Luke Bornn
@LukeBornn
Followers
10,269
Following
47
Media
11
Statuses
51
Co-Founder and Chief Scientist, @ZelusAnalytics . Formerly @SacramentoKings , @SFU , @ASRomaEN , @Harvard
California
Joined September 2011
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
#MetaCensorship
• 384456 Tweets
NABJ
• 295430 Tweets
Black Journalists
• 149476 Tweets
Léon Marchand
• 144793 Tweets
Social Security
• 119624 Tweets
Marta
• 91183 Tweets
Kartel
• 66932 Tweets
Bungie
• 53903 Tweets
RIGI
• 51712 Tweets
Jamaica
• 46816 Tweets
billie
• 46323 Tweets
Alcaraz
• 41568 Tweets
Rachel Scott
• 40595 Tweets
対象2作品
• 39322 Tweets
#alertemédaille
• 35524 Tweets
Powell
• 35505 Tweets
38 ZAM
• 34661 Tweets
South Sudan
• 33879 Tweets
Katie Ledecky
• 33645 Tweets
Phelps
• 32566 Tweets
Sony
• 32290 Tweets
Collins
• 32192 Tweets
Tatum
• 30759 Tweets
Curry
• 28588 Tweets
Espanha
• 24558 Tweets
Kitu
• 21257 Tweets
Harris Faulkner
• 16003 Tweets
Summerville
• 13589 Tweets
岡慎之助
• 12694 Tweets
Kerr
• 11289 Tweets
Tapia
• 10656 Tweets
A reminder, blatantly plagiarized from
@stat_sam
, of why radar plots are misleading. Eye focuses on area, not length.
38
869
1K
Here's a handful of free online resource/textbooks to get you started:
*
*
*
*
*
*
*
*
With
@LukeBornn
's call for sports analysts to broaden their technical skills in mind, which are the best online courses (or textbooks etc.) for learning quantitative skills which are useful across a wide range of domains?
1
0
11
6
114
760
Sports scientists -- Here's evidence that the scientific literature may be grossly over-estimating the value of acute:chronic workload ratios in predicting injuries:
Talk:
Paper:
Code:
7
88
489
My lab has had 11 Sloan papers over the last 5 years:
'14: EPV
'15: Counterpoints, Move or Die
'16: Pressing Game, Court Realty
'17: Possession Sketches, Scorekeeper Bias
'18: Open Spaces, NFL Injury, NBA Replay, Deep Learning Trajectories
here's a summary thread of them all:
4
63
342
Simpson's Paradox in Basketball: Shooting improves as defenders guard closer (!!). Must condition on shot location. h/t Nate +
@afranks53
7
78
219
Decided in the fall that this will be my last year authoring papers at Sloan. As such, this thread is a great (and complete) resource for the 18 papers we’ve authored there since 2014. SSAC has been really good to me and my students — v thankful for doors opened and friends made!
My lab has had 11 Sloan papers over the last 5 years:
'14: EPV
'15: Counterpoints, Move or Die
'16: Pressing Game, Court Realty
'17: Possession Sketches, Scorekeeper Bias
'18: Open Spaces, NFL Injury, NBA Replay, Deep Learning Trajectories
here's a summary thread of them all:
4
63
342
6
20
204
People usually point to collection bias (events never "exactly on the line") and rounding, but there's more. Tagging locations on a 105x68 pitch is really hard: 2-4m errors are normal. But lines provide a reference point, so tagged locations become much more accurate around them.
I was looking at the Premier League event data and noticed that lines are almost visible to naked eye. I think I have seen someone mentioning this before, but it should be a data entry issue, right? No reason for players to avoid lines. Can't find the reference for it.
12
2
102
1
8
112
While watching this 0-0 draw, thought I'd try visualizing some personal golf data. The intersection of golfers and sports analysts couldn't fill a foursome, but perhaps a couple of you are interested. First up, a round at Bandon Dunes from earlier in the summer. 1/5
6
0
39
2019a: Soccer EPV (w/
@JaviOnData
,
@dcervone0
). We can measure at every instance the expected value of the current possession by embedding deep learning (to capture 22-player spatio-temporal dynamics) within Markov models (for explainability).
1
7
39
2018a: Open Spaces (w/
@JaviOnData
). Using tracking data, we can measure how soccer players create space both for themselves and for their teammates.
#SSAC18
1
5
27
And lastly, credit for these projects belongs with the students who lead them:
@dcervone0
,
@alexdamour
,
@afranks53
,
@_amiller_
,
@IavorBojinov
,
@MattvanBommel
,
@JaviOnData
,
@OSPpatrick
, Nate, Yatao, Nazanin. Also, to
@Harvard
,
@SFU
,
@SacramentoKings
for supporting the research.
1
1
25
2018b: NFL Injury (w/
@OSPpatrick
). Turns out predicting injuries is hard. We frame the problem differently, and see that both overall player load and impacts are predictive of injury in the NFL.
#SSAC18
1
8
23
2018d: Deep Learning Trajectories (w/ Y Zhong, N Mehrasa, G Mori). We can use deep learning to extract meaningful features from player tracking data. Can identify which NBA team is on the court (using just tracking data) with 95% accuracy.
