Blog Entry


Posted on: December 13, 2011 1:20 am

Week 14 brought us some separation as the top teams in the AFC (NE, BAL, HOU, PIT) all won and the Jets grabbed a Wild Card spot for now with a badly needed win since all 5 of NYJ's losses are conference losses.  And then there's DEN-BOW.  Seriously?  This story just keeps on getting more and more incredible each week!

On the NFC side, NYG pulled out a huge late win at DAL and secured the lead in the NFC East (DAL could have clinched division title in wk 15 if they had won).  GB and NO continued their positive push forward with NO gaining a playoff berth and GB getting a first round bye, but the field goal happy 49ers stumbled against ARI.  DET and ATL kept up their Wild Card positions, but the Tebow-ed CHI squad is now on the outside looking in.

Playoff clinching scenarios for Week 15 are below.  Looked at DET and ATL potential clinching scenarios, but since both teams are playing AFC opponents they could still end up with 6-6 conference records (not typical for playoff teams) and DET can still be caught by CHI and lost to ATL H2H and ATL lost to CHI H2H...both teams have to wait at least another week for playoff qualification.

Also...since the only scenario keeping PIT from having already clinched a playoff berth is a PIT-TEN-DEN Wild Card tie at 10-6 that goes to Strength of Victory, we looked at whether any combination of game results during Wk 15 could clinch that SOV for PIT and there is none.  So PIT must rely on the scenarios below.



Clinched: HOUSTON (AFC South Champ)
Eliminated: IND (Wk 12), JAC (Wk 13), BUF (Wk 14), CLE (Wk 14), MIA (Wk 14)  

NEW ENGLAND clinches division title with:
1) WIN
2) TIE + NYJ loss/tie
3) NYJ loss

NEW ENGLAND clinches playoff berth with:
1) TIE
2) CIN loss/tie + TEN loss/tie + OAK loss/tie

BALTIMORE clinches playoff berth with:
1) WIN or TIE
2) NYJ loss + TEN loss/tie
3) NYJ loss + OAK loss/tie
4) TEN loss/tie + OAK loss/tie

PITTSBURGH clinches playoff berth with:
1) WIN or TIE
2) NYJ loss
3) TEN loss/tie
4) OAK loss/tie
5) DEN loss


Clinched: GREEN BAY (NFC North Champ + 1st Round Playoff Bye), SAN FRANCISCO (NFC West Champ), NEW ORLEANS (Playoff)
Eliminated: MIN (Wk 12), STL (Wk 12), CAR (Wk 14), TB (Wk 14), WAS (Wk 14) 

GREEN BAY clinches home-field advantage with:
1) WIN or TIE
2) SF loss/tie

NEW ORLEANS clinches division title with:
1) WIN + ATL loss/tie
2) TIE + ATL loss


Since: Jan 8, 2010
Posted on: December 18, 2011 2:49 am



I truly appreciate the evolution of the H2H wildcard tiebreaker that you so elegantly layed out.  In many ways it allows us to to perform a forensic analysis on what the tie-breaking founders intent was for this H2H tiebreaker.  Your research on this matter further justifies that the intent has always been to compare teams based on best combined H2H winning percentage.

Allow me to explain.  You see the key piece to this puzzle is in the 1972 rewording.  I imagine some smart folks along the way shortly after the inception of the wildcard realized that comparing teams based on H2H pct would not be a fair metric in certain circumstances, particularily those cases when all teams had not played one another.  At the time and still to this day this problem does not arise when comparing H2H pct in the division.  I'm willing to bet that those in charge of the procedure realized the problem of not all teams playing one another, thus prompting the rewording in 1972.  But pay particular attention to the wording.  @thenflrules, I'll quote you here:

"Best percentage in head-to-head competition, when applicable.  This is applicable in a three-way tie only if one club beat both others or one club lost to both others."

