SCOAR: Week 2 Update & Predictions

Even with the disappointing start to the season for the Seahawks, it feels good to have football back. Now that Week 1 is in the books and we have some actual 2015 data, the outliers are starting to come back to earth and give us a more realistic view of the quality of teams. There was A LOT of movement this week, and this will probably continue through the 2014-2015 transition period (and to a lesser extent a few weeks after that ends as we expand the dataset). It will take a while before we have enough information to really lock teams into slots without each new week causing turmoil in the ratings.

For those of you who still aren't familiar with SCOAR...

SCOAR (SCaled Opponent-Adjusted point Ratio) is explained in detail here, but I'll provide a quick recap. Opponent-adjusted point ratio is the ratio of a team's points to the average points of that team's opponents, normalized with a logarithm. This value is calculated for a team's points for and points allowed then combined for a total rating. These results are then scaled using the average points scored across the league in order to give a result that corresponds directly to points.

Looking Back

Despite the promising results using 2013's data to predict the early 2014 games, this method had a rough go of it to start this year. Week 1 saw an unimpressive 9-7 picking games straight up, and an even worse 7-9 against the spread. This is far from the 65%/58% results obtained last season. Hopefully this will pick up as the season goes on (it should) and I can figure out a better way to predict the early weeks before next season.

This year I will make note of the best/worst performing teams each week, as well as the best/worst ratings from the season up to that point. Keep in mind that these numbers will change weekly as we add more data for the season. The 49ers victory will not look as impressive if they turn out to be a good team and the Vikings end up at the bottom of the league.

• Week 1 Best Single Game SCOAR - San Francisco 49ers: 38.54
• Week 1 Worst Single Game SCOAR - Minnesota Vikings: -41.80

On to the Numbers!

Despite an out of character high scoring loss in week 1, the Seahawks remain the top team going into Week 2. They were still able to post a positive SCOAR for the week by putting up 31 against the Rams. Their otherworldly 22.95 SCOAR fell back to earth and currently sits at 12.14, a much more realistic number for the top spot. There are two other teams in double digits: the Steelers (10.78) and the Panthers (10.37). The top 5 is rounded out with the Cowboys in 4^th and the Patriots in 5^th. These numbers still look strange based on the week 1 results as they are watered down by weeks 14-17 of 2014. Last week's #2 team (the Rams) fell all the way to 11^th despite their impressive win over the Seahawks due to their 2014 week 13 shutout (52-0 over the Raiders) being wiped off the books.

Notable Rankings

• Biggest Climb - Chiefs/Jets, moving up 8 spots to #10/14
• Deepest Fall - Eagles/Vikings, dropping 13 spots to #16/24
• Highest Loser - Seahawks at #1
• Lowest Winner - Titans at #28
• Most Neutral - Chargers (#17) with a SCOAR of 0.24

Week 2 SCOAR predictions

After a rough start, SCOAR looks to get back on track in Week 2. Another tie is predicted, so I'm not all that optimistic. I do not recommend using these for betting purposes (although you're more than welcome to send me money if they help you).

A reminder that SCOAR doesn't know what a normal football score looks like, so it's bound to put up some very unlikely point totals.