Daryl's Notebook: What I learned from the 2025 MIT Sloan Conference
It's the sports analytics conference season! And when you're living Down Under, you have to resort to watching talks through YouTube instead of attending them live.
This will be a different Notebook edition from the other ones that I have done in the past. Because it is the sports analytics conference season! All of the best data practitioners from different sports have gathered in the US and across Europe to deliver and share their own stories and experiences from working within the industry with the main goal of inspire the next generation of data practitioners and exchange information from different aspects and organisations within the industry.
But…while all of those talks happened/are happening live up north of the world, living in Australia means that I have to settle with watching those talks via YouTube instead of attending them in person. And fortunately for me, more talks have been recorded and posted onto YouTube, which means I can still listen to those talks from within the comfort of my own home. These next couple of Notebooks will also be the place where I write down my thoughts and recaps of those talks, which will include bits of insights and my own opinions about the topic. I won’t cover all of the talks, however, since there are too many and I can only pick the ones that are interesting to me. So, if you are interested in listening to the talks/panels that I did not cover, they are available on YouTube!
The first one on the table for this week’s Notebook will be the MIT Sloan Sports Analytics conference, which took place at the beginning of March of this year (2025, for those reading in the future) in Boston, Massachusetts. All presentations and panels can be found via 42 Analytics’ YouTube channel here!
The Analytics Driven Franchise: Building a Modern NBA Team
Panel by:
Dean Oliver (former Denver Nuggets & Sacramento Kings GM of Analytics, currently a data scientist at ESPN),
Monte McNair (GM of Basketball Operations for the Sacramento Kings),
Shane Battier (former Miami Heat forward & current NBA consultant),
Ariana Andonian (VP of Player Personnel for the Philadelphia 76ers) &
John Hollinger (former VP of Basketball Operations for the Memphis Grizzlies & current NBA columnist for The Athletic/The New York Times)
Why did I pick this panel?
This panel feels quite relevant to me as I am currently (unofficially) working with [redacted] and we are in the process of establishing a data culture within an organisation. It is a different role compared to what I have done before, and while I am still responsible for the technical (coding) stuff related to data, I am also basically acting as a consultant where I give advice on building a tech stack, integrating tech into the workflow, and those sort of stuff. Can’t say much right now because the role is still unofficial, but I am hoping that this talk will give me a glimpse into how top people at NBA teams, a sport that has been taken over by data analytics in the past 10-15 years or so, build an analytics-driven team and what can I learn from how they establish things.
My notes!
MMN (Monte): “Every time we think we catch up, we get new data.”
SB (Shane): “We all have a dataset that is roughly similar. Talk about how you think differently about the data, that maybe is more important than the actual data itself, and how you think differently than the competition.” - Such a good question from Shane here.
MMN: “You can have the data advantage, which is pretty much going away, (because) the (NBA) has standardised a lot of that. You can’t really have more or different data, at least for the bigger stuff. Then there’s how well you’re gonna analyse it, I think there’s certainly some advantage there. Then the throughput. If you have great data, you analyse it great, (but) you can’t make any decisions that’s put on the floor, that’s an issue.”
AA: “A lot of time we do have the similar data, but you can use it to make different arguments. […] (Monty and I) are using the data, but we’re using it and manipulating it to make our arguments stronger, and that gets the interesting debate as well.”
DO: “There’s just more steps to get from numbers to applications on the coaching side, or on the medical side. […] It’s more indirect, but you can have people in your organisation that are particularly good at (communicating data to the coaches and medical staff), and that becomes an advantage.”
Regarding predictive models (draft models, All-Star percentages, etc):
DO: “There’s still all of this other stuff in terms of player development that you need to model and use to build better players, that then fit that 68% (chance of being an All-Star). Yes, 68% they’re gonna be (an All-Star), but if you have the wrong player development, they’re gonna be on the 32%.”
