NFL Power Rankings: An Alternative View

Now that the draft is over and mini-camps have started around the league, the next onslaught of information coming your way will be the never-ending stream of “NFL Power Rankings”.

In some ways, PossessionPoints.com is no different than other information outlets in that we will be presenting our own version of Power Rankings weekly during the season. However, that is where the similarity ends.

We believe that when people look at the lists of Power Rankings they want to see how their favorite team fares in the eyes of those who construct these Power Rankings and if their team has a chance of making it to the playoffs or the Super Bowl.

We would like to state right off the bat that a team’s talent is just one of the factors that determines a team’s win-loss record and a possible playoff berth.

In this article, we would like to highlight four factors that contribute to a team’s final win-loss record. Those factors are randomness, schedule, home field advantage and team quality (traditional Power Ranking if you will). We have provided a lot of the background information for this article on a page on our site.

Randomness:

We know that over the course of a season, a team’s final win-loss record is more than random luck. Much of a team’s success does depend on talent. However, imagine if you will, a scenario where all 32 teams are completely equal in the talent department and there is no such thing as a home field advantage. If these conditions existed, what kind of records do you think you would see?

Do you think that almost all teams would have win-loss records between 7-9 and 9-7?

We put together a quick little spreadsheet and used the 2009 schedule to come up with random results for all 256 games on the schedule. We posted some of the “random season” results on our site, and the results surprised us in that typically, the top teams had 12-4 records, and the bottom teams finished around 4-12.

It may seem strange, until you consider the fact that the chance of winning or losing 12 out of 16 coin tosses is about one in 36, and 11 wins or losses of 16 coin tosses is about one in 15. So, applying this same principle with 32 teams each season, it does make sense that just by random chance, some teams would go 12-4, and some would go 4-12.

We always expect extremes to be rare, and the odds of a random 16-0 or 0-16 record are around 1 in 65,000. As we were working with our spreadsheet, we did see this extreme occur, and it happened on our spreadsheet to the NY Jets. We did not purposely focus on the Jets. It could have been any of the teams, and we could have used any season schedule to get similar results.

We will grant you that the football sometimes takes funny bounces, and this can be a factor in some games during the season making luck or chance a legitimate portion in a team’s win-loss record. It is a totally unpredictable factor, and therefore, it cannot be incorporated into analysis and forecasting equations. Since equations are our thing, we do not want to spend that much time talking about randomness, but we also do not want you to dismiss the fact that it can affect a team’s win-loss record..

Schedule:

Take any ranking system. We do not care if it is an expert’s Power Rankings or last year’s win-loss rankings. You can pick whatever. Play these rankings against this year’s schedule, and you may be surprised at what you find.

To illustrate this point, we took a look at some 2009 Power Rankings that have already been published. For what we are doing in this article, it does not matter if you accept this order as realistic or not. In the example, we chose to focus on the 32 teams were ranked:

1	Patriots	9	Bears	17	Jets	25	Chiefs
2	Steelers	10	Dolphins	18	Broncos	26	Bucs
3	Giants	11	Vikings	19	Texans	27	Bengals
4	Panthers	12	Eagles	20	Saints	28	Browns
5	Colts	13	Packers	21	Redskins	29	Seahawks
6	Falcons	14	Cardinals	22	Bills	30	Raiders
7	Ravens	15	Chargers	23	49ers	31	Rams
8	Titans	16	Cowboys	24	Jaguars	32	Lions

We played these rankings against this year’s schedule neglecting home field advantage or any randomness.

In this scenario, if a team plays a higher-ranked team, the lower-ranked team gets the loss while the higher-ranked wins.

When we played these rankings against the 2009 schedule in this manner, it was no surprise that the top nine teams would become playoff teams. However, the No. 10 Dolphins would miss the playoffs with a record of 9-7 while the No. 11 Vikings would also miss the playoffs with a record of 10-6.

The next playoff team this list for the 2009 season would be the No.12 Eagles who play 11 teams that rank lower than the Eagles and only five teams that rank higher. This would give the Eagles an 11-5 record and a wildcard playoff spot.

Based on these rankings and their schedule, the No. 13 Packers, like the Vikings, would also miss the NFC playoffs with a 10-6 record.

The No. 14 Cardinals and No. 15 Chargers would win their divisions and make the playoffs with 9-7 and 10-6 records respectively.

One would like to think that the 12 best teams make the playoffs each year, but that just is not possible with the way the NFL schedule works, even if every team plays perfectly to its capability each week.

Home Field Advantage:

Now, we are going to get a little more complicated (yes, it is true) but stay with us. To be able to work with home field advantage variations, this we had to quantify how much better one team was relative to another. To keep this as simple as possible, we assumed the difference between each team was a constant.

