Excluding home-runs, one can not drive in runners that aren't on-base, and cannot drive himself in from the bases. A player that tends to get on base will most likely end up scoring more runs, but has no power over whether he gets driven in or not, unless of course the runner advances via speed, which very few players regularly do. The control of the runner's fate is practically only in the hands of one of the hitters that comes after you. I just don't understand how fans can measure a player's offensive production through RBI's/Runs, when those statistics are actually a judgement of a collection of consecutive hitters in a lineup. The only way I'd even care to look at RBI's is if every hitter had the same amount of chances with RISP. Not giving them that constant, is basically measuring to see which plant out of two grows taller, when one plant is getting more sunlight/water. Even if each hitter were given the same amount of chances, I'd still be hesitant to put any value into it, considering how much production with RISP fluctuates. I'll admit, there are cases of hitters who were consistently good at hitting with RISP. But, most of the time, those hitters were already good to start with. There are too many times when a hitter hits amazing with RISP one year, and is simply an average clutch hitter the next year. Take Brandon Phillips, who is essentially having a bad year (.266/.314/.412), as an example. He is hitting .360/.416/.511 with RISP, and has 95 RBI's through 120 games. But in 2010, he hit .246/.364/.381 with RISP, and only contributed 59 RBI's through 155 games. How is it that, although 2010 Brandon Phillips and 2013 Brandon Phillips had very similar years, but had such different numbers in clutch situations? More often than not, small sample sizes(RISP, playoff hitting), that are able to collect sufficient data, will even out to a player's actually talent level. You say Brian Roberts is good at getting RBI's, when really he is only a slightly better hitter in clutch situations than he is in all situations, as is Matt Wieters, Chris Davis, Nate McLouth, and Nick Markakis. OBP is a better indicator than runs, because it only measures the individual's performance. Why should a hitter get credit for scoring one time he walked, and not get credit for not scoring one time he walked? Why should Chris Davis not get credit for something Adam Jones failed to do?
If a team was full of players identically to Brian Roberts, that team would have many old, fragile, slow, one-positioned, defensively challenged players who don't hit for average, power, or get on base.