NFL Overtime

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
24,376
I just went through every game from 2016-2018, and there were 45 overtime games over the three seasons (including playoffs). Here were the results:

- 10 times (22.2%) the team who got the ball first in OT immediately drove for the GW touchdown
- 14 times (31.1%) the team who got the ball second scored either a FG or TD to win the first time they touched the ball
- 2 times (4.4%) the team who got the ball first scored a FG and then stopped the other team from scoring
- 15 times (33.3%) the teams traded multiple possessions before a winner was determined
- 4 times (8.9%) the teams couldn't score and ended in a tie

So only 22.2% of the games ended without the other team touching the ball. The team touching it second won 4 more times on their initial drive than the number of times the team won the game on the opening drive of OT.

Three seasons may seem like a small sample size, but it's 45 games, probably large enough to determine if it's a fair system. Given the actual results, it's hard for anyone to claim that the team that gets the ball first has an unfair advantage.
 

joe dokes

Member
SoSH Member
Jul 18, 2005
30,243
I just went through every game from 2016-2018, and there were 45 overtime games over the three seasons (including playoffs). Here were the results:

- 10 times (22.2%) the team who got the ball first in OT immediately drove for the GW touchdown
- 14 times (31.1%) the team who got the ball second scored either a FG or TD to win the first time they touched the ball
- 2 times (4.4%) the team who got the ball first scored a FG and then stopped the other team from scoring
- 15 times (33.3%) the teams traded multiple possessions before a winner was determined
- 4 times (8.9%) the teams couldn't score and ended in a tie

So only 22.2% of the games ended without the other team touching the ball. The team touching it second won 4 more times on their initial drive than the number of times the team won the game on the opening drive of OT.

Three seasons may seem like a small sample size, but it's 45 games, probably large enough to determine if it's a fair system. Given the actual results, it's hard for anyone to claim that the team that gets the ball first has an unfair advantage.
Their only answer is that its such an abomination for one team to score without the other getting possession is that "10 times is 10 times too many. A different plan will make that number zero."

(I'm surprised there were only two 1st drive FGs that held up to win).
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
24,376
Their only answer is that its such an abomination for one team to score without the other getting possession is that "10 times is 10 times too many. A different plan will make that number zero."

(I'm surprised there were only two 1st drive FGs that held up to win).
Yeah I was not sure what I'd find when I started doing the research. I was surprised too. And you may be right that that's the rejoinder. But if we are trying to create a FAIR system, well, clearly the data shows that this system is very fair and adequate. You're actually more likely to win the game the first time you touch the ball *if you get it second, not first*.
 

SirPsychoSquints

Member
SoSH Member
Jul 13, 2005
5,013
Pittsburgh, PA
This analysis seems to ignore the advantage to getting the ball 3rd, 5th and 7th instead of 4th, 6th and 8th. How do multiple possessions forths break down in terms of winning % for getting the ball first? How many of the ties had an uneven number of possessions?
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
24,376
This analysis seems to ignore the advantage to getting the ball 3rd, 5th and 7th instead of 4th, 6th and 8th. How do multiple possessions forths break down in terms of winning % for getting the ball first? How many of the ties had an uneven number of possessions?
Why is it an advantage getting the ball 3rd? If you get it first and fail to score, the advantage immediately shifts to the other team, who only needs a FG to win. And your question (while a good one) doesn't take into account field position or whatever. My purpose wasn't to examine every single variable that could come into play in the current OT structure. Just to really ask whether it's "fair". I mean, SOMEONE has to get the ball first.

For what it's worth, in those 45 games:

- The team that got the ball first won 21 times
- The team that got the ball second won 20 times
- There were 4 ties

So that's about as even as it gets, right?
 

DennyDoyle'sBoil

Found no thrill on Blueberry Hill
SoSH Member
Sep 9, 2008
42,284
AZ
I think when they were talking about the new rule, the idea was that the chances of winning would flip significantly to the second team unless the first team scored a TD. 22 percent seems a little higher than I would have expected but part of that might have to do with the combination of changing the rule and changing the presumptive starting point to the 25 in the last couple of years.

My view on changing overtime rules has always been pretty simple. You should try to make it a 50/50 proposition without dramatically changing the nature of the contest like the college rules do just a bit too much to me. I think with enough research and even experimentation, you could come up with a system where most of the times most teams would be in equipoise about whether or not take the ball first.

Just by way of example, suppose the winner of flip could chose either to take the ball on the 2 yard line and to start the drive on second down, or to make their opponent do so. I would think that would flip the balance significantly the other way and nobody would want to take the ball. By the same token, I think giving a team the chance to win with a kick off return and making the 25 yard line the presumptive start point is likely, over time, to sway things too much to the first team that has the ball. I respect that the numbers are close to 50/50 at this point but I have a feeling that the 25 is too much and that this is going to flip back.