These two statements are not compatible.  Their first sentence clearly outlines their intent to apply H2H pct when applicable.  But the second sentence is a false statement from a mathematical perspective do to the use of "only if".  They correctly realized that in situations where one team had beaten the other two or lost to the other two that you could apply H2H pct, but they over-reached in their proclamation that these were the only instances where H2H pct is applicable.  You see... H2H percentage is always applicable when all tied teams have played one another. We need to look no further than the Division H2H tiebreaker for proof of this, since a tied team in the division can win the H2H tiebreaker without the neccessity of a sweep.  So their definition of when H2H was applicable was incomplete from the inception of this 1972 rewording.  Born out of this definition of the H2H rule was the concept of a sweep being needed in order for H2H to be applicable at the conference level.

Eventually along the way the first piece of this rule was dropped and replaced with a more concise definition of the second sentence being "H2H sweep".  By dropping the first sentence the we've lost touch with the intent of the rule which is to apply Best percentage in head-to-head competition, when applicable.

Also, eventually along the way smarter heads prevailed and realized that H2H sweep is not the only instance in which H2H can be applied, which is why tiebreaking experts like Joe Ferreira have given us guidance to apply best combined H2H record when a full round robin exists.  And why would he suggest that we do that, because it's applicable just like how it is applicable when performing H2H tiebreaks within the division.

Regarding the relative rarity of when this H2H thing might come into play, I have not done an audit of every single clinching scenario and all possible ties that a team could encounter for every week of every season going back.  So it is impossible for me to say how rare or unrare H2H nauance really is or how often it could make a difference in the clinching scenarios.  That is less important to me than knowing we are correctly applying the tiebreakers consistent with everyone else's understanding of them. Having a season where a coin flip would be needed is rare too, but I still think we should have an understanding of when the coin flip should be applied.  I see this as being no different.  We should have proper understanding of when H2H pct is applicable.  An lastly we should cement that understanding into official rule.  The fact that our persistence for improving the procedure led to the updated note on when to revert is promising.  The discussions that we have here are worthwhile and meaningful.

Also congrats to Brett on the new degree, wish I was in a position to offer you an awesome job, but i know you'll find something.  Best of luck to you.


Since: Dec 12, 2006
Posted on: December 18, 2011 12:39 am


Brett, congratulations on your graduation and I wish you much success!

With the inception of the wildcard in 1970, the head-to-head tiebreaker to determine the fourth team to qualify for the playoffs was originally written as follows:

"Best won-lost percentage in those games played between or among the involved clubs"

The following year, this was added:

"Best won-lost percentage in those games played between or among the involved clubs; providing, however, that this clause shall not be applicable unless each of the involved teams has played at least one of the other involved teams."  Further, it said, "If A, B and C are tied and team C has played neither A nor B, head-to-head cannot be applied."

Then came the clarification the League made in a press release I mentioned earlier that came out during the 1972 season.  In my research, this was the first instance I had found that mentioned a sweep:

"Best percentage in head-to-head competition, when applicable.  This is applicable in a three-way tie only if one club beat both others or one club lost to both others." 

Since then (and this now being the 40th season), the tiebreaking language for H2H to determine the Wildcard teams has not changed and ALWAYS mandated that it be a sweep (a team defeated all others in the set and advanced OR a team lost to all others in the set and dropped out).  In fact, in looking over clinching scenarios over the years I have not found any instance where a team would advance on the H2H tiebreaker for three or more clubs from different divisions based on H2H winning pct. or a H2H round robin.  If there was no sweep, you moved on to the next step. 

I still would like to know what the "framers" would say about 4-way ties...

In those days we could only have a 4-way tie in a division (or possibly even a 5-way in the AFC East or NFC East).  Since 1970, for division ties, it's always been best percentage in games among the clubs.  You could only have a three-way tie before applying the conference/wildcard tiebreakers.  Much like we do today, whenever two or more clubs were in the same division, we applied the division tiebreakers first to narrow the field to one club per division.  It wouldn't be until 1978 when 4-way, 5-way ties, etc. were possible to determine the Wildcard teams.  And with there now being two wildcards, as I had mentioned previously, in those multiple team ties, whenever one or more teams dropped out in a tiebreaking step, according to the original framers as noted in the League by-laws, you always reverted.  