AA: “It’s interesting that different organisations use the models as a piece of the decision-making, as the final of the decision-making. How, and at what stage do we pull that information in, and at what stage do we start talking about it and match it with the eye test, I think that part has gotten better. […] The level of accuracy isn’t hard to say, but how we’re using the information has probably improved in that time.”
JH (John): “What are the challenges on using analytics to actually help keep the players healthy?”
DO: “One of the things that I have seen is, there is evidence that players actually don’t learn a lot during the season, they learn more between the seasons when they have the mental downtime to actually absorb [data].”
MMN: “There is a big nap culture in the NBA and it is a great way for (the players) to rejuvenate their bodies with these unique schedule, that you almost always have the time to nap and get your body recover.”
JH: “What are the remaining roadblocks of getting to the final barrier of implementation [of analytics] within an organisation?”
DO: “You want everybody in an organisation to have some [data] literacy. There are going to be people who know much better and who know it less. But, having an overall culture throughout a team really really helps so that you don’t have barrier of communication.”
SB: “There are a ton of capable people in (the conference room) who can do the work, they can cut the data, they can glean insights. If you are unable to communicate that to a coach, to a GM, or to a scout, you are wasting your time. So if you are in school now, looking to break in, mix in a communication class. That will differentiate you in the job market.”
AA: (From a scouting side) “You have to spend the time off to learn as much as you can, you are not going to create the data in the same way as an analyst is. But, it is some level on us who are not from a quantitative background, (spend time learning) to at least feel comfortable understanding what is being output from the data, and what the different terminology is. It is both sides meeting in the middle, and it is not just fully on the analysts.”
Regarding how to communicate data and insights to players:
DO (Dean): “There’s a danger of information overload, for the players, no doubt about that. […] You want the players to go in and react on instinct, very well-trained instincts. And with less practice time, you have to make sure that, you build those instincts in practice, if you can.”
DO: “During the early part of the season, give (the players) some information, get a sense of who can absorb everything, versus the other guys who can only take so much at a time.”
AA (Ariana): “Different players have different calibrations for how much they want to discuss it in a formal way, and in an informal way. […] Maybe you’re using slightly different terminology, but (the players) know what you’re talking about, they’re capable of understanding what you’re talking about because they’ve seen (data) on ESPN.”
MMN: “We have to think about what to say to players and how that is going to affect them. You are not going to sit down with a player and talk to them because you may think that you are helping, but now the player is doubting themselves. That has to be a conducive effort to figure out not just what to say, but when and how.”
AA: (continuing Monte’s point above) “And what the purpose of it, right? What are you trying to get out of sharing this (information)? What is the purpose and the next steps of you sharing the information?”
AA: (regarding what comes next in basketball analytics) “Mental potential and training. There are different tools out there right now that helping you be as ready as you are late in the game and at the start of the game, and that is related to the sport science side.”
JH: “Analytics vs the Eye Test. How much of our evaluation today are relied on analytics and how much are still from the eye test?”
MMN: “Luke Bornn mentioned this at an earlier soccer panel is ‘the one thing that the new data has given us is you are speaking the same language much much more with your coaches and scouts’. You can (show the data) and then pull the videos immediately. I think that part has been a huge benefit, both the data that you are giving and to show the videos and speak their language, especially with the coaches.”
AA: “At the basic level, if you are not finding a way to be in the middle and use (analytics and eye test) together, it is not really seen in the NBA anymore. You are having a conversation, and maybe you are trying to isolate the eye test and the analytics, so that you can see what is being missed and what are the holes in between them, and how can you find something within that disconnect.”
Decisions on Ice: The Next Frontier of Hockey Analytics
Panel by:
Arda Öcal (ESPN SportsCenter host),
Jeremy Rogalski (Director of Hockey Analytics of the Boston Bruins),
Meghan Chayka (Co-founder and CEO of Stathletics) &
Philippe Desaulniers (Head of Hockey Analytics Technology, Montreal Canadiens)
Why did I pick this panel?
