2016 NFL draft running backs analyzed with machine learning

The following is a list of the 2016 running backs drafted in the order they were taken. After the name of the team that took them, we compare the draft picks with out Machine Learning running back analysis system to see what teams took “yes” players and which teams took “no” players. The players are rated yes or no if they would have been drafted and the probability the system thinks they fit in that category.

  • Ezekiel Elliot –  Cowboys – Yes – 99.862 %
  • Derrick Henry – Titans – Yes – 99.91 %
  • Howard Jordan – Bears – Yes – 99.992 %
  • Devontae Booker – Broncos – Yes – 99.994%
  • Kenneth Dixon – Ravens – Yes – 99.79%
  • Paul Perkins – Giants – Yes – 99.98%
  • CJ Prosise – Seahawks – No – 97.16%
  • Alex Collins – Seahawks – Yes – 99.991%
  • Kenyan Drake – Dolphins – No – 99.28%
  • Jonathan Williams – Bills – No – 97.98%
  • Tyler Ervin – Texans – No – 98.44%
  • Kelvin Taylor – 49ers – No – 98.90%
  • Keith Marshall – Redskins – No – 98.79%
  • Daniel Lasco – Saints – No – 98.30%
  • DeAndre Washington – Yes – 99.993%
  • Andy Janovich – Broncos – No – 97.38%
  • Wendell Smallwood – Eagles – No – 97.16%
  • Dwayne Washington – Lions – No – 99.22%
  • Dan Vitale – Bucs – No – 98.68%
  • Derek Watt – Chargers – No – 96.89%
  • Keenan Reynolds – Ravens – Yes – 99.18%
  • Zac Brooks – Seahawks – No – 97.79%
  • Darius Jackson – Cowboys – No – 98.15%

Machine Learning and NFL running backs draft analysis

NFL Running Back Analysis with Machine Learning

The following is a list of the early round (mostly 1st and 2nd) draft picks, at running back, drafted between 2009 and 2014. Our machine learning running back analysis system was applied to determine what picks were “yes draft” players and which were “no draft” players based only on college stats. The players are rated yes or no if they would have been drafted and the probability the system thinks they fit in that category.

In testing, our system was 86% accurate in finding productive NFL running backs.

*** THIS ANALYSIS IS BASED SOLEY ON COLLEGE STATISTICS OF THE PLAYERS BELOW.

If you picked “yes draft” backs for 2009-2014, you would have had an 86% chance of selecting a productive back.

If you picked “no draft” backs for 2009 – 2014, you would have had a 41% chance of selecting a productive back.

RB data read successfully!

name predict prob

    bernard|giovani [yes] 0.999922
      moreno|knowshon [yes] 0.999879
      wells|chris [yes] 0.999879
      mason|tre [yes] 0.999856
      bell|lavian [yes] 0.999843
      mccoy|lesean [yes] 0.999813
      hillman|ronnie [yes] 0.999795
      sanky|bishop [yes] 0.999777
      hyde|carlos [yes] 0.999777
      hill|jeremy [yes] 0.999757
      williams|ryan [yes] 0.999711
      martin|doug [yes] 0.999622
      james|lamichael [yes] 0.99962
      richardson|trent [yes] 0.999551
      mathews|ryan [yes] 0.999362
      ingram|mark [yes] 0.99917
      ball|montee [yes] 0.996881

    name predict prob

        spiller|cj [no] 0.997899
        sims|charles [no] 0.997508
        brown|donald [no] 0.995403
        lacy|eddie [no] 0.994026
        best|jahvid [no] 0.993481
        vereen|shane [no] 0.993481
        leshoure|mikel [no] 0.98901
        green|shonn [no] 0.988013
        wilson|david [no] 0.983055
        pead|isiah [no] 0.981617
        mccluster|dexter [no] 0.97806
        coffee|glen [no] 0.976095

      Below you can visual where the draft and no draft players stand relative to each other by year and “yes draft” – “no draft”.

      unknown-1

      unknown-2

      unknown-5

      unknown-6

      unknown-3

      unknown-7

      If you picked “yes draft” backs for 2009-2014, you would have had a 86% chance of selecting a productive back.

      If you picked “no draft” backs for 2009 – 2014, you would have had a 41% chance of selecting a productive back.

      We will soon release our analysis of the 2016 running backs that were drafted.

