Thursday, January 3, 2013

Bad Quarterback League: Scoring Analysis

Despite our league's earlier tweaks to the BQBL draft and weekly matchup structure, we used the Grantland standard scoring rules, mostly because Grantland's weekly posts tabulate the score for us--plus, none of us wanted to go through the exercise of coming up with an alternate scoring system, and we wouldn't know how to "fix" it even if the balance were off. Throughout this season, though, a few imbalances in the scores became especially clear.

1. Too much difference between "interception" and "interception return for TD"

Maybe a receiver runs a wrong route, maybe he just gets beat, or maybe he catches a bad break, but there's some element of bad quarterback play in just about every interception. But whether a pick turns into a pick-six isn't just on the quarterback. It's a function of very many things including field position, proximity of other players, and speed/size/ball-handling skill of the intercepting defender, not to mention a bit of luck.

In other words, the quarterback has basically nothing to do whether an interception is returned for a touchdown or for no gain whatsoever. It's not as if the egregiously terrible throws result in defensive touchdowns while the merely bad throws result in simple interceptions.

This is the Bad Quarterback League, though, and we do want to reward miscues, poor decisions, and errors by the QB that hurt the team. There's no question that a pick-six hurts a quarterback's team more than a regular-flavor pick--and there's no question that it's more fun to watch Brady Quinn put points on the board for the other guys than for Kansas City--so some premium for "interception return for TD" over "interception" makes sense. But the standard scoring for an INT-TD, 25 points, is much higher than the score for an INT, 5 points. That seems like too much of a premium for something largely out of a quarterback's control.

2. Too much difference between "interception return for TD" and "fumble return for TD"

An interception and a fumble are functionally the same: both are rooted at least in part in poor decision-making or bad ball control, and both result in your team not having the ball anymore. That's bad, and it results in 5 BQBL points in both cases. An INT-TD and fumble returned for a touchdown are functionally the same, too: both cost your team six points.

In standard BQBL scoring, though, the INT-TD is worth 25 points, while the fumble-TD is worth a mere 10. They're equally harmful to your team, and they result from equally bad quarterbacking, so it only makes sense that they should have the same BQBL value too.

3. Sometimes you earn more points for fumbling than for keeping the ball

Mark Sanchez is backed up against his own end zone. He drops back, looking for some receiver downfield but doesn't find one. Eventually, he hears the foreboding footsteps of a 280-pound defensive lineman bearing down on him, and he knows he has no play. Sanchez takes the sack, goes down in the end zone, and gives up two points. He earns your BQBL team 20 points!

In another version of this sad tale, Sanchez tries to make a play, pitching out to Shonn Greene, but dropping the ball on the ground instead of completing the lateral. The astute defensive lineman falls on the fumble, and Sanchez has now given up six points. But he only earns your BQBL team 10.

Like we've just talked about, a fumble returned for a touchdown is worth 10 points in BQBL, regardless of how long that return happens to be. Getting sacked for a safety in the end zone is worth 20. That's an especially bad quarterback play, because a quarterback on top of his game should always throw the ball away rather than take the sack and the safety. But here's another case where the BQBL score doesn't necessarily accurately reflect just how badly a quarterback has been playing. Giving away six points is always worse than giving away two, so a lost fumble in the end zone should always be worth more than getting sacked for a safety.

4. Completion percentage doesn't always tell the whole story

There's a nice little BQBL windfall (5 points) for completing less than half of your passes, and even more points for completion percentages under 40 (15 points) or 30 (25 points). And clearly, there needs to be some cutoff if the BQBL is going to award points for poor completion percentage. But is a quarterback who completes 18 of 35 passes really having a better day than the one who completes 17 of 35? How about the one who completes 4 of 10?

Consider Sam Bradford's Week-5 brilliance, in which he completed 7 of 21 passes. Seven. On twenty-one attempts. Bradford did get the rare sub-40-percent bonus, but so would a quarterback who completed 17 of 43, which is awful but somehow doesn't seem as bad. The 17-for-43 guy at least tried the whole game. Mr. 7-for-21 might not have.







5. "Benched" needs a more consistent definition

Sometimes, "benching" is obvious: John Skelton's 4 interceptions and 18.2 quarterback rating earn Ryan Lindley a ticket into the game. That's worth a cool 35 BQBL points. The Seahawks' 35-point lead in the same game earns Russell Wilson a trip to the bench and Matt Flynn some playing time; that's worth zero points.

It's not always as clear, though. Aaron Rodgers' 219 yards and 81.9 rating were probably among the worst in his career, but pulling him with 5 minutes to go in the game had as much to do with the Packers' offensive futility and not wanting him to get injured as with his mediocre play. Blaine Gabbert's 7-for-19 and a whopping 53 yards was a truly miserable performance, but his giving way to Chad Henne in the 4th quarter had at least as much to do with a legitimate injury as with his stats.

Most decisions in the BQBL are obvious because they're quantitative, but this one is a little subjective, so awarding it is a tougher call. For a player to earn "benched" points, he must have been pulled by the standard of sufficiently poor play. That requires a judgment call, and it's important that the judgment be applied in the same way in every case.

Recommendations for future BQBL seasons
  • Reduce INT-TD scoring to 10 points. This brings it in line with fumble-TD and reduces a scoring discrepancy based largely on factors out of the QB's control.
  • Keep "sacked in the end zone for a safety" scoring at 20 points, but add the condition that if a quarterback fumbles as he's sacked, he gets the full 20 points anyway. This eliminates the QB "incentive" for losing the football.
  • Add a scoring mechanism for completing a sufficiently small number of passes, perhaps "fewer than 10 completions: 10 points".
  • Retain the 35-point score for benching, but don't award it in cases of injury or when the benched QB's team has no chance of winning the game anyway. Maybe "benched before the 4th quarter" or something similar.
Did anyone else out there run a BQBL this year, with standard or modified rules? Any other recommendations or improvements you've identified?

No comments:

Post a Comment