Aikman Efficiency Ratings

[Coffeedrinker]

Some may remember but I am a big fan of the Aiman Efficiency Ratings.  I think his formula does a very good job of assigning points based on the items that are very important when determine the effectiveness of an offense or defense. 

 

Anyway the past couple of years the rankings have been too depressing to post but this year the Colts are trending up on both offense and defense.

 

Here is the link to the ratings.  Keep in mind it takes until about week 4 or 5 for things to level out and really be a good snapshot of the team's efficiency.

 

http://www.aikman.com/AikmanEfficiencyRatings/tabid/274/articleType/ArticleView/articleId/132/Aikman-Efficiency-Ratings-2018-As-of-Week-2.aspx

[BlueShoe]

I am a big fan of Aikman's ratings too.

 

For what it is worth, I developed my own rating system a few years ago, and this year the Colts are scoring well in my rating system too.

 

My system currently simulates the Colts to finish the season with a 10-6 record. 

[2006Coltsbestever]

Troy was all about Efficiency when he played. I buy his system. He is a Top 20 QB of all-time IMO and 3 time SB winner. 

[DrWhom]

I appreciate the OP posting this link. I am constantly amazed at the breadth of knowledge that posters bring to the forum... 

 

I do have one question: if PHI is ranked so much higher than us in the AER standings (combined scores being considered), does that imply that we should lose to them this weekend? Seems like a logical extension of the AER analysis, doesn't it?

 

 

[Coffeedrinker]

8 hours ago, DrWhom said:

I appreciate the OP posting this link. I am constantly amazed at the breadth of knowledge that posters bring to the forum... 

 

I do have one question: if PHI is ranked so much higher than us in the AER standings (combined scores being considered), does that imply that we should lose to them this weekend? Seems like a logical extension of the AER analysis, doesn't it?

 

 

Yes and no.  I used to use the AER for the office pool but it's not a very accurate indicator until about week 5 or 6.

 

Just like the Redskins have an overall AER than the Colts but the Colts just beat them.

[compuls1v3]

17 hours ago, Coffeedrinker said:

Some may remember but I am a big fan of the Aiman Efficiency Ratings.  I think his formula does a very good job of assigning points based on the items that are very important when determine the effectiveness of an offense or defense. 

 

Anyway the past couple of years the rankings have been too depressing to post but this year the Colts are trending up on both offense and defense.

 

Here is the link to the ratings.  Keep in mind it takes until about week 4 or 5 for things to level out and really be a good snapshot of the team's efficiency.

 

http://www.aikman.com/AikmanEfficiencyRatings/tabid/274/articleType/ArticleView/articleId/132/Aikman-Efficiency-Ratings-2018-As-of-Week-2.aspx

Thanks for sharing Coffeedrinker!  I've never seen this before and it is interesting!

[lollygagger8]

I like that he rated Indy higher than pats

[Coffeedrinker]

There are definitely some interesting things regarding the AERs.

 

The 108.x for the Chiefs is the highest offensive number I have ever seen for a team.  Conversely, the 49.1 defensive rating is the lowest I have seen for a defense.

 

If the early scores are any indication the Rams are as tough as everyone talked about in the off season.

 

For the Colts, it's interesting that the offense is quite a bit higher in the efficiency ratings than the NFL (yards) rankings and a few points lower on the defensive side than the NFL.

[BlueShoe]

3 hours ago, Coffeedrinker said:

Yes and no.  I used to use the AER for the office pool but it's not a very accurate indicator until about week 5 or 6.

 

Just like the Redskins have an overall AER than the Colts but the Colts just beat them.

 

I have the same issue with my system, which is why I use a manual override for specific teams in the first few weeks. The manual override is where I can add my objective opinion to override the statistics. After a month or so has been played then my system predictability starts to become very accurate and I remove the manual overrides. 

[chad72]

9 minutes ago, BlueShoe said:

 

I have the same issue with my system, which is why I use a manual override for specific teams in the first few weeks. The manual override is where I can add my objective opinion to override the statistics. After a month or so has been played then my system predictability starts to become very accurate and I remove the manual overrides. 

