Magic The Gathering has had it's own share of mistakes, some almost killed the game at a certain point and they worry for the competitive scene more than anyone else.
Magic could actually be a good source of insight. Unlike Game Freak, which has a very opaque development system and almost never provides information about its creative process, Wizards of the Coast is very transparent. Individual designers and developers have blogs where they provide tons of anecdotes about where cards come from. They even admit when decisions have to do with sales, like sometimes making more powerful cards rarer than weaker ones.
Pokemon, like Magic, has much higher sales among casual players than among competitive ones. As several people in this thread have referenced, Generation IV sold around 15 million copies worldwide, while the number of those who have ever visited Smogon numbers in the thousands. However, most Wizards employees really love their game and enjoy the competitive aspect of it. They sometimes produce cards aimed directly at competitive play, even as counters for specific tournament decks, because they want to make the game more fun. It costs them nothing as long as they make the game fun for children, and it can make a big difference to serious players. It would not surprise me at all if, for example, Infernape developers specifically discussed disrupting Skarmbliss.
In short, the first priority is making the game fun for the middle schoolers who keep it on the map. But I would bet that several of Game Freak's 70 or so employees lurk regularly on these forums. Some may have even tried Shoddy (note that Wizards of the Coast Creative Director Randy Buehler had a famous run, while holding that job, on the Vintage Proxy tournament scene, a Magic sub-subculture with a similar "under the radar" outlaw-factor to Shoddy). Like Magic R&D at Wizards, GF is a small developer with just a few eyes on each job, and those eyes belong to real people who are aware of what a cool job they have. They think about this stuff a lot.
If Game Freak were truly interested in making a balanced game then each new Pokemon made would be truly viable for competitive play.
Somewhat counterintuitively, this is not actually true. Neither are above anecdotes about Regigigas being bad as a counterexample of concern over competitive play, etc. Actually, numerous ludological studies show that competitive play is more enjoyable when there are some "bad" alternatives which must be discarded. In non-customizable games like chess and Go, this comes in the form of extremely large decision trees, much of which consist of very bad options (like placing yourself in checkmate). In customizable games like Magic and Pokemon, this comes at least partially from the existence of good and bad options in the customization process.
For example, suppose we gave Arceus the moveset of Smeargle, with no species clause. Every Pokemon is equally viable, and the metagame would have the option to be completely diverse. Wizards has published data (and no doubt Game Freak has acquired similar insight) suggesting that people would enjoy this much less.
Furthermore, analysis of professional-level Magic tournaments (which are far more competitive than Pokemon, given the higher stakes) has shown that consistently successful players do the best in moderately diverse metagames, while minimally diverse (only one or two viable decks) and completely diverse (extremely large number of viable decks) have more "random" results relative to rating, record, etc. In other words, for Pokemon, the best metagame, both for fun and for competition, is one in which there are a large number of Pokemon, many but not all of which are competitively viable. In other words, exactly what we have. I find it impossible to believe that this is a coincidence.