So many recent Super Bowls have featured teams that have won multiple Super Bowls (namely the Patriots) that it occurred to me that this year’s (LVI) is something of an exception. The Bengals are winless in two trips to the Super Bowl, and the Rams have won only one of their four appearances. This led me to wonder if their combined 1-5 record is the worst among two teams appearing in a Super Bowl. I decided to dig into this, and in the process unearthed a bunch of equally trivial facts about Super Bowl performance.
Let’s start with the worst combined won-lost Super Bowl records of teams playing in a Super Bowl. If you limit yourself to a minimum of five games, then yes, the Bengals vs. Rams 1-5 record (0.167) is the worst, followed by the Broncos vs. Washington in Super Bowl XXII (1-4), Broncos vs. Falcons in Super Bowl XXXIII (also 1-4), and Broncos vs. Panthers in Super Bowl 50 (2-6).
Honorable mention goes to the Raiders vs. Vikings in Super Bowl XI, who had a combined 0-4 record, the only time two winless teams had more than two total losses entering the game. This is also the only Super Bowl that has featured two winless teams with at least one loss.
The most combined losses for any two teams playing in a Super Bowl belongs to the Patriots vs. Rams in Super Bowl LIII, with 7, against 6 combined wins.
There have been twelve Super Bowls featuring two teams that had never won a previous Super Bowl. Unsurprisingly, six of the first seven Super Bowls and eight of the first eleven fall into this category, but there have been only four such Super Bowls since. The most recent one was Super Bowl XXXIV, Rams vs. Titans, and before that Super Bowl XXI, Broncos vs. Giants.
Now let’s pivot to the games featuring the teams with the best Super Bowl records. Leading the way is Super Bowl XLVII, 49ers vs. Ravens, who entered the game with a combined record of 6-0. This is also the only Super Bowl featuring two undefeated teams with at least one win. Next is Super Bowl XLIII, Cardinals vs. Steelers at 5-1, though as it was the Cardinals first (and only, to date) appearance that record belongs entirely to the Steelers. Third is Super Bowl XLV, Packers vs. Steelers at 9-2, which is also the most combined wins for any two teams.
Next, let’s talk about the performance of teams appearing in their first Super Bowls. There have only been four Super Bowls where both teams were making their first appearances: I (of course), Chiefs vs. Packers, III, Colts vs. Jets, XVI, Bengals vs. 49ers, and XX, Bears vs. Patriots. Those of you hoping for a Browns (or Jaguars, or Texans) vs. Lions matchup are probably going to be waiting for a while. The most recent NFC team to debut in the Super Bowl is the Saints, in XLIV; the most recent AFC team to debut is the Ravens, in XXXV. Coincidentally, both of them won, but prior to the Saints, the previous three NFC teams to debut had lost, and prior to the Ravens, the previous six AFC team to debut had lost.
Overall teams have a 9-19 record in the first Super Bowls. The AFC teams are 3-9, and the NFC teams are 6-10. Note that this includes the Colts in the latter category, since they were in the NFL when they made their debut in Super Bowl III. The records would be 3-10 and 6-9 respectively if you count the Colts as an AFC team.
Finally, let’s consider the Super Bowls featuring the longest drought since a previous victory by either team. It has been 22 years since the (then St. Louis) Rams’ sole Super Bowl victory in XXXIV, which is a long time, but only good enough for fourth longest drought. The three longer droughts are: 1) Super Bowl XXXI, Packers vs. Patriots, 29 years after the Packers’ last victory (at that point, the Patriots had lost their only previous appearance), 2) Super Bowl XL, Seahawks vs. Steelers, 26 years after the Steelers’ fourth Super Bowl victory of the 1970s, and 3) Super Bowl LIV, Chiefs vs. 49ers, 25 years after the 49ers’ last victory (and fifty years after the Chiefs’).
In the meantime, I’ll hold out hope that someday the Jets will face the Lions, or Cardinals, or Falcons, or Panthers, or Vikings, and smash this drought record. And for Super Bowl LVI this year, I’ll be rooting for the Bengals to become the 21st team to earn their first victory in the big game.
When I wrote How I Play “Real” Fantasy Football last year, I mentioned my core criterion for determining whether I consider a particular weekend of NFL football a success:
It started years ago with a simple rule of thumb that I employed for a long time: A successful NFL weekend for me was one in which the Seahawks, Jets, and Giants all won. The Seahawks because I’ve lived in the Seattle area for most of my adult life and have actively rooted for them for most of that time, the Jets because of my historical and long-suffering allegiance to them, and the Giants because of my legacy loyalty to all things New York.
This confluence of victories, which a friend of mine has labeled the “Danny Trifecta”, has been rare in recent years due to the consistently poor play of the Jets and Giants, both of whom sported identical 18-49 won-loss records from the beginning of the 2017 season through week 3 of the 2021 season. So it was an unexpected pleasure yesterday when the Jets and Giants — both seven-point underdogs — came back from deficits in the fourth quarter and ultimately won in overtime, capped by the Seahawks breaking a two-game losing streak to defeat a tough divisional opponent on the road.
