 |
APBRmetrics
The statistical revolution will not be televised.
|
| View previous topic :: View next topic |
| Author |
Message |
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 979
Location: Seattle
|
 Posted: Sat Jul 23, 2005 2:07 am Post subject: Dan's latest |
 |
|
Part two on the use of capologists and stats experts (but mostly the latter).
| Quote: | This makes it hard for basketball people to see the value of serious stats work, especially when they repeatedly get free offers from academics or retired millionaires from Microsoft to do work for them for free. In such an environment all stats people look the same, so it seems ludicrous to invest time and money into any one expert. But as I argued in Part I, it is that kind of investment that will result in stats experts helping teams.
But this should not all be put on the teams. We in the basketball stats community need to do a better job communicating. We need to come up with convincing arguments for why we are relevant, why we are not a threat to basketball people, and how we can help basketball people be better at what they do. It will take time, but eventually I think we can do this. |
Thoughts? |
|
| Back to top |
|
 |
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 979
Location: Seattle
|
| Posted: Sat Jul 23, 2005 2:10 am Post subject: |
|
|
To me, the arguing why we're relevant is still the most important. When I interviewed Rick Sund for the Sonics Play Moneyball series, he asked me whether I was a stats guy (not common knowledge on the basketball side of things) and then why.
I didn't really have a good reason to give him. Part of it was that I didn't want to get into myself too deeply, but it's partially I don't have a really good answer. I would say my answer is basically that I know I'm subjective about many things and I like the objectivity that stats provide, but that's not very strong. |
|
| Back to top |
|
|
NickS
Joined: 30 Dec 2004
Posts: 384
|
| Posted: Sat Jul 23, 2005 2:17 pm Post subject: |
|
|
The reason I'm a stat guy is because I don't have time to watch hundreds of hours of basketball. It's a way for me to follow the game and the players. So I would find stats interesting regardless of whether they provide anything that can't be gotten by watching the game. But I believe that they do.
The reason to use stats in any field is because humans are poor at evaluating probability. We tend to see patterns where there aren't, overestimate the probability of low frequency events and, most importantly, have a tendency towards confimation bias -- looking for evidence that confirm our preexisting beliefs.
One of the things that's said in defense of stats in baseball is that you can't tell the difference between a .260 hitter and a .280 hitter by watching one game or one series. The difference amounts to one extra hit every 2 weeks. Similarly is there any way to tell just by watching whether Eddy Curry is more or less prone to turnovers than Yao Ming?
Similarly I think that one of the best uses of stats is to provoke questions and try to map out ways in which questions can be answered. How can we tell if a team is shooting "too many" or "too few" three-pointers? Do shot-blockers have an "intimidation" effect? How valuable are "scoring" point guards compared to "traditional" point guards? Are specialists more or less valuable than generalists? How valuable is it to have guards who can rebound or big men who can pass? What separates a good shooter from a great shooter? Stats can't answer all of those questions but they can rule out some wrong answers that have intuitive appeal and focus attention on possibilities that are more likely to be correct.
Finally I would link to DeanO's old post on the power of context |
|
| Back to top |
|
|
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
|
| Posted: Sun Jul 24, 2005 3:56 pm Post subject: |
|
|
| NickS wrote: | The reason I'm a stat guy is because I don't have time to watch hundreds of hours of basketball. It's a way for me to follow the game and the players. So I would find stats interesting regardless of whether they provide anything that can't be gotten by watching the game. But I believe that they do.
|
Definitely a good reason. One of the things I can definitely say when I'm doing work for a team is that my stats "see" all the games, whereas scouts can only see a few... But the entertainment value has been there in reading boxscores since I was 10 years old.
| NickS wrote: | The reason to use stats in any field is because humans are poor at evaluating probability. We tend to see patterns where there aren't, overestimate the probability of low frequency events and, most importantly, have a tendency towards confimation bias -- looking for evidence that confirm our preexisting beliefs.
|
Confirmation bias is definitely a big deal in basketball. I see it a lot. Because of that, I am very careful in how I communicate what I know and don't know. I have to communicate in such a way to minimize the risk of confirmation bias.
| NickS wrote: | One of the things that's said in defense of stats in baseball is that you can't tell the difference between a .260 hitter and a .280 hitter by watching one game or one series. The difference amounts to one extra hit every 2 weeks. Similarly is there any way to tell just by watching whether Eddy Curry is more or less prone to turnovers than Yao Ming?
|
This is one of my favorite stories. We think qualitatively of the difference between a .300 hitter and a .260 hitter and it's HUGE. But it really isn't all that much.
| NickS wrote: | Similarly I think that one of the best uses of stats is to provoke questions and try to map out ways in which questions can be answered. How can we tell if a team is shooting "too many" or "too few" three-pointers? Do shot-blockers have an "intimidation" effect? How valuable are "scoring" point guards compared to "traditional" point guards? Are specialists more or less valuable than generalists? How valuable is it to have guards who can rebound or big men who can pass? What separates a good shooter from a great shooter? Stats can't answer all of those questions but they can rule out some wrong answers that have intuitive appeal and focus attention on possibilities that are more likely to be correct.
|
Having been a coach for our summer league team, I also remember that one of the best uses is to identify how to improve players. Sitting with players in a game, I could give them advice based on stats and it was very satisfying to see them go out, do the things, and have success. It was definitely a nice complement to the x's and o's that our other coaches had. The things weren't usually revolutionary (though I think one or two were especially insightful), but it was good to see how stats and the philosophies behind a lot of stat models really can help make players better, not just evaluate them...