I've posted before about a Miami-Buffalo tiebreaker (1998?) that came down to point differential. Do you have the details on that one?
At quick glance, it looks like Miami won that tiebreaker but Buffalo had a better common games record? (tied h2h, div, and conf)

Brett, sorry I'm just getting to this now.  This is what I have for 1998 MIA/BUF:

1) H2H: 1-1
2) Div: MIA 4-4, BUF 4-4
3) Conf: MIA 7-5, BUF 7-5
4) CG: MIA 6-4, BUF 6-4
5) Div Net Points: MIA PF 148-PA 142=6, BUF PF 158-PF 158=0

I think I was thinking of the Seattle-Dallas SoV tiebreaker in 2003 that was decided by only 3 games. I imagine that was decided on the final weekend and therefore it was possible that one of the SoV determining games was played on Sunday night (though it obviously didn't come down to that final game).


You are correct.  DAL had clinched a playoff berth in Week 16.  SEA then clinched in the final week and won SOV tiebreaker over DAL for the fifth seed after the results of the Week 17 afternoon games and did not come down to PIT at BAL game on Sunday night.


Since: Nov 21, 2011
Posted on: December 18, 2011 12:15 am


Btw, Just to point out, the 1 in 6.2 million results in it being 50% likely every 4 million years approx, and 90% likely about every 15 million years. Almost exactly once in a lifetime...

Since: Nov 21, 2011
Posted on: December 17, 2011 11:48 pm


Brett: Great thinking! Trust me that I'm good enough with the math; I just lost my head again and did it for division not 3 way tie in my "method". Division may be 1/100,000 (though with that mistake I'm not sure what else I messed up), but that's irrelevant as it just goes to H2H record. Thinking it through is definitely the hard part. I'm not certain how you arrived at the odds for each part, but I do think all the parts come together properly. Complex stuff! I guess I'm losing a timeout, huh!

Jeff: I would certainly say yes, but since it is Saturday night, I think I should wait till after tomorrow's games and you should do a new one. I'll gladly check that. 

Since: Nov 20, 2007
Posted on: December 17, 2011 7:18 pm


All I can say is WOW to the last few pages of the Week 15 Scenarios blog.  I read it all(of course), as I read all the entries, but WOW indeed.  Intense. 

And just to go back to some other stuff that we talk about at times here on the blog.

theNFLrules said:

Also, you'll notice that the book does not include the 2012 Opponents.  Those were released in October to all clubs and was updated in the pdf version as well.  If you need those, I posted the 2012 Opponents in the "Week 12 Playoff Scenarios" thread.  It's on page 2.
Ivan put this out earlier this year, after I requested it and thanks again very much Ivan.  The only things that were missing from the 2012 schedules was the non-common opponents schedules for the teams from the same divisions.  Teams from the same divisions always have 12 common games at season's end and 2 non-common games and of course 2 head to head games.

We now have 1 game scheduled next year(of 32) for the non-common opponents games.  The AFC North and AFC South aren't scheduled to play each other next year(they are all playing each other this year), but the Cleveland Browns and Indianapolis Colts have each locked up 4th place in their respective divisions and thus will be playing next year in a non-common opponent game.

The AFC South teams are hosting the AFC North teams, in the non-common game matchups, in 2012 and thus we have the 1st non-common game scheduled for 2012 and it's.......drumroll please......the Cleveland Browns @ Indianapolis Colts.  It's therefore the 2nd season in a row that the Colts will be hosting the Browns.  I am sure we'll see quite a few more non-common game matchups determined after this weekend's game are over.

Giants fan:

You corrected me and let me realize that the G-men can only get as far as the 4 seed.  Thank you for pointing that out.  So, can you do me a favor and tell me if all the other "highest" seed potentials that I had listed, were correct or incorrect?  That's not including the Thursday game, but from the actual end of Week 14 standings.  Thanks.................Jeff