This led me to wonder the last time I enjoyed such a trifecta. For that, I have to go all the back to Week 14 of the 2018 season, when both the Jets and Giants defeated divisional opponents on Sunday, December 9th and the Seahawks defeated Minnesota the following night. After two years, nine months, and 24 days, it’s worthy of celebration.
In my previous post, I wondered how much retrospective analysis of NFL draft picks is done, and devised a shortcut method of determining this: Looking at where NFL players elected to the Pro Football Hall of Fame who were draft-eligible from 1967 to the present were drafted. My analysis led me to construct a trivia quiz.
Here is the answer key to the quiz:
56%. Of the 146 Hall of Fame players who were draft-eligible in 1967 or later, 82 were selected in the first round.
5%. Of those 146 players, 8 were undrafted. They are: Larry Little (1967), Jim Langer (1970), Cliff Harris (1970), Drew Pearson (1973), Donnie Shell (1974), Warren Moon (1984), John Randle (1990), Kurt Warner (1994). Little and Langer were offensive linemates on the dominant Dolphins teams of the early 1970s, and Harris and Pearson were teammates on the dominant Cowboys teams of the mid-to-late 1970s.
#4. 11 Hall of Famers were drafted in the #4 position in the draft. They are: Bob Griese (1967), Joe Greene (1969), John Hannah (1973), Walter Payton (1975), Dan Hampton (1979), Kenny Easley (1981), Chris Doleman (1985), Derrick Thomas (1989), Jonathan Ogden (1996), Charles Woodson (1998), Edgerrin James (1999).
11. See above.
10. Ron Yary (1968), O.J. Simpson (1969), Terry Bradshaw (1970), Lee Roy Selmon (1976), Earl Campbell (1978), John Elway (1983), Bruce Smith (1985), Troy Aikman (1989), Orlando Pace (1997), Peyton Manning (1998), Tim Couch (1999). OK, not Tim Couch.
2. Per above, 1968-1970 and 1997-1998. Give yourself credit for 3, but only if you chose to count 1968-1969 and 1969-1970 separately.
#33. 4 Hall of Famers were drafted #33, though none of them when it was the first pick of the second round, as it is today. They are: Ted Hendricks (1969), Fred Dean (1975), Brett Favre (1991), Isaac Bruce (1994).
#214. Ken Houston, drafted by Houston in the 9th round of the 1967 draft. Choice #198 was a nod to Tom Brady being selected #199 in 2000 draft, and choice #321 was a nod to Giants Hall of Fame OT Rosey Brown, who was drafted #322 by the Giants in the 27th round of 1953 draft.
1967, with 9 (including one undrafted).They are: Bob Griese (#4), Floyd Little (#6), Alan Page (#15), Gene Upshaw (#17), Lem Barney (#34), Willie Lanier (#50), Rayfield Wright (#182), Ken Houston (#214), and Larry Little (undrafted). Note that Wikipedia reports 10 Hall of Famers for this draft, as they also include Jan Stenerud. I exclude Stenerud because he was originally selected in the 1966 AFL draft. The runner-up drafts are 1983 with 8, and 1968 and 1981 with 7 each. We’ll see if any of the more recent drafts will eventually challenge these numbers; currently no draft after 1996 has more than four Hall of Famers.
1992. 1972 and 1977 have one each (Franco Harris and Tony Dorsett, respectively). Tony Dungy is also a Hall of Famer who was draft-eligible in 1977 (though undrafted), but he’s in as a coach, not a player. Some people consider the 1984 draft to be one with no Hall of Famers, but I count the three Hall of Famers who were selected in the 1984 supplemental draft: Steve Young, Gary Zimmerman, and Reggie White.
What can we conclude from all of this? First, it’s incredibly difficult to predict when drafting who is going to be a future Hall of Famer. If we constrain ourselves to the years 1967 through 2000, 29% (= 10/34) of the #1 draft picks are in the Hall of Fame. If you look at the top five draft picks from each year, it’s 21% (35 players out of 170 picks). And while we don’t know for certain how many Hall of Fame players more recent drafts will turn out, it’s not looking promising for the top picks. Of the fifteen players chosen #1 from 2001-2015, the best are Eli Manning, Mario Williams, and Jake Long, and it drops off pretty quickly after that, so maintaining the 29% hit rate seems unlikely.
A major thing that has changed in the NFL over the past twenty years is the greater emphasis of the passing game, and the central role of the quarterback to that. Teams are increasingly spending their top draft picks in search of franchise QBs. And while I don’t have hard evidence for this assertion, it seems like it’s easier to predict that an outstanding college lineman will be HoF caliber than a college quarterback. I conjecture that the skills and attributes that make a lineman great in college translate to the NFL more readily than for a QB.
From 1981 through 2000, 7 of the 20 (35%) #1 picks were QBs, and three of them (Elway, Aikman, and P. Manning) are in the Hall of Fame. (The four who aren’t? Vinny Testaverde, Jeff George, Drew Bledsoe, and Tim Couch.) From 2001 to 2015, 11 of the 15 (73%) #1 picks were QBs, and from 2016 to 2021, 5 of 6 (83%).