In short, I think you could come up with the equipoise point. No kickoff but you have to choose to start on the 16 if you take the ball? Starts to get to be a harder choice, right? Somewhere, there's the right balance, whether it's the 15, 18, 22, whatever.

You could perhaps even go back to first score wins if you made the consequence of taking the ball significant enough. Like, first points wins, but if you choose to take the ball you start on the 15 at first and 15.

Just playing around, but something like that.

This analysis seems to ignore the advantage to getting the ball 3rd, 5th and 7th instead of 4th, 6th and 8th. How do multiple possessions forths break down in terms of winning % for getting the ball first? How many of the ties had an uneven number of possessions?
It starts to be a pretty small sample size but yes, for those games that go longer than two possessions there is likely to be an advantage to team that had the ball first in that they will always get a chance to win before the other team does in all subsequent series of alternating possessions. But it's so negligible as to not really be much of an issue if you come up with a system that otherwise generally produces 50/50 outcomes.
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
24,376
Well, maybe 45 games is too small a sample, but when the outcome is 21-20 with 4 ties, that's virtually 50-50 outcomes. So it's hard to argue that there's a real advantage given to the team with the ball first.
 

dcmissle

Deflatigator
Lifetime Member
SoSH Member
Aug 4, 2005
28,269
I’m ok with current regime.

I would play full periods in the playoffs.

“Oh — but you’re going to confer an advantage on whoever plays the winner!” Ok, but you put yourself in that spot by not being able to close out the game in regulation, so deal with it.
 

DennyDoyle'sBoil

Found no thrill on Blueberry Hill
SoSH Member
Sep 9, 2008
42,284
AZ
Well, maybe 45 games is too small a sample, but when the outcome is 21-20 with 4 ties, that's virtually 50-50 outcomes. So it's hard to argue that there's a real advantage given to the team with the ball first.
I understand your point, but I also don't think those who disagree are all just doing it because it has worked out twice in big games for the Patriots or whatever. It's not just fans but also teams themselves that perceive that there is an advantage to winning the toss.

Even if the numbers work out right over a relatively small period, it's still a problem if everyone wants to take the ball. Maybe it's a perception problem. Or maybe it's a small sample size problem and the teams really do know where the advantage is. Or maybe they are making stupid decisions.

As a fan, I really want my team to win the toss. This is especially true in a game like the Super Bowl or the AFCCG where I know the teams are likely to be more tired and where the adrenaline has worn off or the game is longer.

That's a problem worth trying to fix. If there is a simple change that can be made where the perception is that it's not really a huge deal whether or not you win the toss, then I don't have a problem with it. If it turns out teams are being stupid it will work itself out and the only ones that will be punished are the stupid ones.
 

SirPsychoSquints

Member
SoSH Member
Jul 13, 2005
5,013
Pittsburgh, PA
Why is it an advantage getting the ball 3rd? If you get it first and fail to score, the advantage immediately shifts to the other team, who only needs a FG to win. And your question (while a good one) doesn't take into account field position or whatever. My purpose wasn't to examine every single variable that could come into play in the current OT structure. Just to really ask whether it's "fair". I mean, SOMEONE has to get the ball first.
If each team fails to score, the team that got the ball first can win with a FG. It's definitely an advantage to get first crack at the win each set of possessions. I don't follow your point about field possession. If starting with the ball gives you superior field position on future drives on average, that would be an advantage. If starting with the ball gives you inferior field position on future drives on average, that would cut into the advantage of starting with the ball.

While someone has to get the ball first, that needn't necessarily be the recipient of a standard kickoff (bidding systems have been discussed, the college rules have been discussed, BB has suggested you simply continue from where regulation left off), the overtime needn't necessarily be sudden death (BB's suggested overtime of another 10 minutes, the college rules).

To answer my own question, of the 4 ties, 1 had equal possessions and 3 had more possessions for the team with the ball first.

Of the 15 multiple possessions games, 7 were won by the original ball receiving team in their 2nd possession, 4 were won by the kicking team in their 2nd possession, 3 were won by the receiving team in their 3rd possession, and 1 was won by the kicking team in their 3rd possession. This, the receiving team went 10-5 in these games, bringing their overall record to 22-19-4.

That's pretty damn close - winning percentage of only 53.7%!
 

SirPsychoSquints

Member
SoSH Member
Jul 13, 2005
5,013
Pittsburgh, PA
I see there was a lot of posting since I started replying - I basically agree with everyone in the thread. It's close enough for me, and I don't necessarily think everything needs to be 100%, it's more important to be 100% unbiased.
 

