There is still a slight knock on on-base percentage in some circles that it is a sabermetric or "Moneyball" phenomenon. (Although "Moneyball" was simply about Oakland's methods of finding market inefficiencies than the specific inefficiencies themselves, but I digress.) On-base is viewed as a statistic for lead-off hitters or a measure primarily of plate discipline and walks.
In some sense, I can understand this sentiment. In a technical sense it is true. Without walks, you are essentially left with batting average instead of on-base percentage. But honestly, when it comes down to it, no one should be concerned about walks versus hits when it comes to measuring something that either on-base percentage or batting average generally portends to measure.
What we want out of a stat like batting average or on-base percentage, fundamentally, is how often a hitter succeeds versus fails. Failure in baseball, for a hitter, is making an out. Success is avoiding an out and getting on base. On-base percentage is the measure we are looking for since it takes into account all manners of getting on base and avoiding outs, namely both hits and walks.
Some may argue that we should pay more attention to batting average because hits are less valuable than walks. A hitter's discipline is secondary to how well he can get hits. To that point, yes, I would rather have a guy arrive at a .400 on-base percentage via a .400 batting average. There are distinct advantages to a hitter getting hits versus walks. With a decent base-runner on second, a hitter can drive in the runner with a hit but not with a walk.
One problem: It's extremely difficult in today's game for a hitter to base-hit his way to a good on-base percentage. The best single-season batting averages of the last 15 years or so are in the .330 to .360 range. In a couple of seasons, players have reached the .370s. Left only to batting average, these players would get on base at okay rates but nothing special. And these are the best single season batting averages. Imagine how hard it is for most hitters, even the ones who post batting averages in the .300 to .320 range, to get on base at good rates.
The skills of defensive players, defensive positioning, glove technology and the advent of hit charts make it more difficult than ever for players to base-hit their way to respectability at the plate. Hitters still must be able to hit the ball around, through and over defenses in order to post respectable batting averages, but they also must draw their share of walks so that they get on base enough to be valuable to their teams.
Back in the 19-teens or 1920s and even into the 1930s and 1940s, many hitters could largely base-hit their way to respectability in the on-base/out-avoidance department. Back then no one had charts they could pull up to tell them where a hitter is most likely to hit the ball. If a fielder had a glove, it was a far cry from today's technology. Plus human beings are just bigger, stronger and faster while the ground that players have to cover hasn't changed much over the years.
So while we should only be concerned about the end result – how often a hitter avoids outs/gets on base – and we should be concerned about whether he arrives there with hits or walks, a player posting a good on-base percentage via hits only is more or less possible only in theory. A player needs some semblance of respectable plate discipline to get on base and avoid outs often enough to be valuable to his team.