Those 1981-2000 quarterbacks combined for 13 Super Bowl appearances and a 7-6 record. (You can increase those numbers to 14 and 8-6 if you want to give Bledsoe credit for his role on the 2001 Patriots.)
If you look at the eleven 2001-2015 #1 pick quarterbacks, you have three combined Super Bowl appearances to date: Two wins by Eli Manning and a loss by Cam Newton. (If you choose to include backups, you can count David Carr’s win in 2011 and Alex Smith’s loss in 2013.) Now you can argue that comparing Super Bowl records is unfair in the Tom Brady era, but let’s look at who is a reasonably candidate for the Hall of Fame from this cohort. Eli Manning has a good shot, but is not a slam dunk. Cam Newton has won an MVP award, but seems like a long shot. Matthew Stafford has had a long career and put up some good numbers, but nothing HoF-worthy, esp. without a playoff victory. Still, Stafford’s not out of the question if he can lead the Rams to success. Carson Palmer and Andrew Luck had impressive if injury-riddled careers, but neither seems like a plausible HoF candidate. That leaves Michael Vick, Carr, Smith, JaMarcus Russell, Sam Bradford, and Jameis Winston. The likely future HoF quarterbacks from this era, Ben Roethlisberger and Aaron Rodgers, were first-round picks, but not in the top ten.
It’s way too early to tell how the QBs picked #1 in the last five years will fare, but given how much top college QBs have been front-loaded in recent years, it seems likely that future Hall of Fame QBs from this era will be first-round picks, if not #1. An open question is whether there are likely to be later-round QB picks who have HoF-caliber careers, the way Tom Brady (6th round in 2000) and Russell Wilson (3rd round in 2012, and still only a potential HoFer) have.
Broadening the analysis beyond quarterbacks picked #1, we can see that there is an apparent correlation between draft round and eventual Hall of Fame membership:
Pro Football Hall of Fame players by NFL draft round, 1967-present
More than half of the future Hall of Famers were chosen in the first round, and more than 80% in the first three rounds. So even if the top picks are hit-and-miss, especially recently at quarterback, the top of the draft doesn’t seem to miss to many Hall of Fame-caliber players.
Here’s the breakdown of these players by position. I could have included some questions about this on the quiz, but I didn’t do this analysis until later.
Pro Football Hall of Fame players by position, 1967-present
While it’s noteworthy that the quarterback position is near the bottom of this list, it’s tricky to compare this to other positions where there are typically two or more starters per team. If you break down the position groups further, for example, you find that there are only five centers and six tight ends in this group. So fourteen QBs is actually high by that measure.
Narrowing the analysis to only Hall of Fame players probably doesn’t do a great job of determining the draft’s ability to identify the overwhelming number of NFL players who have solid but not top-tier careers. Even broadening the list to All-Pro selections would give a more meaningful measure. I’ll leave that on my to-do list for a later date.
The 2021 NFL Draft is now in the history books, and with it, the interminable pre-draft analysis has concluded and been replaced by interminable post-draft analysis. Given how closely scrutinized the process is, how many millions of dollars are at stake, and how much analytics have taken over the game of NFL football, you’d think that there would be more research put into how effective teams are at drafting players and how to optimize one’s draft picks. But even the best NFL analysts’ take can be roughly summarized as, “Drafting players is a crapshoot, so the best approach is to draft as many players as possible to increase your team’s chances of finding NFL-caliber players.”
It’s easy to look at recent high-profile draft picks and spot the apparent randomness. 2012: Andrew Luck and Robert Griffin III drafted #1 and #2; Russell Wilson drafted #75. 2015: Jameis Winston and Marcus Mariota drafted #1 and #2. 2016: Jared Goff and Carson Wentz drafted #1 and #2. 2017: Mitchell Trubisky drafted #2, Patrick Mahomes #10, Deshaun Watson #12. None of the six quarterbacks drafted #1 between 2009 and 2016 (Matthew Stafford, Sam Bradford, Cam Newton, Luck, Winston, and Goff) are still with their original teams, including two drafted by the Rams. The same goes for the five quarterbacks drafted #2 and #3 from 2012 to 2017 (RG3, Blake Bortles, Mariota, Wentz, and Trubisky).
This belies the notion that using a top-three draft pick to select a potential franchise quarterback is a solid bet. Unless teams have suddenly gotten smarter or luckier, and given that the four most recent #1 picks and the top three picks in 2021 were all quarterbacks, there’s a good chance this trend will continue.
Because of injuries and team quality, there is always going to be a significant element of fortune with the ultimate performance of top draft picks, as exemplified by Luck and RG3. But I wonder how much serious analysis has been done in this regard. For example, how strongly do PFF grades for NFL players correlate with their original draft order? On average, do first rounders grade higher than second rounders, and if so, by how much? Does the relative performance of certain position groups correlate more with draft order than others (e.g. offensive linemen versus quarterbacks)? Is there a strong correlation between performance and draft order at different phases of a career (years 1-3 vs. 4-6), and is that correlated by position group (e.g. do running backs start stronger and fade faster while linemen start more slowly but continue to improve)?
