NFL 26-27

One thing that bothers me is the use of win-loss records to gauge SOS. Why?

Trust me, there are far better ways to judge SOS; I don't. 

How many of you trust win-loss records to gauge strength?

Later today I'll post my end-of-season SOS, which does not include week 18; it's each team's strength after playing their 16th game.

1.  Ari

1.  Ari

There are various other ways that do it by other variables.  Like power ratings, ELOs, etc.  Some will break it down for H/A, etc.

But the final season ending SoS would look something like this:

Detroit Lions — 1522

New York Giants — 1517

Kansas City Chiefs — 1517

Minnesota Vikings — 1514

Washington Commanders — 1513

Tennessee Titans — 1511

Cleveland Browns — 1509

Los Angeles Rams — 1509

Arizona Cardinals — 1508

Indianapolis Colts — 1505

Green Bay Packers — 1504

Seattle Seahawks — 1503

Dallas Cowboys — 1503

Las Vegas Raiders — 1503

Chicago Bears — 1502

Baltimore Ravens — 1500

Philadelphia Eagles — 1500

San Francisco 49ers — 1499

Houston Texans — 1499

Tampa Bay Buccaneers — 1498

Denver Broncos — 1498

New York Jets — 1498

Cincinnati Bengals — 1497

Atlanta Falcons — 1494

Los Angeles Chargers — 1493

Jacksonville Jaguars — 1490

Carolina Panthers — 1488

New Orleans Saints — 1486

Buffalo Bills — 1485

Pittsburgh Steelers — 1485

Miami Dolphins — 1475

New England Patriots — 1462

My SOS looks much different.

I had Ariz having the hardest sked.

Easiest: Denver.

I break down a team's SOS for their wins and losses. 

That's why I made a killing in the Super Bowl; I just didn't believe in N.E. Why?

Seattle's base PR 12.14969 and NE 11.86263.

It's a road game for both so i'd make Sea like a 1/2 point fav.

When I compared their sow/sol it wasn't close .....in Seattle's wins, their SOS was -1.592, and NE was -5.079. League avg win SOS was -3.419.

Now, as for their losses, not many for either team, but enough for me to make a sound case for Seattle.

Seattle had a SOL rating of 4.582 and NE was -2.21 BIG difference- while league avg for team losses was 3.419.

So basically, during the season Seattle beat better teams and lost to better teams.

@jets96

I had Ariz having the hardest sked.

Easiest: Denver. 

Exactly how do you get to that point? All wins count the same and all losses the same?

That is why the ELO is decent -- it evaluates how each team's competition was that they performed against.

Having AZ and DEN that far from where others do is curious as to how you break it down to put them at the very top and very bottom.

@Raiders22

I basically use a team's opponent base rating, weighted by game results and averaged across all games.

And what does that get you? is there a number?

Of course, it gives me a number. It goes into my ratings, updated every week after a game is played; it's weighted so it doesn't blow up my base rating. And I made a mistake: I made NE the easiest at -4.54, Den was 2nd at -3.71, and Ari remains the hardest, sos at 2.83.

Sorry about that.

So, basically you had the same SoS as the W/L of the teams that they beat and lost to gave -- Arizona and New England.  I thought that you were looking for something more in-depth.

That is why I say the other ways look at how well they performed against other team's ratings.

For example, the SoS that uses just W/L would have a team that went 10-7 that lost very close games to elite teams OR it could be a team that went 10-7 and had been much stronger than another team that won many one-score games. 

This version of SoS treat both teams the same.

An 11-6 team that has an ELO of 1750 AND a 11-6 team that has an ELO of 1610 would recognize one team is better.  

Traditional SoS would consider them the same.  An ELO-based system would recognize that one team is better and one team is worse.

Then when you compare the SoS based on the opponent's ELO you get a clearer picture of how good they were compared to their opponent's power rating and not just compared to their opponent's W/L record.

Sometimes a game-by-game ELO comparison is better; sometimes a rolling power rating as the season progresses is better.

Suppose Team A finishes 11-6 -- but against the #30 ranked ELO schedule.

Suppose Team B finishes 9-8 -- but against the #1 ranked ELO schedule.

A simple SoS favors team A.  But an ELO system favors team B because they proved themselves against better competition.

Etc., etc. 

@Raiders22

 My SOS is using the opponent's power ratings.

I wish i could upload my week 17 power ratings to you here, but no way of doing it.

@jets96

@Raiders22

 It does amaze me how many people post picks on preseason games. You can't win in-season, but now you think you can win in the preseason,

Preseason is a time to watch and learn; the season doesn't really start until week 3 or 4. 

Even RAS is 2-8. EVERYBODY knows to watch the QB rotations; that's fine and good if you know how many snaps they were taking, and you don't.

See you during the season.

..

Yeah, this might be the wisest words said in this thread... 

Model did well last week taking Mia, Cle, and n.o and the last two winning outright.

Week 3 

464 pit +3.5

488 den +2.5

Luck