Since: Jan 8, 2010
Posted on: December 17, 2011 2:54 pm


Noob, you said "Imagine all four teams within a division tied at 8-8. Team A is 4-2 in the division, but 0-2 vs. D Teams B and C are 3-3 Team D is 2-4, but 2-0 vs. A Now  A advances, although there is a sufficient mathematical basis that A is inferior to D."... The key thing to remember is that all the tiebreakers are trying to assess who best among the collection of teams not who is dominant over each team in the collection individually. The only tie-breaker that sort of deviates from that philosophy is the H2H sweep and the only reason for that is because in many cases there isn't going to be a mathematical basis for comparison. So the definition of a sweep guarantees that the it only gets applied when there is sufficient mathematical basis to say that one team is better than the rest. The other key thing to remember is Best COMBINED record is what matters at the H2H level. I have no problem with Team A advancing over team D even though Team D has the better direct H2H record against Team A. Team A is still the best Team when looking at the collection as a whole. And granted if the collection size were to ever be reduced to just Team A and Team D then team D ought to advance before Team A. This is why we undergo the practice of reverting. Best COMBINED is more of a philosophical debate then one of illogical appliction. The powers that be have decided that what matters most at the H2H level in a division is COMBINED record since in the division a full round robin will always exist and therefore mathematical basis always exists. They could have decided that the Team that sold the most Hotdogs would be the winner. Fortunately cooler heads realized that COMBINED record was a much more fair metric, even though there could be instances where an advancing team might be individually worse by said metric. Remember always try to decide who is best among the collection first. If the collection size reduces to two teams eventually then so be it. If it doesn't then also so bet it. I hope that helps. -Cheers -Jerry

Since: Nov 6, 2007
Posted on: December 17, 2011 2:45 pm


I know this is random, and unrelated to Week 15, but what do you guys think has a better chance of happening??

A. Every team in the NFL finishes 8-8
B. 3 or more teams are tied for a spot in the playoffs, and are tied in every tiebreaker, and they are forced to go into a 3 or more team Coin Toss

Before you say (definitely B), think about it, those 3 teams would have to not only be tied in conference record, common games record (or not enough games) SOV, SOS, but they would also have to be tied in, ranking in points scored vs points allowed (both nfl and conference), net points (in conference games), net TD's, and net points in all games, that would just be ridiculous 

Since: Dec 18, 2008
Posted on: December 17, 2011 2:01 pm


nygsb42: Not convinced by your method. lol.

Here is what I did: although it admittedly has a huge gaping hole in that it assumes the scheduling format will stay the same for next 6.2 million years! lol.

Note: each subsequent probability in the string of probabilities is determined/estimated given that everything above it (or to its left) has already occurred.

P(1-0-1, 1-1-0, 0-1-1) = a*b*c*d*e*f*g, where

a = P(a year in which there is at least one tied game)
b = P(3 teams, from different divisions, are tied for a Wild Card spot, seeds #1 to #3, or seeds #2 to #4)
c = P(a complete set of 3 head-to-head games among the 3 teams)
d = P(the tied game is among the set of 3 head-to-head games)
e = P(a second tied game exists)
f = P(the second tied game is one of 14 qualifying games)
- ie. the team not involved in the head-to-head tied game has 14 games vs. teams outside of the 3-way tie.
g = P(no head-to-head sweep)
- ie. the head-to-head records are not [2-0-0, 0-1-1, 0-1-1] or [1-0-1, 1-0-1, 0-2-0])

P(1-0-1, 1-1-0, 0-1-1) = a*b*c*d*e*f*g
P(1-0-1, 1-1-0, 0-1-1) = (1/5)*(1/5)*(1/16)*(3/256)*(1/5)*(14/255)*(1/2)
P(1-0-1, 1-1-0, 0-1-1) = approximately 1.6 x 10^(-7) = 0.00000016
1/0.00000016 = approximately 6.2 million. Therefore, for a given year, the odds are 1 in 6.2 million.

Since: Dec 18, 2008
Posted on: December 17, 2011 1:06 pm


This is what I think is going on with Joe, Elias, etc. regarding the head-to-head stuff.

When Joe said last year (and/or 2 years ago) that a complete set of head-to-head games would qualify in the Wild Card tie-breakers - he was speaking from his personal experience of roughly 10 years handling the league's tie-breakers. The head-to-head complete set for Wild Card tie-breakers was one of our "logical inferences" that we agreed must be made, despite the language not specifically referencing it.

