Anybody read this book and have reactions?
-
gdavis8848
- Posts: 1
- Joined: Mon Aug 22, 2011 10:58 pm
Re: Anybody read this book and have reactions?
Crow,
I think this book is great. It gives you a behind-the-scenes look at how analytics are being used in the NBA today. It is a quick read. I highly recommend it.
Greg
I think this book is great. It gives you a behind-the-scenes look at how analytics are being used in the NBA today. It is a quick read. I highly recommend it.
Greg
Re: Anybody read this book and have reactions?
Thanks for the response. Did anything in particular surprise you? Reveal new techniques or level of detail not previously disclosed to the public? Anything notable said about struggle to get coaches and players (and GMs and owners) to use the information and analysis?
Did you get the sense that teams are fully utilizing what they have or generating data & analysis that more than not lays there understudied and acted upon? For every hour put into building and maintain databases and spitting out & beautifying outputs how many hours go into studying the data hard player by player, analytic piece by piece and beneath them and proposing and selling actions? Should it be 5-10? At least 2? Could it even be 0.5 or less? Does this book give any basis for answering this question? Any former or current insider willing to address this general question? Say a team has 2 data analysts who work 60 hours a week. That would be 6,000 hours a year. How much went to the production process vs. reflection / recommendation based on the output? I doubt it is 50% but tell me I am wrong or wrong to expect or want that ratio. I'd hope it was at least 25%, but I wonder. Could be even less?
How many hours did top managers spend with any of it? 2 hours a day for 5 guys for a year would be 3,500 hours. Or if you said that every hour of their work was informed by it, then you claim 15,000 hours of use. Are there 5 guys per team who use advanced data analytics "a lot", "all the time" or beyond what a rabid analytic fan does in their spare time and who actually influence play and personnel? Is that the norm now or still sortof of the exception with most not doing that much more than they've always done with "data" even though they have lots more available and a far greater variety? With all the other things GMs, coaches and players do, what is a realistic estimate for use?
Do teams need more help on the backend of analytics? Almost all the hires seem to be on the front end production or at least based on front end production resume skills. I can think of a few team analytic staff hires where highly successful USE of analytics probably played a big role in the selection but in most cases it seems like hires are mostly or entirely based on analytic production.
To focus on one technique, RPM produced by ESPN consultants or a rough equivalent produced in house, how many hours of direct study and informed use are teams devoting to it? Does this book address this technique and question at all? If it does, I will be more inclined to read it. The scale of use of RPM by teams may range from none to casual use to intense use. But even in relatively intense use, how intense is it? I could imagine RPM estimates for own players or targets being studied, sliced and diced, calibrated, compared to other one metrics and discrete stats for dozens, hundreds even 1,000 plus hours especially for key players and impact estimates that vary a lot by reference tool. Compared to the multi-million dollar and win impact significance of these issues spending $50,000 in staff time PER CASE Just For This Piece of a comprehensive analysis set would not seem unreasonable. Spending less might be.
Would you say the book reveals anything that breaks the veil of team secrecy? A lot or a little?
Calling it a quick read makes me concerned that it might be a book length recap of what has essentially already been in print. There is a place for that but that isn't what would interest me most. I'd prefer thick, long chapters based on stuff nobody else has gotten and we weren't supposed to get. The title of the book is suggestive of the latter but some of the Amazon reviews give the impression of mostly summary of what was already out there.
Did you get the sense that teams are fully utilizing what they have or generating data & analysis that more than not lays there understudied and acted upon? For every hour put into building and maintain databases and spitting out & beautifying outputs how many hours go into studying the data hard player by player, analytic piece by piece and beneath them and proposing and selling actions? Should it be 5-10? At least 2? Could it even be 0.5 or less? Does this book give any basis for answering this question? Any former or current insider willing to address this general question? Say a team has 2 data analysts who work 60 hours a week. That would be 6,000 hours a year. How much went to the production process vs. reflection / recommendation based on the output? I doubt it is 50% but tell me I am wrong or wrong to expect or want that ratio. I'd hope it was at least 25%, but I wonder. Could be even less?