I could research all of this myself, but it would take a lot of time. So I took a major shortcut by looking only at the players elected to the Pro Football Hall of Fame who were drafted (or draft eligible) starting in 1967. I chose 1967 because it’s the first year of that the NFL and AFL merged their drafts. Including the 2021 HoF class, there are 146 players in this cohort.
Before I get into the details and what did and did not surprise me, I thought it would be fun to construct a quiz based on the analysis. Ten questions, multiple choice. Remember we’re only looking at Hall of Fame players who were first draft-eligible in 1967 or later. Let’s see how good your guesses are:
What percentage of Hall of Famers were drafted in the first round?
What percentage of Hall of Famers were undrafted?
Which draft position has the most Hall of Famers?
How many Hall of Famers were drafted in the position referenced in the previous question?
How many Hall of Famers were drafted #1?
How many times have Hall of Famers been drafted #1 in consecutive years?
Outside of the first round, which draft position has the most Hall of Famers?
Excluding undrafted and supplemental picks, what is the lowest draft position of a Hall of Famer?
From 1967 through 2000, which draft year has the most Hall of Famers?
There is only one draft year from 1967-2000 with no Hall of Famers. Which is it?
Once you’ve attempted to answer these questions, go to the follow-up post to read the answer key and some additional analysis.
Last week marked the fortieth anniversary of the release of the album Trust by Elvis Costello and The Attractions. It was the last in a string of five remarkable Costello albums produced by Nick Lowe, going back to My Aim is True, his debut. It was a transitional album, one that primarily featured the new wave rock for which he was best known but also foreshadowed the upcoming forays into Country & Western and Popular music that became staples of his broadened repertoire in the 1980s and beyond. And it is personally significant for me as the album that introduced me to Costello and began my enduring appreciation for his work.
While I had a passing awareness of Costello in the late 1970s and knew a couple of his songs, back then I thought of him mostly as the angry English guy with the old-fashioned, chunky, black glasses. At the start, I was drawn to Trust not by Costello but by Glenn Tilbrook‘s vocals on the song From a Whisper to a Scream. (I’d been a big fan of Squeeze from the moment I first heard the song If I Didn’t Love You on the radio.) It was July 1981 when I really got into From a Whisper to a Scream, which led me to discovering the rest of Trust, and from there, immersion in Costello’s back catalog. In the summer of 1983, I saw him perform live for the first time, in concert with The Attractions on a Hudson River pier in Manhattan, touring in support of Punch the Clock(confession: not one of my favorite Costello albums). He wore red shoes.
One sign of my affection for Trust is that it’s one of the few albums I own in both LP and CD formats. I think I used to own it on prerecorded cassette, too, but I’m not sure, as I said farewell to my cassette collection several years ago and I can no longer recall everything I purchased in the Columbia House-fueled buying binges of my college years. It remains one of my favorite Costello albums, and while I don’t like all of its songs, several — Clubland, Watch Your Step, Different Finger, Shot With His Own Gun, and yes, From a Whisper to a Scream — fill me with the same joy and energy today that they did when I was a teenager.
Four years ago this month in their February 2017 issue, Car and Driver published a column by Aaron Robinson that addressed the transition from internal combustion engine vehicles to battery-electric vehicles in an extremely clever alternate-universe manner. I’ve shared this column with numerous people over the years, but for some reason C&D has never seen fit to make it available on their website, even though contemporaneous columns from the same author are there. The column is also inexplicably impervious to Google and other search engines.
So as a public service to other fans of this column, I am reprinting it here, both in its original format and as searchable text. Should C&D or Mr. Robinson ever choose to make it available online, I will cheerfully remove this copy and redirect readers to their site.
Upfront by Aaron Robinson Car and Driver, February 2017
Report to the Future Tech Committee, Society of Vehicle Engineers: As there has been much discussion regarding a new form of propulsion being proposed for motor vehicles, we have been tasked with compiling this report on the technology and its prospects for practical and commercial applications. Here is the executive summary:
As has been widely reported, the proposed technology, combustion of hydrocarbon fuels in a closed cylinder, represents a dramatic departure from the battery-electric powertrains that currently power 99 percent of our nation’s vehicle fleet. Given the radical upheaval this would cause to both the automotive-manufacturing industry and our all-electric recharging infrastructure, a thorough examination of internal-combustion technology is in order before further investment should proceed.
Early prototypes of the new “engine” appear as a heavy and bulky metal casing called a “block,” usually made of iron or aluminum, in which one or several reciprocating pistons are connected to a common crankshaft with rods. The hydrocarbon fuel and air are introduced separately to each cylinder via manifold vacuum or directly under pressure and ignited by a sparking device, whereupon the rapid heating and expansion of the gases displace the piston(s). The process is repeated serially to create continuous crankshaft rotation.
Several technical and market challenges are apparent. The number of moving parts in the “engine” that must be manufactured and machined to fine tolerances is many times that of our current electric motors, which have a single rotating assembly. Also, the best designs are only about 40 percent efficient as waste heat is lost through friction, transfer to the cooling system, and exhaust. Additionally, unlike electric motors, which make peak torque just above zero rpm, the new engine’s torque delivery is by comparison delayed, as it must first develop significant crankshaft rotational speed.
