<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Badr Albreiki on Medium]]></title>
        <description><![CDATA[Stories by Badr Albreiki on Medium]]></description>
        <link>https://medium.com/@badr.albreiki12?source=rss-107ed08aadd7------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*x-RZBPK7PlUUNPqN</url>
            <title>Stories by Badr Albreiki on Medium</title>
            <link>https://medium.com/@badr.albreiki12?source=rss-107ed08aadd7------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 07:54:48 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@badr.albreiki12/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Can AI Spot Changes in a Basketball Player Before They Become a Bigger Problem?]]></title>
            <link>https://medium.com/@badr.albreiki12/can-ai-spot-changes-in-a-basketball-player-before-they-become-a-bigger-problem-40256f26800a?source=rss-107ed08aadd7------2</link>
            <guid isPermaLink="false">https://medium.com/p/40256f26800a</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[business]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[health]]></category>
            <dc:creator><![CDATA[Badr Albreiki]]></dc:creator>
            <pubDate>Sat, 24 Oct 2026 08:04:48 GMT</pubDate>
            <atom:updated>2026-10-24T08:04:48.691Z</atom:updated>
            <content:encoded><![CDATA[<p>By Badr | TKS Learn Article V2</p><p>The problem: every player is different</p><p>Basketball is a sport where players constantly jump, land, sprint, stop suddenly and change direction. These movements are part of the game, but training and playing also put stress on the body. If a player is not recovering well or their movement starts changing, it is worth paying attention to.</p><p>This made me think about a question: could AI help identify these changes before they turn into a bigger problem?</p><p>At first, comparing players seems like a good way to understand performance. But a player who naturally jumps lower than another player is not automatically less healthy. Every athlete has different abilities, habits and recovery needs.</p><p>My main research question is: Can AI detect meaningful changes in a basketball player’s own movement, workload and recovery patterns before those changes become a bigger problem?</p><p>Instead of trying to predict the exact moment someone will get injured, I want to investigate whether technology can recognise when something has changed from what is normal for that specific player.</p><ol><li>What signals should AI actually track?</li></ol><p>AI needs information to find patterns. However, collecting more data does not automatically make a system better. The important question is which signals provide useful information and how reliable they are.</p><p>Training workload</p><p>Workload is the amount of training and playing an athlete does. In basketball, this could include training duration, playing time, number of jumps, accelerations and sudden changes of direction.</p><p>For example, imagine a player usually has moderate training sessions but suddenly has several very intense sessions close together. That change might be worth reviewing, especially if the player is also reporting tiredness or soreness.</p><p>Catapult, a sports technology company, already provides basketball tracking systems that measure movement and workload. Its technology can categorise movements such as running, jumping and changing direction.</p><p>However, a high workload does not automatically mean an injury is coming. Players can handle different amounts of training, and workload needs to be understood alongside recovery, training history and other information.</p><p>Sleep and recovery</p><p>Sleep is another possible signal because athletes need recovery between sessions. A player might record how long they slept and rate how rested they feel.</p><p>But the evidence is not as simple as saying that less sleep means an injury will happen. A 2021 systematic review of sleep and sports injuries examined 12 prospective studies and found limited evidence linking poor sleep to injuries in adult athletic populations. The researchers said more studies were needed.</p><p>This is important because it shows why an AI system should not treat one short night of sleep as proof that a player is at high risk. Sleep data may help explain how someone feels, but it should be considered with other signals.</p><p>Jumping and landing mechanics</p><p>Basketball players repeatedly jump for rebounds, blocks and shots. How a player lands can be studied using movement analysis, which measures how body parts move during an action.</p><p>Some systems use cameras and AI to estimate body positions without attaching markers to the athlete. This is called markerless motion capture.</p><p>A systematic review published in 2025 examined 53 studies of markerless motion capture during activities such as jumping, landing, running and cutting. It found that reliability and accuracy varied depending on the movement, body joint and technology used.</p><p>This means movement tracking has potential, but it is not perfect. If a camera system makes a measurement error, AI might mistake that error for a real change in the athlete.</p><p>Which signals matter most?</p><p>I would start with three main categories:</p><ul><li>Workload: How much and how intensely is the player training?</li><li>* Recovery: How well rested and energetic does the player feel?</li><li>* Movement: Are jumping and landing patterns changing over time?