I always felt to have a good indicator or 'feel' for how the running game or passing game is really working as a practical matter, it was best to take the total yards and attempts and take out the top 10% and bottom 10% and recompute the average so that the 'statistical average' really reflects the average (think - "typical" as a definition). In other words, throw out the outliers and look at the core data. By way of quick example, if we had a set of running plays with gains as follows: -1, 2, 3, 4, 6, 7, 9, 10, 22, and 45, you come up a mathematical average of 10.7 yards per carry on the ten runs. However, if you throw out the bottom 10% being the -1 and 2 yard carries and the top ten % being of course the two long runs, then you come up with a 6.5 yard per carry mathematical average. While both are pretty good averages, the 10.7 arguably is bumped way up by just one single play being a 45 yard gain. Likewise, a big loss of say 19 yards on a QB sack can make otherwise decent running yards by the quarterback's designed runs look poor. Obviously, we can say that with a sample size of just ten carries, each carry can have a substantial effect on the statistical average if it is an outlier (really high or low). Likewise, a team with only a dozen passes per game and 6 completions will have a modest 50% completion rate but if two of those passes are long TDs, the yards per attempt and per catch yards can be very high. The larger the number of attempts (whether runs or passes), the more typical or average will be the statistical averages. To be most meaningful and indicative of the relative success of the running game, a team would need to attempt perhaps 30 or more runs per game and similarly for passes. Osborne's teams often did not pass enough to fairly indicate the strength of the passing game, although most also agreed that NU passing was quite effective in generating yardage and keeping the defense somewhat honest in their schemes.
We all would surely agree that Riley's offense is predominantly a passing attack with significant number of running plays mixed in. Osborne was the opposite by general desription with the bulk being runs and a limited number of passes mixed in. Ultimately, the more effective the passes are for Riley, the better the runs should become as the defense will attempt to adjust and adapt itself to limit the passing attack and become more susceptible to runs. It was vice versa with Osborne's run oriented attack as defneses focused on crowding the line and loading up the box to try to overplay the run. Beck's idea was to keep the defense guessing constantly by sometimes over passing and sometimes over running, seemingly on a randum basis. I believe Beck's plays were often called as randomly as drawing them out of a hat rather than sequenced plays where one or more plays are intended to set up one or more other plays that follow. Random is NOT, in my view, an ideal way to call plays, although it may be more unpredictable than any other way to select plays. One can keep the defense honest by random play calls IF the defense fears your playmaking ability. Often, it seems defenses are not afraid of our offensive power.