Should NFL Teams Sit or Play Rookie Quarterbacks

PHILADELPHIA — Andy Reid’s plan was fairly simple. After taking Donovan McNabb with the No. 2 pick in the 1999 draft, Reid didn’t want his young quarterback to play until he was ready.

And until the team around him was ready. That was a big part of the plan, too. That’s why Reid signed free-agent quarterback Doug Pederson to start while McNabb learned. Pederson knew the offense from his time with Reid in Green Bay, and he could make sure the team lined up correctly and that everyone was on the same page.

Pederson started the first seven games of the 1999 season. McNabb took over after that. He remained the No. 1 quarterback for a decade.SIt or Play Rookie QBS

2016 NFL Draft Worst Quarterback Value Picks

Applying our machine learning system to the quarterbacks in the 2016 NFL draft, the worst values are as follows. The system ranks quarterbacks as yes or no draft and assigns a probability based only on the players college statistics.

1. Jets – Christian Hackenberg – no (73%)
New York Jest took Christian Hackenberg as the 4th quarterback in the 2016 draft. Our machine learning QB system rated him a “no” with an 73% probability.

2. Patriots – Jacoby Brissett – no (98%)
New England selected Jacoby Brissett as the 5th quarterback in the draft. Brissett had our systems highest no ranking at 98%. We suspect he may be utilized in other than a QB capacity. We will see…

3. Raiders – Connor Cook – no (84%)
Oakland drafted Connor Cook as the 7th QB taken yet he was a “no” draft with an 84% ranking.

Here are some more numbers to consider from our system:

Drafted only 23 of 32 starting QBs for 2015
12 playoff teams in 2015, the system drafted 10 QBs (83%)
2015 Divisional Round the system drafted 7 of 8 QBs (87%)
2015 Conference Finals, the system drafted all 4 QBs (100%)
Super Bowl 50 teams, the system drafted both QBs. (100%)
Drafted every 2015 division winning quarterback except one – Hoyer. (87%)

2016 NFL Draft Best Quarterback Value picks

Applying our machine learning system to the quarterbacks in the 2016 NFL draft, the best values are as follows. The system ranks quarterbacks as yes or no draft and assigns a probability based only on the players college statistics.

1. Dolphins – Brandon Doughtey – yes (86%)
Miami took Brandon Doughtey as the last quarterback in the 2016 draft. Our machine learning QB system rated him a “yes” with an 86% probability.

2. Bills – Cardale Jones – yes (97%)
Buffalo selected Cardale Jones as the 9th quarterback in the draft. Jones had our systems highest yes ranking at 97%.

3. Chiefs – Kevin Hogan – yes (82%)
Kansas City drafted Kevin Hogan as the 10th QB taken yet he was a yes draft with an 82% ranking.

4. Lions – Jake Ruddock – yes (83%)
Detroit took Michigan QB Jake Ruddock as the 12th quarterback in the draft. Our system had Ruddock as a yes with an 83% prorbability.

Here are some more numbers to consider from our system:

Drafted only 23 of 32 starting QBs for 2015
12 playoff teams in 2015, the system drafted 10 QBs (83%)
2015 Divisional Round the system drafted 7 of 8 QBs (87%)
2015 Conference Finals, the system drafted all 4 QBs (100%)
Super Bowl 50 teams, the system drafted both QBs. (100%)
Drafted every 2015 division winning quarterback except one – Hoyer. (87%)

2016 NFL Draft – Quarterbacks Analyzed with Machine Learning

The following is a list of the 2016 Quarterbacks drafted in the order they were taken. After the name of the team that took them, we compare the draft picks with out Machine Learning QB analysis system to see what teams took “yes” players and which teams took “no” players. The players are rated yes or no if they would have been drafted and the probability the system thinks they fit in that category.

Jared Goff – Los Angeles – yes – (89%)
Carson Wentz – Philadelphia – yes – (83%)
Paxton Lynch – Denver – yes – (95%)
Christian Hackenberg – NY Jets – no – (73%)
Jacoby Brissett – New England – no – (98%)
Cody Kessler – Cleveland – yes – (90%)
Connor Cook – Oakland – no – (84%)
Dak Prescott – Dallas – yes – (56%)
Cardale Jones – Buffalo – yes – (97%)
Kevin Hogan – Kansas City – yes (82%)
Nate Sudfeld – Washington – no – (81%)
Jake Rudock – Detroit – yes – (83%)
Brandon Allen – Jacksonville – no – (95%)
Jeff Driskel – San Francisco – yes – (60%)
Brandon Doughty – Miami – yes – (86%)

As we watch the class of 2016 quarterbacks progress, it will be interesting to look back at this list and see where they stood at the day of the draft according to our Machine Learning QB analysis system

Machine Learning Picks Superbowl Quarterbacks

Our Machine Learning System Helps Identify Superbowl winning Quarterbacks based only college statistics.

