Basketball Analytics, aka, different ways to tell us information that we already sort of know. From PER, to WS48, to plus/minus, to real plus/minus, to shot charts, to shot charts with smaller dots, to David Berri, to John Hollinger, to whoever the latest dude is who figured out you could enter basketball data in R and get it to spit out stuff. It’s all been quite fun and worthy of a few clicks here and there, but when are you guys actually going to tell us something useful? I mean, actual useful information that can be actioned and not just casually and sparingly glanced over in a long-form article only to be forgotten a week later.

Here are a few basketball problems that need addressing, or at the very least, should be talked about more:

Risk of Injury
Tell me when a player is likely to be injured, and tell me early. Figure out what his bone density is, what force (mass times acceleration) his joints can take, and then relate that information to his on-court movements to project how long he’ll last. For example, you can tell that the way Derrick Rose players he’s prone to injury – quantify that for every player, preferably before he’s drafted.

Effort Measure
Come up with a mathematical way of measuring effort without falling back to vanilla statistics like distance ran and speed. Figure out what the player’s “capacity” is using tests, and then determine how much of that capacity he’s using during different parts of the game. Pro Tip: If Andrea Bargnani isn’t dead last in this category, your work is wrong.

Grading the Coach
We measure players a lot, we don’t measure coaches enough. Much like how baseball is a game of probabilities, so is basketball, the math just happens to be a little more complex then pitching a lefty against a lefty. At the very least, there needs to be a binary indicator of whether the coach made the statistically correct move in late-game situations. For example, given the inbounder, the four other offensive guys on the court, the side at which the ball was being inbounded, the time on the clock, the score in the game, the venue, is it a statistically correct decision to not pressure the inbounder and sag?

Quantifying High Basketball IQ
We got an IQ test, how come we don’t have a basketball IQ test? All the data is there and it’s a question of sifting through it. Any decision on the court can be – based on all the other happenings on the court – retrospectively graded as right or wrong. A player with a highest basketball IQ might make the highest percentage of good decisions, as opposed to bad ones. I want to see if Landry Fields is what he’s cranked up to be. Or Matt Bonner for that matter.

Floor Stretchability
This idea that having John Salmons camp out by the fire in the corner means he’s stretching the floor needs to be validated. How much does a person’s reputation and past/in-game performance influence the defense’s opinion of him? How many threes does James Johnson have to hit before he causes a shift in the defense? By what factor does a particular player stretch the floor and how does his in-game play change that factor? I would think this is a straightforward one to figure out.

What is the Perfect Free Throw?
Based on a person’s height, what’s the ideal release point, force, angle, trajectory? Figure it out for every player and give them their particulars – watch them shoot 90%+. Seriously, I’m shocked that there are professional basketball players that shoot 50% from the free-throw line, that’s got to change.

Strategic Technical Fouls
Profile a player to determine what series of events cause them to pick up a technical. Then try to replicate that sequence in hopes of throwing him off his game. This can be extended to essentially influencing a player’s mental state negatively through pure basketball play. And poking them in the right places when nobody’s watching.

Figuring out Full Court Pressure
Teams get burned by full court pressure all the time. In the hopes of forcing a turnover, they end up creating a hole in the defense which ironically puts the offense at an advantage. Given a set of offensive players, their approximately positions, and the inbounder and ball-handler, there has to be an optimal defensive configuration that is most likely to cause a turnover while minimizing risk of defensive breakdown. Figure that out. As a bonus, come up with suggested defensive configurations which increase the chance of causing a turnover, while understandably increasing risk of defensive breakdown.

Optimizing Minutes Played
What is the ideal rest pattern for a player so that they can last an 82-game season and the playoffs? If they’re extended in one game for 44 minutes, how should that influence their playing time in the next game so as to not cause harm over the course of, say a roadtrip. Obviously, a player’s body needs to be analyzed, then correlated with the type of movements on the court, the pressure exerted, the effort expended, and likely a host of other factors. Put them all together, shake the box, and out you get a number telling you how many minutes DeMar DeRozan should be playing on the second night of a back-to-back having played 42 minutes the night before.

Psychological State
Players have money, money gets you shiny things, it also brings distractions and problems, which can influence a player’s on-court state of mind. Remember Keon Clark? Yeah, money did him no good. There’s got to be a way to measure a player’s psychological state before a game, and then see what, if any, impact that has on his game. Once you got that, adjust lifestyle to tune psychological state, thus improving his game. I’m talking girl problems, alimony, size of posse, everything. That all affects a player’s psychological state.

So there you have it – some questions that I always ask myself, knowing that very well that I can’t answer them, so maybe #analytics can?

  • jjdynomite

    Other than “money did him on good” instead of “no good” with reference to Keon Clark (on a related note, one can also measure proximity to weed either legally — Denver Nuggets — or otherwise) this was a kickass article.

    Three in a row, Zarar — the Zach Lowe of RaptorsRepublic.

