Thanks buddy! I'm happy to hear people liked it. There's a stigma against list articles that's out there. I try to ensure that mine are actually unique and worth reading. They're anything but low effort, so it's nice to hear that people are liking them.
I think the only game this season that could join the list is the Broncos beating the Bills. The Rams have played the Lions close in the past two seasons and beat the Vikings so I don't think a Rams run would be too surprising; A rookie QB beating the MVP favorite and 13 win Bills team on the road who may have their best chance to get out of the conference would be big.
The Chiefs losing before the conference championship could be up there, but I think a good number of people realize they have 15 wins because of luck in once score games.
The Broncos beating the Bills would not make it. Not even close. People don't really realise it, because they've been doing it under people's noses, but the Broncos have put up a positive 114 point differential this season, and played just seven one possession games. That's fewer than the Lions have played. They've also become the clear 2010 Chargers award winner in the final weeks of the season, as they've finished the year with a 1-6 record in their one score games. Yes, that means they have a 9-1 record in multi-possession games. This Bronco team finished 10-7, but it would behoove you to think of them more like a 12.5 win team. They're the real deal.
Teams like this are why I have both measures of upset vibe and measures of actual quality baked into this formula. Broncos vs Bills may pass the upset vibe test, but in terms of the actual difference in team quality, the point differential difference is only 43 in favour of Buffalo. The QB difference is only about ten spots. I actually think the Bills this year are like the Cowboys last year. They've been totally hosed having to take out this Bronco team in the first round. They're still the better team, and I'd bet on the Bills, but if the Broncos played a lesser team like the Chiefs right now (the real ones), I don't know who I'd bet on. That's how good they are.
The Rams are quite the opposite. If they somehow beat the Lions it may fail the vibe test, but it certainly would pass the team quality test. The Rams have played a really strong schedule, but against that schedule they've played 13 one possession games. Their 8-5 record in those one possession games is all that's gotten them here. Their pittance of a 2-2 record in multi-possession games lets me know that this is little more than a .500 team in a bad disguise. Their point differential is -19 (close enough to zero to treat it as zero) and their SRS is pretty much exactly zero. They're the second worse playoff team this season (Texans), and the AFC has no team of the Lions' calibre. That's why the NFC can post an upset of top ten quality this season and the AFC can't.
If the Rams were to play the Lions, the point differential difference would be 179. The SRS difference would be 13.9. Let's treat the spread as eight points, indicative of very little upset vibe. The QB difference would be about ten spots. According to my formula, that would make it the third biggest upset of all time. If the upset vibe is bigger than eight points, it might begin to challenge for number two, although the top two are pretty untouchable in my opinion.
No Chiefs game can make the list this year. Their point differential is a scant 59, smaller than every AFC playoff team except Pittsburgh and Houston. If Houston were to beat KC (the only possibility that would get the 2024 Chiefs on here), it would depend a lot on the upset vibe, as the SRS difference would only be 4.9. I personally doubt even the best possible matchup would have any chance of landing here.
At some point I may run through all the Super Bowl era playoff tournaments and figure out the biggest upsets that even could have happened. I have a feeling that NE-NYG may still be number one even on that list. Very few teams in NFL history have had an SRS as good as 16.8, let alone being 16.8 BETTER than their opponents. That'd make me wonder if we're just having bad luck in the 2020s, or if upsets are legitimately less likely now than they used to be.
I guess I was leaning too heavily into the vibes area. I think I was going more for what would make the "most surprising list" based on perception rather than quality.
Boy, would be a lot to go through here, and brings back a lot of memories, as I had a streak as a kid where I never missed a playoff game.
As a Giants fan, I don't delude myself that they had any business beating the Patriots after '07, or the Packers after '11. I think God has us still paying.