</li></ul><p>Heart rate could also be useful, but adding more measurements would only make sense if they provide reliable information. For a first version, I think a smaller set of useful signals is better than collecting everything possible.</p><p>2. Why a personal baseline matters</p><p>A personal baseline is a picture of what is normal for an individual athlete. It could include their usual training workload, sleep patterns, energy ratings and movement measurements.</p><p>Imagine two basketball players. One normally completes high-intensity training sessions and feels fine afterwards. Another normally trains less and needs more recovery. Giving both players the same warning because they trained for the same number of minutes would ignore their differences.</p><p>A system based on personal baselines would compare each player with their own previous data.</p><p>However, a baseline cannot be created perfectly from one or two check-ins. A few days may not represent normal behaviour, especially when training changes between practice days, games and rest days.</p><p>I would collect data regularly over several weeks as an initial testing period. This is a starting design choice, not a scientifically proven minimum. The system would need to keep learning as the player’s training and fitness changed.</p><p>It would also need to account for the type of day. A player’s movement and energy after a match might naturally differ from their movement on a recovery day.</p><p>Visual 1 — How a personal-baseline system could work</p><p>Suggested diagram to create:</p><p>Player data collected over time → Compare with the player’s usual pattern → Identify a possible unusual change → Coach or qualified health professional reviews it.</p><p>Caption: A possible AI monitoring process. The system identifies changes for review rather than declaring that an injury is about to happen.</p><p>Source: This is my proposed workflow, informed by research on AI-based sports injury prediction.</p><p>3. How can AI tell a real change from a normal one?</p><p>This is one of the hardest parts of the problem.</p><p>A player might sleep less than usual one night, have a difficult training session or land differently on one jump. These changes might mean something, but they could also be normal variation.</p><p>A useful system should look for patterns rather than react to every unusual measurement. For example, it could check whether several signals have changed around the same time: the player’s workload has increased, their energy ratings have dropped, and their movement measurements have changed.</p><p>That combination might deserve attention. However, it still would not prove that an injury will happen.</p><p>The quality of the data matters too. If a camera angle changes or a wearable device records inaccurate information, the system could produce a false alarm. A false alarm happens when a system flags a problem that is not actually there.</p><p>There is also the opposite problem: a system might miss a genuine issue. This is why AI should support human judgement rather than replace it.</p><p>A 2025 review of machine-learning approaches to sports injury prediction reviewed 38 relevant studies. It found that results were affected by differences in datasets, injury definitions and research methods. Some models reported strong performance, but that did not always mean they would be useful in real-world medical decisions.</p><p>This is a major limitation. A model can appear accurate in a research study but still struggle when used with different players, teams or training conditions.</p><p>4. What technology already exists?</p><p>I found two examples that show how parts of this idea are already being developed.</p><p>Catapult: measuring basketball workload</p><p>Catapult’s basketball technology uses athlete-monitoring systems and movement data to help teams understand training demands and player performance.</p><p>This is useful for my research because it shows that tracking basketball-specific movement is already possible. A future system could use this kind of information as one part of a player’s personal baseline.</p><p>However, measuring workload is not the same as accurately predicting injuries. The data still needs to be interpreted in context.</p><p>Uplift Labs: analysing movement with cameras</p><p>Uplift Labs offers AI-powered movement analysis using cameras such as iPhones and iPads to capture movement and create 3D analysis.</p><p>This is interesting because it shows how movement analysis might become more accessible without needing a traditional laboratory setup.</p><p>For my idea, similar technology could potentially compare a player’s jumping or landing mechanics over time. But the research on markerless motion capture shows that accuracy varies between tasks and systems. Before using these measurements to make important health decisions, they would need to be tested carefully.</p><p>Visual 2 — Two parts of the possible solution</p><p>Suggested visual: A split image showing a basketball player wearing an athlete-monitoring sensor on one side and a player being recorded by cameras during a jump on the other.</p><p>Caption: Wearable sensors can help measure workload, while camera-based systems can analyse movement. Combining them could provide a wider picture, but only if the measurements are reliable.</p><p>Sources: Catapult basketball technology and Uplift Labs sports performance. Use images from these sources only if their reuse is permitted; otherwise, create your own diagram.