BigJimEd

Member
SoSH Member
Jan 4, 2002
4,432
In short, I think you could come up with the equipoise point. No kickoff but you have to choose to start on the 16 if you take the ball? Starts to get to be a harder choice, right? Somewhere, there's the right balance, whether it's the 15, 18, 22, whatever.

You could perhaps even go back to first score wins if you made the consequence of taking the ball significant enough. Like, first points wins, but if you choose to take the ball you start on the 15 at first and 15.
I think you are misreading the problem. Most complaints, virtually all that I've seen, are about teams NOT touching the ball and your solution is to make that happen more often.
Yes, you might change the calculus on wanting to win the coin toss but I think you make ther underlying issue worse.
 

InstaFace

The Ultimate One
SoSH Member
Sep 27, 2016
21,766
Pittsburgh, PA
What percent of NFL drives, in general, end in touchdowns? The last 3 years: 23.2%, 19.7%, and 21.7% in 2016.

How about in overtime?
- 10 times (22.2%) the team who got the ball first in OT immediately drove for the GW touchdown
The only thing surprising about this data is that it is surprising some people.
 

DennyDoyle'sBoil

Found no thrill on Blueberry Hill
SoSH Member
Sep 9, 2008
42,284
AZ
I think you are misreading the problem. Most complaints, virtually all that I've seen, are about teams NOT touching the ball and your solution is to make that happen more often.
Yes, you might change the calculus on wanting to win the coin toss but I think you make ther underlying issue worse.
I don't see that. If you moved the starting point of the first team to possess to the 16 or whatever then you'd significantly increase how often the second team touches the ball.

Edit: Also, there's my main point. The harder you make the choice for the team that wins the toss the more often the team that wins the toss is going to not take the ball, preferring instead to take the field position advantage and the opportunity to win with a FG instead of a TD. If you find the point at which teams are in equipoise about whether to take the ball or whether to start on defense, then teams will choose to start on defense half the time. And those teams can hardly complain, if they do, that they should have been given the chance to possess the ball since they were given the chance to possess the ball but chose not to.
 

Mystic Merlin

Member
SoSH Member
Sep 21, 2007
46,769
Hartford, CT
I’m ok with current regime.

I would play full periods in the playoffs.

“Oh — but you’re going to confer an advantage on whoever plays the winner!” Ok, but you put yourself in that spot by not being able to close out the game in regulation, so deal with it.
I’m supportive of this approach.

I don’t think it ever happens because: TV.
 

BigJimEd

Member
SoSH Member
Jan 4, 2002
4,432
I don't see that. If you moved the starting point of the first team to possess to the 16 or whatever then you'd significantly increase how often the second team touches the ball.

Edit: Also, there's my main point. The harder you make the choice for the team that wins the toss the more often the team that wins the toss is going to not take the ball, preferring instead to take the field position advantage and the opportunity to win with a FG instead of a TD. If you find the point at which teams are in equipoise about whether to take the ball or whether to start on defense, then teams will choose to start on defense half the time. And those teams can hardly complain, if they do, that they should have been given the chance to possess the ball since they were given the chance to possess the ball but chose not to.
Sorry I was confusing a couple posts and points. Yes on your suggestion teams may choose defense more often but I think in order to reach that point, you'd likely make it far more likely the team who touched the ball 2nd wins. Which I think would make it worse.
Also seems gimmicky and takes more away from the game the college system, imo.

I think, as stated above, if the goal is for both teams to possess the ball than they should go to full- quarters. Of course that could lead to some long games. You could limit it to one quarter and then go to the current rules. That would likely get rid of some of ther critics but not all.
You could also just do it for the playoffs.
 

BigJimEd

Member
SoSH Member
Jan 4, 2002
4,432
I’m supportive of this approach.

I don’t think it ever happens because: TV.
I don't think TV would complain too much if it was just the playoffs and if they limit it to 1 or 2 quarters max. I don't think you'd want players doing more than that anyway
 

DennyDoyle'sBoil

Found no thrill on Blueberry Hill
SoSH Member
Sep 9, 2008
42,284
AZ
Yes on your suggestion teams may choose defense more often but I think in order to reach that point, you'd likely make it far more likely the team who touched the ball 2nd wins.
So, isn't that basically saying that teams would do dumb stuff against their own self interest? I tend to think that actually teams would figure it out.

Or, more likely, what would happen is the league would pick the starting point and see what Belichick did and if he started deferring then everyone would. I'm only half joking, but still . . . .
 

