Mathmatical Margin for Error to make the playoffs

YosemiteSam

Unfriendly and Aloof!
Messages
45,944
Reaction score
22,322
I was thinking about when people say, a couple of early losses is not a big deal. This is something that just doesn't sit so well with me because in any season, you generally have a limited amount of games you can lose before you will miss the playoffs. Therefore every loss counts no matter when that loss occurs and to say otherwise is a bunch of hooey!

So, for every loss you have, it raises the importance of winning all unplayed games. I think I have the makings of general guideline to give people a generic idea of a teams margin for error to making the playoffs based on how many losses they have.

First thing I want to do is obviously point out that all situations aren't the same, so there is no perfect formula. The only thing we can do is get a guesstimation based on sound numbers.

Secondly, if you a anal-retentive statistical stickler; this isn't for you. If you want to complain how mathematically some things aren't accounted for, then just move along. This isn't about creating perfect numbers, it's about comprehending a teams situation. This isn't a percentage. (it doesn't add up that way) It's more of a margin for error.

Alright, the formula starts with a constant. That constant I've decided will be 10 and 6. 10 and 6 is what I consider the general season record required to be a playoff team. Yes, sometimes a better is required and sometimes even more losses will get you in, but we can't consider all possibilities when you're dealing with a constant. (obviously, you can adjust the constant to how you see fit)

If a 10 and 6 record is required; how does each loss affect your margin for error to make the playoffs? Well, that depends on how many games you have left to play. The more games you have left, the higher impact each loss puts on the rest of your games therefore lowering your margin for error.

We will start at the beginning of the season. You have 16 games remaining to be played and 6 acceptable losses to give.

The equation is Margin for Error = Total Acceptable Losses Remaining / Games Remaining.

1.000 means you've locked in a playoff spot. (ie, you can error (lose) the rest of your games)

So, 6 (Acceptable Losses remaining) / 6 (games remaining) = 1.000

Obviously, if you win more than 10 games, that Margin for error will go well over 1.000. I believe anything below .300 creates an extremely poor outlook for a team making the playoffs. Of course that can change if the team starts wining.

At the start of the season, 6 losses / 16 games remaining = .375 margin for error.

The Cowboys started 0-2. 4 / 14 = .285. That puts the Cowboys on the wrong side of the river. With the Win over the Texans, it barely righted the ship returning to a .307. If the Cowboys lose to the Titans, it will put them in a very bad spot at .250. Winning would put them at .333.

If the Cowboys go on a 3 game winning streak starting this weekend? .400. Losing streak? .100. Basically, if the Cowboys had 5 losses with 10 games to go. You should probably just start looking at next season.

Here is a a few different scenarios.
  • 8-0 with 8 games left: .750
  • 6-2 with 8 games left: .500
  • 4-4 with 8 games left: .250
  • 9-3 with 4 games left: .750
  • 9-5 with 2 games left: .500
  • 5-5 with 6 games left: .166

Once you've lost 6 games, no matter how many are left, you will always have a .000 margin for error.

Now, people say that early losses aren't that important and this in a sense actually suggests the same thing because early losses move the margin for error only slightly early on while losses later in the season move it in much larger chunks. I think what is missed here though is the fact that the effect of early losses start having a compounding effect on the margin for error as the season goes on. In that case, I think people only perceive that early losses aren't that big of a deal because they are seeing it from a distance. IMO, every loss counts and you will feel the effect of each loss as you move closer and closer to seasons end. In the end, a loss is a loss and it has the exact same effect no matter if it comes early or late. The only difference is the effect of an early loss is stretched out over the course of a season rather than applied in bulk all at once.

Anyway, just something I thought about on the train ride to work. I think it looks pretty good, but always looking to improve it. You see anything that would make this better?
 
8-8 might get you into the playoffs this year in the NFC. Typically you don't count on that, but the point for most people after the first two games was that we were just one game out with 14 games left, so you do have to, at some level, take into account what's happening around you.
 
This year 9-7 may win the East. With the winner the only rep from the East.
 
AdamJT13;3599930 said:

If there is one thing I've learned, it's that probability is a complete fail when applied to the NFL.

As I tell everyone, wagering on the NFL is foolish. The term, "on any Sunday" is a staple of truth. Anything can happen. Probability predicts the most likely to win the Superbowl and many times those teams don't even make the playoffs. Probability, predicts who the worst teams will be and many of those are the ones that end up in the playoffs. This is why I didn't even attempt to create actually probability charts for the NFL. The only thing probable is, the improbable.

Margin for error doesn't predict anything, it defines how steep the hill a team must climb is.
 
nyc;3599910 said:
I was thinking about when people say, a couple of early losses is not a big deal. This is something that just doesn't sit so well with me because in any season, you generally have a limited amount of games you can lose before you will miss the playoffs. Therefore every loss counts no matter when that loss occurs and to say otherwise is a bunch of hooey!

So, for every loss you have, it raises the importance of winning all unplayed games. I think I have the makings of general guideline to give people a generic idea of a teams margin for error to making the playoffs based on how many losses they have.

First thing I want to do is obviously point out that all situations aren't the same, so there is no perfect formula. The only thing we can do is get a guesstimation based on sound numbers.

Secondly, if you a anal-retentive statistical stickler; this isn't for you. If you want to complain how mathematically some things aren't accounted for, then just move along. This isn't about creating perfect numbers, it's about comprehending a teams situation. This isn't a percentage. (it doesn't add up that way) It's more of a margin for error.

Alright, the formula starts with a constant. That constant I've decided will be 10 and 6. 10 and 6 is what I consider the general season record required to be a playoff team. Yes, sometimes a better is required and sometimes even more losses will get you in, but we can't consider all possibilities when you're dealing with a constant. (obviously, you can adjust the constant to how you see fit)

If a 10 and 6 record is required; how does each loss affect your margin for error to make the playoffs? Well, that depends on how many games you have left to play. The more games you have left, the higher impact each loss puts on the rest of your games therefore lowering your margin for error.

