Blog Entry

NFL Tiebreaking Rules and Procedures Discussion

Posted on: January 3, 2012 1:51 am
 

Due to incredible demand (okay...maybe it's just a few vocal individuals)...I'd like to float this blog entry to discuss NFL Tiebreaking Rules and Procedures.

Discussion topics should include:

- Any questions on the Rules and how they are applied in figuring out NFL scenarios prior to the end of the regular season (my job for last 18 seasons)

- Historical Anecdotes on NFL Tiebreaking Rules and past examples of the application of the Rules on Prior seasons

- Potential ways for the NFL to improve the Tiebreaker Rules

- Why can't we all just get along?  (kidding...sort of)

- Anything else that comes up NFL Tiebreaker, Draft order, Scheduling Formula related


Hopefully we'll get our regulars here along with some new voices to chime in...and I will check here as often as possible to answer any questions directed at me.


Let the party begin!!! 

 

Comments

Since: Dec 20, 2010
Posted on: January 6, 2012 10:20 am
 

NFL Tiebreaking Rules and Procedures Discussion

Okay not gonna belabor my point after this, but just a few quick thoughts from Brett's response:

Not so, Pillbox. Oakland was 1-3 vs. the AFC East, while Denver and San Diego were each 2-2. Oakland would drop out, then Denver would win the AFC West with a 2-2 record vs. the NFC North. (San Diego was 1-3 vs. the NFC North.)


Okay I WAS wrong about Oakland (which I believe agn pointed out in the earlier thread).  But I think you missed my point: I don't think interconfernce games should ever be used, period, in the tiebreaking steps ... so the NFC North records are irrelevent. So San Diego would have won the divsion since --- as you stated -- each were 2-2 against the AFC East but San Diego was 2-0 in their other AFC games and Denver was 1-1. As I stated -- San Diego --at the start of the season -- had a supposed 2nd-place schedule in non-common games, Denver had a supposed 4th place schedule in non-common games. So currently, Denver is rewarded for performing poorer against what the NFL deemed was a weaker schedule at the start of the season. I get that at season's end, the actual records of the non-common teams were better for the Broncos' opponents! But that's not my point. They still did not take advantage of their supposed easier schedule, and yet were rewarded for that. Remember back in the day when "the 5th place schedule" was talked about so much? It's not nearly so lopsided now, but it's certainly something that's been brought up in the past.

Now I totally get yours', Joe's, Vito's, Matt's rationalization for the common games and why they are used -- I am just looking at it through a diffeernt prism. I could give other reasons about why I think interconference games should not be counted (ie... they play only once every 4 years, and visit the opposing team's stadium once every eight years), etc. etc. It's just my opinion. It's pretty clear it's not going to change.

- You claim that the subset of games increases at each tiebreaking step. However, this is not the case due to common games being a larger subset than conference games. So, how can you simultaneously claim that the inclusion or exclusion of head-to-head games is manipulated to fit this paradigm? - a paradigm which doesn't exist! 

Again, I think you are missing my point! First off, I never claimed one thing or another. I was just reacting to Joe's posting of "14 common games" and then, later, his explanation of Denver's division win with a better record based on "10 common games." I wasn't the one simultaneoulsy exlcuding and including them!
But even given Joe's later explanation, you can't argue that division games aren't [always] counted in determining the common games record, but are *always* counted in determining conference record.
Can you?


 



Since: Dec 18, 2008
Posted on: January 6, 2012 8:00 am
 

NFL Tiebreaking Rules and Procedures Discussion


Wildcard Team A  10-6-0
Wildcard Team B  10-6-0
Division Leader C 10-6-0
[Brett said]
It bears mentioning that this can only be true if Team A and Team B are from different divisions. (Division Leader C could be from any of the 4 divisions). Then team B could use Division leader C to leapfrog team A.
I should clarify further:

Assuming that Teams A and B are from different divisions, then:

- If Division Leader C is in the same division as A, then B can use C to leapfrog A.
Team A could hold tiebreaker over C, but not over Team B.

- If Division Leader C is in the same division as B, then A can use C to leapfrog B.
Team B could hold tiebreaker over C, but not over Team A.

- If Division Leader C is not in either A's or B's division, then A or B can use C to leapfrog the other.
Head-to-head between A and B might not apply with C in the mix.