  • puffer

    Important questions (well…interesting ones at least) that need to be answered. And some of them seem like it wouldn’t be a huge stretch to answer them. Grading the coach, BBall IQ and floor stretchability seem like they could be done by anyone with a basic knowledge of statistics and a willingness to put in a modicum of effort (not me).

  • KuH

    The problem isn’t statistics. Statistics has already figured (some!) of these things out. The problem is getting players and coaches to adapt the insights of statistics.

    What is the perfect free throw? We know this. For people who shoot less than 80% already, the simple answer is throw it underhand. Now get an NBA player to adopt this!

    Floor stretchability. We know the answer here too. It’s called 3 point shooting percentage. Coaches will drill the players to play a certain defensive scheme, cheating or not away from the open player in the corner. If the player in the corner gets passes and hits the shot, the defense may change, or it may not. Want it to change? Hit more of those shots.

    Effort. This is a simple one. It’s called percent of available rebounds, modified by your height. It’s the ultimate trench warfare, where only the tough survive. And yes, Bargnani comes pretty close to dead last in this one.

    Grading the coach. This is a real damn tough one. You can grade the coach on whether or not they play the best players on their bench, and how many minutes they play them (I’m feeling like a little smoked salmon right now). But once you go beyond putting the best players on the floor, what else is in your control? Did the players not rotate on defense because they’re a bunch of putzes, or because you didn’t train them right? Did the team degenerate to iso-ball because you didn’t teach them how to pass, or because they freeze when they get the ball and can only think of shooting? How much do you grade the coach for what the players do in game? Is it all on the coach? All on the players? Figure out how to slice that pie and you can grade coaches …

    Oh, and as for things like risk of injury, optimal minutes played per player, and psychological state, all sounds great. You need to get the data on the players to do the analytics on. That would be fun to do, but the team will never give it to you … if they even measure it themselves. But if you measure it, it can be analyzed!

    • webfeat

      >>RE: Grading the coach

      Add up the expected win share ratings of the players on a team, and compare that to actual wins. The delta is the coach’s rating.

  • LuckyMystery

    Although interesting to think about most of these ideas are pretty dumb. I’m not saying that analytics shouldn’t be used for circumstances that can actually be quantified. However to try and suggest that we use analytics to find out whether a player will get injured, or to determine players psychological states before every game is just silly. Obviously teams wouldn’t draft players if they knew they were Greg Oden, Sometimes players in what appears to be a fragile mental state play their best ball. Ron Artest and Rodman come to mind.

    Every team would love to find the perfect basketball player, but I can’t think of any analytics that would determine that Lebron is the best player on the planet, he just is. You see it every game. You don’t need someone to tell you if a player works hard, yiou can watch with your own eyes. No one is going to be fooled into thinking Bargnani is a hard worker just because some calculation suggests he works a little harder then we think.

    Basketball and for that matter hockey and football don’t lend themselves to stats the way baseball does. Baseball is beautiful because the stats don’t lie. What makes sports great is that players can get injured, players can have talent and not live up to them, players can be better or worse depending on the system they play in. If we could just pre determine all of this their would be no intrigue to the draft, or really the games themselves.

    • KuH

      Yeah, the whole article read like it was pulling the leg of the stats guys. I just decided to take it seriously, for the fun of it. Looks like I wasn’t the only one …

  • ckh26

    You know when the dog days of summer are officially in full swing when the articles on analytics and advanced statistics begin to surface and start to be the staple of conversation amongst the faithful. A valiant effort big Z and credit to you for finding something to publish. We should insert a clock similar to the “days to the draft clock” to count down the start of training camp.

  • Trini

    Welcome to mid-July, Raptor fans…

  • Chatham_Cathy

    “poking them in the right places when nobody’s watching.” Laughed out loud at that line Arse. I’m with you. Those are the types of factors I am pondering game by game. Case in point, when the core guys all started having kids I’m thinking “well, shit, there goes another season. ” Because as anyone with kids can attest, you ain’t getting no sleep for the next 3 years. Hence, poor performance from players with young kids. I could go on about that particular analytic bit I gotta go ten to a screaming kid…

    • http://skritch.blogspot.com/ Nathan Ng

      Interesting points re: kids. Consider however an alternate viewpoint: having a family can settle down and mature a player. They’re no longer footloose and fancy free, clubbing every night, chasing … This was alluded to in the media as *contributing* to the Raptors team chemistry (e.g. http://www.citynews.ca/2014/04/30/band-of-daddies-behind-raptors-chemistry/) rather than disrupting it.

      Also, at that income level, pretty sure you can have a full time nanny around to handle the kid when you don’t feel like it.

      Though — as you pointed out, I would still want to see what the research and stats say.

    • JayTeam

      Or have players go the Shawn Kemp route. None of his kids kept him up at night.

