RChung

There's a group of relatively low data quality but relatively inexpensive (nothing is inexpensive -- just relatively so) products on the market. A friend who's a coach calls them fine gateway drugs.

Nothing's perfect, and every power meter is trying to fill a different niche, so there's room for lots of alternatives, including gateway drugs. There are also some products that have relatively high data quality but are relatively inexpensive. I put the P2M and the PT hub in that group (I haven't analyzed data from the PT chainring yet so don't know whether it would fall into that group or not). Some people, even racers, will be perfectly happy with the Stages. Some people, even non-racers, won't. Racing category actually isn't a terribly good predictor of what you'll want or need or will be able to use.

To get back to the OP, he's got two relatively low data quality devices, the Stages and a trainer, and there's no simple way for him to know which is worse -- or if the problem is his legs. In his case, the discrepancy is large enough to be obvious. There are cases where the discrepancies may be small most of the time but large under certain circumstances, and not obvious. Lots of people appear to think that if something isn't obvious it can't be important.