Furthermore, unlike our common electric motors, constant lubrication of the engine’s moving parts is required by a separate supply of hydrocarbon lubricant. This lubricant has a limited life due to contamination and heat cycling and must be replaced periodically. Will today’s motorists, accustomed to nearly maintenance-free electric powertrains, even accept a vehicle that has frequent and possibly costly service intervals in which the used lubricant, a slick and staining material laced with toxic heavy metals, must be safely disposed of?
Further, hydrocarbon combustion in the presence of the two main atmospheric components of nitrogen and oxygen produces substantial noise that will have to be greatly suppressed to be acceptable to both drivers as well as communities accustomed to hearing nothing from a motor vehicle but a faint whine. Also, the chemical reaction produces compounds that some medical experts believe to be unhealthy.
There is also the combustible nature of the fuel. Unlike electricity, which does not leak or evaporate and which has a proven infrastructure for home delivery, hydrocarbon fuel, and specifically its most commercially viable form, gasoline, both leaks and evaporates and is extremely flammable as well as toxic, the odors alone inducing rapid nausea. While batteries can overheat, that is a gin fizz compared with what gasoline does when lit. And to have any meaningful range, vehicles will be required to carry up to 20 gallons of it, enough explosive power to easily destroy the vehicle, its occupants, and surrounding structures.
Thus, the issue of gasoline refueling raises many questions. Obviously, consumers cannot be allowed to refuel at home as they currently do with free electricity from their rooftop solar panels. They will have to drive to a licensed commercial operation outfitted with the requisite specialized equipment. The SVE Safety Committee is already studying the matter and, in consultation with our lawyers, has developed some initial recommendations, such as requiring the driver to leave the vehicle with a trained technician who conducts the refueling in an open-air pit of reinforced concrete wearing some form of blast-proof garment. On the positive side, vast sources of crude oil, the raw form of gasoline, are said to lie in a wide range of locations, from the Alaskan tundra to the coastal waters of California, though the most accessible pockets are beneath the sands of the Middle East. The State Department has noted that increased trade resulting from our bulk purchases of crude oil can only help further cement friendly relations with our many allies in the region.
In summary: The market penetration of the internal-combustion engine is handicapped by several technical hurdles. A small market is possible among machinery enthusiasts of the type who prefer complex mechanical watches to simple and reliable digital timepieces. However, estimating the size of this market would be, at this point, purely conjecture.
Ten months ago, when I wrote My Super Bowl Rooting Record, the coronavirus we now know as Covid-19 was just starting to appear in the United States. The first U.S. death from the virus had occurred a few days earlier, though we didn’t learn this for months. The following post, which is a long overdue prequel to that February post, comes at a time when the virus has irrevocably altered lives around the world, and the positive case, hospitalization, and fatality numbers in the U.S. have reached record highs. Despite the NFL’s efforts, its 2020 season has been continually disrupted, and at this moment it’s not entirely clear that the full season will be completed as planned. While I do not wish to minimize the tragedies that this pandemic has brought upon us, I welcome NFL football as a distraction from them. And so I share this.
I understand why the NFL and its fans like fantasy football. The gamification of player performance helps keep fans engaged and paying attention even when their team isn’t playing, or when their team is playing but is hopelessly inept. Personally, however, I don’t care for fantasy football for this very reason: It causes fans focus on and care about factors that are not directly related to winning and losing games.
Traditional point-spread-style betting is better because it focuses on the outcomes of the games themselves, though still not strictly on the winning and losing. For example, if a team is favored to win by five points and you pick them to cover the spread, you’re disappointed if they win by a field goal, even though from the team’s (and the team’s real fans’) perspective, the victory is all that matters. Last weekend’s Seahawks vs. Eagles game illustrates the point nicely.
In reaction to all of this, I’ve developed my own system for picking games, which I refer to as “Real Fantasy Football.” Simply put, it’s my method of choosing which team to root for in each NFL game. This raises the questions, A) how do I pick the desired winner of each game, and B) how do I measure my success?
It started years ago with a simple rule of thumb that I employed for a long time: A successful NFL weekend for me was one in which the Seahawks, Jets, and Giants all won. The Seahawks because I’ve lived in the Seattle area for most of my adult life and have actively rooted for them for most of that time, the Jets because of my historical and long-suffering allegiance to them, and the Giants because of my legacy loyalty to all things New York.
You might wonder what happens when two of these teams play each other, and the answer is my default hierarchy is Seahawks over Jets over Giants. In rare cases I will flip that order when playoff considerations are at stake. In fact, the only time I ever rooted against the Seahawks in person at a Seahawks home game was in late December 2008, when they hosted the Bret Favre-quarterbacked Jets. The Seahawks were in their final season under coach Mike Holmgren and at 3-11 were already out of playoff contention, while the 9-5 Jets were fighting for a playoff spot. Of course, the Seahawks won. In the snow. The Jets ended up losing the following week as well, missing the playoffs, and firing coach Eric Mangini. In other words, a typical Jets season.