 

Good to see you posting here again, man!!! 

 

I agree with the sentiment. For example, the Dolphins, not sure how much to make of their wins over the Titans and Jets except the fact that their defense played stout in both those games. Winning in the NFL is often decided by a few plays in close games and some teams/players make it and some do not.

 

Plus, there is the consistency factor. Chargers in 2010 were Top 10 in O and D but did not make the playoffs. They were very good last year too, did not make the playoffs. They just do not put together a consistent set of performances week in and week out even though statistically, they may have a high ceiling that shows up in enough games thus keeping their averages high.

[BlueShoe]

Just now, chad72 said:

 

Good to see you posting here again, man!!! 

 

Thank you for the kind words! I am thankful to be blessed with so much to keep me busy. 

 

Family, the day job, coaching football, and my many other football hobbies are fulfilling my life more than ever. 

 

I make an attempt to peek in here from time-to-time to take the temperature, and see what the main topics are. However, I have learned not to stay long. I have found that a lot of my good energy can quickly turn into bad energy on this board, which then ends with me contributing to the bad energy. I think it stems from most members having so much passion for the Colts and football.

 

This place can become a melting pot and I have come to the conclusion that it is best to distance myself from Colts boards, and use my energy on something that is actually productive.

[BlueShoe]

41 minutes ago, chad72 said:

I agree with the sentiment. For example, the Dolphins, not sure how much to make of their wins over the Titans and Jets except the fact that their defense played stout in both those games. Winning in the NFL is often decided by a few plays in close games and some teams/players make it and some do not.

 

Plus, there is the consistency factor. Chargers in 2010 were Top 10 in O and D but did not make the playoffs. They were very good last year too, did not make the playoffs. They just do not put together a consistent set of performances week in and week out even though statistically, they may have a high ceiling that shows up in enough games thus keeping their averages high.

 

I agree. Something else to keep an eye on is the Patriots.... They are not scoring well in my system, and I had to actually give them a manual override to bump them up. They beat the Texans by a touchdown and soundly lost to the Jaguars. The Jaguars are my systems number 1 ranked team, and it is very early so this could go either way for the Patriots. They play the Dolphins (2-0), then the Colts (we could easily be 2-0), and then the Chiefs (2-0). 

 

This could be a bad start for the Patriots. I think they are feeling it too, which is why they are taking fliers on guys like Corey Coleman and Josh Gordon. 

[Coffeedrinker]

2 hours ago, BlueShoe said:

 

I agree. Something else to keep an eye on is the Patriots.... They are not scoring well in my system, and I had to actually give them a manual override to bump them up. They beat the Texans by a touchdown and soundly lost to the Jaguars. The Jaguars are my systems number 1 ranked team, and it is very early so this could go either way for the Patriots. They play the Dolphins (2-0), then the Colts (we could easily be 2-0), and then the Chiefs (2-0). 

 

This could be a bad start for the Patriots. I think they are feeling it too, which is why they are taking fliers on guys like Corey Coleman and Josh Gordon. 

Sounds like an interesting formula.  Do you care to share or is it something you want to keep private?

[jvan1973]

2 hours ago, BlueShoe said:

 

I agree. Something else to keep an eye on is the Patriots.... They are not scoring well in my system, and I had to actually give them a manual override to bump them up. They beat the Texans by a touchdown and soundly lost to the Jaguars. The Jaguars are my systems number 1 ranked team, and it is very early so this could go either way for the Patriots. They play the Dolphins (2-0), then the Colts (we could easily be 2-0), and then the Chiefs (2-0). 

 

This could be a bad start for the Patriots. I think they are feeling it too, which is why they are taking fliers on guys like Corey Coleman and Josh Gordon. 

If it's a true system to rank teams,   why would you manually adjust it?

[DougDew]

33 minutes ago, jvan1973 said:

If it's a true system to rank teams,   why would you manually adjust it?