Morgan's Magic Snowplow

Member
SoSH Member
Jul 2, 2006
22,345
Philadelphia
I just went through every game from 2016-2018, and there were 45 overtime games over the three seasons (including playoffs). Here were the results:

- 10 times (22.2%) the team who got the ball first in OT immediately drove for the GW touchdown
- 14 times (31.1%) the team who got the ball second scored either a FG or TD to win the first time they touched the ball
- 2 times (4.4%) the team who got the ball first scored a FG and then stopped the other team from scoring
- 15 times (33.3%) the teams traded multiple possessions before a winner was determined
- 4 times (8.9%) the teams couldn't score and ended in a tie

So only 22.2% of the games ended without the other team touching the ball. The team touching it second won 4 more times on their initial drive than the number of times the team won the game on the opening drive of OT.

Three seasons may seem like a small sample size, but it's 45 games, probably large enough to determine if it's a fair system. Given the actual results, it's hard for anyone to claim that the team that gets the ball first has an unfair advantage.
What percentage of games were won by the team receiving the ball first? The difference between that number and 50% (in a large enough sample) is ultimately the most important question.
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
24,376
What percentage of games were won by the team receiving the ball first? The difference between that number and 50% (in a large enough sample) is ultimately the most important question.
Of the 45 OT games the past three years, the team getting the ball first won 21 games, the team getting it second won 20 games, and there were 4 ties. About as close to 50-50 as it gets.
 

BigJimEd

Member
SoSH Member
Jan 4, 2002
4,432
So, isn't that basically saying that teams would do dumb stuff against their own self interest? I tend to think that actually teams would figure it out.
To some degree yes but it wouldn't be so blatant or obvious. Teams do not want to risk not getting the ball.
Right now the % is pretty darn close to 50% (21 to 20). You can't get closer than that but short of extenuating circumstances, teams are taking the ball. So how big of an advantage do teams need in order for them to risk not getting the ball? Win 53%? 55%? 60%? More?
And you'd still have instances where teams didn't get it and they'd complain.
 

Morgan's Magic Snowplow

Member
SoSH Member
Jul 2, 2006
22,345
Philadelphia
Of the 45 OT games the past three years, the team getting the ball first won 21 games, the team getting it second won 20 games, and there were 4 ties. About as close to 50-50 as it gets.
Some people might question whether the sample is large enough but that to me is the strongest evidence against making any changes. If the current rules are not randomly giving one team a meaningfully higher chance to win then who cares?
 

InstaFace

The Ultimate One
SoSH Member
Sep 27, 2016
21,766
Pittsburgh, PA
I’m supportive of this approach.

I don’t think it ever happens because: TV.
You think the TV networks would be upset about the possibility of getting additional playoff football?

Hell, I think the NFL would charge them more once the rule change goes through, just on the off chance that a game goes to overtime every year or two.

But like you, I'm supportive of that approach for playoffs. I'd say: Play one full additional quarter, starting with a coin toss and kickoff. If still tied at the end of that quarter, play continues after changing ends, and next team to score wins.
 

Bellhorn

Lumiere
SoSH Member
Aug 22, 2006
2,328
Brighton, MA
Of the 45 OT games the past three years, the team getting the ball first won 21 games, the team getting it second won 20 games, and there were 4 ties. About as close to 50-50 as it gets.
This is totally meaningless. You appear to have no clue what error bars are.

Here is a better way to look at it: as a ballpark estimate, an NFL possession ends in an offensive score roughly 35% of the time. If you treat possessions as independent, identically-distributed events (and I have yet to hear a good reason not to), this means that under the old OT rules, the team receiving the kickoff (assuming that both teams are average) ends up as approximately a 60-40 favorite.

All that changes under the new rules is that the win probability associated with scoring a FG on the first possession changes from 100% to something much closer to 50% (this is very hard to pin down from first principles alone, as it leads to a rare situation where the second team is forced to play with four downs over the entire field, but is under no time pressure). But given that well under 50% of NFL drives result in points, I think we can safely put 50% as a lower bound on the first team's win probability given an opening-possession FG.

It is also hard to pin down the probability of the first team scoring a FG on their opening possession, as they now face an unusually incentive to play for a TD. But given the observed drive outcome data, we can put the upper bound at 15%.

This means we have an upper bound of 7.5% on the amount by which the coin flip advantage is reduced under the new rules. This means that they are, at the very least, still a 52.5% favorite.

Since one team has to win the coin toss, this means that we still have a 5% win probability swing depending on an event that has absolutely nothing to do with football skill. While this is much better than 20%, as under the old rules, it is still a legitimate source of objection.
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
24,376
This is totally meaningless. You appear to have no clue what error bars are.

Here is a better way to look at it: as a ballpark estimate, an NFL possession ends in an offensive score roughly 35% of the time. If you treat possessions as independent, identically-distributed events (and I have yet to hear a good reason not to), this means that under the old OT rules, the team receiving the kickoff (assuming that both teams are average) ends up as approximately a 60-40 favorite.