How many hours did top managers spend with any of it? 2 hours a day for 5 guys for a year would be 3,500 hours. Or if you said that every hour of their work was informed by it, then you claim 15,000 hours of use. Are there 5 guys per team who use advanced data analytics "a lot", "all the time" or beyond what a rabid analytic fan does in their spare time and who actually influence play and personnel? Is that the norm now or still sortof of the exception with most not doing that much more than they've always done with "data" even though they have lots more available and a far greater variety? With all the other things GMs, coaches and players do, what is a realistic estimate for use?
Do teams need more help on the backend of analytics? Almost all the hires seem to be on the front end production or at least based on front end production resume skills. I can think of a few team analytic staff hires where highly successful USE of analytics probably played a big role in the selection but in most cases it seems like hires are mostly or entirely based on analytic production.
To focus on one technique, RPM produced by ESPN consultants or a rough equivalent produced in house, how many hours of direct study and informed use are teams devoting to it? Does this book address this technique and question at all? If it does, I will be more inclined to read it. The scale of use of RPM by teams may range from none to casual use to intense use. But even in relatively intense use, how intense is it? I could imagine RPM estimates for own players or targets being studied, sliced and diced, calibrated, compared to other one metrics and discrete stats for dozens, hundreds even 1,000 plus hours especially for key players and impact estimates that vary a lot by reference tool. Compared to the multi-million dollar and win impact significance of these issues spending $50,000 in staff time PER CASE Just For This Piece of a comprehensive analysis set would not seem unreasonable. Spending less might be.
Would you say the book reveals anything that breaks the veil of team secrecy? A lot or a little?
Calling it a quick read makes me concerned that it might be a book length recap of what has essentially already been in print. There is a place for that but that isn't what would interest me most. I'd prefer thick, long chapters based on stuff nobody else has gotten and we weren't supposed to get. The title of the book is suggestive of the latter but some of the Amazon reviews give the impression of mostly summary of what was already out there.
Re: Anybody read this book and have reactions?
Suppose somebody went to a team (or a forum...) and suggested adding 100 people to basketball operations (across all sub-departments / disciplines) above team current or NBA average. I'd guess you could do that for $10-15 million a year. How many GMs and owners would even consider it? Has any done it to that scale? Why not on trial basis and then judge value? 50 adds? 25?
Re: Anybody read this book and have reactions?
The follow response I actually posted on my blog - since it's something I frequently lament about but rarely mention there:
One example, the Dallas Mavs (luv ya Mark, don't unfollow me or cross me off your possible future hire list) & the Rajon Rondo debacle. The time of their trade for him, I died (as a fan) a little inside. I knew it was a horrible trade for the Mavs, just horrible. I told everyone who would listen how I felt- just not a ton on Twitter for, honestly, fear of losing said billionaire follow I previously mentioned. Would a team of quality analysts and basketball people have traded for an expiring Rondo contract (while the possibility of resigning him would have been tied to an almost max offer - think about that, offering a MAX contract to Rajon Rondo) - and to get Rondo give up Jae Crowder (! - already a poor man's Draymond Green, but the Mavs hadn't fully figured that out yet), Brandan Wright (also very underrated, but more in a post box score stuffer way than Jae), Jameer Nelson, a FIRST round pick, and a 2nd round pick? Remember - they did this at a time when the Mavs offense was #1 in the league in efficiency by a solid margin.
I know, the Mavs also got Dwight Powell. Yay.
The assets the Mavs gave up in that trade will probably cost them many, many millions of dollars over the next number of years. A more depleted roster, a roster that will be tough to rebuild with those future draft picks gone. Would a team of analysts (honest analysts who aren't sycophants) allowed that trade without screaming at the top of their lungs "please, PLEASE, no!"? Just in this one case - spending greater resources in basketball operations would have SAVED the Mavs millions in lost revenue (ie, lost wins) the next number of years.
I don't think it's a stretch to say that the Boston Celtics probably have put more focus and money into basketball ops and their analytics department than the Mavs or most other teams.
Main point, teams need to WIN games. Having much greater amounts of salary paid to a sales department (trying to sell tickets & get sponsors) than an analytics team devoted helping with proper roster construction, game planning, and player usage (help keep players fresh & avoid injuries) makes little sense to me.
My final point - maybe I'm wrong. Maybe many of these teams are already putting, say, more than $10 million a year into their analytics department? Maybe they are spending more money on analytics than they are their sales department? Honestly, these are mainly (slightly educated) assumptions I am making, from what I've been told by others (some more in the know) and from what I see (media reports).