Its not.

 

Algorithms are imperfect and so are rating systems, as wells as any type of statistically based model.  You simply cannot collect enough data to make them perfect, so manual overlays have to be an integral part of any model.  The goal is to try to have as few of them as possible, and to make the overlays objective and without bias. 

[jvan1973]

11 minutes ago, DougDew said:

Its not.

 

Algorithms are imperfect and so are rating systems, as wells as any type of statistically based model.  You simply cannot collect enough data to make them perfect, so manual overlays have to be an integral part of any model.  The goal is to try to have as few of them as possible, and to make the overlays objective and without bias. 

 So you know how his system works? 

[Coffeedrinker]

6 minutes ago, jvan1973 said:

 So you know how his system works? 

He doesn't have to know how BlueShoe's system works, what he stated is a known for any statistical or data based formula.  

[jvan1973]

3 minutes ago, Coffeedrinker said:

He doesn't have to know how BlueShoe's system works, what he stated is a known for any statistical or data based formula.  

Are there manual overrides in Aikmans system 

[DougDew]

4 minutes ago, jvan1973 said:

 So you know how his system works? 

Its how all statistical models work.  So, yes, as long as his rating system falls into the category of "statistical based", it has a human judgement component to it. 

 

The system could be totally human judgment.  And could be much better than any algorithm.

 

He's got a small pool of players to work with, so the data for the stats will be imperfect, increasing the likehood he needs human input to make the ratings make sense. 

 

Its not like its an insurance company with statistical data used to compile behavioral patterns and tendencies to assign car insurance rates with millions of people as its universe.  Even at that, those companies will make adjustments for how far you are from your work, where you live, etc. so the universe of data gets smaller.  Then they bump your rates up for individual behavior, like speeding tickets or accidents.  Every time you narrow the field, the increasing individuality and uniqueness of the data requires human judgement to make a decision.

 

Until we all become Google robots programmed with the same software and live exactly the same lives.

[BlueShoe]

59 minutes ago, jvan1973 said:

If it's a true system to rank teams,   why would you manually adjust it?

 

I have already answered that, but you didn’t read it. 

 

There is not enough data after only a couple of weeks. After about a month or so the system becomes more accurate. 

 

Its not perfect anyway. Just something that is right more often than it is wrong. And something I like to play around with. 

[BlueShoe]

4 minutes ago, jvan1973 said:

Are there manual overrides in Aikmans system 

 

If you use his data after only 2 weeks there will need to be. 

 

Again, this was pointed out earlier in the thread. 

[2006Coltsbestever]

3 minutes ago, jvan1973 said:

Are there manual overrides in Aikmans system 

He probably does have them because every statistically formula is flawed. I love coming up with stuff like Aikman has and have found any  system I have tried has flaws in them.

[BlueShoe]

3 minutes ago, DougDew said:

Its how all statistical models work.  So, yes, as long as his rating system falls into the category of "statistical based", it has a human judgement component to it. 

 

The system could be totally human judgment.  And could be much better than any algorithm.

 

He's got a small pool of players to work with, so the data for the stats will be imperfect, increasing the likehood he needs human input to make the ratings make sense. 

 

Its not like its an insurance company with statistical data used to compile behavioral patterns and tendencies to assign car insurance rates with millions of people as its universe.  Even at that, those companies will make adjustments for how far you are from your work, where you live, etc. so the universe of data gets smaller.  Then they bump your rates up for individual behavior, like speeding tickets or accidents.  Every time you narrow the field, the increasing individuality and uniqueness of the data requires human judgement to make a decision.

 

Until we all become Google robots programmed with the same software and live exactly the same lives.

 

 Well said!

[Coffeedrinker]

13 minutes ago, jvan1973 said:

Are there manual overrides in Aikmans system 

I'm not sure why you are being obtuse about this but I will play along.  Aikman's ratings is a formula and his purpose is not to predict but to quantify current results.  So there are no overrides because all it does is calculate information that has already happened.