All that changes under the new rules is that the win probability associated with scoring a FG on the first possession changes from 100% to something much closer to 50% (this is very hard to pin down from first principles alone, as it leads to a rare situation where the second team is forced to play with four downs over the entire field, but is under no time pressure). But given that well under 50% of NFL drives result in points, I think we can safely put 50% as a lower bound on the first team's win probability given an opening-possession FG.

It is also hard to pin down the probability of the first team scoring a FG on their opening possession, as they now face an unusually incentive to play for a TD. But given the observed drive outcome data, we can put the upper bound at 15%.

This means we have an upper bound of 7.5% on the amount by which the coin flip advantage is reduced under the new rules. This means that they are, at the very least, still a 52.5% favorite.

Since one team has to win the coin toss, this means that we still have a 5% win probability swing depending on an event that has absolutely nothing to do with football skill. While this is much better than 20%, as under the old rules, it is still a legitimate source of objection.
Well, I was answering a specific question from Morgan's Magic Snowplow, which was: "What percentage of games were won by the team receiving the ball first?" So I answered it with data.

But you're right. I don't understand what most of that means. You give a 52.5% favorite figure ("at the very least") to the team that gets the ball first. If you factor out the four ties, in the data sample of the past three years' worth of overtime games, when there's been a winner, the team with the ball first has actually (not just based on probability but in actuality) won 21 of the 41 times. 21 out of 41 is 51.2%. Compare that number to your 52.5% figure and it seems like unless that's a massive coincidence, it seems pretty much right. It's impossible to have a system that is exactly 50-50, since humans play the game and it's not fair coins being flipped. So 51.2% or 52.5% seem pretty reasonable to me.
 

The Gray Eagle

Member
SoSH Member
Aug 1, 2001
16,725
I personally dislike having a coin toss for OT at all.

They should just make a rule that either the home team always gets the ball to start OT, or the road team always does, or the team that kicked off the second half does, whatever-- just a rule so that both teams know in advance who will get the ball first if the game goes to OT, and they can plan their strategy in the 4th quarter accordingly.

I also agree that playoff OT should be a full quarter.
 

BusRaker

Member
SoSH Member
Aug 11, 2006
2,371
For what it's worth, in those 45 games:

- The team that got the ball first won 21 times
- The team that got the ball second won 20 times
- There were 4 ties

So that's about as even as it gets, right?
No it should have been 20-20-4. That one time was the Patriots against the Chiefs so it's obviously unfair /s
 

bankshot1

Member
SoSH Member
Feb 12, 2003
24,652
where I was last at
I'm not sure why playing defense to prevent a first score has been discarded, like the Rams did against the Saints in the NFC title game, but If first possession in OT is just a modest advantage (1-2%) make scoring on the first possession a little harder. If the opening kick-off is downed for a touchback start at the 15 or 20 rather than the 25.

Or play defense and force a stop.

And I like the element of "sort of sudden victory" in OT, as opposed to playing an additional quarter (or 30 minutes) as they do in soccer.
 

DJnVa

Dorito Dawg
SoSH Member
Dec 16, 2010
53,850
Add in XPs in OT. Team scores TD on first possession they can choose to kick XP or go for 2. If they go for 2 and convert the game is over. If not, other team gets the ball.


Or a tweak that if team scores TD on first possession they must go for 2 (or hit 50 yard XP or something). If they convert, game is over. If not, other team gets ball and we go from there.
 

joe dokes

Member
SoSH Member
Jul 18, 2005
30,243
This is totally meaningless. You appear to have no clue what error bars are.

Here is a better way to look at it: as a ballpark estimate, an NFL possession ends in an offensive score roughly 35% of the time. If you treat possessions as independent, identically-distributed events (and I have yet to hear a good reason not to), this means that under the old OT rules, the team receiving the kickoff (assuming that both teams are average) ends up as approximately a 60-40 favorite.

All that changes under the new rules is that the win probability associated with scoring a FG on the first possession changes from 100% to something much closer to 50% (this is very hard to pin down from first principles alone, as it leads to a rare situation where the second team is forced to play with four downs over the entire field, but is under no time pressure). But given that well under 50% of NFL drives result in points, I think we can safely put 50% as a lower bound on the first team's win probability given an opening-possession FG.

It is also hard to pin down the probability of the first team scoring a FG on their opening possession, as they now face an unusually incentive to play for a TD. But given the observed drive outcome data, we can put the upper bound at 15%.

This means we have an upper bound of 7.5% on the amount by which the coin flip advantage is reduced under the new rules. This means that they are, at the very least, still a 52.5% favorite.

Since one team has to win the coin toss, this means that we still have a 5% win probability swing depending on an event that has absolutely nothing to do with football skill. While this is much better than 20%, as under the old rules, it is still a legitimate source of objection.

