HOW to predict football results | Big Data or Big Fail?

Predicting football is hard enough for the average fan, but what if you had access to years worth of data, statistical models and even AI technology? Analytics company Stats Perform and investment bank Goldman Sachs attempted to forecast Euro 2020 using their vast resources. Some of their predictions were spot on, others fell flat. Will data and AI ever get to the point where it can accurately predict football? Or are there simply too many variables?

Report: Janek Speight
Editing: Christian Kahle

▸ subscribe: https://goo.gl/Y799RK
▸ help us create subtitles: https://goo.gl/fz2L61
▸ our channel: http://youtube.com/dwkickoff

More football stories here:
Facebook: http://facebook.com/dw.kickoff
Twitter: http://twitter.com/dw_sports
Website: http://dw.com/sports

We kindly ask viewers to read and stick to the DW netiquette policy on our channel: https://p.dw.com/p/MF1G

George Chahrouri and Eric Eager come on to talk about the Rams being Super Bowl Champs, Bengals blown lead, and more

0:00 – Intro
0:58 – Who should be the Super Bowl MVP?
6:12 – Take the L: The Rams win it all
8:56 – Did the Rams win it or the Bengals lose it?
18:35 – Sean McVay is a top ___ coach
25:08 – Where do the Bengals go from here?
28:30 – Power Ranking the AFC teams heading in to 2022
33:40 – What will the NFC look like next season?
38:08 – Biggest questions going into next season
40:46 – Who would have been MVP for the Bengals?
42:25 – Power Ranking the NFC next season
52:29 – Who can make a Bengals-like leap next season?
57:20 – What should teams without established QBs do next season?
1:02:10 – Super Bowl records set tonight
1:06:51 – What do the Vikings do next season?
1:12:30 – Top 5 QBs going into next season

MANSCAPED: Get 20% Off and Free Shipping with the code PFF at http://Manscaped.com

WESTERN & SOUTHERN: Ask Cris Collinsworth anything! Submit your question for a chance to win a catered party for February’s big game: http://WesternSouthern.com/AskCris

DRAFTKINGS: Download the DraftKings Sportsbook app, use code PFF, bet FIVE DOLLARS on the Super Bowl and win 0 IN FREE BETS if your team is victorious: http://bit.ly/2VuVi4i

🏈2021 Fantasy Football Rankings: https://www.pff.com/fantasy/rankings
🚀2021 Fantasy Football Projections: https://www.pff.com/fantasy/projections
​​💥Fantasy Football Stats & Advanced Metrics: https://www.pff.com/fantasy/stats
​​💯PFF NFL Player Grades: https://www.pff.com/grades​
🕶 PFF Premium Stats: https://premium.pff.com/nfl/teams?sea…
Video Rating: / 5

40 Comments

  1. Beaskrishna S

    Nobody predicted corner taken quickly Origi, 6-1,8-2,its balotelli Agueroooo, thats the beauty of the game

  2. Natsu Kiiroi

    Human : Bla bla bla Data Bla bla bla Statistics Bla bla bla Probability
    Paul the Octopus : Hold me Tentacle mate 😂😂😂😂😂😂

  3. Prem Rai

    cup football has, and always will give unpredicted results

  4. Firelayer1

    Big data in millions of years with a Jupiter brain that can simulate the whole history and future (at least the next billions of years) of the universe perfectly: My technology is above your understanding

  5. Golden Fighter

    6:04 Someone is a fan of Dragon Ball

  6. GBGF

    well its like in every other part of life: if you have all information you can say whats going to happen.

  7. Blancos

    Well is obvious, how can you pridict that let's say "chiesa girlfriend cheated on him". That alone would have an effect on the result. There are so many things to consider and for me let's call it "Game data" is 5/10% of a game. Because if after 20 minutes Sergio Ramos breaks Salah's shoulder the game would be billions of times different, not because Salah is jesus but because his presence changes everything.

  8. Amir Limodehi

    Who here came for Leicester city but got trashed

  9. Aurelian

    Humans are irrational. We are not robots. Of course we wont be predicted by mere data. Football may be a team game, but dont forget its made up of individuals.

  10. Padraig McGrath

    The most reliable method of attempting to predict the results of football-matches is the eye-test. You will have a better idea what the strengths and weaknesses of any team are by WATCHING THEIR RECENT MATCHES, NOT BY ANALYZING THEIR STATS. Analyzing stats is just a just a way of simulating reality. To make predictions about future contingents in reality, you need more than a simulation – you need to look directly at the real thing. So your best method is really to just watch games, and analyze them tactically, based upon your tactical understanding of the game itself.
    Then you can also factor in questions like "Are Team A tactically better suited or tactically worse suited than most other teams to exploit Team B's particular weaknesses?"
    Data-driven football will turn out to be pseudo-scientific in the end. The best predictive method is simply to watch and analyze games visually.

  11. light case

    Data is hard for Football because of how injuries and the age of players can affect how creative and well they respond to difficulties, football isnt based on only set pieces its so much more random compared to other sports because of how free-flowing and creative the game is, so I think the data is doing their best with such a unpredictability.

    Belgium unexpectedly had De Bruyne and Hazard injured right before the game so it was unusual for them, but made them crash out, as well as fouls on Lukaku being ignored by the ref because he wouldnt exaggerate fouls

  12. poweroftheriff

    Let's read the darter

  13. ⠀ ⠀

    It’s easier to predict a live game than pre match

  14. James Li

    Goldman Sachs can’t even predict the share price of any random company literally of the next minute, so… these are just guesstimates, no “predictions“.