 

But when I use AER to try and determine which team will win the game, I use "manual over-rides"; I factor in home field advantage, key injuries I will even average the last 5 weeks compared to the overall to see if a team is rising or falling in efficiency.

 

 

[BlueShoe]

1 hour ago, Coffeedrinker said:

Sounds like an interesting formula.  Do you care to share or is it something you want to keep private?

 

There is a lot to it, and I tweak it from time to time.

 

Offensive YPP and Defensive YPP is important. Coaches are always talking about the importance of chunk plays and I think they look at YPP on both sides as being very important. 

 

NFL.com just started releasing blocking stats and I use that vs ESPNs new defensive stats which include quarterback hits and tackles for loss. I think this can tell a bigger story than sacks. A sack can be a flash play where a player went unblocked on just one play. However when a team has 9.5 tackles for loss like Jacksonville did once this year then in my mind, that tells a story about dominant a defense (especially if they do this at a high rate).

 

The stats I use come from NFL.com and ESPN. I hook into their team tables to get most of the team data.

 

I don’t completely buy into a team getting more turnovers. Obviously a team forcing more turnovers is likely to win the game, but it is too sporadic to predict. I see turnovers as more of a hindsight way of looking at a game rather than a predictable stat. That said, I do give the stat some weight because some teams are just good at forcing turnovers while others are not. 

 

Obviously the wins and losses have some weight and so does the strength of schedule. 

 

Yards per game has some weight too, but this can easily be skewed by garbage yards. 

 

Points also play a role, but again can be easily skewed by garbage points.

 

Aikman is using a lot of the same stats that I am. We don’t weigh everything the same though. 

 

I think YPP, OL stats, and DL stats tell a story that helps to make games somewhat predictable.

 

Also, Aikman's efficiency stats inspired me to come up with my own system as it sounds like he did for you too.

[jvan1973]

53 minutes ago, Coffeedrinker said:

I'm not sure why you are being obtuse about this but I will play along.  Aikman's ratings is a formula and his purpose is not to predict but to quantify current results.  So there are no overrides because all it does is calculate information that has already happened.

 

But when I use AER to try and determine which team will win the game, I use "manual over-rides"; I factor in home field advantage, key injuries I will even average the last 5 weeks compared to the overall to see if a t8eam is rising or falling in efficiency.

 

 

It was a question.   I'm not being obtuse about anything.   The way i understand the aikman system is completely stat based.    I was asking how his system was similar if you can manually change the actual outcomes.   

[Coffeedrinker]

Here is the updated AER through week 6.

 

http://www.aikman.com/AikmanEfficiencyRatings/tabid/274/articleType/ArticleView/articleId/134/Aikman-Efficiency-Ratings-2018-As-of-Week-6.aspx

 

Colts are overall 17.  17 offense and 23 defense.  The O is a bit lower than I expected and the D is a bit higher, I guess they were helped a bit by the jets kicking 6 FGs.

 

This affirms what many on this forum say, the Colts are the best 1-5 team in the league and a couple of bounces/breaks here and there they could be 3-3 at least.

[Jdubu]

1 hour ago, Coffeedrinker said:

Here is the updated AER through week 6.

 

http://www.aikman.com/AikmanEfficiencyRatings/tabid/274/articleType/ArticleView/articleId/134/Aikman-Efficiency-Ratings-2018-As-of-Week-6.aspx

 

Colts are overall 17.  17 offense and 23 defense.  The O is a bit lower than I expected and the D is a bit higher, I guess they were helped a bit by the jets kicking 6 FGs.

 

This affirms what many on this forum say, the Colts are the best 1-5 team in the league and a couple of bounces/breaks here and there they could be 3-3 at least.

It’s really not difficult to imagine this team as 3-3, last year, it would have been. 

[LJpalmbeacher2]

1 hour ago, Coffeedrinker said:

Here is the updated AER through week 6.