So, the top ranked team was assigned a +8 (just our own statistical yardstick), the second team a +7.5, the third team a +7. Each team was graded 0.5 away from the one next to it, and the value of 0 was skipped so the bottom team had a -8. With this tight a ranking in our system, very small changes in the value of home field make a big difference.

Now, we can start modifying how much home field is worth and observe what happens to win -loss records and playoff scenarios. In the extreme adjustment of home field advantage, all teams would be 8-8, wining all their home games and losing all their road games. This would be boring. Who wants the home team to always win?

Once again, we used the same example rankings. When we made the home field advantage factor worth just one percent, the playoff picture changed very little. The No. 12 Eagles fell into a tie for the final NFC wildcard playoff spot as they saw their record fall to 10-6. However, it was not the No. 11 Vikings with whom they tied. It was the No. 13 Packers.

When we bumped up home field to two percent, the Packers’ record rose to 12-4, and they tied the Bears for the division. The Vikings remained behind them with a record of 10-6. At this level, the Eagles miss the playoffs with a 9-7 record. The Cardinals would still win the NFC west but now with an 11-5 record. The AFC playoff picture would remain unchanged.

At the four percent level, the Bears, Vikings and Packers all make the playoffs with 11-5 records. The Panthers and Falcons would tie for the NFC South title with 10-6 records.

The only point of this little exercise is to illustrate what many people take as a given. Schedule combined with a home field advantage does play a role in a team’s season and win-loss record. Yet, many Power Rankings omit these factors. Even with the constant distribution of team quality that we used, you can see how the value of home field can change win-loss records and the playoff picture.

Team Quality (Traditional Power Ranking):

Most “Power Rankings” authors are trying to enumerate teams in order of quality. In other words, best teams are ranked highest and the not-as-talented teams are ranked at the bottom. The question we always come up with is: How much better is the No. 1 team over the No. 2 team or even the No. 10 team?

An even distribution like we used in the “home field” analysis above is not a very realistic distribution. The top team may be well ahead of the remainder of the league, as it was with the 2007 Patriots, or the top five teams may be very close together like last season was by our measure. The same is true at the bottom. By our measure, the 2008 Lions fell well below the rest of the league.

To illustrate this, we took the same rankings and gave the top 12 teams relative performance grades between 34 and 45 (Again, our statistical yardstick). The bottom 12 teams got values between -34 and -45 while the middle eight teams were given values between 4 and -4.

Of course with no home field advantage, the records are the same as we described in the “schedule” section. We need to vary home field by larger increments now since the separation between the 32 teams is much greater being between 45 and -45 as opposed to 8 and -8 in the example above.

At a four percent adjustment, the Chargers no longer make the playoffs and are replaced by the Broncos who had a “Power Ranking” of 18. The Bears and Vikings make the playoffs while the Packers do not.

At a 15% adjustment, the Chargers are back in with a 9-7 record while the Broncos fall out with an 8-8 record. The Packers move back into contention for a playoff spot as their record equals the Vikings at 11-5. A tie breaker would determine who made the playoffs.

So What?

You are probably thinking, “So what, if you play with numbers enough you can make them say anything you want.” Well, not really, even given their relatively low power ranking of 14 there was virtually nothing we could do to make the Cardinals miss the playoffs. The NFC playoff picture fluctuated much more than the AFC playoff picture did.

Even if an all-knowing supernatural NFL power provided a perfectly ordered list of best to worst teams, the chances that the top 12 teams on this list make the playoffs would be very slim for the reasons we have explained.

A Different List:

As we said at the start of the article, we also publish a list that resembles the “Power Rankings” list. We call it our “Performance Rankings” because it is based on how teams perform in our PossessionPoints stat.

We are not going to explain all that goes into our stat in this article because we would bore you to death, but the end result is a positive or negative number we call the “Relative Performance Measure” or RPM. Our RPM list is just the ranking of teams from the largest RPM to the smallest. But with the RPM, you can see how close one team is to another. In fact, two adjacent teams may either be a fraction or several numbers apart. Typically the 32 teams will be spread over a 90-point or more range.

Below is a Performance Rankings chart. The last column, on the right shows the final RPM from last season (with one exception we will talk about later). It should be no surprise to note that the Steelers are on the top of the RPM list, and the Lions are on the bottom. Where some teams fell in the middle might be a surprise to some fans, but that is where the stat put them.

Some of you might be thinking, “Last year is history, why are you showing this now?”