  15. JUUG LORD

    The Dragon Ball OST really caught me off guard

  16. Dwipayana -

    Run the simulation hundreds thousands times?? Well, try 14.000.605 times next time (if you know, you know)😏😜

  17. sid

    The beautiful game ⚽❤️

  18. bobon123

    Predict or forecast does not mean that you get to a 100% accuracy. What you can get is falsifiable unbiased probabilities: if Italy had a 30% chance of winning against Belgium, an Italy win is not surprising, is not a mistake of the model. Post-hoc, you can check if your model is unbiased by looking at the overall distribution of results, and not just one. This also means that in a model, even if you say "we predict Belgium will win", what it actually means is that Belgium had the highest probability to win – let's say 20%, a decisive advantage given the 4% average win probability for a random team. If you are able to say that a team has a 20% probability of winning while the average team has 4%, this is not an unpredictable game: there is a clear signal in the data noise. However 20% of Belgium winning means that it is still less likely that the prediction is wrong, it is more likely that the most likely team to win does not win. Unless Italy had 0 probability to win accordingly to their model, Italy winning is not a mistake. Low probability events, with probabilities of 4-5%, happens all the time.

    Of course you can get sharper probabilities, and therefore less "surprises", with better models and data. But this does not mean that it is not predictable, it is a matter of degrees of predictability according to quantitative metrics of predictability.

  19. your service

    High probability is not to be confused with exact prediction, that is why many people misunderstand weather forecasting. In weather forecasting a certain condition is compared to past situations with same or similar conditions and their development. And it was weather simulation that led Edward N. Lorenz to the conclusion that only systems with cyclical behavior can be predicted, which doesn't really exist in nature. Nonetheless, when it comes to betting in football, simulations can help to raise the probability of picking the right team. For scientists football is rather a good field to test their models because there is so much data on matches and players.

  20. Lwando Madikizela

    Predicting football is like predicting the weather accurately. It's impossible.

  21. John Blaze

    Diversity @ it’s best. Wokeness lost again.

  22. WaylenWaddle #FinsUp

    Burrow is not a top 5 QB

  23. James Varney

    As far as who could be next year's Bengals, I'd say the Browns, but I am higher on Baker than most at this point.

  24. Aidan MacMullin

    You guys should give the viewers bets from other sports during the off-season

  25. Mike H

    Cam Akers doesn't stink, the Rams' run blocking was absolutely pathetic the whole game. No Tyler Higbee hurt in more ways than missing his presence in the pass game. Bengals lived in the backfield on run calls and only Barry Sanders could've made something out of what was a lot of nowhere to run situations

  26. Godvernment

    Wow a bunch of white guys fat shaming a minority. I’m literally shaking. Why are there so few minorities at PFF? Are you saying minorities couldn’t do just as good of a job as all these WHITE people?

  27. Matthew Teague

    I get that by trade you guys are just looking at stats/analytics… but don’t we have a data point of Belichick eating McVay for lunch as a coach?

  28. George Steiner

    Baseball does suck

  29. Geo McD

    Brady to the 49ers would be ridiculous

  30. Geo McD

    KC have gotten worse each of the last 2 seasons yet they are ranked 2nd in the AFC?? Mahomes wasn't even in the top 5 QBs this season

  31. MustLoveBeards84

    Is Al Michaels retiring? Right before the trophy ceremony it cut to Al and Collinsworth in the booth, and Al said, “Well, it’s been 13 years for us doing the Super Bowl”. And then he clearly teared up. Collinsworth seemed not to notice and talked about something else. Al was clearly moved in that moment.

    He’s so freaking underrated— he’s just one of those guys that is so good at his job, makes it look so easy, it goes unnoticed.

    Anyone got any data on Al?

  32. Tomahawk_Nole

    MVP should go to the refs for calling that BS PI on 3rd and goal 🤦‍♂️🤦‍♂️

  33. David Butler

    Glad to see the long suffering Stafford win a championship. Gave the best years of his career to one of the worst organizations in the NFL. Though the Lions overall have won 4 NFL titles, they haven’t been very successful since the 90s. Despite all the losing, Stafford quietly did his job and honored his contracts. Given the opportunity, in this his 13th year to QB a Rams team, built specifically to win a NFL title, he delivered. Whatever, u say about Stafford, the man has paid his dues in triplets, and deserves this title.

  34. joe short

    Outro song anybody?

  35. joe short

    I do not understand the constant chargers hype. Meanwhile oakland gets treated like garbage. Meanwhile chargers lost to the Raiders for a chance to go to the playoffs. Mind you, raiders lost a very good coach, their best wr, waller was banged up and Jacobs has been banged up for 2 yrs. What wr Corp would you rather have? Raiders or chargers? D line and o line are a wash but I'll argue chargers have a better roster. Chargers need to prove their what you keep saying they are by winning something that matters and for the love of god give raiders some respect. Now with the coaching changes I wouldn't put raiders in the top 10 but I damn sure wouldn't put chargers there either… They've won nothing and proved nothing. Then again steelers keep making the playoffs and they keep getting ignored also so 😆

  36. semicharmedlife311

    Well, Kupp was chosen MVP, George. In my mind, he was the most deserving and congrats to him.

  37. Rodney Spears

    You guys completely forgot about Tomlin

  38. Aryaan Khan

    Tell Eric to stop breathing into the mic so heavily when he's not speaking. Sounds like he just ran a marathon. All love though.

  39. Qwerty Asdfg

    What a bullshit call on 3rd and goal……..

  40. RamsnDodgersfan

    I would not put Wilson as high as they did, he’s regressed on the backside of his career. Stafford should be 6.

Comments are closed