 

http://www.aikman.com/AikmanEfficiencyRatings/tabid/274/articleType/ArticleView/articleId/134/Aikman-Efficiency-Ratings-2018-As-of-Week-6.aspx

 

Colts are overall 17.  17 offense and 23 defense.  The O is a bit lower than I expected and the D is a bit higher, I guess they were helped a bit by the jets kicking 6 FGs.

 

This affirms what many on this forum say, the Colts are the best 1-5 team in the league and a couple of bounces/breaks here and there they could be 3-3 at least.

 

...and a Force that will need to be reckoned with starting this week!

[LockeDown]

This is gonna sound crazy, but I think we can get 16 and 22.  

[Blueblazes]

Been a Colts fan for a long time and try to be optimistic, but we're stretching here guys.  We are 1-5, losing to some pretty bad teams. At some point, Bill Parecells was right, you are what your record is, but I guess trying to look at anything semi-promising is good at this point.

[Coffeedrinker]

AER's updated after week 7.  Colts shoot up to #8 overall with the impressive offensive and defensive efficiency in the Bills game.

 

One of the things I like is Aikmann puts the NFL ranking (based on yards) next to the AER.  

 

Offensively, Colts are now 12th in AER (17th in NFL).  Defensively the Colts sky rocket to 13 (up 8 points from week 6) and 21st in the NFL (up from 23rd after week 6).

[Valpo2004]

How is the AER calculated? 

[DraftMaster]

You know what my favorite stat is? Wins. SB wins to be exact.

[Coffeedrinker]

49 minutes ago, Valpo2004 said:

How is the AER calculated? 

Here is a pretty good article about how it came to be.

 

https://www.nytimes.com/2005/12/13/sports/tv-sports-in-numbers-game-aikmans-add-up.html

 

Here is another good breakdown written about the Cowboys.

 

https://www.bloggingtheboys.com/2014/9/25/6843557/stats-the-cowboys-will-watch-closely-this-season-the-aikman-efficiency-rating

 

 

[Coffeedrinker]

Aikmann finally posted the update after Week 8

 

After the impressive offensive performance against Oakland, the Colts move further up to #6 offensively with a 91.2 (87.7 after week 7).

 

Defensively, as expected, the Colts dropped a couple to #15 at 67.2 (69.8 the previous week)

 

In the overall the Colts move to #7 with a combined AER of 158.8.  Up from #8 the week before with a combined AER of 157.4.

 

For the game this week against the Jags.  The Jags have a combined 139.0 and are ranked 27th overall.  They have a 69.7 offensive score (well below the NFL average of 88) and a 69.3 defensive score (slightly above the NFL average of 66.8)

[Coffeedrinker]

here are the AERs after Week 10

 

Position wise, the Colts remain at #6 offensively with a 91.9 (up 0.7 from Week 8)

 

Defensively though the Colts drop down to #16 with a 66.6 rating (down 0.6 from week 8)  So that is two games in a row where they defense is trending down.  No surprise to anyone that has watched the games.

 

Overall, the Colts fall to number 8 with a combined 158.5.

 

To relate it to the Titants game.  The Titans are 11th overall with a 156.5.  This is with a 25th rated offense at 79.4 and a 3rd rated defense at 77.0.

 

For a break down Colts O vs Titans D.

 

The AER average offensive rating is 84.2.  The Colts offense is 91.9 so the Colts are +7.7 points above the average.

The AER average defensive rating is 65.6.  The Titans D is 77 for a difference of +11.4 above the average.  

 

So the edge goes to the Titans D.

 

Colts D vs Titans O.

At 79.4 the Titans offense is - 4.8 below the average

At 66.6 the Colts are +1 above the average.  So this would should the edge to the Colts D.

 

Some other things to consider.  The Colts O has been trending up for several weeks now, while the Titans D has stayed pretty much the same.  On the other hand the Colts D has been trending down the last couple of weeks, while the Tenn O has been trending up.