In my current system, I start by dividing each week’s games into three tiers. The top tier includes the games involving those three teams. The middle tier includes games that affect the playoff chances of those three teams. (In recent seasons this has meant the Seahawks’ playoff chances, though in 2020 the weakness of the NFC East means that the games involving the Giants’ division rivals are now in this tier.)
The bottom tier includes all games that don’t fall into either of first two tiers, and typically represents about half the total games. My determination in these games is based on my personal feelings about the teams involved, with several factors involved. Here they are, in no particular order, with some examples:
Preference for teams who have gone the longest without a championship or against those who have won many championships. Helps: Cleveland, Buffalo, Detroit, Minnesota, Philadelphia until 2017, Kansas City until 2019. Hurts: New England, Pittsburgh, Green Bay, Dallas, San Francisco.
Preference against bitter division rivals of my top tier teams. Hurts: San Francisco, New England, Miami, Dallas, Washington.
Preference against teams with offensive nicknames and logos. Hurts: Washington before 2020.
Preference for/against owners. Helps: Jacksonville. Hurts: Washington, Dallas, Las Vegas.
Preference for/against current GMs/head coaches. Helps: Denver, Kansas City, Atlanta. Hurts: New England, Las Vegas.
Note that I do not take betting line favorites into consideration at all when making my picks.
In addition to the tiers, I rank all games in a given week from 1 to N. Typically the bottom few games each week fall into the “I don’t really care who wins, I just want the game to be exciting” category, but even in those games I force myself to pick a preferred winner.
The first three picks were a straightforward application of my top tier. Games 4 and 5 were to help Seattle reclaim first place in the NFC West, and games 6-8 were to help Seattle against its top-ranked conference rivals. Game 9 was to help the Giants claim first place in the NFC East. Games 10 and 11 were based on giving preference to Buffalo and Cleveland, though in this case both of their opponents happened to also be in the “never won a Super Bowl” club. Game 12 was rooting against Las Vegas. The remaining games were based on the mildest of preferences.
How do I measure my success? The simplest manner is total number of successful picks. Also, because my methodology leads me to disproportionately root for underdogs, I compare my results against the designated favorites for each game (going 7-9 when I’ve picked six favorites is better than going 9-7 when I’ve picked ten favorites). Finally, I calculate a weighted score by giving 3, 2, and 1 points per game to each respective tier. (Having the Jets and Giants go a collective 4-18 through the first twelve weeks of the 2020 season has not aided this last metric.)
It turns out that Week 12 was not a representative sample of my record. At 11-5, it was my best week of the season, not only in absolute terms but by having picked five more winners than favorites. (Contrast it with Week 7, in which Cleveland was my only successful pick.) Here is my week-by-week record to date for 2020:
So at the beginning of the season I fell quickly behind the favorites tally. This isn’t entirely surprising, as the who the favorites are gets more accurate when there is more recent regular season game data, especially in 2020, with its absence of preseason games and reduced home field advantage. With a strong Week 12 showing I’ve nearly caught back up.
When I started this system several years ago, I waited until around Thanksgiving to start picking games, figuring that the games didn’t matter until the playoff stakes were evident. But starting last season (2019), I’ve begun with Week l and continued through the playoffs and the Super Bowl. It may appear from this detailed description that I spend a lot of time on this, but in practice it occupies less than fifteen minutes per week. I especially value that I now start each NFL weekend with a clear sense of which games I should care most about.
Every year I revise this system, so 2021 will see some tweaks. I’m considering changing how I calculate the weighted score to give more relative weight to all Seahawks games — realistically, I care a lot more about how the Seahawks do each week than all of the bottom tier games put together — and/or playoff-affecting games that occur later in the season.
Like most NFL fans, especially those who root for perennially mismanaged franchises, I have to decide prior to nearly every Super Bowl which of the conference champions to root for. While I’m old enough to have watched nearly all of them live, I’ve had a local team to root for fewer than ten times. Even if I include Super Bowl III which, while I don’t remember watching live, I have viewed many times on tape and would have had a clear favorite, I can cite at most nine Super Bowls in which I felt a strong pre-game stake in the outcome: 1 Jets, 3 Seahawks, and 5 Giants.
When I don’t have one of those teams to root for, I usually pick based on a combination of the following: Which team has won the fewest Super Bowls, hasn’t won for the longest time, or is the underdog. For example, this year it was easy to root for Kansas City based on the first two criteria, and for a long time it has been easy to root against New England. (My most recent exception to these guidelines is Super Bowl 50, when I chose Denver over Carolina based on what I correctly assumed would be Peyton Manning’s final game.) Based on these and my memories of having rooted for a lot of losing teams in Super Bowl blow-outs (hello, Buffalo), I would have guessed that my preferred team has lost far more often than it has won. But I had never gone back and counted, until now.
So I reviewed the complete list of Super Bowls and reconstructed which ones I watched and for which teams I was rooting. The results were somewhat surprising to me.
