Winning In The NFL Through Luck And Skill


Unlike to the Premier League, the length of the NFL season allows "luck" to have a greater impact. As the NFL comes to London, today on the blog Mark Taylor discusses luck and skill in the NFL.

It is easy to jump to seemingly sensible, yet ultimately hasty and flawed judgements surrounding the relative merits of sporting teams, based on very little evidence. However compelling the evidence appears to be or reasonable the subsequent conclusion, it is almost always better to wait until many more matches have been played before making definitive assertions.

Team Success In The Premier League

On opening weekend of the Premier League season, two sides that can reasonably expect to finish at opposite ends of the table, enjoyed unremarkable results. Manchester City had to wait until Monday night to start their campaign, but won comfortably at home to Newcastle, courtesy of two goals in each half, helped by a Newcastle red card to smooth their way to three points.

In contrast, newly promoted Cardiff fell tamely to West Ham by two goals to nil on opening Saturday, leaving their manager, Malky Mackay ruing his side’s failure to take their chances.

So, the early results appeared to confirm the general opinion about the relative merits of the two Cities in advance of their meeting in Cardiff the following week. Pre match estimates of Manchester City’s superiority in week two’s meeting, gave them in excess of a 70% chance of returning with three points.

Over the course of the 90 minutes, the recent Premiership champions, dominated possession, had more shots, led early in the second period, but ultimately lost as Cardiff showed a ruthless finishing streak, by claiming three close range goals, two from corners, to edge a five goal thriller.

It is all too easy to forget the results from week one and weave an equally plausible narrative, whereby Manchester are a pale imitation of the 2011/12 side and Cardiff will cruise to mid table respectability, come May. The truth is probably less dramatic. Manchester City is almost certainly a championship contender and Cardiff remains likely candidates for relegation. But on one match day in late August the way in which Cardiff’s limited shots turned into more goals than Manchester managed to produce from their much more numerous attempts gave them an unexpected victory to savour.

Team Success In The NFL

Three thousand miles away and a few weeks later, the NFL season also kicked off its 2013 campaign on Thursday, with Super Bowl winners, Baltimore taking a sound beating at the hands of arguably last year’s best overall team, Denver. The season continued on Sunday through to Monday and, as with the EPL the matches produced a few upsets. Buffalo very nearly hung on as 10 point home underdogs against New England and Tennessee did manage to upset home favourites, Pittsburgh.

Manchester City will rue the loss to Cardiff, but the 38 game format of the EPL will allow them many chances to recover the lost ground and allow opportunities for near rivals to similarly slip up, but by contrast, Pittsburgh and Baltimore have already used up over 6% of the total regular season games in a losing effort. In short, random chance can decide matches in any sport, but in some the length and format of the competition gives “luck” a much greater say in deciding the ultimate winner.

Over the lifetime of the EPL just five sides have lifted the trophy, both Manchester sides, Arsenal, Chelsea and Blackburn, compared to the NFL, where twelve different sides have lifted the Vince Lombardi Trophy over the same timescale. The more competitively balanced NFL owes much to the draft system, but also the division and conference format, followed by a playoff decider. To win the EPL you simply need to be one of the best four teams in the league, to win the Super bowl you need to be good, but also you have to be “lucky”.

To illustrate the grip the format has on deciding the best NFL side for each season, we can simulate entirely seasons many times, while incorporating the likely differences in team quality, as well as the unbalanced schedule. In a competition comprising 32 teams, each side only plays 13 unique opponents during regular season play.

An Analysis: Baltimore Ravens 2012

Baltimore, as the current holders of the Super Bowl is the natural team to begin with. Their 10-6 record was just good enough to win the AFC North ahead of Cincinnati. The number four seeding meant they hosted a game on wildcard weekend, won an improbable victory in double overtime at Denver, the side with the best regular season record and then beat New England on the road and finally survived San Francisco camping in their red zone as time expired in the Super Bowl.