Do have some questions about your low rating of the '87 Vikings. I think the point spread in the game against the Saints partly reflects that they were quite well regarded at that time, which is my memory. I think there was a dropped pass by Darrin Nelson in the championship game against Washington, or they could have made the Super Bowl. Note they were 8-4 in the "non scab" games. Note also that, in 1986, they had outscored opponents by 125 points, then were 173 points in the black for the '88 regular season, and 76 points in the black for '89. The '87 regular season is the aberration, and take out the scab games, and they're still +36 over a 12-game season then, anyway.
You know David, you may have me on this one. I simply made the assumption that football reference was like every other major football stats website, and simply entirely excluded the scab games from their SRS measurements. Considering 10% of this list takes place in 1987, I should've looked into this deeper, but thankfully, it doesn't hurt too badly.
I plugged into the formula what would've happened if the Vikings had a point differential 36 points better and an SRS of zero (instead of -2.6). This causes the Saints game to drop by three spots, but remain in the top 20, and does not alter the second spot at all, so really no harm no foul, as the rankings themselves are never the true point of my ranking lists anyway. I just need an excuse to talk about these games. As long as no game was robbed of its rightful place in the top 20, I'm not truly bothered.
So while yes the Vikings' point differential was more positive than it seems on paper, I still don't know if this changes my assertion that this was the worst playoff team ever as of 1987. It depends on how much you care about SoS. In their 12 games against non-scab opposition, just four of them came against teams with eight wins or more in 1987, with Monday Night Football against the Broncos being the only win in those four games.
Against this incredibly weak schedule (with the clear majority of the games coming against teams 7-8 or worse), the Vikings played eight one possession contests in the span of just 12 games. Their 5-3 record in those contests was not unsustainable, but just playing them against such weak opposition (in almost every instance) tells us all we need to know about a team. I'm not saying they didn't deserve their spot in a 1987 NFC that was crazily weak, and in the modern big playoff era, a team worse than this makes the playoffs most years, but in their time this was an extremely bad playoff team, one that normally would've been the victim of one of those 40-10 blowouts the 49ers were so fond of putting on people.
This is why I had the measures of team quality baked into this list. This could've been a highly regarded team. I wasn't there. I can't go back and feel the vibes, but I don't see why they would've been in retrospect, and based on what you tell me, they were probably overrated based upon how good they were in 1986. This was a prototypical .500 team, and even though they ended up 8-4, that's probably a bit of an overestimation of what their abilities were.
Of course, it turned out not to be that way, as this -2.5 DSRS 1987 Viking defence somehow turned back into their 70s selves for these playoffs, making easy work of a top ten QB three rounds in a row, in Bobby Hebert, Joe Montana, and Doug Williams. They played their best of all three rounds against Williams and the Skins, but like you say, it's the offence that let them down.
In many ways, they're a lot like the 2007 Giants, in that they're a defence that has no business catching fire at all, finding their flame and burning so hot that not even Joe Montana or Bobby Hebert (or Tom Brady or Brett Favre) can move the ball on them. Could that have been predicted in 1987 or in 2007? Perhaps, but I certainly wouldn't have done it.
The Vikings had 6 Pro Bowlers that year. The 49ers had 5, the Saints had 6. In 1988, the Vikings had 9 Pro Bowlers, tying with the Bengals for the most in the NFL. I have no issue with you creating an intelligent formula and applying it, but making larger claims about the ability of a team based on this rote application of one-season regular season data is suspect. For one thing, I suspect a season is never a large enough sample size for near precision, and that is one reason why we see change from season to season, beyond just roster change and an extra off-season. Then, there are injuries during a season. There are strategic shifts your team takes, and ones that the opponent takes in regard to you. And couldn't it be argued that the extent of one playoff upset has to be lessened if the same team engineered later ones? I could see deciding that either way, but is something to think about. You see the problems with a static approach.
I understand they had good players David. Once again, I was not there, so I can't tell you why they had three All-Pro players (plus a Pro Bowl MLB) on a defence that performed so poorly over the course of the 1987 regular season, but they did. I wish we had more modern statistics like offensive and defensive success rates to apply back then. It would clear up a lot of these issues. More modern stats would also help me figure out what the heck was going on with Anthony Carter, whose 24.3 yards per catch on 38 catches boggles the mind. Something like that should not be possible as late in NFL history as 1987, but he did it.