Side note:
For a 3-way tie within a division, head-to-head records of [2-1-1, 2-2-0, 1-2-1] qualify.
That, to me, is completely analagous to the [1-0-1, 1-1-0, 0-1-1] case for a 3-way Wild Card tie, which now we are being told does not qualify.
Elias, on the other hand, is an organization that is hired to follow the rules as they are written. Thus, sometimes Elias's interpretation may not exactly match up with Joe's. I'm guessing it was pressure fom Elias to get the "reverting language" clarified. Otherwise, they may be stuck in a pickle when the language tells them to do one thing and the gurus tell them to do another.

If we have a question about the tiebreaking rules, it should be directed to the original framers (the people that thenflrules mentions). If nothing else, the framers should at least be consulted on any tiebreaking issue that arises. After that, it is up to the Competition Committee to decide if a change to the written document is warranted. A change could be to...
a) Correct the langauge, so that it more accurately reflects the true procedure as it is commonly implemented,
b) Correct the language, so that it more accurately reflects the intentions of the framers,
C) Modify the language, so that it says what the Competition Committee decides is in the best interest of all the NFL teams.

Once the Competition Committee reaches a decision, it must pass a vote by "the membership." I'm quoting Joe here because I don't understand what "the membership" is or what it consists of. Now that I think of it, I also do not know who is on the Competition Committee. Here is that quote from Joe:

On Dec 1 [2010] at 2:22am, Joe said...

II'll check with them again to see if the League should change the tiebreaker language to reflect change from 3 division conferences to 4 division conferences.  That requires a recommendation from the Competition Committee and vote by the membership.

thenflrules: I love that you found the following from 1972!

I had looked at a 1972 League press release which clarified the 1st tiebreaker for conference/wildcard tiebreakers, "Best percentage in head-to-head competition, when applicable" which also added:  "This is applicable in a three-way tie only if one club beat both others or one club lost to both others."  No ties were to be included.  If there was no sweep, we just moved on to conference record.

Back then these ties could include teams from the same division such that a pair of teams in the 3-way tie could have two head-to-head games. If that is true, then head-to-head records could have been 2-1, 1-1, 0-1 (A splits with B, A beats C, and B didn't play C). Then, according to the above statement, team A wins the Wild Card despite no mathematical basis for saying that team A performed better than team B.  I think my point is that there will always be a case that simply is not thought of when writing down the rules (or not checked afterwards to make sure the rule, as written, handles the case properly).

I still would like to know what the "framers" would say about 4-way ties, since these ties have the 2-1, 2-1, 1-2, 1-2 (head-to-head) case involving no tied games. (The note above, from 1972, could refer only to 3-way ties.)

I also would like to know if the "framers" actually discussed the 1-0-1, 1-1-0, 0-1-1, 3-way case. To me, it's not evident one way or the other.


Jerry and anyone else upset over this head-to-head confusion should relax knowing that the odds of these scenarios ever occuring are astronomical. Jerry, how hard could it be to modify the code? Search for the relevant string of code and for every occurrence, comment it out! Next time, if you are unsure of a rule, draft 2 separate versions of the code. (sorry if that sounded harsh)

Keep in mind that neither Elias nor Joe (nor anyone else) has never made the answer to our question OFFICIAL - due to these examples never having actually occurred! As Noob and nygsb42 have alluded to, they may never occur! Even the 2-1, 2-1, 1-2, 1-2 case would be something like 1 in 3400 years!

Since: Dec 26, 2009
Posted on: December 17, 2011 12:14 pm


We all think that is illogical, including Jerry. Jerry was not saying the 1-0-1, 1-1-0, 0-1-1 case should not qualify. He was saying your hypothetical 4-way case with [2-0-1, ...] with an incomplete set of head-to-head games should not qualify.

Imagine all four teams within a division tied at 8-8.
Team A is 4-2 in the division, but 0-2 vs. D
Teams B and C are 3-3
Team D is 2-4, but 2-0 vs. A
Now  A advances, although there is a sufficient mathematical basis that A is inferior to D.

But in case of an incomplete set of four teams tied within in a conference, where A can be proven to be at least not inferior to any other team involved, they should not advance. That's where I don't get it.

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