybe even if the dyno doesn't tell the whole story. Here's something else to consider. Before we had easy access to dynos (mid-to-late 90s), we used the factory OBD II tool from Nissan to measure MAF voltage gains; it was a reliable way to see power on the move. We'd get this rig going along with our Horiba diagnostics, bung up an O2 sensor and go for a ride! We were looking for voltage gains and near-enough stoichiometric AFR. We made some fast cars like that. While the company is a shadow of its former self, JWT and Clark Steppler were amazing tuners for all things Nissan. They were so good that they tuned wildly different makes. One famous Honda drag race car with a Nissan ECU (16 bit setup and converted to MAF) comes to mind here.