Since: Dec 18, 2008
Posted on: January 6, 2012 7:36 am
 

NFL Tiebreaking Rules and Procedures Discussion


Brett said:
1)  Jerry's 6 "other procedural clarifications" each need a header that alerts the reader in what situations the clarification is needed.
Jerry responded:
Not necessary.  Each clarification is obvious where it is needed.  Can you give me an example of one of those 6 clarifications that a user would be confused as to which steps it is applicable to.  In my opinion after reading each clarification, I see know way in which there could be confusion as to which steps the said clarification should be applicable.
It's my belief that the casual fan should be able to look at the tiebreaking steps and apply them, without having to read and digest the entire document. When I made my comment above, I hadn't yet read any of the 6 clarifications, nor did I feel like I wanted to read any of them unless they pertained to whatever tiebreak situation I was currently looking at.

Here are my suggestions for the Other Procedural Clarifications (henceforth refered to as OPC's).

OPC #1: To me, this is essential (not "other"). I think it should be located immediately before step #1. One cannot apply the steps properly without the knowledge contained here.

OPC #2: Also essential to me. This and OPC #1 are the only 2 things I have above my tiebreaking steps. Everything else, to me, is secondary or "other". Additionally, OPC #2 is more of an explanation for your note (2) above the tiebreaking steps. It does not instruct how to properly implement the tiebreaking steps (unless note (2) is removed).

OPC  #3, #4, #5, and #6 each apply to a single tiebreaking step. They each should have a header indicating such.
Other Procedureal Clarifications:

a) Procedure reverts to step (1) when one or more teams are eliminated:
...

b) Winning Percentages are used for all records
....

c) Strength of Victory:
...

d) Strength of Schedule:
...

e) Ranking in Points Scored and Points Allowed:
...

f) Coin Toss:
...
The headers make this section easier to read, and alert the reader to what he or she may need to read more carefully at a later time.


Brett said:
2)  Also not a fan of the 2 large paragraphs preceeding the tiebreaking steps.
Jerry responded:
It ought to be this way.  I layed it out this way because I don't like how in the existing procedure you have to look all over the place to gain insights into how to apply the procedure. ...
I agree that in the existing procedure you have to look all over the place to gain insights into how to apply the procedure, and I agree that this can be fixed. However, I disagree with placing all this information ahead of the tiebreaking steps. I think we can fix this with better organization of the information and the usage of headers to direct the reader where to find the needed information.

Brett asked:
3)  Jerry, why on earth do you still have this: "The original seeding within a division always remains the same for all subsequent applications of the procedure that are necessary to identify the Wild-Card participants"?
Jerry responded:
.... For now, it is still there as an insurance policy to ensure that a user never thinks to re-rank the remaining tied division teams simply beacuse another team within that same division has been awared a Wild-Card spot. ...
I suppose an insurance policy is not a bad thing. However:

An underlying (perhaps hidden) principle in the tiebreaking procedure is to "re-rank" the remaining tied division teams. Allow me to explain:

When we rank 3 or more tied teams in the same division, Joe is quick to point out that we are not "re-ranking". However, in the one and only "ranking" of the teams, we are "re-applying the tiebreaking steps" each time a winner is determined. To me, "reapplying the tiebreaking steps with the remaining unranked teams" and "re-ranking the remaining teams" are essentially the same thing.

Therefore, to include such phrasing as that above, indicates to me that teams are intitially ranked in some way other than the way that they actually are. Thus, to me, it is a very misleading statement that should be removed altogether.


Brett asked:
Lastly...are we focused on cleaning up the existing language, or are we focused on implenting new/changed steps?
Jerry responded:
I don't want to speak for others, but my goal is to get this version of the procedure approved, which should in effect cleanup all the existing language hiccups of the current procedure as well as cement the H2H rule to account for group sweeps and round robins in the absense of a sweep.
I think we have a better chance at accomplishing these two goals separately. As well, I think our best bet is to seek approval for each new/changed step separately. If we gain the trust of the Competition Committee, then each additional change should be easier to accomplish.

After thinking about this, I think right now we should be discussing everything and anything. Then, when all ideas are exhausted, we should decide what we care about most and first seek approval only for that. (For me, this would be "cleaning up the existing language".)



Since: Dec 18, 2008
Posted on: January 6, 2012 6:05 am
 

NFL Tiebreaking Rules and Procedures Discussion



Jerry said:
I know a lot of you might think it would be common sense to not include the division leaders when breaking ties amongst wild card teams but [including the division leaders] actually makes a lot of sense.  It's possible that you can get different results by including the Division Leaders in the process.  For example suppose you have two teams wild card eligible and one division leader all with the same record.  

Wildcard Team A  10-6-0
Wildcard Team B  10-6-0
Division Leader C 10-6-0

...