You seem like you're a big fan of the 1987 Vikings David. I am too, because I think this curse against threepeats (but not repeats) is a good thing for the league, and they and they alone are the ones that kept this curse going. It cannot be denied that they had one of the better defensive performances in playoff history for three consecutive weeks, but perhaps we ought to think in different terms.
I challenge you David to make a case for somebody else. Let's say in the merger era (1970-1987), name me a playoff team that was worse than the 1987 Minnesota Vikings. There are a few teams with a shot. Don't get me wrong. The 1977 Minnesota team that also appears on this list is a strong candidate, but even if I am underrating the 1987 team by calling them the absolute worst, which you are convincing me of, and I am willing to admit, I need me a counterexample. Just one will do.
I did consider applying a penalty to upsets in which a team also commandeered an upset in later rounds, but I decided against it based on the vibe test. That Minnesota took out New Orleans the round before did not seem to alter the public opinion that SF would slaughter them. Same goes for the 2007 Giants, whose second round upset over Dallas that year nearly landed in the top 20. It did not temper the feeling that they had no chance against the Patriots. I see your point that perhaps it could be a hidden team quality indicator, and perhaps if I revise this formula at all I will include it, but I don't feel too bad about omitting it, because this is an upset list. Vibes are the most important thing, and prior upsets don't appear to change the public's feelings about future upsets.
I also thought about implementing results from prior and future seasons into the formula, but that arose issues like 2001 Tom Brady vs 2001 Kurt Warner being a commanding advantage for the Patriots, no matter how much I faded their influence (because Kurt Warner would become the league's worst QB in 2002, and a backup in 2003), which just wasn't the reality of the time, so I got rid of the idea of including the future pretty quickly.
As far as including the past, I got pretty far into the making of this formula with that idea in mind, but then I ran into teams like the 1983 Miami Dolphins, whom had made the Super Bowl the year before, but were a completely different character of team in 1982 with no Marino than they were in 1983 with Dan. How do you wrap in back to back seasons with such differing playstyles? Then, there arises problems with a team like the 1998 Atlanta Falcons, who had been quite bad in 1997 despite the same Dan Reeves, same Chris Chandler, and same Jamal Anderson. Should I have attached this to the 1998 team with a 153 point differential and ten point SRS? I didn't feel that would be fair to a really good Atlanta team.
It's a point that can be argued, but I prefer to believe there are no fluke teams in the NFL. One year wonders are a real thing, but great teams are never flukes. Great players can be, because their sample size is so small (often less than 100 for non-QB offensive skill players, and sample sizes get weird for defensive players), but teams have sample sizes of 2000 plays or more. It's hard to fluke your way through a sample that big. It's possible to win a lot of games without deserving it (i.e. 2022 Vikings, 2024 Chiefs), but that's why the real record was the least important variable in this formula. You cannot trick the peripheral numbers in a sample that big.
That's why the 2022 Vikings had a PD of -3 (indicative of an eight win team), and the 2024 Chiefs had just a positive 59 (indicative of a ten win team). The W-L record can be fooled, but these surrounding numbers cannot be fooled. One season is not a big enough sample to tell us how good teams are going to be in the future. Fair enough, but I choose to believe that a team's sample of about 2000 plays is enough to tell us how good they are right now, which is why I thought one season's worth of data was all that needed implanted into my upset formula.
Pro Football Reference does not remove the scab games from SRS for 1987. I also don’t think other sites do either. Sadly, those were official games. The ‘87 Vikings went 0-3 in those games which bring their SRS down.
I was thinking more along the lines of The DVOA/VOA crowd, which does not include the scab games, but like I said to David, even if we treat their SRS as three points better, and their point differential as 36 (their point differential in non-scab games) instead of a mere positive one, number two does not change, and the Saints game falls down a few spots but not out of the top 20, so in this case it's no harm. No foul.
Loved this one Robbie! Great stuff!
