Here is why the QBR is worthless

[Rambos]

By Mike Sando

San Francisco's Colin Kaepernick wasn't nearly as good during the *****' 16-13 defeat at St. Louis. A case could be made that he cost his team the game with a poor decision leading to a safety and a botched pitch. That is why I was quite surprised to see QBR reward Kaepernick with an 82.4 score Sunday.

Dean Oliver, Albert Larcada and Michelle Sastri of ESPN's analytics team pointed to positive plays Kaepernick made and limitations in the NFL's official game charting system in explaining the surprising figure. Those limitations affect very few plays to this degree, but they significantly limited how much blame Kaepernick received for his role in a costly errant pitch to Ted Ginn Jr.

Among the highlights from their responses:

Scramble huge: Kaepernick's 50-yard scramble to the St. Louis 14-yard line increased the *****' win probability from 57.1 percent to 80.9 percent. His in-game QBR score spiked from 67.0 to 89.3 because the situation was so important and because Kaepernick, not a receiver, accounted for the gain. This was the fourth-longest scramble by a quarterback since 2008.
Still expected to win: The fumble and ensuing touchdown return by the Rams' Janoris Jenkins dropped the *****' win probability from 90.3 percent to 61.7 percent. The ***** were still favored to win because the Rams needed a two-point try to tie.
Shared blame: Kaepernick didn't make critical mistakes after the 50-yard scramble. The holding penalty against tight end Delanie Walker was a killer play. That penalty helped the Rams get the ball back with 1:34 remaining. A penalty against the *****' Dashon Goldson for unnecessary roughness with 1:07 remaining was also critical because the Rams' kicker, Greg Zuerlein, is such a threat from long range.
Strange situation: QBR relies, in part, on official play-by-play data from the NFL. The league scored Kaepernick's fumble as an aborted play, disregarding the role Kaepernick's errant pitch played in the turnover. As a result, QBR did not "blame" Kaepernick as much as it would have blamed him ideally. This will happen in rare cases, skewing the QBR score for a single game.
Similar situations: The NFL play-by-play accounted for Kaepernick's fumble the same way it accounted for a Week 3 fumble on a pitch from Washington's Robert Griffin III to Brandon Banks. Kaepernick's pitch was off-target. The one Griffin delivered appeared perfect. Kaepernick was mostly to blame for the *****' fumble. Banks was mostly to blame for the Redskins' fumble. Yet, the official play-by-play accounted for those plays in the same manner, as aborted plays.


I would expect ESPN's analytics department to seek ways around these sorts of abnormalities.

"We are looking into additional tracking going backward and forward to correct this illogicality, but don’t have it in place yet," Larcada said.

He took a safety and that played started at the 12 yard-line, his decision to drop all the way back into the endzone was all him

[TheBritishRam]

He also ran out of bounds in the last two minutes. The QBR stat is absolute garbage.

[AvengerRam]

Mark Sanchez registered a QBR of 127.8 yesterday. He recieved 92 points for "strategic holding of clipboard."

[Varg6]

^ That made me laugh out loud.