It bears mentioning that this can only be true if Team A and Team B are from different divisions. (Division Leader C could be from any of the 4 divisions). Then team B could use Division leader C to leapfrog team A.

Jerry brought this up last year and because of him I included "non-division winning" or "non division winner" when necessary in my version of the language. I think it's a very valid point which happens to be very easy to address properly in an update to the language.

Jerry said:
Vice versa, we should also be explicitly instructing users to exclude non-division-winners when seeding the division titlists.
This is first time this has been brought to my attention (or the first time I've noticed it mentioned by anyone). I am still digesting it, but, as of now, I'm not convinced this is necessary. It's true that including non-division-winners could change the seeding among division winners, but it just doesn't make sense to me to include a team in a tiebreak process that would just be "thrown out" if selected.

I feel the same way about including division winners in a Wild Card tiebreak process - to me, it just doesn't make sense to do this. But, I still include "non division winner" in my version of the language. Without these words, I don't believe the sentence in which the words appear would be correct. Compare:
a) Amongst a group of teams tied for a wildcard spot, select only the highest ranking team (but non-division winner) from each division. Only then, proceed to step (1).

b) Amongst a group of teams tied for a wildcard spot, select only the highest ranking team from each division. Only then, proceed to step (1).

Choice (b), on the surface, sounds like it's open to selecting division winners before proceeding to step (1). On the other hand, if a team is "tied for a wildcard spot", then I think by definition this team cannot be a division winner. So "(but non-division winner)" may, in fact, be unnecessary.

Jerry said:
[Other Procedural Clarification #6, Coin Toss:]
If the coin toss is needed to break ties among 3 or more clubs, each club will have one league assigned representative flip a coin at the same time as the other representatives. The winner is the club whose coin flip result is different from each of the other clubs.  If this condition is not met, such as HHH or TTHH, then this process is continually repeated until a winner is reached.
I think your idea is a great one. Here's another idea: a lottery system where each tied team has exactly one lottery ball in a lottery tank. Or, does anyone remember Jeff's "scramble" idea? I think it was Jeff, the Colts fan (not the Packers fan).







Since: Jan 8, 2010
Posted on: January 6, 2012 5:25 am
 

NFL Tiebreaking Rules and Procedures Discussion

Brett,
1)  Jerry's 6 "other procedural clarifications" each need a header that alerts the reader in what situations the clarification is needed.
Not necessary.  Each clarification is obvious where it is needed.  Can you give me an example of one of those 6 clarifications that a user would be confused as to which steps it is applicable to.  In my opinion after reading each clarification, I see know way in which there could be confusion as to which steps the said clarification should be applicable.

2)  Also not a fan of the 2 large paragraphs preceeding the tiebreaking steps.
It needs ought to be this way.  I layed it out this way because I don't like how in the existing procedure you have to look all over the place to gain insights into how to apply the procedure.  Those 2 paragraphs pertain to the flow and applicability of the procedure, which in my opinion is higher priority than the steps themselves.  Those 2 paragraphs very quickly educate a new user on all the ways the procedure shall be used. One of my goals in drafting this consolidated procedure was to reduce the need for the user to have to read ahead to understand how to even use the procedure, as is currently the case with seeding the Division Leaders.

3)  Jerry, why on earth do you still have this: "The original seeding within a division always remains the same for all subsequent applications of the procedure that are necessary to identify the Wild-Card participants"?
The wording here is not identical to the existing procedure.  Nonetheless, I too considered removing this clause.  For now, it is still there as an insurance policy to ensure that a user never thinks to re-rank the remaining tied division teams simply beacuse another team within that same division has been awared a Wild-Card spot.  Though, I am working on some draft re-wordings of this note.

Lastly...are we focused on cleaning up the existing language, or are we focused on implenting new/changed steps?
I don't want to speak for others, but my goal is to get this version of the procedure approved, which should in effect cleanup all the existing language hiccups of the current procedure as well as cement the H2H rule to account for group sweeps and round robins in the absense of a sweep.  I don't really have an opinion at this time regarding the Common Games / Conference Games step order.  I haven't seen or recall a very strong arguement to change those existing steps.  I know we've talked about Common / Conference a lot but to be honest I'm not even sure what everyone's stance is on the issue.  Are there any among us that are adament about making alterations to those steps?

-Cheers
-Jerry 



Since: Dec 18, 2008
Posted on: January 6, 2012 3:54 am
 

NFL Tiebreaking Rules and Procedures Discussion

A few replies to comments by Pillbox, Agn, and Vito at the end of the week 17 blog.

