Go… Advanced Analytics? The Glorious Minutiae of Data in Pro Sports

A talk about why advanced analytics in sports are super-cool, and how far numbers have come in helping us analyse sports and make both athletes and teams better, as well as helping us wina rguments on the internet!

Transcript

1. Go… Advanced Analytics? The Glorious Minutiae of Data in Pro Sports @IvanaMcConnell More cool metrics, interesting reading, and people to follow at the end!
4. advanced analytics
5. advanced stats analytics
6. advanced metrics stats analytics
7. saber metrics metrics stats analytics advanced
8. empirical analysis of sports metrics saber metrics stats analytics advanced
9. What makes them advanced? They matter. They (try to) answer the complicated questions!
10. Objectivity & depth Not just goals/points scored: • “Which player contributed most positively to their team's offense?” • “Which NBA point guard contributes most to teammates’ success?” • “How many touchdowns is this player likely to score this year?”
11. Evaluation • “Based on the past few years, what kind of season should I expect from my favourite team this year?” • “What should the Vegas odds be for this season?”
12. Validation • “Which player should I love/hate/offer a terrible contract to?” • “Which player is secretly great?” • “Which player looks great, but is actually awful?” • “How do I prove this other fan objectively wrong?”
13. & SportVU Records the X and Y position of players and X, Y, Z position of the ball (in real time), 25 times a second!
14. Win Shares How many W’s is each player responsible for?
15. Win Shares • 95-96 Chicago Bulls: 72 wins, 10 losses! • Michael Jordan: 20.4 WS • Ron Harper: 6.3 WS http://www.basketball-reference.com/
16. WS/48 • Win Shares per 48 minutes played! • Michael Jordan: 20.4 WS, .317 WS/48 • Ron Harper: 6.3 WS, .161 WS/48 http://www.basketball-reference.com/
17. WS: The great… • LeBron James and the 2015-16 Cleveland Cavaliers: 57-25 • James’ WS: 13.6 (.242 WS/48) • Without him, no playoffs! • In the playoffs? 4.7 WS (.274 WS/48) http://www.basketball-reference.com/
18. …and the terrible! • Emmanuel Mudiay, Denver Nuggets point guard • 2015-16 WS: -2.1 • He made his team objectively worse. Ouch. http://www.basketball-reference.com/
19. EFG% Effective field goal percentage
 Takes into account that three-pointers are more valuable!
20. ALL THE THINGS The power of combining advanced metrics
21. https://projects.fivethirtyeight.com/carmelo/
22. DVOA Defense-adjusted Value Over Average
23. DVOA • Gives proper credit for scoring points and winning games • How much better (%) are you than average, in a given play? http://footballoutsiders.com/
24. DVOA: 2015-16 NFL season • Carson Palmer (the good): 34.4%, (1st) • Brock Osweiler (the bad): -3.2% (21st) http://footballoutsiders.com/
25. During the offseason…
26. DVOA: 2016-17 NFL season • Brock Osweiler (the unequivocally terrible): -27.1% 
 
 
 
 
 
 • Rock bottom :( http://footballoutsiders.com/
27. & Statcast Radar equipment tracks every object and person on the diamond ⚾
28. Launch angle The vertical angle at which the ball leaves a player's bat after being hit. ⚾
30. Wins Above Replacement “wins over the perceived value of a replacement-level* player at each position” ⚾ * replacement-level = a player added to the team with minimal cost and effort
31. Wins Above Replacement • “a player’s total contribution to the team” • Solid starter = 2-3 WAR/season • Superstar = 5-6 WAR/season • Mike Trout, MVP 2015-16: 10.5 WAR ⚾
32. Weighted runs created (wRC+) “This player was worth X runs to the team this year, after controlling for ballpark” ⚾
33. Weighted runs created (wRC+) • A player’s total offensive value measured by run, controlled for ballpark • League average = 100 • An amazing player has a wRC+ of < 160 • An awful player has a wRC+ of < 50 ⚾
34. ⚽ Expected goals (xG) How many goals your team should expect to score based on its actions on the field
35. ⚽ Expected goals (ExpG) • Just because a team creates lots of chances doesn’t mean it creates good ones! • Expected goal difference = ExpG scored - ExpG allowed. • The greater the difference, the better the team! http://www.sportsnet.ca/soccer/soccer-analytics-expcect-goals-advanced-stats-premier-league/ https://www.theguardian.com/football/2017/mar/30/expected-goals-big-football-data-leicester-city-norwich
36. ⚾ Pythagorean Record How many games “should” your team have won?
37. Arguments against it Not everyone loves advanced stats (hard to believe, I know!)
39. Moderation They’re not meant to replace the game!
40. Let’s revisit that Steph Curry shot! What do advanced metrics say about it?
42. How’s this? He makes three-pointers just as well with a defender 2 feet away… …as an average NBA shooter does with the nearest defender 12 feet away. http://fivethirtyeight.com/features/stephen-curry-is-the-revolution/
43. Or this? He shoots better as the clock winds down! http://fivethirtyeight.com/features/stephen-curry-is-the-revolution/
44. One more! He shoots the notoriously difficult pull-up three at a rate of 42%. The rest of the league? 28%. The rest of the league’s average on regular three-pointers? 35%. http://fivethirtyeight.com/features/stephen-curry-is-the-revolution/
46. Thank you!
47. Want to read more? • https://sports.vice.com/en_ca/article/one-jungle-share-at-a-time-meet-the- bill-james-of-esports • http://grantland.com/features/expected-value-possession-nba-analytics/ • http://www.beyondtheboxscore.com/2014/5/26/5743956/sabermetrics-stats- offense-learn-sabermetrics • http://bleacherreport.com/articles/1813902-advanced-nba-stats-for- dummies-how-to-understand-the-new-hoops-math • https://fivethirtyeight.com/features/the-new-science-of-hitting/ • https://www.hudl.com/blog/sports-analytics-101-an-intro-to-advanced-stats
48. Some more cool metrics: • Basketball: Points per possession (PPP), offensive & defensive rating (ORtg/DRtg), player efficiency rating (PER) - Basketball • Soccer: Total shots ratio (TSR), key passes • Baseball: On-base plus slugging percentage (OPS), batting average on balls in play (BABIP) • NFL: Quarterback rating (QBR)
49. Some awesome sports-numbers folks: • Basketball: Kirk Goldsberry, Zach Lowe, Kevin Pelton, Kevin Arnovitz • Soccer: Michael Caley, Mike Goodman • NFL Football: Bill Barnwell, Peter King • Baseball: Nate Silver, Ben Lindbergh
50. Writing-places and stats sources: • FiveThirtyEight • Grantland (RIP) • The Ringer • FanGraphs • FootballOutsiders • Statsbomb