Having followed Meghan Chayka for a while now, I have learned that hockey also went through an analytics boom not so long ago, and Meghan was at the forefront of that movement. As she described (and I am paraphrasing her words from memory), hockey went from a traditional game to now having data and technology being implemented into almost every aspects of the game. This sounds…very similar to what football is going through right now, which means it is a good opportunity to gain experiences from the people who have been through a data revelation and bring them back to football.
It is also a good opportunity to take a step back from football and listen to what other sports have to offer with their analytics. Focusing on only the sport that you are working on can limit one’s creativity and innovation as there can only be so much that one can do, and drawing inspirations from other sports is the best way to expand the horizon. There have been many football analytics innovations that drew inspirations from other sports, like the xG timeline or the match momentum graph that we now familiar with. And there are always new things to learn, so I am always into that!
My notes!
AO (Arda): “What are some of the things that have changed the most in (hockey analytics)?”
PD (Philippe): “When we first started, the dataset that was available for hockey was minimal. […] Since then, the introduction of tracking data in recent years is taking another step in revolution.”
MC (Meghan): “It’s not one company or client. It’s people that are willing to take risks and put money into tech and into different spaces.”
JR (Jeremy): “One, just the way to describe the game. Seeing how pre-scout for a game evolves from coaches watching 3 games on video and look for some common trends in that and maybe do a deep dive on the season, but now you can describe how a team’s power play formation in so much details. And the second thing is, if a team’s internal platform breaks on the trade deadline, you’re in a lot of trouble. There’s a lot of information in there that you are not going to be able to get out. So, getting teams investing in technology to bring that information to them has been a massive change in the past 5 years.”
AO: “How has (communication) evolve in recent times at all levels of hockey?”
JR: “In terms of building a relationship with a coach, understand what their schedule looks like, what their needs are, whether they have windows of time to work through and educate on the data, and also package (the data) in a digestible way. You’re still, in some way, also do it with the coach. If there’s a point you’re trying to make, there’s the time for facts. But sometimes you have to package it in a narrative that helps them understand how they can use that fact to improve the team and give it to the player.”
PD: “I really see three things mainly that are important in order to have communication between our group and the coaches:
First thing is, it’s really great to have a translator. You’ll want to have someone on the team who’s able to speak the language with the coaches and who understand on the analytics side.
The second thing that we’ve found effective is, we’ve started creating some concepts that we try to keep as simple and as high level as possible, but that then can become part of the vocabulary that we have with the coaching staff.
The third thing that we’ve done is get the coaches on board. Work on a way of organising the metrics based on what they’re looking for, what their understanding of the game is, and organise the information that way, so when (the coaches) start digging in, they can actually see what they’re looking for, and you’ll get the buy-in from them too because they worked on it to build that structure.”
AO, expanding and generalising MC’s question: “From different experience level of coaches in the NHL or otherwise, how have you found their acceptance and how they use (analytics) in their day-to-day job?”
MC: “It can be (coaches’) background too. People come to coaching in so many different ways, so they have such a different way of understanding and ease with (analytics), and then they can just implement it, and know what’s assumptions, what’s noise, and what’s actually important to them and to the job that they need to do. So it’s just being realistic with your personnel.”
MC: “I will say that I have noticed, […] young people now expect data, no matter whether it’s a coach or a player coming up. It’s not ‘should I use analytics?’, it’s like ‘why aren’t we using analytics? why isn’t my agent doing that for me? why isn’t my team doing that for me?’. That both empowers the people working in tech, but also empowers the person who that data is being used to dictate their career. Giving them some control and oversight and insight into what they actually have to improve.”
AO: “What are some of the things that you are finding that (either data or video analytics) is doing better than the other?”
JR: “As I watch game and I’m seeing things that are happening in a play, and I’m so eager to get to the point where (those actions) are quantifiable (PD: and how can we measure that)”.
PD: “You will see when watching the videos or the live game, you will definitely see details that are not captured with the data at this point. And seeing (the data) is definitely important in understanding how the players are playing and how humans are behaving, and it’s something that we want to get to at some point. We’ll probably never get to the end of it, so it’s always a compliment.”