So their post season run was both magnificent and unlikely, but simulations of their regular season emphasises that a side in a competitive division such as the AFC North was in 2012, also needs the random variation present in single match outcomes to fall in their favour. Simulations for the AFC North see Baltimore averaging just over 9 season wins, likely not enough to even make the playoffs in most years. Chicago in the NFC conference won ten games and failed to make the playoffs in 2012. Around 3% of the simulations saw Baltimore win 13 matches, good enough for top seeding, but at the other extreme, they also won just four matches around once every 200 simulations. They weren’t even the most likely winners of the division. Pittsburgh was and even Cleveland, the weakest of the four AFC North sides, won the division around 3% of the time.

So even with the benefit of inputs that reflect team ability, the ultimate winner at the end of a mere 16 game season, followed by a three or four game post season, may not even have been the most likely long-term winner of their division.

Multiple Simulations of the AFC North in 2012.

 Team Percentage of Simulated Divisional Titles Won for 2012
Pittsburgh  41%
Baltimore  38%
Cincinnati  18%
Cleveland  3%

In addition to demonstrating how the best side is far from guaranteed to top a four team, 16 match table, other staples of sporting contests, such as home field advantage also occasionally throws up less likely outcomes if we simulate enough seasons.

Home field advantage in the NFL is worth around three points and over all simulations of the AFC North using team strengths from 2012, the win percentage for all home sides was around 57%. This is perfectly in keeping with a three point home field advantage. However, in nearly 20% of seasons, a side actually won more away games than they did home ones. Another case of random variation hinting at an ability, (in this case, a side appearing to be more able to win more often on the road), when all we are really seeing is an unusual, but expected run of results being produced over the truncated scale of just 16 matches.

Additional Analysis: Denver Broncos 2012

The picture was slightly more optimistic for Denver, the AFC’s top seed with a 13-3 regular season record in 2012, playing in the relatively uncompetitive AFC West. Despite being the favoured side in around 13 of their 16 matches, they still only “won” their division a shade over 70% of the time over the course of the simulations.

Multiple Simulations of the AFC West in 2012.

 Team Percentage of Simulated Divisional Titles Won for 2012
Denver  71%
San Diego  20%
Oakland  6%
Kansas City  3%

Their average number of regular season wins was a more impressive eleven, but they too managed to record around 16% of seasons where their eight wins or fewer wouldn’t have merited post season play. A single near perfect 15-1 year every 200 simulations was balanced by a lowly 4-12 effort in the same timeframe.

Although the likely distribution of regular season wins clearly indicates that Denver were, long-term the better side than AFC West opponents, Kansas City, the tails representing Kansas’ best efforts and Denver’s worse also overlap. Indicating that Manchester City at Cardiff style random variation in the NFL could, over a short, 16 sequence of matches, conceivably elevate the division’s least competent side above its best.

Concluding Thoughts

The lack of a threat of relegation obviously softens the blow that the impact random variation can have on individual seasons in the NFL. Skill only gets its partial just deserts, but the upside is the relative unpredictability of the spectacle that keeps fans watching and the opportunity for rewards to be spread around the teams. It is inconceivable that Manchester United will finish outside of the top three or four in the EPL or Cardiff really will prove superior to Manchester City. 38 matches, coupled with the differences in talent will almost always guarantee that. But closer matched, divisional dead weights can go from last to first in the NFL.

Betting on the lesser Premiership teams to challenge for the ultimate title is a futile exercise, but betting on divisional or Super Bowl winning potential is much less of a long shot, if you are prepared to accept that many may have failed to look past last year’s standings and see the continued presence of the “luck”, both good and bad, amongst the more obvious talent.



Looking for NFL betting tips? Check our NFL Betting Tips Board for all our community's NFL betting picks right throughout the season.



Read more of Mark's work on his The Power Of Goals blog

And follow Mark of Twitter: @MarkTaylor0

NFL and football fan. I've seen my two favourite sides, Stoke and the San Diego Chargers play at the new Wembley....and both lost.