I dont understand the math, so I cant disagree with it. But at a fundamental level, those that want to change things have to have some sort of reason. Those folks will not present your math. They will say, "not fair.....1st team with ball has too much of an advantage." So far, they haven't had an advantage, so that's a tough argument.

As a practical matter, the determining factor will be whether the NFL wants to reduce to zero the number of times the loser doesn't get the ball, because that's the only thing it is "concerned' about. That's a matter of optics as much as it is some notion of "fairness."
 

DennyDoyle'sBoil

Found no thrill on Blueberry Hill
SoSH Member
Sep 9, 2008
42,284
AZ
I dont understand the math, so I cant disagree with it. But at a fundamental level, those that want to change things have to have some sort of reason. Those folks will not present your math. They will say, "not fair.....1st team with ball has too much of an advantage." So far, they haven't had an advantage, so that's a tough argument.

As a practical matter, the determining factor will be whether the NFL wants to reduce to zero the number of times the loser doesn't get the ball, because that's the only thing it is "concerned' about. That's a matter of optics as much as it is some notion of "fairness."
Here's a stat I'd like to know -- how often does the team that wins the toss take the ball. Because at the end of the day, I think fairness point comes from a disconnect between the perceived and actual advantage. Perception shouldn't matter too much, but when the perception is causing all the teams to act in a way that they all, almost uniformly, care who wins the toss, then it's worth potentially changing things. I mean, we like to rag on them, but if all 32 head coaches think there's an advantage under current rules, then that's a problem, even if the sample size of 45 doesn't back them up.
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
24,376
Here's a stat I'd like to know -- how often does the team that wins the toss take the ball. Because at the end of the day, I think fairness point comes from a disconnect between the perceived and actual advantage. Perception shouldn't matter too much, but when the perception is causing all the teams to act in a way that they all, almost uniformly, care who wins the toss, then it's worth potentially changing things. I mean, we like to rag on them, but if all 32 head coaches think there's an advantage under current rules, then that's a problem, even if the sample size of 45 doesn't back them up.
Well even if the advantage is only 51.2% to 48.8%, that's an advantage. Every coach would prefer 51.2% odds over 48.8% odds, and rightly so. Meaning on only the very rare occasion would a coach go counter to that - such as BB did on a couple of occasions when the weather conditions were really extreme.
 

bankshot1

Member
SoSH Member
Feb 12, 2003
24,652
where I was last at
If you had to start at the 15 yard line (or some other arbitrary less favorable field position-than the 25- to remove the slight advantage and become closer to 50/50) one might assume a coach would prefer to play D and a field position strategy in OT.
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
24,376
If you had to start at the 15 yard line (or some other arbitrary less favorable field position-than the 25- to remove the slight advantage and become closer to 50/50) one might assume a coach would prefer to play D and a field position strategy in OT.
I'm sure there's a point in field position where the odds go from above 50% to below 50%. Moving it back to the 20 might just be right about at that tipping point. Maybe we should let Bellhorn figure out the odds.
 

Al Zarilla

Member
SoSH Member
Dec 8, 2005
58,878
San Andreas Fault
If you had to start at the 15 yard line (or some other arbitrary less favorable field position-than the 25- to remove the slight advantage and become closer to 50/50) one might assume a coach would prefer to play D and a field position strategy in OT.
Depends on how regulation time went. For example, in SB 51 and the AFCCG against KC, the Patriots had the huge advantage in time of possession, and so the opposing defenses were probably gassed. There is no way you wouldn’t give the ball to Brady, even at the 15, in both cases.
 

tims4wins

PN23's replacement
SoSH Member
Jul 15, 2005
37,060
Hingham, MA
Depends on how regulation time went. For example, in SB 51 and the AFCCG against KC, the Patriots had the huge advantage in time of possession, and so the opposing defenses were probably gassed. There is no way you wouldn’t give the ball to Brady, even at the 15, in both cases.
Exactly. The Pats went 92 yards to tie the game. I’d have taken the ball on the 5.

Edit: talking about LI, not KC
 
Last edited:

GrandSlamPozo

New Member
May 16, 2017
105
Make OT a continuation of the 4th quarter in terms of possession and keep the sudden death rules the same. Completely fair and even adds an element of strategy in situations where a team is on the cusp of field goal range at the end of regulation, giving them the option to choose between kicking a long field goal to win in regulation or keeping possession in OT but needing to finish the drive for a TD to win outright. Thoughts?
 

SirPsychoSquints

Member
SoSH Member
Jul 13, 2005
5,013
Pittsburgh, PA
I personally dislike having a coin toss for OT at all.

They should just make a rule that either the home team always gets the ball to start OT, or the road team always does, or the team that kicked off the second half does, whatever-- just a rule so that both teams know in advance who will get the ball first if the game goes to OT, and they can plan their strategy in the 4th quarter accordingly.