</p><p>Why is the NBA interested?</p><p>The NBA Launchpad programme identifies player health and wellbeing as an area for new technology. Its listed priorities include using AI and computer vision to help identify and reduce injury risk, improve recovery and support health decisions.</p><p>This does not prove that a system like mine would work. It does show that player health technology is a real area of interest in basketball, rather than just a futuristic idea.</p><p>5. What is still missing?</p><p>After looking at the research and existing technologies, I think the biggest challenge is not simply collecting data. It is deciding what the data actually means.</p><p>A useful system would need to answer several questions:</p><ul><li>How much data is enough? A short record may not represent a player’s normal patterns, but there is no single number of days that works for every athlete.</li><li>* Which measurements are reliable? A system must understand its own measurement errors before treating a change as meaningful.</li><li>* How should signals be combined? A change in sleep alone might not mean much, but several unusual changes together could be worth reviewing.</li><li>* How can false alarms be reduced? The system needs to avoid worrying players and coaches about every small difference.</li><li>* Can it work for different players? A model should be tested with different athletes and real basketball conditions, not just the data it was trained on.</li></ul><p>My main takeaway is that an AI system should be judged on whether it provides useful, trustworthy information, not just whether it can produce a prediction.</p><p>6. My possible direction: CourtHealth</p><p>Based on this research, I want to explore a concept called CourtHealth.</p><p>The idea would be a basketball health dashboard that helps a player track their own patterns over time. A first version could record training minutes, sleep, energy and simple movement observations. It could display charts and highlight changes that may deserve a closer look.</p><p>For example, if a player’s energy ratings are lower than usual for several check-ins, the dashboard could flag the pattern for review. It would not say that the player is injured or will definitely get injured.</p><p>I would start with simple tracking and visual comparisons before adding more advanced AI. Later, movement data could potentially be included if I could access a reliable way to measure it.</p><p>This is only a proposed direction, not a validated injury-prevention product. To prove that it works, it would need testing, reliable data and input from qualified sports and medical professionals.</p><p>Visual 3 — A possible CourtHealth dashboard</p><p>Suggested mock-up: Three simple charts for training workload, sleep and energy, with a message such as “Your recent pattern has changed — review your check-ins.”</p><p>Caption: An early concept for showing personal trends. Any sample data should be labelled as fictional, and alerts should not be presented as medical diagnoses.</p><p>Source: Original concept by me, based on the research discussed in this article.</p><p>My hot take</p><p>I think the future of sports AI should focus less on claiming it can predict every injury and more on helping athletes understand their own bodies.</p><p>A system does not need to know exactly what will happen next to be useful. If it can reliably notice a change, explain what information led to that result and help a coach ask the right questions, that could be valuable.</p><p>But I would rather have a system that honestly says “this pattern looks different; someone should review it” than one that confidently gives a prediction it cannot prove.</p><p>Conclusion</p><p>My research suggests that AI could help basketball teams monitor changes in workload, recovery and movement. Existing technologies already measure some of these signals, but the evidence also shows important limitations in measurement accuracy and injury prediction.</p><p>The key is understanding what is normal for each athlete and identifying changes without confusing them with everyday variation. Personal baselines could be a useful starting point, but they need enough reliable data and regular review.</p><p>CourtHealth is my possible next step: a simple dashboard that explores this idea in practice. Before building advanced predictions, I want to focus on tracking patterns clearly and responsibly.</p><p>The question is not whether AI can predict every injury. It is whether AI can help people notice meaningful changes earlier and make better-informed decisions.</p><p>Sources / References</p><ol><li>Leckey, C. et al. (2025). Machine learning approaches to injury risk prediction in sport: a scoping review with evidence synthesis. British Journal of Sports Medicine. Read the study</li><li>2. The Association Between Poor Sleep and the Incidence of Sport and Physical Training-Related Injuries in Adult Athletic Populations: A Systematic Review (2021). Read the study</li><li>3. Reliability and validity of lower extremity and trunk kinematics measured with markerless motion capture during sports-related and functional tasks: A systematic review (2025). Read the study</li><li>4. Catapult. Basketball Technology. Official technology page</li><li>5. Uplift Labs. Sports Performance. Official technology page</li><li>6. NBA Launchpad. Priority Areas. Official programme page</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=40256f26800a" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>