Before sharing the results, I should emphasize two points:
I haven’t considered which teams I predicted would win. I’m much more interested in who I wanted to win. (Hello again, Buffalo.)
I’ve tracked who I was rooting for to win at that time, which definitely isn’t always who I would root for based on current knowledge or if the teams played today. For example, I rooted for Washington three times in the 1980s (they went 2-1 in those games), but in retrospect I wouldn’t have rooted for them then given how I feel about the team now. And I rooted for New England in their first four Super Bowls (the two pre-Brady and the first two with Brady), but haven’t rooted for them since.
Of the 54 Super Bowls to date, there are seven that I either didn’t watch or have no memory of watching. Super Bowl V is the first one I have a clear memory of watching live, and starting with Super Bowl VIII I’ve watched all of them with the exception of XVI (the first Cincinnati-San Francisco game). Adding the Jets game as a special case means I’ve watched 48 Super Bowls. In those games, I have a 22-26 record. This is somewhat better than I expected.
As interesting to me as the total record are the recent and the longest streaks. Starting with Super Bowl XLVIII, i.e. the past seven games, there haven’t been two consecutive years in which I’ve ended up on the same side of the desired outcome (largely due to New England’s recent appearances). Super Bowl XLVIII was also the end of my longest streak of successful picks, at five, which started with New Orleans’ win (XLIV). My longest streak of unsuccessful picks is six in the 1990s, starting with Buffalo’s second of four straight Super Bowl losses (XXVI) and continuing through Green Bay’s victory over New England (XXXI). (Those six coincided with the end of the NFC’s record streak of 13 consecutive Super Bowl victories.) I’ve also noticed that in the first eight Super Bowls I watched consecutively as a youth (VIII through XV), I was 1-7, which probably explains why I have stronger memories of rooting for the losing teams.
Here is the full list. (Due to most browsers’ table rendering, you might need to scroll side-to-side to view all of the columns.) For further reference, consult Wikipedia’s list of Super Bowl champions.
TL;DR While the Seahawks’ season-ending game against the 49ers is the most significant factor by far in who will win the NFC West, if you’re a Seahawks fan, you should be rooting this weekend and next for Atlanta, Minnesota, and Philadelphia to win and for Green Bay, New Orleans, and Washington to lose.
It’s time for my late autumn foray into NFL playoff mathematics.
It is well established that the season-ending game between San Francisco and Seattle at CenturyLink Field will in all likelihood decide the championship of the NFC West division. While both teams have clinched playoff berths, the winner of the NFC West will hold at least the third seed in the conference, which guarantees hosting at least one playoff game. This winner also has an excellent shot at either the first or second seed, which would give them a first-round bye. The runner-up would at best earn the fifth seed, meaning no bye and an extremely slim chance of hosting a playoff game. So these teams’ regular season finishes matter a lot.
[N.B. For the purposes of the following discussion, I am ignoring the possibility of tie games.]
Both teams currently sit at 11-3 with two games remaining, and Seattle holds the tie-breaking edge over San Francisco because of its victory in their first matchup last month. As a result, if Seattle defeats San Francisco in their December 29th rematch, Seattle wins the NFC West, regardless of what happens this weekend. For example, in the case where Seattle loses and San Francisco wins this weekend, both teams would finish 12-4, and Seattle would win the NFC West by virtue of a 2-0 record in their head-to-head games.
If either Seattle loses or San Francisco wins this weekend (or both), and then San Francisco defeats Seattle in Week 17, San Francisco finishes at least one game ahead of Seattle and claims the NFC West with a 12-4 or 13-3 record.
This leaves one remaining scenario: Seattle wins and San Francisco loses this weekend, and then San Francisco defeats Seattle in Week 17. Both teams would finish 12-4, and would have split the season’s head-to-head series. Determining the NFC West winner in this situation requires diving deeply into both other teams’ records and the NFL’s tiebreaking rules.
This scenario also raises the following question: Is it possible for Seattle to clinch the NFC West this weekend, irrespective of the outcome of its rematch with San Francisco? The fact that it’s not mentioned in the coverage of this weekend’s games suggests that the answer is no, but since it’s not addressed explicitly, I decided to analyze it myself.
Since we’ve already determined that Seattle and San Francisco have matching won-lost and head-to-head records, we look at the division record tiebreaker: Both teams would be 4-2. Next up is won-lost percentage in common games, where both teams would be 9-3, and then conference record, where both teams would also be 9-3.
The next tiebreakers are strength of victory and then strength of schedule, where things get considerably more unpredictable, so buckle up.
To calculate the strength of victory, you can exclude the teams that both Seattle and San Francisco defeated and their victories against each other, which collectively account for nine of their twelve victories. This leaves the following six teams to consider: Atlanta, Minnesota, and Philadelphia for Seattle, and Green Bay, New Orleans, and Washington for San Francisco. Currently that totals to 22-20 for Seattle and 25-17 for San Francisco. While that seems like a significant margin in San Francisco’s favor, it could break either way, from Seattle finishing with a 28-20 vs. 25-23 edge to San Francisco finishing with a 31-17 to 22-26 edge. (Incidentally, only one of the remaining games, Green Bay at Minnesota this Monday night, involves two of the above teams.) More relevant to this discussion, even if all of this weekend’s games break in Seattle’s favor, the two teams’ strength of victory calculations would be an identical 25-20, so there’s no way that Seattle could take the insurmountable lead in strength of victory required to clinch the NFC West prior to Week 17’s games. Alas, we conclude that even with help from other teams, there are no NFC West-clinching scenarios for Seattle this weekend.