I also agree that playoff OT should be a full quarter.
I don't like that you've replaced a random bias discovered between regulation and OT with a known bias at the beginning. That's a huge advantage to give to either the home team or the road team, or to the team that won the opening coin flip instead!

I much prefer the "continue from end of regulation" option.
 

SirPsychoSquints

Member
SoSH Member
Jul 13, 2005
5,013
Pittsburgh, PA
One thing to keep in mind about simply applying regulation stats to overtime - the environment they're operating in is different. For one thing, as pointed out above, you might have gassed defenses. For another, as discussed on the Bet The Process podcast by Jeff Ma and Rufus Peabody, NFL coaches are generally not aggressive enough. The overtime incentives will force them to be more aggressive (like when getting the ball second if the other team has kicked a field goal already).

I don't know how to tease out the differences, but I think it's folly to ignore them.
 

Van Everyman

Member
SoSH Member
Apr 30, 2009
26,993
Newton
Possibly missing something here but why not just say both teams get at least one possession and if the score is no monger tied after each team possessed the ball once, the game is over?

Is there a reason that the game has to *end* if the first team scores a TD?
 

Papelbon's Poutine

Homeland Security
SoSH Member
Dec 4, 2005
19,615
Portsmouth, NH
I much prefer the "continue from end of regulation" option.
Only issue I see there is if it's like the transition from 1/2 and 3/4 quarter or if it's like halftime. IF it's the former - and you are at same down and distance - then maybe? If it's like halftime and you just get the over time kickoff, I don't think there's anyway we would see that.

Teams would get the ball and just stall, maybe even kneeling if the other team was out of timeouts. Would make for pretty boring viewing.

I don't think there's any issue with the new system. If you want to argue that each team should get a possession no matter what, I can see that, but at the risk of repeating myself, I'd stick with what they have or go to the college system.
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
24,376
I'll go on record as saying that I like the system just the way it is. I think it's fair. And as to the "coin flip" randomness, hey football does contain some luck so there you go. And so far, the current system has produced nearly as 50-50 a ratio as you could possibly hope for in terms of which team wins, so even if we come up with a more "fair" system, it'll be hard pressed to actually produce more "fair" results.

I hate the college system. And I don't like the "first team to score wins" system. I think the current system is pretty solid and have yet to see a compelling (well, compelling to ME, as if my opinion means anything) reason to change it.
 

BigJimEd

Member
SoSH Member
Jan 4, 2002
4,432
Possibly missing something here but why not just say both teams get at least one possession and if the score is no monger tied after each team possessed the ball once, the game is over?

Is there a reason that the game has to *end* if the first team scores a TD?
Few things. First, the length of game has to be considered. How often and much will that increase game length.
Also, there is something to be said for the excitement and drama of possibility being the game winning TD.

Perhaps most importantly, that gives an advantage to the 2nd team. They know what they need. They know to go for it on 4th down or if they can settle for the FG rather than the TD.

I'm fine with the rule as is. I think it's pretty even. Not a big fan of ther college rules.

However I can see the NFL moving this way. I believe some teams are more concerned with both teams possessing the ball than they are making it "fair".
 

SirPsychoSquints

Member
SoSH Member
Jul 13, 2005
5,013
Pittsburgh, PA
Only issue I see there is if it's like the transition from 1/2 and 3/4 quarter or if it's like halftime. IF it's the former - and you are at same down and distance - then maybe? If it's like halftime and you just get the over time kickoff, I don't think there's anyway we would see that.

Teams would get the ball and just stall, maybe even kneeling if the other team was out of timeouts. Would make for pretty boring viewing.

I don't think there's any issue with the new system. If you want to argue that each team should get a possession no matter what, I can see that, but at the risk of repeating myself, I'd stick with what they have or go to the college system.
The idea I'm referring to is like the 1/2 and 3/4 quarter transitions. It has the advantage of everyone knowing what the end game looks like ahead of time, teams have the ability to manage time and field position with their eyes open.

I hate the college system because it's less and less like football.
 

Bellhorn

Lumiere
SoSH Member
Aug 22, 2006
2,328
Brighton, MA
Make OT a continuation of the 4th quarter in terms of possession and keep the sudden death rules the same. Completely fair and even adds an element of strategy in situations where a team is on the cusp of field goal range at the end of regulation, giving them the option to choose between kicking a long field goal to win in regulation or keeping possession in OT but needing to finish the drive for a TD to win outright. Thoughts?
Love it. I wrote a blog post many years ago about NFL overtime issues, and arrived at the same conclusion. This was when the old rules were still in effect, so I did not consider the TD requirement wrinkle, but I agree that that makes it even better.