Rank	Team	W	L	Relative Performance measure
1	Steelers	15	1	37.66
2	Vikings	14	2	36.31
3	Eagles	14	2	35.77
4	Ravens	13	3	35.66
5	Giants	14	2	34.15
6	Patriots	16	0	32.43
7	Panthers	14	2	29.71
8	Packers	12	4	26.05
9	Chargers	12	4	24.94
10	Jets	10	6	19.24
11	Texans	12	4	15.72
12	Titans	11	5	15.54
13	Falcons	8	8	14.02
14	Saints	8	8	12.63
15	Cowboys	7	9	11.85
16	Dolphins	8	8	11.84
17	Cardinals	11	5	10.93
18	Bucs	5	11	5.42
19	Jaguars	7	9	-5.77
20	Colts	6	10	-8.22
21	Redskins	5	11	-9.19
22	49ers	6	10	-14.59
23	Broncos	6	10	-17.70
24	Bills	2	14	-17.76
25	Browns	5	11	-23.48
26	Bears	6	10	-24.12
27	Chiefs	1	15	-39.24
28	Raiders	2	14	-42.77
29	Bengals	2	14	-44.38
30	Rams	2	14	-46.06
31	Seahawks	2	14	-48.49
32	Lions	0	16	-60.94

Well, if you look at the chart and each team’s win-loss record, you will note that these records are not the win-loss records from 2008. What these win-loss columns show is a projection of each team for 2009 if each of the teams played their 2009 schedule at the same RPM level as 2008.

Do Not Panic! This is at best a partial look at what’s ahead. We have not made our modifications for offseason trades and drafts.

The one adjustment we talked about comes into play here. The Panthers’ 2008 RPM was not good. We adjusted them so they would fall with the leaders up at No. 7, within 8 points of the top ranked Steelers. Even with that adjustment, when we played the Panthers revised RPM against their schedule with a four percent home field adjustment, Carolina ends up with a 9-7 record. (Four percent home field adjustment is the best historical adjustment for our RPM).

Conversely, look at the No. 17 Cardinals. They had an RPM of 10.93 (the Super Bowl brought them down some). However, when we play that RPM against their 2009 schedule, we project a 12-4 record.

We could find a lot of interesting facts in this chart, but we are just going to focus on one more. The No. 13 Falcons and the No. 14 Saints are in the same division and had very similar 2008 RPMs of 12 and 14. Yet, if the No. 14 Saints play at the same level of their 2008 RPM in 2009, we project that they would have a 10-6 record while the No.13 Falcons would go to just 8-8.

Our list ranks teams based on how they perform in our stat. History has shown us that this is a pretty good way to look at which teams can ultimately win in the playoffs. Our list would be sorted much differently if we sorted teams just by projected wins.

In a nutshell, we believe that Performance Rankings are more than just opinion-based lists. Factors such as home field advantage and strength of schedule have to come into play as well if fans are going to get a real feel for how their team’s chances at success. This season we will add “Projected Win-Loss Records” to the RPM value when we publish our weekly rankings. This way readers will get a feel for relative quality from the RPM number as well as how it would appear to project to wins and losses.

(For more background on how the RPM shows itself in the Playoffs you may want to read an article we wrote during last year’s playoffs “NFL Playoff Upsets: What Upsets?”)

Should Mark Sanchez Be the New York Jets' 2009 Starter?

What can Jets’ fans expect this season from rookie QB Mark Sanchez? With all the hype and exposure surrounding the “new” New York Jet, we, at PossessionPoints.com, wanted to try to keep our expectations as grounded as possible.

We decided to look at the top ten quarterbacks of last season (by QB rating) and see how their immersion into NFL play prepared them for success. Which quarterbacks got to wade into the professional football waters slowly, and which ones were thrown off the pier without much notice and expected to swim?

Here they are:


Philip Rivers – 2008 QB Rating: 105.5

In the 2004 and 2005 seasons, Rivers only appeared in two games, and he only attempted 30 passes. These stats put him in what we call the “let’s let him get used to the water slowly” category.

Chad Pennington – 2008 QB Rating: 97.4

In 2000, Pennington appeared in one game, and in 2001, he appeared in two. He attempted only 25 passes in the two years. With these numbers, Pennington is also another toe dipper in the NFL waters.

Kurt Warner – 2008 QB Rating: 96.9

Warner's NFL career started late. Although signed as a free agent out of Northern Iowa by the Packers in 1994, Warner never saw opening day. He was released in the preseason. Stints in Europe and the Arena Football League preceded his final arrival into the NFL. Even though Warner had vast football experience behind him before he put on a Rams’ uniform, Warner still played in just one game in 1998 and attempted only 11 passes. He, like Pennington and Rivers, is also a slow wader.

