What's in a playoff team?

Suffice it to say, we at There's a Catch don't think much of unbalanced schedules. It is an ugly stain on Major League Baseball. The NFL, on the other hand, is stuck with it. So long as you only play 16 games in a 32-team league, you're going to have some teams getting an easier time.

But the fact that each team has to play six games against the three other teams in its division, and only ten games against other foes, really swings things askew. If we subtract off the 12-12 record that every division has within itself and look only at the 40 games that each division plays against outside opponents, we see that there is significant disparity:

Division	W	L	Pct
AFC East	23	17	.575
AFC South	22	18	.550
AFC West	22	18	.550
AFC North	21	19	.525
NFC East	20	20	.500
NFC North	18	22	.450
NFC South	17	23	.425
NFC West	17	23	.425

In other words, some divisions are better than others, and so some teams have more difficult intra-division schedules than others. Further, if you're the best team in your division, you get an easier schedule than the worst team. This is backwards from the inter-division games. (At least, how they are intended. Schedule difficulty is set based on the previous year's outcome, which occasionally backfires.)

It is important to remember that the 40 games each division plays are still unbalanced. However, averaging over four teams we achieve more balance than while looking at teams individually.

Why does schedule imbalance matter? Because it affects which teams make the playoffs. First of all, when choosing division winners, we consider overall record, which is clearly affected by schedule difficulty. This issue could be addressed by using only intra-division games to decide division titles. Second, wild card teams are also chosen using overall record, which is even more imbalanced since different teams also play different division schedules. This could be addressed by normalizing the influence of divisional foes. I'll discuss both of these possible solutions in more detail momentarily.

But first, what criteria should we use to choose playoff teams to begin with? This is a tricky question with many possible answers. The current NFL strategy, and for professional sports in general, is to honor the won-lost record above all else. (College football has a very different strategy.) Another possible answer is to choose the teams that have the best chance of winning in the postseason, whatever that means.

Rewarding won-lost record is, of course, the main reason why unbalanced schedules are problematic. Even with schedule balance, it can still be problematic, because teams play at different levels at different times, including because of injured players. Beyond these problems, it is also insufficient to determine the postseason field. In the NFL, teams often finish with the same record and tie-breaking rules have to be invoked to decide division winners or wild card teams. This year, Denver and Kansas City both finished 9-7 and tied vying for the second and final AFC wild card spot. The New York Giants qualified over three other 8-8 teams for the final NFC spot (Green Bay, Carolina, and St. Louis).

The rules for deciding tiebreakers are complicated to say the least. It seems absurd to expect players to take them into account when they're out on the field deciding how hard they need to try to win. They are also more or less arbitrary. Head-to-head record makes sense intuitively, especially if it shows up in a division setting where two games are involved. But considered negatively, if one team has the advantage head-to-head, then the other team has the advantage against the rest of the league. We can play this same game with any possible criterion.

Unfortunately, the only solutions I can propose for fixing the tiebreaker issue are rather silly:

1. Don't break ties involving playoff qualification. Expand the playoffs ad hoc to include all tied teams.
2. Break ties involving seeding with a coin toss.
3. I forget the other one.

However, I do have potential solutions for schedule imbalance, as I suggested above.

Division winners should be decided by intra-division games only. This throws away the imbalance created by playing different extra-division opponents. Note that, for the 2006 season, this would affect (only?) three of eight divisions, assuming overall record breaks ties. Tennessee (4-2) would have won the AFC South instead of Indianapolis (3-3), Carolina (5-1) would have won the NFC South instead of New Orleans (4-2), and, wildly, Arizona (4-2) would have won the NFC West instead of Seattle (3-3). More on this oddity later. Notice that all three of these teams failed to qualify even for the wild card.

At least at present, all members of a division play the same four opponents in the other conference (from a single division), and all four teams in another division in the same conference. Those eight games could also be used to determine division winners, if you think six games is not enough. I won't discuss this possibility further.

Wild card winners should be decided by subtracting off half the effect of each team's intra-division games. This would normalize the won-lost record to a total of 13 games against 13 unique opponents, rather than 16 games with 3 opponents played twice. Thus, teams with weak divisions, or with good match-ups against their division rivals, would lose some of their advantage, and teams with strong divisions or bad match-ups would gain. Again, this doesn't prevent the incidence of ties, which must continue to be broken or ignored some way or other. Here's how the standings would come out under this criterion (current division winners—not those under my proposal—are in bold):

AFC	W	L
San Diego	11.5	1.5
Baltimore	10.5	2.5
Indianapolis	10.5	2.5
New England	10	3
New York Jets	8	5
Denver	7.5	5.5
Kansas City	7	6
Jacksonville	7	6
Pittsburgh	6.5	6.5
Cincinnati	6	7
Tennessee	6	7
Buffalo	5.5	7.5
Miami	5.5	7.5
Houston	4.5	8.5
Cleveland	4	9
Oakland	2	11
NFC	W	L
Chicago	10.5	2.5
New Orleans	8	5
Dallas	8	5
Philadelphia	7.5	5.5
Seattle	7.5	5.5
St. Louis	7	6
New York Giants	6	7
Green Bay	5.5	7.5
Carolina	5.5	7.5
Atlanta	5.5	7.5
San Francisco	5.5	7.5
Minnesota	5	8
Washington	4.5	8.5
Tampa Bay	4	9
Arizona	3	10
Detroit	3	10

This would have put Denver and St. Louis into the playoffs in place of Kansas City and the New York Giants. Since the Chiefs and Giants both lost in the Wild Card round, this doesn't seem too bad. And there are no ties, instead of a two-way and a four-way one.

If we adopted both new rules, then the AFC would have Tennessee as a division winner, and Indianapolis and the New York Jets as the wild card teams, with the other three division winners the same. The NFC would have Carolina and Arizona as division winners and New Orleans and Dallas as the wild cards, bumping Seattle and the New York Giants from the playoffs. The Giants have already lost, so that's no big deal. But where did the Cardinals come from?

Arizona had a rough year, including an eight-game losing streak after winning their opener against the 49ers. Outside their 4-2 NFC West schedule, they were 1-9, including 0-4 against the AFC. But were they really that bad? If we look at their close losses, 16-14 versus St. Louis in week three could have had them 5-1 in the division. They also lost to Kansas City and Chicago by three points or fewer. All of their victories were by six points or more. So maybe they were an 8-8 team suffering from some very bad luck? In which case the postseason doesn't seem so far fetched. Or maybe they really were a bad team. Their only inter-division victory was against Detroit. They lost to lowly Oakland. (To be fair, they also got stuck with games against KC, Denver, and San Diego, the formidable remainder of the AFC West. The other three teams in the NFC West each also lost to the Chargers and Chiefs, but beat the Broncos.) But if the Cardinals were bad, why did they give the other NFC West teams so much trouble? I have no idea.