Pillbox said:
It seems to me (and I guess only me) that the way division games are counted is shoe-horned to fit the model that allows common opponents to supercede conference record in breaking divisional ties.
- You claim that the subset of games increases at each tiebreaking step. However, this is not the case due to common games being a larger subset than conference games. So, how can you simultaneously claim that the inclusion or exclusion of head-to-head games is manipulated to fit this paradigm? - a paradigm which doesn't exist!

- If, as you suggest, head-to-head games are excluded from division records, how would this affect your theory of "increasing subset size"? It doesn't seem to affect it at all.

- Lastly, what agenda would the Competition Committee have for puting common games ahead of conference record? (other than the one Joe pointed out.) In other words, for what agenda would they even consider manipulating the rules/definitions?

Pillbox said:
Oh, I LOVE agn's idea of counting conference common games, if common games have to be before conference record. I believe Oakland would have won the AFC West this year if his system was utilized.
Not so, Pillbox. Oakland was 1-3 vs. the AFC East, while Denver and San Diego were each 2-2. Oakland would drop out, then Denver would win the AFC West with a 2-2 record vs. the NFC North. (San Diego was 1-3 vs. the NFC North.)

Agn said:
For 4 teams 6 H2H games, 0 more common conference games, followed by either 4 more non-confernce games (as it is stands now, the eight common games are grouped), or 2 more un-common conference games.
You meant: "For 4 teams 6 H2H games, 0 more division games, 4 more common conference games, followed by...."

Agn said:
If one is "rewarded" for winning one type of game over another (H2H, division, conference), then that reward should be consistently applied. In the division TB common game step, a team is rewarded for winning a non-conf game ahead of a conf game, and this is inconsistent with every other TB step in the other processes.
This is a very keen observation. However, the only reason out-of-conference games are not counted when breaking ties between teams in different divisions is because each team has played an entirely different set of these games. Since teams in the same division play the same out-of-conference games, then, in my opinion, these games should be used to break a divisional tie (and used ahead of the 2 un-common conference games).

Vito said:
But if the scheduling formula changed, one could imagine non-conf games being common in a wc tiebreak.
Very astute. However, according to Agn's "consistency theory", this would be ok. Acceptable since conference records would still be compared before common games (in a wc tiebreak among teams from different divisions).








Since: Dec 18, 2008
Posted on: January 5, 2012 5:49 pm
 

NFL Tiebreaking Rules and Procedures Discussion

I haven't had time yet to read through everything and formulate all of my thoughts. However, I want to say a few quick things:

1)  Jerry's 6 "other procedural clarifications" each need a header that alerts the reader in what situations the clarification is needed.

2)  Also not a fan of the 2 large paragraphs preceeding the tiebreaking steps.

3)  Jerry, why on earth do you still have this: "The original seeding within a division always remains the same for all subsequent applications of the procedure that are necessary to identify the Wild-Card participants"?

Lastly...are we focused on cleaning up the existing language, or are we focused on implenting new/changed steps?




Since: Dec 27, 2006
Posted on: January 5, 2012 10:23 am
 

NFL Tiebreaking Rules and Procedures Discussion

    1. Head-to-head sweep (applicable if a group of one or more clubs has each defeated or each lost to each of the remaining group of clubs)
    2. Best combined head-to-head record among tied clubs (applicable only if each tied club has played each other tied club)
    3. If all tied clubs are from same division, best record in games played within the division.
    4. If all tied clubs are from same division, best record in common games.
    5. Best record in games played within the conference.
    6. If all tied clubs are not from same division, best record in common games  (applicable if each team has played a minimum of 4 games among common opponents)



Since: Dec 27, 2006
Posted on: January 5, 2012 9:52 am
 

NFL Tiebreaking Rules and Procedures Discussion

I think it should be clarified outside the list of steps itself that common games is only used once.
If the change I suggest is implemented, this become more clear.






Since: Nov 21, 2011
Posted on: January 5, 2012 8:56 am
 

NFL Tiebreaking Rules and Procedures Discussion

Database, great job with your rewritten procedures! I was actually able to read through it without a headache. And I must say, everything is pretty clear as far as I could tell. Two minor critiques/suggestions: first, I think it should be clarified outside the list of steps itself that common games is only used once. That way it will eliminate any confusion caused by consolidating into one list. Second, in explaining SoS, you use the word "matchups". I think "opponents" or "clubs" makes more sense in the context, and is the consistent language used in the procedure.


The views expressed in this blog are solely those of the author and do not reflect the views of CBS Sports or CBSSports.com