Drew Brees – 2008 QB Rating: 96.2

Drafted by the Chargers in 2001, Brees played in just one game and attempted 27 passes in his first season. In his second season, he played in all 16 games for the Chargers. After an impressive 4-0start, Brees ended his second season with a mediocre 8-8 record and a QB rating of only 76. The Chargers benched him during the 2003 season after he played in 11 games in favor of veteran Doug Flutie. The combined efforts of these two QBs gave San Diego a miserable 4-12 record. Brees’ QB rating sat at 67 for those 11 games. For us, Brees fell into the “you can throw him in the water, but better give him a life preserver to survive.” category. After his near drowning in San Diego, a change of scenery put Brees on the path to quarterback success.
Peyton Manning – 2008 QB Rating: 95

In 1998, Peyton Manning was drafted No. 1 for the Colts and proceeded to play 16 games while attempting 575 passes. His QB rating was the lowest of his career at 71. We categorize Manning as a quarterback who was thrown off the pier and expected to excel. While he has excelled, his first year was his worst and the Colts ended their season with a horrible 3-13 record.

Aaron Rodgers – 2008 QB Rating: 93.8

In 2005, 2006 and 2007, Rodgers played in a total of seven games. He attempted just 59 passes in those three seasons prior to his starting in 2008 in place of the relocated Brett Favre. Rodgers is definitely another toe dipper in the NFL quarterback pool. Rodgers is like many NFL quarterbacks who appear to thrive after two-to-three seasons of sitting. .

Matt Schaub – 2008 QB Rating: 92.7

In 2004, 2005 and 2006, Schaub attempted a total of just 161 passes for the Falcons. In 2008 alone, he attempted 380 for the Texans. Schaub, a three-year backup QB to Michael Vick, is another example of teaching a QB to swim before putting him into action.

Tony Romo – 2008 QB Rating: 91.4

Signed as an undrafted free agent with the Cowboys in 2003, Romo did not throw a pass for them until three years later in 2006. Yes, Romo only held a clipboard for his first three years, but his slow immersion gave him the confidence and skills to succeed.

Jeff Garcia – 2008 QB Rating: 90.2

Does anyone not appreciate Jeff Garcia’s trek to the NFL? Garcia played in Canada for four years prior to his NFL experience. In his first NFL season, he did play in 13 games and attempted 375 passes. We really have a tough time putting him in the “just throw him in the water and see if he stays afloat” category as he seems to have swam pretty well up North. So, we are going to say that his four years in Canada is the equivalent of the three years as an NFL backup.

Matt Cassel – 2008 QB Rating: 89.4

Our vote for “Cinderella Quarterback”. As a backup from 2005-2007, Cassel attempted just 45 passes in three years. We all know about 2008 and what fate dealt to both Tom Brady and Matt Cassel. For Cassel, who also served as the backup at USC, he might have easily assumed that his NFL uniform would always include a pair of headphones and a clipboard. However, he learned well in those three-plus years and proved himself when given the opportunity. We will mercifully put him in the toe dipper category, but to be honest, he probably falls more into the “stuck on the beach and watching everyone else have a good time” category.
That’s the top ten from 2008. If you notice, the big rookies who made all the headlines last year are not in that list. So, how did they perform?

While not in the top 10, Matt Ryan came close and ranked No. 11 with a QB rating of 87.7. Impressive for a rookie who we feel literally had the fate of the Falcons faithful on his young shoulders. His coach had little choice but to start him from the beginning, since they also traded away Matt Schaub. However, head coach Mike Smith did give Ryan breathing room by supplying him with a copious number of running plays especially in the earlier part of the season.

Not in the top ten and sitting at 22 with a QB rating of 80.3 is Joe Flacco of the Ravens. In the preseason, few thought Flacco would be the starter, but injuries and strange virus that sidelined Troy Smith, put Flacco from wader to jumper. Give him credit for accomplishing what he did. The Ravens formidable defense was Flacco’s life preserver.

What we take from our little study is this: If Rex Ryan follows conventional QB development philosophy, he will start veteran Kellen Clemens and leave Sanchez carrying a clipboard for a year or two. Ryan may believe that he can have the same success with Sanchez as the Falcons have with Ryan and the Ravens have with Flacco. Hopefully, somewhere in his coaching strategy is the knowledge that even the most gifted rookie quarterback cannot carry a team.

In our view, the Jets’ success or lack thereof is more dependent on Rex Ryan than it is on Sanchez. If Ryan does start Sanchez and puts him in a position where he makes mistakes that cost the Jets wins, we would put the blame on Rex Ryan and his staff..

One more note, you may ask why we focused on Sanchez and the Jets instead of on Matt Stafford and the Lions. The answer is simply this: The Lions can only go up from that 0-16 record. Stafford could start, without setting the NFL on fire, and still improve the Lions. We think Stafford should be fine as long as Detroit’s coaching staff keeps him away from excessive sacks which could affect his confidence and future performances.