If that already twice mentioned certain follow reads this - I bet there's a chance I may get lambasted in DM if my assumptions are off. That'll be interesting. Hell, if ANY NBA admin DM me (there are a number who follow) - I am ALL ears. If better info is presented to me, I'm more than willing to add any future illumination or clarification to this write up if need be.
But, if that is true (teams are spending way more resources on basketball ops than I assume), my 17 year draft model retrodiction results are even scarier. One guy with ZERO financial resources, using only college box score data (ie, a fraction of info available to teams), can project players that correlate to real future NBA success (tested historically) MUCH better than actual NBA teams of GMs, coaches, scouts, and analysts can? How is that possible?
Similar to what I've mentioned before - these teams are, theoretically, billion dollar companies. Having such an insanely small (relative) amount of paid salary going out to basketball operations & analytic departments seems pretty short sighted to me. A quality full TEAM of analysts/basketball people of varying approaches and abilities working with the gm, coaches, & has to make for better decisions. A single bad decision (let alone MANY bad decisions) can cost teams many, many MILLIONS of dollars. Putting these decisions in the hands of a couple people (think, Vlade Divac being one of a couple) - with maybe some input from some scouts & maybe a little input from an analyst - just seems SO risky long term.Crow wrote:Suppose somebody went to a team (or a forum...) and suggested adding 100 people to basketball operations (across all sub-departments / disciplines) above team current or NBA average. I'd guess you could do that for $10-15 million a year. How many GMs and owners would even consider it? Has any done it to that scale? Why not on trial basis and then judge value? 50 adds? 25?
One example, the Dallas Mavs (luv ya Mark, don't unfollow me or cross me off your possible future hire list) & the Rajon Rondo debacle. The time of their trade for him, I died (as a fan) a little inside. I knew it was a horrible trade for the Mavs, just horrible. I told everyone who would listen how I felt- just not a ton on Twitter for, honestly, fear of losing said billionaire follow I previously mentioned. Would a team of quality analysts and basketball people have traded for an expiring Rondo contract (while the possibility of resigning him would have been tied to an almost max offer - think about that, offering a MAX contract to Rajon Rondo) - and to get Rondo give up Jae Crowder (! - already a poor man's Draymond Green, but the Mavs hadn't fully figured that out yet), Brandan Wright (also very underrated, but more in a post box score stuffer way than Jae), Jameer Nelson, a FIRST round pick, and a 2nd round pick? Remember - they did this at a time when the Mavs offense was #1 in the league in efficiency by a solid margin.
I know, the Mavs also got Dwight Powell. Yay.
The assets the Mavs gave up in that trade will probably cost them many, many millions of dollars over the next number of years. A more depleted roster, a roster that will be tough to rebuild with those future draft picks gone. Would a team of analysts (honest analysts who aren't sycophants) allowed that trade without screaming at the top of their lungs "please, PLEASE, no!"? Just in this one case - spending greater resources in basketball operations would have SAVED the Mavs millions in lost revenue (ie, lost wins) the next number of years.
I don't think it's a stretch to say that the Boston Celtics probably have put more focus and money into basketball ops and their analytics department than the Mavs or most other teams.
Main point, teams need to WIN games. Having much greater amounts of salary paid to a sales department (trying to sell tickets & get sponsors) than an analytics team devoted helping with proper roster construction, game planning, and player usage (help keep players fresh & avoid injuries) makes little sense to me.
My final point - maybe I'm wrong. Maybe many of these teams are already putting, say, more than $10 million a year into their analytics department? Maybe they are spending more money on analytics than they are their sales department? Honestly, these are mainly (slightly educated) assumptions I am making, from what I've been told by others (some more in the know) and from what I see (media reports).
If that already twice mentioned certain follow reads this - I bet there's a chance I may get lambasted in DM if my assumptions are off. That'll be interesting. Hell, if ANY NBA admin DM me (there are a number who follow) - I am ALL ears. If better info is presented to me, I'm more than willing to add any future illumination or clarification to this write up if need be.
But, if that is true (teams are spending way more resources on basketball ops than I assume), my 17 year draft model retrodiction results are even scarier. One guy with ZERO financial resources, using only college box score data (ie, a fraction of info available to teams), can project players that correlate to real future NBA success (tested historically) MUCH better than actual NBA teams of GMs, coaches, scouts, and analysts can? How is that possible?