The following is a list of the Super Bowl winning teams and the quarterbacks for the last 21 years. We were unable to find sufficient college data to analyze Kurt Warner.

Super Bowl 30. Troy Aikman
Super Bowl 31. Brett Favre
Super Bowl 32. John Elway
Super Bowl 33. John Elway
Super Bowl 34. Kurt Warner (insufficient college stats)
Super Bowl 35. Trent Dilfer
Super Bowl 36. Tom Brady
Super Bowl 37. Brad Johnson
Super Bowl 38. Tom Brady
Super Bowl 39. Tom Brady
Super Bowl 40. Ben Roethlisberger
Super Bowl 41. Peyton Manning
Super Bowl 42. Eli Manning
Super Bowl 43: Ben Roethlisberger
Super Bowl 44: Drew Brees
Super Bowl 45: Aaron Rogers
Super Bowl 46: Eli Manning
Super Bowl 47: Joe Flacco
Super Bowl 48: Russell Wilson
Super Bowl 49: Tom Brady
Super Bowl 50: Peyton Manning

We analyzed all of these quarterbacks with our machine learning system. The system considers ONLY college statistics of the quarterbacks. It classifies the QB as a draft or no draft and assigns a probability. The results are amazing (see below):

Yes Draft
aikman|troy – [yes] 68.04%
elway|john – [yes] 87.82%
brady|tom – [yes] 51.46%
roethlesberger|ben – [yes] 79.81%
manning|peyton – [yes] 56.32%
manning|eli – [yes] 84.23%
brees|drew – [yes] 99.82%
flacco|joe – [yes] 78.36%
wilson|russel – [yes] 81.45%

No Draft
dilfer|trent – [no] 70.58%
johnson|brad – [no] 98.53%
favre|brett – [no] 81.53%
rodgers|aaron – [no] 62.96% ***

***Left college after 2 years affecting stats

So of the 20 Superbowl wins, the yes picks of the system had 16 of the 20 or 80% of the wins. The no picks had 4 wins and there were no repeat winners. The no picks had 20% of the Superbowl wins in the last 20. We would not bet against Aaron Rodgers repeating as his college stats are skewed by leaving college after 2 year.

NFL Player Seeking Math PhD from MIT

The National Football League offseason is supposed to be a time for players to relax, recover from the brutality of the prior season, and prepare for the next one.
Not for Baltimore Ravens player John Urschel. The 6’3”, 305-pound offensive lineman will begin a PhD in mathematics at the Massachusetts Institute of Technology this year. The Hulk-like math geek, who graduated from Penn State with a 4.0 grade point average, will study spectral graph theory, numerical linear algebra, and machine learning.

NFL Player MIT PhD

Machine Learning and big data

Machine Learning and big data

Machine learning is a fundamental tool in creating a world that can sense and react to dynamic, distributed phenomena. The number of variables and factors that can be taken into consideration by this methodology is unlimited.

It weaves together real-time data collection with the automation of business processes, and is ideally suited to deal with complex, disparate data sources and the high number of variables involved in data sets that are large, diverse and fast changing. In other words, machine learning is primed to handle big data.

Wherein traditional analytics tools are limited by data volume and the need for human interaction to specify program execution, machine learning offers the scale, speed and accuracy needed to truly uncover the full value of big data.

– See more at: http://www.information-age.com/it-management/strategy-and-innovation/123460772/machine-learning-set-unlock-power-big-data#sthash.2LEaz6Pz.dpuf

Machine learning trending on Google

A recent query using Google Trends shows an interesting level of interest in machine learning over time (see figure below). There was an emergence in hype around the 2005 time-frame and led to a cooling off period, but once big data started heating up around 2010, the upward swing in interest continues until today. The good thing is that “machine learning” really is just a confluence of related disciplines like computer science, statistics and mathematics. These fields aren’t going anywhere and neither is machine learning. Statistical learning is here to stay!

Machine Learning Trending on Google

Machine Learning NFL Quarterback and Draft Prediction & Analysis