Thanks buddy! I'm happy to hear people liked it. There's a stigma against list articles that's out there. I try to ensure that mine are actually unique and worth reading. They're anything but low effort, so it's nice to hear that people are liking them.
I think the only game this season that could join the list is the Broncos beating the Bills. The Rams have played the Lions close in the past two seasons and beat the Vikings so I don't think a Rams run would be too surprising; A rookie QB beating the MVP favorite and 13 win Bills team on the road who may have their best chance to get out of the conference would be big.
The Chiefs losing before the conference championship could be up there, but I think a good number of people realize they have 15 wins because of luck in once score games.
The Broncos beating the Bills would not make it. Not even close. People don't really realise it, because they've been doing it under people's noses, but the Broncos have put up a positive 114 point differential this season, and played just seven one possession games. That's fewer than the Lions have played. They've also become the clear 2010 Chargers award winner in the final weeks of the season, as they've finished the year with a 1-6 record in their one score games. Yes, that means they have a 9-1 record in multi-possession games. This Bronco team finished 10-7, but it would behoove you to think of them more like a 12.5 win team. They're the real deal.
Teams like this are why I have both measures of upset vibe and measures of actual quality baked into this formula. Broncos vs Bills may pass the upset vibe test, but in terms of the actual difference in team quality, the point differential difference is only 43 in favour of Buffalo. The QB difference is only about ten spots. I actually think the Bills this year are like the Cowboys last year. They've been totally hosed having to take out this Bronco team in the first round. They're still the better team, and I'd bet on the Bills, but if the Broncos played a lesser team like the Chiefs right now (the real ones), I don't know who I'd bet on. That's how good they are.
The Rams are quite the opposite. If they somehow beat the Lions it may fail the vibe test, but it certainly would pass the team quality test. The Rams have played a really strong schedule, but against that schedule they've played 13 one possession games. Their 8-5 record in those one possession games is all that's gotten them here. Their pittance of a 2-2 record in multi-possession games lets me know that this is little more than a .500 team in a bad disguise. Their point differential is -19 (close enough to zero to treat it as zero) and their SRS is pretty much exactly zero. They're the second worse playoff team this season (Texans), and the AFC has no team of the Lions' calibre. That's why the NFC can post an upset of top ten quality this season and the AFC can't.
If the Rams were to play the Lions, the point differential difference would be 179. The SRS difference would be 13.9. Let's treat the spread as eight points, indicative of very little upset vibe. The QB difference would be about ten spots. According to my formula, that would make it the third biggest upset of all time. If the upset vibe is bigger than eight points, it might begin to challenge for number two, although the top two are pretty untouchable in my opinion.
No Chiefs game can make the list this year. Their point differential is a scant 59, smaller than every AFC playoff team except Pittsburgh and Houston. If Houston were to beat KC (the only possibility that would get the 2024 Chiefs on here), it would depend a lot on the upset vibe, as the SRS difference would only be 4.9. I personally doubt even the best possible matchup would have any chance of landing here.
At some point I may run through all the Super Bowl era playoff tournaments and figure out the biggest upsets that even could have happened. I have a feeling that NE-NYG may still be number one even on that list. Very few teams in NFL history have had an SRS as good as 16.8, let alone being 16.8 BETTER than their opponents. That'd make me wonder if we're just having bad luck in the 2020s, or if upsets are legitimately less likely now than they used to be.
I guess I was leaning too heavily into the vibes area. I think I was going more for what would make the "most surprising list" based on perception rather than quality.
Boy, would be a lot to go through here, and brings back a lot of memories, as I had a streak as a kid where I never missed a playoff game.
As a Giants fan, I don't delude myself that they had any business beating the Patriots after '07, or the Packers after '11. I think God has us still paying.