In the rare event that Seattle and San Francisco end the season exactly tied in strength of victory, the strength of schedule tiebreaker favors Seattle. The NFL’s scheduling formula dictates that any pair of teams in the same division have exactly two games with non-common opponents: This season, they’re Minnesota and Philadelphia for Seattle and Green Bay and Washington for San Francisco. Currently this gives Seattle a 17-11 to 14-14 edge, so with the right combination of outcomes in this weekend’s games, Seattle could attain an insurmountable margin in strength of schedule. This is useful, because the next several tiebreakers after strength of schedule are based on differentials in points scored which, due to Seattle’s predilection for winning close games, gives San Francisco a huge advantage.
What all of this means is that if you want Seattle to have the best possible chance to win the NFC West, you are rooting for Atlanta, Minnesota and Philadelphia to win both of their remaining games, and for Green Bay, New Orleans, and Washington to lose both of their remaining games. The one additional complication is that if you want to maximize Seattle’s chances of getting the top playoff seed in the NFC, you want Green Bay and New Orleans to have identical records — I’m not going to explain why here — which means that if New Orleans defeats Tennessee on Sunday, you might want to root for Green Bay to defeat Minnesota on Monday.
Or you can ignore all of this and focus your energy on rooting for Seattle to win its remaining two games.
During my Microsoft career I had the opportunity to work on a number of projects that resulted in patents for which I am listed as an inventor. Setting aside the issues surrounding software patents and patent trolling in general, what pride I possess in any patent attached to my name has almost nothing to do with the patent itself, the filing and granting of which are largely the work of attorneys. The pride comes from utility of the work that spurred the patent application and my personal contributions to that work.
I care little about the several patents where I am listed as a co-inventor along with Bill Gates — for example, 8,341,405: Access management in an off-premise environment — despite whatever reflected glory there is from my association with him in the patent records. These patents were part of a “forward patenting” project into which I put minimal effort, so I deserve little if any credit for any inventions that ensued.
I’ve been thinking about this patent recently because today is the twentieth anniversary of the launch of MSN Messenger 1.0, the product in which the typing indicator debuted. While MSN Messenger is no longer in use, the typing indicator was quickly and widely copied and is still a feature of nearly every chat and messaging app. Whether you use Facebook Messenger, iMessage, WhatsApp, Skype, etc., or a customer support website’s chat plug-in, you encounter the animating dots or “[name] is typing…” message that traces its origin to this work.
David Auerbach explained the typing indicator’s history in detail in Slate five years ago, so I won’t repeat it all here. The summary is that prior to MSN Messenger, real-time chat apps either didn’t indicate to you whether the user on the other side was responding to your message (IRC or AIM), or else they showed you every character that the user typed in real time (Unix talk and ICQ). When we were developing MSN Messenger, we believed we could do something better; something that would give you real-time feedback about whether the other user was typing, while being relatively network efficient and providing the other user a semblance of privacy with their thoughts and typos. I designed and coded the implementation for the detection and communication of the typing indicator, and slapped on a rough UI to create a proof of concept. Once we confirmed with self-hosting that it worked well, David and others designed and implemented the polished UI that shipped in MSN Messenger 1.0.
Since the typing indicator patent covers only the detection and communication mechanisms and not the user interface, David and others who worked on the initial implementation are not listed as inventors on the patent. This does not diminish their contributions to the work. MSN Messenger 1.0 became better known for Microsoft’s IM interoperability battle with AOL, which David has described in his essay Chat Wars.
The reason that the typing indicator patent is significant today — July 22, 2019 — is that the initial application was filed on July 21, 1999, the day before Messenger’s launch, and a US patent typically expires twenty years after filing. This should mean that the original patent protection is no longer in effect as of today, except that Microsoft’s lawyers filed several continuation patents, most recently 7,418,495, so the typing indicator patent has a little bit of life left in it.
That said, Microsoft has never pursued anyone for infringing this patent as far as I know (which is not far). In many cases Microsoft has patent cross-license agreements with the the other products’ owners (e.g. Apple in the case of iMessage) that render the issue moot. Also, it’s possible that newer implementations of the feature don’t infringe on the specific claims made by this family of patents, but I have no idea if anyone has bothered to verify this.
Several years after MSN Messenger first launched, still in the pre-smartphone era, I employed the typing indicator as an example when I visited my child’s first-grade class for one of those “What your parent does for a living” sessions. I created these mock-ups of an IM session to demonstrate to six-year-olds how it worked, and printed them poster-sized:
[The text of the typing indicator was then “is writing” because pen support had been added to Messenger.]
The quality of these mock-ups should illuminate why someone other than me implemented the shipping version of the typing indicator UI.