It wouldn't be a stretch to say that Football Outsiders was created for the express purpose of showing that the running game is dependent on winning more than winning is dependent on the running game. Indeed, the first article that Aaron Schatz went live with on our site was titled "The Establishment Clause" and used NFL rushing statistics to make that very argument.
The article was a statement about the general tendency of teams to run with a lead at the end of a game than to run in pursuit of a lead at the beginning of a game. However, it made no claims about specific teams outside of using anecdotal evidence to reinforce this point. Furthermore, it didn't say anything about a given team's tendency to use a specific running back based on whether it is ahead or behind in a game.
In fantasy football, this latter consideration (i.e., running back usage patterns) is of particular interest because one of the biggest letdowns for an owner is to see his No. 1 scorer disappear from the game plan by virtue of that back's team being on the wrong end of a blowout. Adrian Peterson owners should know this feeling well, as that's exactly what they experienced this week while watching "Monday Night Football."
Therefore, suffice it to say that it's useful to know which running backs are more reliant than others on whether their team is ahead or behind in the game. Ideally, you'd want to have backs that are involved in the offense an equal amount regardless. However, as we'll see, those types of backs are pretty rare. Alternatively, for the vast majority of running backs, you'll want to identify their lead-dependent usage patterns and then compare that to how you think the game is going to go, given the matchup between teams.
To discern lead-dependent usage patterns for fantasy running backs, I went into Football Outsiders' play-by-play database and looked at how many times each running back was either given the ball on a running play or was the intended target of a pass in each of our "score gaps." I then added up how many total times his team was in that particular score-gap situation during the games in which he played. Finally, I divided the running back-related number by the team-related number and came up with a rate that tells you how involved each running back has been in his team's offense depending on whether his team was ahead or behind (and by how much).
The four score-gap situations are (1) "winning big," which is a lead of eight or more points; (2) "winning small," which is a lead of one to seven points; (3) "tied or losing small," which is anything from a tie score to a seven-point deficit; and (4) "losing big," which is a deficit of eight or more points.
Below is a table showing the score-related usage rates for the top 24 fantasy running backs this season, excluding Michael Bush, Ben Tate, DeMarco Murray and Willis McGahee. The first three are backups who have excelled when given the opportunity, and although Murray will almost certainly remain the starter in Dallas after Felix Jones returns from injury, it's unclear exactly how the touches will be split between the two. McGahee is omitted because his play participation has been a huge outlier given that he started as a backup, missed two full games to injury and left early in two others.