Do have some questions about your low rating of the '87 Vikings. I think the point spread in the game against the Saints partly reflects that they were quite well regarded at that time, which is my memory. I think there was a dropped pass by Darrin Nelson in the championship game against Washington, or they could have made the Super Bowl. Note they were 8-4 in the "non scab" games. Note also that, in 1986, they had outscored opponents by 125 points, then were 173 points in the black for the '88 regular season, and 76 points in the black for '89. The '87 regular season is the aberration, and take out the scab games, and they're still +36 over a 12-game season then, anyway.
You know David, you may have me on this one. I simply made the assumption that football reference was like every other major football stats website, and simply entirely excluded the scab games from their SRS measurements. Considering 10% of this list takes place in 1987, I should've looked into this deeper, but thankfully, it doesn't hurt too badly.
I plugged into the formula what would've happened if the Vikings had a point differential 36 points better and an SRS of zero (instead of -2.6). This causes the Saints game to drop by three spots, but remain in the top 20, and does not alter the second spot at all, so really no harm no foul, as the rankings themselves are never the true point of my ranking lists anyway. I just need an excuse to talk about these games. As long as no game was robbed of its rightful place in the top 20, I'm not truly bothered.
So while yes the Vikings' point differential was more positive than it seems on paper, I still don't know if this changes my assertion that this was the worst playoff team ever as of 1987. It depends on how much you care about SoS. In their 12 games against non-scab opposition, just four of them came against teams with eight wins or more in 1987, with Monday Night Football against the Broncos being the only win in those four games.
Against this incredibly weak schedule (with the clear majority of the games coming against teams 7-8 or worse), the Vikings played eight one possession contests in the span of just 12 games. Their 5-3 record in those contests was not unsustainable, but just playing them against such weak opposition (in almost every instance) tells us all we need to know about a team. I'm not saying they didn't deserve their spot in a 1987 NFC that was crazily weak, and in the modern big playoff era, a team worse than this makes the playoffs most years, but in their time this was an extremely bad playoff team, one that normally would've been the victim of one of those 40-10 blowouts the 49ers were so fond of putting on people.
This is why I had the measures of team quality baked into this list. This could've been a highly regarded team. I wasn't there. I can't go back and feel the vibes, but I don't see why they would've been in retrospect, and based on what you tell me, they were probably overrated based upon how good they were in 1986. This was a prototypical .500 team, and even though they ended up 8-4, that's probably a bit of an overestimation of what their abilities were.
Of course, it turned out not to be that way, as this -2.5 DSRS 1987 Viking defence somehow turned back into their 70s selves for these playoffs, making easy work of a top ten QB three rounds in a row, in Bobby Hebert, Joe Montana, and Doug Williams. They played their best of all three rounds against Williams and the Skins, but like you say, it's the offence that let them down.
In many ways, they're a lot like the 2007 Giants, in that they're a defence that has no business catching fire at all, finding their flame and burning so hot that not even Joe Montana or Bobby Hebert (or Tom Brady or Brett Favre) can move the ball on them. Could that have been predicted in 1987 or in 2007? Perhaps, but I certainly wouldn't have done it.
The Vikings had 6 Pro Bowlers that year. The 49ers had 5, the Saints had 6. In 1988, the Vikings had 9 Pro Bowlers, tying with the Bengals for the most in the NFL. I have no issue with you creating an intelligent formula and applying it, but making larger claims about the ability of a team based on this rote application of one-season regular season data is suspect. For one thing, I suspect a season is never a large enough sample size for near precision, and that is one reason why we see change from season to season, beyond just roster change and an extra off-season. Then, there are injuries during a season. There are strategic shifts your team takes, and ones that the opponent takes in regard to you. And couldn't it be argued that the extent of one playoff upset has to be lessened if the same team engineered later ones? I could see deciding that either way, but is something to think about. You see the problems with a static approach.
I understand they had good players David. Once again, I was not there, so I can't tell you why they had three All-Pro players (plus a Pro Bowl MLB) on a defence that performed so poorly over the course of the 1987 regular season, but they did. I wish we had more modern statistics like offensive and defensive success rates to apply back then. It would clear up a lot of these issues. More modern stats would also help me figure out what the heck was going on with Anthony Carter, whose 24.3 yards per catch on 38 catches boggles the mind. Something like that should not be possible as late in NFL history as 1987, but he did it.
