How much does point differential correlate with W-L records?

CSKreager
Posts: 463
Joined: Sun Oct 12, 2014 8:13 pm

How much does point differential correlate with W-L records?

Post by CSKreager »

Last year's Vikings were one of the football teams of all time with 13 wins, but a hilarious point differential- losing big and winning close save for their first/last wins of the season.

In 2004, the Falcons went 11-5 and were the NFC's #2 seed, but had an uninspiring point differential of +3

In 2002, those same Falcons were an unimpressive 9-6-1 and the NFC #6 seed, but had a +88 point differential Take out 2lopsided losses to the SB champion Buccaneers, their other 4 losses were by a combined 16 points

Meanwhile the 1981 Patriots were the strangest bad team ever: 2-14 with a point differential of -48 and never lost a game by more than 14 points. The record was awful, but they never got blown out like most 2-14 teams do (The 2-14 Jets of 2021 by contrast had a -214 point differential

The 2010 Chargers had a +119 point differential that was undone by cataclysmic special teams

The 2003 Patriots' point differential was +110 but save for 2 blowout wins vs the Eagles and Bills, most of their W's were cliffhangers by small margins.
User avatar
74_75_78_79_
Posts: 2331
Joined: Sat Oct 11, 2014 1:25 pm

Re: How much does point differential correlate with W-L reco

Post by 74_75_78_79_ »

The 2012 Colts were 11-5 yet a minus-30 PD!! Football Outsiders call them the worst 11-5 team ever.

Joe Bugel's Phoenix Cardinals in '93 finished 7-9 yet put up a +57 PD!!
ChrisBabcock
Posts: 1725
Joined: Sat Oct 11, 2014 3:36 pm
Location: Tonawanda, NY

Re: How much does point differential correlate with W-L reco

Post by ChrisBabcock »

Joe Bugel's Phoenix Cardinals in '93 finished 7-9 yet put up a +57 PD!!
The exact reverse of that... The 2017 Bills were 9-7 and -57.
Bob Gill
Posts: 559
Joined: Wed Oct 08, 2014 7:16 pm

Re: How much does point differential correlate with W-L reco

Post by Bob Gill »

In the USFL in 1984, George Allen's Arizona Wranglers and Steve Young's Los Angeles Express tied for the Pacific Division title with identical 10-8 records. But LA was minus-35 points for the season (373-338), while Arizona outscored its opponents by a sensational 218 points (502-284).
User avatar
TanksAndSpartans
Posts: 1153
Joined: Wed Nov 18, 2015 1:05 am

Re: How much does point differential correlate with W-L reco

Post by TanksAndSpartans »

I used to have a co-worker that used point differential to predict college fb games. There was more to it than just the total differential - I think a big piece of it was looking at the differential when teams shared common opponents. He didn't care about record at all. I don't know all the details, but I think part of what he was trying to get around was the issue of teams having really different strength of schedules. He would have his own weekly ranking. He of course claimed it was really good. One thing I remember is his system did like the 2004 Utah team when some in the national media wanted to dismiss them. He successfully predicted they would win their bowl game in a blowout.
sheajets
Posts: 1103
Joined: Sun Oct 12, 2014 12:22 am

Re: How much does point differential correlate with W-L reco

Post by sheajets »

74_75_78_79_ wrote:The 2012 Colts were 11-5 yet a minus-30 PD!! Football Outsiders call them the worst 11-5 team ever.

Joe Bugel's Phoenix Cardinals in '93 finished 7-9 yet put up a +57 PD!!
That Colts team late in the year played a game against the 2-13 Chiefs (mired in their darkest of days, just before Reid came in) Indy gives up 352 yards rushing and 507 total yards...and wins 20-13. KC committed 3 turnovers, 9 penalties, and missed a field goal.

That 93 Cardinal team is tough to figure. They gave up the 7th fewest points in the league, but they were 21st in total yards allowed. Looks like a bend but don't break D...They weren't particularly good at taking the ball away...only 9 interceptions, least in the league and 26 turnovers overall which put them 21st. Games where they did get big defensive plays (even defensive TD's, multiple) they ended up losing somehow (vs Philly and N.O.) They didn't get to the QB all that much either.

Overall their offense was above average that season. Looking at them they could've easily ended up 9-7 or 10-6 with a little luck.
rhickok1109
Posts: 1473
Joined: Sun Oct 12, 2014 8:57 am

Re: How much does point differential correlate with W-L reco

Post by rhickok1109 »

These responses focus on real outliers that don't truly address the question.

PFR uses the Pythagorean formula to determine expected winning percentage based on points scored and points against, and it really works quite well.
Bob Gill
Posts: 559
Joined: Wed Oct 08, 2014 7:16 pm

Re: How much does point differential correlate with W-L reco

Post by Bob Gill »

rhickok1109 wrote:These responses focus on real outliers that don't truly address the question.

PFR uses the Pythagorean formula to determine expected winning percentage based on points scored and points against, and it really works quite well.
Sure, but the ones that work out right aren't fun.
JohnH19
Posts: 910
Joined: Mon Oct 13, 2014 6:18 pm

Re: How much does point differential correlate with W-L reco

Post by JohnH19 »

The ‘71 Bengals were 4-10 and +19. The ‘66 Eagles were 9-5 and -14.
CSKreager
Posts: 463
Joined: Sun Oct 12, 2014 8:13 pm

Re: How much does point differential correlate with W-L reco

Post by CSKreager »

74_75_78_79_ wrote: Joe Bugel's Phoenix Cardinals in '93 finished 7-9 yet put up a +57 PD!!
8 of Phoenix's 9 losses in 1993 were by 7 points or less (only the 49ers beat them by double digits)

Bugel's Cardinals were the masters of the moral victory- nobody ever looked better in close losses than those Cards.
Post Reply