  • Jordan

    I’d like to think I’m reasonably well-versed in baseball analytics, but I am only a casual NBA fan. My question is: what kinds of information are available is these databases?
    In baseball, nearly every event is recorded. Since it is a series of discrete events, it is very easy (relatively speaking) to determine the value of individual components of the game. Considering the power of physical PitchFX data or the new proprietary data available from MLBAM, it is easy to wonder how basketball analytics could ever compare to baseball, in terms of accurately assigning value of events.
    How could we ever quantify, for example, the degree to which a defender influences an offensive player’s shot attempt?
    Food for thought, though. I am not asking rhetorically either. I’m hoping someone who is better informed about basketball to answer me.

    • DDayLewis

      Here’s a bit on defensive impact from NBA.com’s SportVU information. It’s kinda the basketball equivalent of PitchFX

      http://stats.nba.com/playerTrackingDefense.html?pageNo=1&rowsPerPage=25

    • LuckyMystery

      I agree as I stated above.

      Some people may find baseball boring and sure at times it can be. But the stats available make baseball more nuanced than just pitch, hit, throw and run. The stats in Baseball allow you to more accurately compare players from different generations. It’s easier to compare a player like Barry Bond’s to Hank Aaron, than it is to say compare MJ to Oscar Robertson.

      I mean let’s be serious by the numbers Oscar Robertson has to be the best player of all time, but he’s hardly even in the conversation. I just don’t see a way you can accurately use analytics, because there are just way to many variables. Take for example Rudy Gay’s numbers in Toronto and then in Sacramento. He was taking 3 less shots a game in Sac, and hitting more shots per game, yet his point total only went up about half a point a game. His rebounds went down but his assists went up, his 2 point shooting % went way up but his 3 point % which was already not good in Toronto went even further down.

      • jjdynomite

        Your points are not accurate as you are offering generational comparisons while Zarar is looking to compare modern day athletes to each other, which is starting to happen, “variables” notwithstanding.

        Many educated commentators hail Oscar Robertson as one of the greatest players of all-time; Chris Webber did so in the recent NBA special on the Big O, and Bill Simmons ranked him #9 of all-time in his 2010 Book of Basketball. The only reason he’s less renowned is that he played in the 60s and early 70s — when black players were held in less esteem due to racism — not in the 80s and 90s like Jordan.

        Moreover, one can only realistically compare Barry Bonds to Hank Aaron if you know the growth hormone compounds at the Bay Area Laboratory Co-operative. Aaron had to deal with terrible racism and the burden it must have caused to his mental well-being as a player and human over his long career. And no steroids to mitigate that fact.

        /multiple comparison fail

        • LuckyMystery

          You truly are a moron. First I am not commenting on Zarar’s article but to a different commenter’s queries.

          #9 all time is not the best of all time. Hence not in the conversation. So thank you for proving my point. I clearly said that you can’t compare Oscar to MJ by the numbers. So again thanks for proving my point.

          Thanks for mentioning the steroid era and the racism during Aaron’s playing days, when I was quite obviously just picking 2 players that were at the top of the sport in different generations. Would it have made you happier if I chose a white player and let’s say Ken Griffey Jr. As I stated earlier you can more accurately compare players in baseball to different generations because of the large amount of stats available.

          Thanks for coming out though

          • jjdynomite

            “You can more accurately compare players in baseball to different generations because of the large amount of stats available.”

            Except for…
            – Deadball era
            – Juiced ball era
            – Steroid era
            – Lowering the pitchers mound
            – Relief pitchers and 5-man rotation
            – The DH impacting pitching stats in the AL

            … and that’s just off the top of my head. But other than that, you’re perfectly right, you incredibly retarded moronic imbecile.

            /fuck off noob

            • LuckyMystery

              I truly don’t know why I’m wasting my time with you. you have no grip on the English language.

              I have been comparing the amount of stats that are readily available in Baseball to basketball this entire time. Please learn how to read, and then take the time to let the info sink in properly, before you make moronic comments, trying to prove yourself.

          • FLUXLAND

            I find Simmons’s writing entertaining, but surely his ranking is not the decisive authority on Robertson’s (aka Player of the Century) place in the GOAT conversation? Also, it’s a conversation that rarely involves PGs; but his stats, alone, are what easily places him right in the middle of it, no?

            • LuckyMystery

              That’s the point I was trying to make. His stats alone should at least put him in top 3 consideration. But he is hardly ever mentioned as a top 5. I am just saying you would think a guy that averaged a triple double would be in that conversation. Seeing as no one else has ever done it. My point wasn’t to say Oscar wasn’t great, but that stats alone need to be taken with a grain of salt if you will. Where as in baseball if you are looking to find out what player is the best at this certain aspect of the game, there will almost certainly be a calculations to get that answer.

  • Pingback: Neon Bets | 10 Questions Basketball Analytics Needs to Answer()

  • Louvens Remy

    I’m just waiting to hear the news that Zahar got shot by Daryl Morey’s executive assistant.

  • Pingback: 10 Questions Basketball Analytics Needs to Answer | Baseball News()