You seem like you're a big fan of the 1987 Vikings David. I am too, because I think this curse against threepeats (but not repeats) is a good thing for the league, and they and they alone are the ones that kept this curse going. It cannot be denied that they had one of the better defensive performances in playoff history for three consecutive weeks, but perhaps we ought to think in different terms.
I challenge you David to make a case for somebody else. Let's say in the merger era (1970-1987), name me a playoff team that was worse than the 1987 Minnesota Vikings. There are a few teams with a shot. Don't get me wrong. The 1977 Minnesota team that also appears on this list is a strong candidate, but even if I am underrating the 1987 team by calling them the absolute worst, which you are convincing me of, and I am willing to admit, I need me a counterexample. Just one will do.
I did consider applying a penalty to upsets in which a team also commandeered an upset in later rounds, but I decided against it based on the vibe test. That Minnesota took out New Orleans the round before did not seem to alter the public opinion that SF would slaughter them. Same goes for the 2007 Giants, whose second round upset over Dallas that year nearly landed in the top 20. It did not temper the feeling that they had no chance against the Patriots. I see your point that perhaps it could be a hidden team quality indicator, and perhaps if I revise this formula at all I will include it, but I don't feel too bad about omitting it, because this is an upset list. Vibes are the most important thing, and prior upsets don't appear to change the public's feelings about future upsets.
I also thought about implementing results from prior and future seasons into the formula, but that arose issues like 2001 Tom Brady vs 2001 Kurt Warner being a commanding advantage for the Patriots, no matter how much I faded their influence (because Kurt Warner would become the league's worst QB in 2002, and a backup in 2003), which just wasn't the reality of the time, so I got rid of the idea of including the future pretty quickly.
As far as including the past, I got pretty far into the making of this formula with that idea in mind, but then I ran into teams like the 1983 Miami Dolphins, whom had made the Super Bowl the year before, but were a completely different character of team in 1982 with no Marino than they were in 1983 with Dan. How do you wrap in back to back seasons with such differing playstyles? Then, there arises problems with a team like the 1998 Atlanta Falcons, who had been quite bad in 1997 despite the same Dan Reeves, same Chris Chandler, and same Jamal Anderson. Should I have attached this to the 1998 team with a 153 point differential and ten point SRS? I didn't feel that would be fair to a really good Atlanta team.
It's a point that can be argued, but I prefer to believe there are no fluke teams in the NFL. One year wonders are a real thing, but great teams are never flukes. Great players can be, because their sample size is so small (often less than 100 for non-QB offensive skill players, and sample sizes get weird for defensive players), but teams have sample sizes of 2000 plays or more. It's hard to fluke your way through a sample that big. It's possible to win a lot of games without deserving it (i.e. 2022 Vikings, 2024 Chiefs), but that's why the real record was the least important variable in this formula. You cannot trick the peripheral numbers in a sample that big.
That's why the 2022 Vikings had a PD of -3 (indicative of an eight win team), and the 2024 Chiefs had just a positive 59 (indicative of a ten win team). The W-L record can be fooled, but these surrounding numbers cannot be fooled. One season is not a big enough sample to tell us how good teams are going to be in the future. Fair enough, but I choose to believe that a team's sample of about 2000 plays is enough to tell us how good they are right now, which is why I thought one season's worth of data was all that needed implanted into my upset formula.
Pro Football Reference does not remove the scab games from SRS for 1987. I also don’t think other sites do either. Sadly, those were official games. The ‘87 Vikings went 0-3 in those games which bring their SRS down.
I was thinking more along the lines of The DVOA/VOA crowd, which does not include the scab games, but like I said to David, even if we treat their SRS as three points better, and their point differential as 36 (their point differential in non-scab games) instead of a mere positive one, number two does not change, and the Saints game falls down a few spots but not out of the top 20, so in this case it's no harm. No foul.
I'll pay closer attention to 1987 in the future.