#SSAC18
1
7
23
2018c: NBA Replay (w/ Nate Sandholtz). By simulating seasons at the micro-second level, we can test hypotheses like "what happens if player X takes less mid-range shots in the first 12 seconds of the shot clock."
#SSAC18
2
3
18
2019e: Fast Not Loose (w/
@yuorme
,
@Chris_LogiQ
, Mehrsan Javan). We take a comprehensive look at pace in hockey from a spatio-temporal perspective, and subsequently study how it correlates with other events of interest.
1
3
16
2014: EPV (w/
@dcervone0
,
@alexdamour
,
@kirkgoldsberry
). We can measure at every instance the expected value of the current play, given all spatio-temporal info. Can be used to value player decision-making.
#SSAC14
1
2
16
2016a: Pressing Game (w/
@IavorBojinov
). Using tracking data, we measure and visualize where soccer teams create defensive pressure throughout a match.
#SSAC16
1
1
14
2015b: Move or Die (w/
@alexdamour
,
@dcervone0
,
@kirkgoldsberry
). We show that predictable offenses generate worse opportunities than entropic offenses do, as measured by their TPM (see EPV).
#SSAC15
1
1
15
2017a: Possession Sketches (w/
@_amiller_
). We can combine functional clustering and topic models to extract semantic meaning from player tracking data, isolating possessions with similar structure.
#SSAC17
1
0
13
2016b: Court Realty (w/
@dcervone0
,
@kirkgoldsberry
). The basketball court can be compared to a real estate market, with values of space derived from players' decisions when in those spaces.
#SSAC16
1
0
12
2019b: ACWR-Injury (w/
@OSPpatrick
,
@DarcyNorman
). Using data in soccer and football, we illustrate that training schedule confounds the relationship between ACWR and injury, calling into question much of the previous literature on the subject.
1
4
13
2019c: Winning Isn't Everything (w/
@SLiQ_Nick94
,
@yuorme
,
@Chris_LogiQ
, Mehrsan Javan). We look beyond just face-off win% in hockey to understand all the mechanisms by which face-off takers add value (clean wins, directionality, handedness, etc.).
1
3
12
I'm going to tack the papers from this year (Soccer EPV, ACWR-Injury, Winning Isn't Everything, Markov-Poisson, Fast Not Loose, Chuckers, Highlight/Lowlight) on to this thread. Again, credit belongs to my various colleagues. Lastly, adding
@Sportlogiq
to the list of supporters.
1
2
12
2017b: Scorekeeper Bias (w/
@MattvanBommel
). Not all assists/blocks in the NBA are made equal, and we can measure it. Using tracking data, we remove contextual features to get scorekeeper generosity and bias.
#SSAC17
1
1
8
2019f: Chuckers (w/
@nsandholtz
,
@jwmortensen
). We introduce the idea of spatial allocative efficiency in basketball lineups using spatial hierarchical rank models, allowing us to understand how lineups lose efficiency through who shoots where.
1
3
9
2019g: Highlight/Lowlight (w/
@Ev_Keane
,
@phildesaulniers
, Mehrsan Javan). We automatically construct highlight and lowlight reels through explainable temporal game models, incorporating levers for producers to fine-tune for multiple purposes.
1
3
10
2015a: Counterpoints (w/
@afranks53
,
@_amiller_
,
@kirkgoldsberry
). We can identify defensive matchups using tracking data, which can then be used to measure a player's ability to prevent shots (and make them worse) in different areas of court.
#SSAC15
3
1
7
2019d: Markov-Poisson (w/
@jwmortensen
). We develop a flexible nonparametric model for player movement based on conditional Poisson processes, allowing us to capture complex dynamics while incorporating player and team hierarchies.
1
1
8
0
0
6
This next one is ugly, but I kinda like it. I spared you by removing all the triple bogies... 3/5
3
0
5
Here's my last 20 rounds (sadly, stretching over 6+ months...). 2/5
1
0
3
A simple one looking at strokes gained against distance, split by shot type. Pretty remarkable how costly penalties and shots into recovery situations are. Turns out the only thing I'm half-competent at is recovery punch-outs. 4/5
1
0
2
@StatsInTheWild
Yes, we expect/hope to accommodate more students than we have scholarships (so those who don't need funding should still apply, noting on the app that they don't require $), but it's possible we get more applications than we have space. Will know more after July 31 deadline.
0
0
1
And lastly, the same but focused on driver. Time for a new driver... 5/5
3
0
2
@ErikSelim
V1 has lots of viz within the app, but these ones are all created in R (mostly via ggplot)
1
0
1
@SeidlOnSports
It will not. Our plan is to host a live stream for the scientific sessions, but the keynote Q&A and panel (TBD) will just be for the in-person audience.
1
0
1
@forrestwastaken
You can find 2016 here () and 2018 here (). 2020 was cancelled due to the pandemic.
0
0
1
@ConnorJungle
@ArccosGolf
V1 Game from
@v1sports
-- I've tried a ton of options but this has been the best so far.
0
0
1