 

How does this help you determine who is going to win?  It doesn't.  It shows that it should be a close, hard fought game.  If you look at just the numbers plus home field for the Colts, the Colts should win it.  But turnovers and ST play can negate any "numbers advantage".

[Coffeedrinker]

Here is the AER for Week 14

 

The Colts O has dropped over the past couple of weeks in efficiency, which is to be expected after the Jags game.  They are the 7th most efficient offense at 89.8 (ranked 8th in offense by yards/game)

 

On D they have been steadily rising, they are now 11th in the AER (11th in the NFL yards allowed as well) at 68.3.

 

So for the Dallas I like to compare the O vs D as it relates to the AER average.

 

Colts offense is at 89.8 which is 5.8 points above the offensive average of 84.0

The Dallas D is at 72.6 (3rd highest AER for the defense) which is 6.8 points above the average of 65.8.  So this would indicate a slight edge to the Dallas D.

 

The Colts D is at 68.3 which is 2.5 points above the average of 65.8.

The Dallas O is at 81.2 (21st in the AER) which is 2.8 points below the average of 84.0.  So this would indicate a pretty significant edge to the Colts D.

 

I have to admit, I was surprised to see the Cowboys D as high as it was and to see the Cowboys O as low as it was.  As most saw during the Eagles game, the Boys O is not very efficient but they do have some playmakers that can score quickly.  

 

Should be a fun game to watch.

[Larry Horseman]

8 minutes ago, Coffeedrinker said:

Here is the AER for Week 14

 

The Colts O has dropped over the past couple of weeks in efficiency, which is to be expected after the Jags game.  They are the 7th most efficient offense at 89.8 (ranked 8th in offense by yards/game)

 

On D they have been steadily rising, they are now 11th in the AER (11th in the NFL yards allowed as well) at 68.3.

 

So for the Dallas I like to compare the O vs D as it relates to the AER average.

 

Colts offense is at 89.8 which is 5.8 points above the offensive average of 84.0

The Dallas D is at 72.6 (3rd highest AER for the defense) which is 6.8 points above the average of 65.8.  So this would indicate a slight edge to the Dallas D.

 

The Colts D is at 68.3 which is 2.5 points above the average of 65.8.

The Dallas O is at 81.2 (21st in the AER) which is 2.8 points below the average of 84.0.  So this would indicate a pretty significant edge to the Colts D.

 

I have to admit, I was surprised to see the Cowboys D as high as it was and to see the Cowboys O as low as it was.  As most saw during the Eagles game, the Boys O is not very efficient but they do have some playmakers that can score quickly.  

 

Should be a fun game to watch.

 

I think it will be a good game - I've been thinking the Cowboys and Colts are very similar teams related to where they rank in the overall NFL - Dallas is stronger in some areas but we have a significant QB advantage. Both teams have also grown and developed over the course of the year. 

 

Promising the Colts are sitting at #7 in these rankings

[Coffeedrinker]

Final update:  AER 2018 Season

 

Colts ended the season #6 overall in the AER.

- Colts are #4 in offensive AER with a 90.8 (leage average 83.4)

- Colts are #9 in defensive AER with a 70.3 (league average 66.5)

 

Now this is why I like the AER and why I think AER is the best overall evaluation of how good teams are.

 

The Top 9 in overall AER are (in order):

Bears, Saints, Rams, Seahawks, Chargers, Colts, Chiefs, Patriots, Ravens. - All playoff teams.

 

The three outliers are the Texans (#13), Eagles (#14) and the Cowboys (#16)

 

#10, 11, 12 are the Steelers, Vikings and Titans.  All teams that were in the playoff hunt until the final whistle blew on the last game in that conference.  

 

The only team in the top half (16 or above) of AER with a losing record is the Packers at #15.

 

Teams ranked 17-32 in AER all have losing records.

 

The Cardinals are ranked #32 in overall AER.  They also have the worst record in the NFL. And the #32 ranked offense in AER.

 

Tampa Bay had the worst defense per AER with a 55.4, using the NFL yardage for ranking Tampa is #27