Blog post is linked below. Apart from being embarrassingly wordy, I think it holds up pretty well.

http://danieljepson.blogspot.com/2009/11/thoughts-on-nfl-overtime.html
 

Bellhorn

Lumiere
SoSH Member
Aug 22, 2006
2,328
Brighton, MA
Well, I was answering a specific question from Morgan's Magic Snowplow, which was: "What percentage of games were won by the team receiving the ball first?" So I answered it with data.

But you're right. I don't understand what most of that means. You give a 52.5% favorite figure ("at the very least") to the team that gets the ball first. If you factor out the four ties, in the data sample of the past three years' worth of overtime games, when there's been a winner, the team with the ball first has actually (not just based on probability but in actuality) won 21 of the 41 times. 21 out of 41 is 51.2%. Compare that number to your 52.5% figure and it seems like unless that's a massive coincidence, it seems pretty much right. It's impossible to have a system that is exactly 50-50, since humans play the game and it's not fair coins being flipped. So 51.2% or 52.5% seem pretty reasonable to me.
Apologies, didn't mean to be rude last time. But the underlying probability is what matters, and the data-based "actuality" that you repeatedly cite throughout this thread (not just in response to MMS) is only relevant if the sample size is large enough to ensure that it converges to the true probability. That is not remotely the case here. Think about it this way: if you flipped an apparently fair coin 45 times, and it came up heads 25 times, would you conclude that the probability of a heads outcome was 55.6%? Of course not - you have a very strong prior that the probability of a fair coin flip is 50%, and since a clustering of 25 heads in 45 50/50 flips is something that you can easily see due to random chance alone, you have no reason to reject that prior assumption. The outcome distribution of a 45-flip sample is sufficiently wide that the result you see (unless it is an extreme outlier) does not tell you much about the true probability of a single flip - your prior assumption of 50% is still a much stronger consideration.

Similarly, a win-loss record over 45 games does not pin down the true probability that the team receiving the ball first will win, to any interesting degree whatsoever. A 21-20-4 record is consistent with a 45% win probability. It is consistent with a 55% win probability. It is inconsistent with a 90% win probability, but who cares, we knew going in that that was not the case.

That is why it is much more effective to model the true probability based on drive outcomes, where the percentages are established to a much higher degree of confidence. We still cannot establish the answer to the same degree of confidence as the 50% probability of a fair coin flip - as someone else noted, the scoring probabilities may change somewhat over the course of a game, and as I noted in my OP, the second team's possession following an opening-possession FG is particularly difficult to evaluate. But we can still show that there is a definite advantage to receiving the ball first. And keep in mind that 52.5% is a lower bound - the real figure is almost certainly higher.
 

BaseballJones

ivanvamp
SoSH Member
Oct 1, 2015
24,376
Apologies, didn't mean to be rude last time. But the underlying probability is what matters, and the data-based "actuality" that you repeatedly cite throughout this thread (not just in response to MMS) is only relevant if the sample size is large enough to ensure that it converges to the true probability. That is not remotely the case here. Think about it this way: if you flipped an apparently fair coin 45 times, and it came up heads 25 times, would you conclude that the probability of a heads outcome was 55.6%? Of course not - you have a very strong prior that the probability of a fair coin flip is 50%, and since a clustering of 25 heads in 45 50/50 flips is something that you can easily see due to random chance alone, you have no reason to reject that prior assumption. The outcome distribution of a 45-flip sample is sufficiently wide that the result you see (unless it is an extreme outlier) does not tell you much about the true probability of a single flip - your prior assumption of 50% is still a much stronger consideration.

Similarly, a win-loss record over 45 games does not pin down the true probability that the team receiving the ball first will win, to any interesting degree whatsoever. A 21-20-4 record is consistent with a 45% win probability. It is consistent with a 55% win probability. It is inconsistent with a 90% win probability, but who cares, we knew going in that that was not the case.

That is why it is much more effective to model the true probability based on drive outcomes, where the percentages are established to a much higher degree of confidence. We still cannot establish the answer to the same degree of confidence as the 50% probability of a fair coin flip - as someone else noted, the scoring probabilities may change somewhat over the course of a game, and as I noted in my OP, the second team's possession following an opening-possession FG is particularly difficult to evaluate. But we can still show that there is a definite advantage to receiving the ball first. And keep in mind that 52.5% is a lower bound - the real figure is almost certainly higher.
1. How many games represent a sufficient enough sample size?
2. What setup is the most "fair" statistically speaking?
 

DennyDoyle'sBoil

Found no thrill on Blueberry Hill
SoSH Member
Sep 9, 2008
42,284
AZ
1. How many games represent a sufficient enough sample size?
I'm not a statistician but I would think it's close to obvious that in a sample size where 1 different result changes your probability 2.5 percent, you really don't have much of a sample.