Following up on the weekly Cornerback performance post (see individual player evaluations here for Week 6) here are some of the games that might be impacted by cornerback play.
Our three metrics are:
- Attack Ratio: Average of Passes Defended/(Passes Defended+Completions Allowed) and Total Targets/Total Snaps
- Bleed Ratio: Total Yards Against/Total Snaps
- Burn Ratio: (Total Interceptions+50% of Forced Fumbles/Total Snaps) minus (Total Touchdowns/Total Snaps)
(There is a bit of change in the Burn Ratio calculation to include the under appreciated Forced Fumbles in the metric.)
We take the Z score of each ratio and compare that to the Z score for that team's week 7 opponent.
Here are a few games that might be interesting:
CLE's below average QBR could be compromised by CIN's above average metrics including Attack (indicative of QB's reluctance to throw on the Bengal's CBs) and Bleed (CIN CB's are not giving away receiving yards). And CLE isn't making it easy on itself with the QB drama. The CLE receivers can be dynamic, but the history is saying CIN CBs shouldn't be victimized this game.
LAR CB's are stingy with the TDs having given up just 2 TDs so far this year. The Bears scariest thing about the Bears right now is their QBR. Rams corners should hold up.
Both teams' CBs give up yards but TEN is exactly 1 standard deviation below the mean while the PIT QBR is nearly 1 deviation above the mean. Big Ben should put up nice numbers today.
You could have seen it coming but no one really wants a piece of the BUF CB's but it's no help when the NYJ QBR is pretty absymal. The difference between the attack high and the QBR low in terms of range is over 3.7 standard deviations. Safe to say you should not start any NYJ receivers this week (like anyone would anyway?) who should have their hands full.
Thanks for hanging in with me through the format change. Look out for the return of WR ROI for College Football next week!
*********************************************************************************
No comments:
Post a Comment