AO: “What are some future applications and trends that will be very important in (hockey analytics)?”
JR: “VR is definitely interesting and has been talked about a lot. Goaltenders are currently using it for their preparation.”
PD: “The tracking data revolution that is happening right now, and we are now bearing the fruits of that, and we are making good ways into using it. Having that tracking data at other levels is something that is coming out and will be very helpful. And the next step is going to be around pose, where you are going to be able to locate players as they were on the ice, where their stick was, is probably the next step for getting the model to do better, getting more precision on everything that we measure.”
AO, with a question from the audience: “10 years from now, how different do you think an NFL box score (scoreboard) will look like in terms of what statistics are listed?”
PD: “One of the things that I’ve noticed in the last 2-3 years now is how much more, at least on the media side, is how much people are talking about xG, and it’s not something that has not made it yet to the standard box score. And I think it’s a really key metric that is super useful to understand what patterns of the game are like, was it a defensive disaster, was it the goalie was a star. Breaking this down is fun and interesting to do with a metric like the xG, it has started to catch on and I think it’s going to continue and will be in one of the box scores.”
AO, with another question from the audience: “How has analytics enable scouting international players where data is sometimes non-existent?”
JR: “[…] It has helped a lot. We’re in constant communication with our scouting group, and video is good too. Because if you are going to make a recommendation to send a guy to France to go watch a player, you would want to have a bit of certainty behind that before you book that ticket.”
MC: “It’s the accuracy constraints by the quality of the video and what the feed looks like. Make (the games) easy to track and make the numbers visible, please!”
AO, with the last question from the audience: “What do you think the next big challenge will be and the requisite skills for the upcoming students?”
PD: “[…] Anything that has to do with spatial-temporal analysis is definitely something that we are looking into, and just in general, understanding software systems.”
MC: “It’s such a buzz words right now, but AI, automation, computer vision, anything that helps work with these extremely large datasets. If you can be in any of those technical realms and get a chance to take on some projects or internships in those areas, there is an explosion of jobs, and very well-paid, but you have to be very well-skilled to do them.”
JR: “[…] Just having the curiosity and a willingness to learn. There’s not one universal coding language or anything that you need to know. […] You just have to be willing to grow and learn and adapt to the needs.”
Recap & reflect
Even though it is just two panels that I got to watch, there are still so many things to unpack.
One of the things that I came away with, and this is mostly on the general level of things, was that data analytics in sport still has a lot of rooms to grow, especially with more investments being made and more technologies are being developed and implemented. But it is also important to balance it out with the traditional eye test since data can be used as an “underline” for any decision-making and finding certainties.
Another thing that both panels seem to be emphasise quite heavily is the importance of translating insights and numbers into something that coaches can use and implement to improve the team. I think Dean Oliver said it well, which is “if you have all of the technical skills to work with data, but you cannot communicate that with the coaches or medical staff, then you are wasting your time”. One of the ways that analytics teams are solving that problem is to have the coaches involved at some stage, either through a translator who can translate terminologies back and forth, or have the coaches directly work and develop the analytics products.
It is also a similar process for when communicating data with the players, and I wish I had learned this earlier on when I started. As a data analyst, I think you are also a data storyteller. Every player that you work with is a different audience that you need to cater your story (visualisations and numbers) towards and they all receive it at a different pace. It’s basically like school, except you are the teacher and the players are your students, even though you might be way younger than some of the players are. The end goal is still helping players to be the better version of themselves, but some players might get there quicker and are more receptive to data, and some might take more time and needs the time and space to process new info.
Lastly, the key to having a successful data analytics team is to establish a data culture within an organisation. It can be through data literacy, where there are different levels of understanding data and analytics, or being able to communicate data well within that environment. I think this builds off the point about communicating with coaches, because at the end of the day, the end product is what happens on the pitch/court/ice ring/whatever surface the sport is played on. If data cannot influence or assist the decisions that helps the coaches, the players, and the staff to get there, then it does not provide much value to a team.