We will start at the beginning of the season. You have 16 games remaining to be played and 6 acceptable losses to give.

The equation is Margin for Error = Total Acceptable Losses Remaining / Games Remaining.

1.000 means you've locked in a playoff spot. (ie, you can error (lose) the rest of your games)

So, 6 (Acceptable Losses remaining) / 6 (games remaining) = 1.000

Obviously, if you win more than 10 games, that Margin for error will go well over 1.000. I believe anything below .300 creates an extremely poor outlook for a team making the playoffs. Of course that can change if the team starts wining.

At the start of the season, 6 losses / 16 games remaining = .375 margin for error.

The Cowboys started 0-2. 4 / 14 = .285. That puts the Cowboys on the wrong side of the river. With the Win over the Texans, it barely righted the ship returning to a .307. If the Cowboys lose to the Titans, it will put them in a very bad spot at .250. Winning would put them at .333.

If the Cowboys go on a 3 game winning streak starting this weekend? .400. Losing streak? .100. Basically, if the Cowboys had 5 losses with 10 games to go. You should probably just start looking at next season.

Here is a a few different scenarios.
  • 8-0 with 8 games left: .750
  • 6-2 with 8 games left: .500
  • 4-4 with 8 games left: .250
  • 9-3 with 4 games left: .750
  • 9-5 with 2 games left: .500
  • 5-5 with 6 games left: .166
Once you've lost 6 games, no matter how many are left, you will always have a .000 margin for error.

Now, people say that early losses aren't that important and this in a sense actually suggests the same thing because early losses move the margin for error only slightly early on while losses later in the season move it in much larger chunks. I think what is missed here though is the fact that the effect of early losses start having a compounding effect on the margin for error as the season goes on. In that case, I think people only perceive that early losses aren't that big of a deal because they are seeing it from a distance. IMO, every loss counts and you will feel the effect of each loss as you move closer and closer to seasons end. In the end, a loss is a loss and it has the exact same effect no matter if it comes early or late. The only difference is the effect of an early loss is stretched out over the course of a season rather than applied in bulk all at once.

Anyway, just something I thought about on the train ride to work. I think it looks pretty good, but always looking to improve it. You see anything that would make this better?


OK

Now go back and compile the results of every NFL season and compare them to this analysis and we can see if it has any validity whatsoever.


I do understand what you are saying but every single 10 and 6 team that has ever made the playoffs lost 6 games. Doesn't matter one bit when during the season they lost them, just that they lost them. Troy talked about this a bit during the broadcast. His point was that the majority of the time if a team starts out 0-3 it is because they are a bad team. Sometimes good teams have bad things happen.

Of course you want to keep your losses in the bank as best you can but the quality of the team is what gets you to the playoffs. Quality teams can and do get off to slow starts sometimes.
 
Any mathematics are irrelevant. If the cowboys start playing the way their are capable of playing (based on talent) then they'll not only make the playoffs, they'll win the division and probably get a 1st round bye.
 
Vtwin;3600020 said:
OK

Now go back and compile the results of every NFL season and compare them to this analysis and we can see if it has any validity whatsoever.


I do understand what you are saying but every single 10 and 6 team that has ever made the playoffs lost 6 games. Doesn't matter one bit when during the season they lost them, just that they lost them. Troy talked about this a bit during the broadcast. His point was that the majority of the time if a team starts out 0-3 it is because they are a bad team. Sometimes good teams have bad things happen.

Of course you want to keep your losses in the bank as best you can but the quality of the team is what gets you to the playoffs. Quality teams can and do get off to slow starts sometimes.

I agree and noted, this isn't a probability chart. You just can't apply probability to the NFL when it comes to wins and losses for a team, it just doesn't work most of the time. Teams don't live up to their probabilities which in turn skews Strength of Schedule for other teams. (which screws off probabilities even more!) It's a margin for error chart. All it defines (as I noted to Adam) is how steep the climb is for a team wanting to make the playoffs. It doesn't matter if they are playing teams that are 0-10 or 10-0. They still only have {X} games they can lose. Since you have {X} amount of losses left and {Y} amount of games left, this gives you the ability to break it down into a single value that determines the difficultly in the teams position to make the playoffs.

Again, it predicts nothing!
 
Rack Bauer;3600035 said:
Any mathematics are irrelevant. If the cowboys start playing the way their are capable of playing (based on talent) then they'll not only make the playoffs, they'll win the division and probably get a 1st round bye.

Right.
1990, Cowboys were 3-7. They win 4 straight to be 7-7. Only needed to win 1 game to make playoffs.
Aikmen gets hurt in Philly. They lose. Laufenberg starts against a horrible Falcons team. We lose.
at 7-9, we still have a chance. If the Saints lose to the Rams.
Monday night football. Saints losing, line up to kick a FG. They miss, but offsides against the Rams. They make the next try and win. Saints are in, we are out.

Thing is at 7-9, we would have played a very beat up and poor Bears team. And we probably would have had Aikmen back.

So stats do not mean anything to me, until that 8th loss. Then depends how many games left.
 
AdamJT13;3599930 said:
Interesting that in the example on that website, a 10-6 record for the Cowboys shows the chance of missing the playoffs at 22% while with a 9-7 record that percentage jumps to 62%.

So, basically, nyc has it just about right....the Cowboys really need to get to 10-6 if they want a good chance of making the playoffs.
 

Staff online

Forum statistics

Threads
474,019
Messages
14,507,010
Members
24,207
Latest member
TomGiantsfan
Back
Top