How Badly Would a Roethlisberger Suspension Hurt the Steelers?

Over his career, Roethlisberger has averaged 6.7 net yards per attempt (NYPA) and a 3.4% interception rate. In 2009, the league average was 6.2 NYPA and 3.1% int rate. According to my game probability model, all things being equal, Roethlisberger's passing stats would turn an average team into one with a 57% chance of beating an average opponent.

If we replace him with a substitute for a few games, it's hard to predict how the Steelers' passing game would do. But just as an exercise, say their passing game drops by one standard deviation. Their NYPA would be 5.7 and their int rate would be 4.5%. All else equal, this would drop them from a 57% team to a 40% team.

If the suspension, which is expected to be announced later today, were for four games, you could think of it as a 4 * 0.17 = 0.7 win penalty for Pittsburgh.

But there's something unusual about Roethlisberger's replacements over the past several years. Since 2006, Steeler backup QBs have averaged 7.4 NYPA and a 2.1% int rate, a substantial improvement over Roethlisberger's numbers! Charlie Batch, Byron Leftwhich, and Dennis Dixon combined for 1930 yards on 243 attempts. They gave up 5 ints and 54 sack yards. The backup QBs differed from Roethlisberger primarily in int rate and sack rate. Roethlisberger was sacked on 9.1% of his drop backs, while his replacements were sacked at a 4.0% clip.

Assuming a backup, whoever it is, continues to perform so well, he would make a 50% team into a 67% team. The backups were sometimes playing in lower leverage situations, so it's not fair to expect they would continue to perform as well as they did. It's also likely they would regress to the mean, but it's hard to say by how much. Still, these numbers might make it easier for the Steelers to consider putting Roethlisberger on the trading block.

For reference, you can see how the Steeler backups compare to Roethlisberger in terms of Win Probability Added (WPA) and Expected Points Added (EPA) over the past several seasons here.

  • Spread The Love
  • Digg This Post
  • Tweet This Post
  • Stumble This Post
  • Submit This Post To Delicious
  • Submit This Post To Reddit
  • Submit This Post To Mixx

12 Responses to “How Badly Would a Roethlisberger Suspension Hurt the Steelers?”

  1. James says:

    This is off-topic, but I think this community will appreciate it. It's from the Sloan Sports Conference with Mark Cuban, Jonathan Kraft, Daryl Morey, Bill Polian, Bill Simmons, Michael Lewis on Moneyball topics. It's really long, about 80 minutes, but I'm enjoying all of it. 4th-and-2 comes up about 38 minutes in.

  2. James says:

    After reading the article, that's impressive! I know Roethlisberger takes way more sacks than is necessary by holding onto the ball so long, but I didn't know his back-ups had been so effective (particularly the interceptions).

  3. Ian says:

    I often think there's a case for having two QBs take the practice snaps throughout the week, especially if they are two different styles of QB.

    Let's say you had a scrambler and a pocket passer, if the defense didn't know who was playing, how could you effectively scheme for that? Maybe that's why the Steelers backups have done so well comparatively.

  4. Jeff Clarke says:

    Maybe we should relabel this:

    The "Matt Cassel" problem. Cassel barely plays since high school. He gets thrown into the middle of one of the greatest offenses ever. He looks like a hall-of-famer. An incredibly mediocre team signs him to a big contract and he looks pretty damn mediocre himself.

    Its really hard to say how much of a QB's success is him and how much is his supporting cast. I remember reading one study (I think it was at pro-football-reference, it might have been here) that said replacing the starter with the backup will cost a team about 3 points a game on average. I think that is a lot lower than most people would think instinctually, but as the Steelers backups and Matt Cassel have shown, QBs with above average stats usually have above average supporting casts and vice versa.

    It also raises the question of how good is a QB really? Nearly everybody would say that Roethlisberger is a much better QB than Brady Quinn, but what would happen if you put Ben on the Browns and Quinn on the Steelers. There is no real way of knowing.

  5. James says:

    Jeff, although it may seem like "only" 3 points, over the past three years teams have scored on average 22 points a game. A 3-point drop is a 14% decrease, or about two extra losses.

  6. Bob Weber says:


    I would take the opposite approach. Why is it that some of these programs that are really good can't get over the hump until the right QB steps in. The Patriots are a good example. Go back a little further to the Drew Bledsoe days. Bledsoe had good skills, but until Brady, who also hadn't played much, stepped in they couldn't win it all.

    Likewise, the Steelers struggled for years until Rothlisberger joined the team. Kordell Stewart and Tommy Maddox couldn't get it done, but the rookie Rothlisberger had a huge impact.

    Batch and Leftwhich have both been starters in the league, the stats say their numbers are better than Rothlisberger's, so why aren't they someplace winning championships?

  7. Brian Burke says:

    I don't think the backups are actually any better. I think the recent success of the Steelers is due primarily to its 'outlier' defense and a slightly above average passing attack. Roethlisberger, however, is not an indispensable part of that attack. I'm sure he's very good, but the line and receivers and coaching are also going to help any QB be fairly successful on that team.

  8. Tim says:

    Backup quarterbacks on the Steelers likely ran more run oriented offenses, or at least were expected by opposing defenses to call a higher percentage of running plays than when Ben Roethlisberger is the quarterback. The expectation (and then the resulting stats) from the passing plays will be higher when playing against a defense planning for running plays.

  9. Anonymous says:

    It's seen in all sports that when a back-up is in, teams often rally around the back up. Is there a metric for this, for all sports, absolutley not. But, think back to how often in hockey when a goalie goes down the team rallies and plays tighter defense. In baseball you see terrible pitchers get more run support than the best pitchers.

    The psychology takes effect as well.

  10. Anonymous says:

    [citation needed]

  11. Brian Burke says:

    I forgot all about this great post from Doug Drinen a couple years ago. He found that replacing a starter with a backup on average hurts a team by between 2 and 3 points per game. This is equivalent to about 1 win per season.

    In the following post, Doug found that replacing a starter with a backup hurts a team's chances of winning a game by an average of 6%. Again, over 16 games, this costs an average of 1 win.

    My estimate of .44 wins for a quarter season may have been too harsh. That would mean about 1.3 wins for an entire season.

  12. Gary Porpora says:

    A stat analysis tells only one side of the story and often doesn't tell enough of it.

    In the last decade--probably the last 40 years--the Steelers have had the best defense in nearly every statistical category.

    ANY back-up would perform better for the Steelers because they have a comfort edge back-ups for other teams don 't. Steeler QBs know they don't have to be Peyton Manning, just be ...well, Byron Leftwich or Charlie Batch--take care of the ball, get a touchdown, a couple field goals and let defense win the game.

    Also, the term "franchise quarterback" is so overused it has no real meaning. I wrote a column on that topic for

Leave a Reply

Note: Only a member of this blog may post a comment.