After almost exactly a year in the business, I am more convinced than ever that the value of statistical analysis in basketball can be huge.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers. |
|
| Back to top |
|
|
Dan Rosenbaum
Joined: 03 Jan 2005
Posts: 541
Location: Greensboro, North Carolina
|
| Posted: Sun Jul 24, 2005 4:05 pm Post subject: |
|
|
| NickS wrote: | The reason I'm a stat guy is because I don't have time to watch hundreds of hours of basketball. It's a way for me to follow the game and the players. So I would find stats interesting regardless of whether they provide anything that can't be gotten by watching the game. But I believe that they do.
The reason to use stats in any field is because humans are poor at evaluating probability. We tend to see patterns where there aren't, overestimate the probability of low frequency events and, most importantly, have a tendency towards confimation bias -- looking for evidence that confirm our preexisting beliefs.
One of the things that's said in defense of stats in baseball is that you can't tell the difference between a .260 hitter and a .280 hitter by watching one game or one series. The difference amounts to one extra hit every 2 weeks. Similarly is there any way to tell just by watching whether Eddy Curry is more or less prone to turnovers than Yao Ming?
Similarly I think that one of the best uses of stats is to provoke questions and try to map out ways in which questions can be answered. How can we tell if a team is shooting "too many" or "too few" three-pointers? Do shot-blockers have an "intimidation" effect? How valuable are "scoring" point guards compared to "traditional" point guards? Are specialists more or less valuable than generalists? How valuable is it to have guards who can rebound or big men who can pass? What separates a good shooter from a great shooter? Stats can't answer all of those questions but they can rule out some wrong answers that have intuitive appeal and focus attention on possibilities that are more likely to be correct.
Finally I would link to DeanO's old post on the power of context |
I love this post. IMO, it is one of the best posts that I have seen in years of visiting message boards. Thank you! I would love to hear what others have to say on this topic, before I chime in. |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3608
Location: Hendersonville, NC
|
| Posted: Sun Jul 24, 2005 6:19 pm Post subject: |
|
|
I'll quickly agree with Dan and Dean that this is a superb bit of commentary. Sometimes I expect thoughtful posts don't get enough (if any) nice round of applause, or such.
And I hope Dean's first year is followed by another, and ... |
|
| Back to top |
|
|
NickS
Joined: 30 Dec 2004
Posts: 384
|
| Posted: Sun Jul 24, 2005 7:58 pm Post subject: |
|
|
Wow, I'm flattered to be praised by such luminaries.
Thanks. |
|
| Back to top |
|
|
kbche
Joined: 19 Jul 2005
Posts: 51
Location: washington d.c.
|
| Posted: Sun Jul 24, 2005 11:13 pm Post subject: |
|
|
The statistical analyses of Dan and Dean are very interesting and add another dimension to basketball interpretation. I think that all teams should pursue any and all avenues to improve the quality of their team and their chances of winning a championship.
I am a novice to basketball statistics. I have researched cNBAStats (current statistical analyses of NBA basketball) including "Basketball on Paper", articles by Dan Rosenbaum, Roland ratings, and articles relating to the WINVAL rating system. NBA executives are interested in the bottom line. Statisticians need to clearly define the added value and the goals of these statistical analyses. The statistics commonly encountered are player based, and they appear to be designed to generally determine the best (most effective) player in the NBA. How can these stats be used from year to year, and team to team?
I am a firm believer that basketball is a science that can be accurately characterized, and the statistical Holy Grail which Dan mentions is within reach.
KAB  |
|
| Back to top |
|
|
SGreenwell
Joined: 12 Feb 2005
Posts: 76
Location: Rhode Island
|
| Posted: Mon Jul 25, 2005 1:11 am Post subject: |
|
|
| kbche wrote: | | Statisticians need to clearly define the added value and the goals of these statistical analyses. |
I think this is the biggest hurdle. When talking to people about stats, whether it be baseball or basketball, I always have to get over the "prove it" hurdle. And it's hard to prove it unless you're a GM, coach or someone else with direct control over a team.
You can cite Theo Epstein or the work of the Supersonics, but I've noticed a stubborness to accept stats unless there is a clear preponderance of evidence. Hell, people still don't believe Oakland has a winning gameplan - "They may compete with a $50 million payroll by god, but they haven't won a World Series!!!" |
|
| Back to top |
|
|
bchaikin
Joined: 27 Jan 2005
Posts: 689
Location: cleveland, ohio
|
| Posted: Tue Jul 26, 2005 5:03 pm Post subject: |
|
|
how about playing devil's advocate...
How can we tell if a team is shooting "too many" or "too few" three-pointers?
depends on how many they are hitting and the score of the game...
Do shot-blockers have an "intimidation" effect?
76ers stats guru harvey pollack used to keep track of what he called "intimidations" - shots that a player alters enough that they are prevented from going in but where the shot was not actually blocked (he kept track of this for players like shawn bradley and manute bol and the numbers were more than each player's actual blocked shots)...
How valuable are "scoring" point guards compared to "traditional" point guards?
depends on who their teammates are...
Are specialists more or less valuable than generalists?
again depends on who their teammates are...
How valuable is it to have guards who can rebound or big men who can pass?
same thing, depends on who they are playing with...
What separates a good shooter from a great shooter?
the one who makes the most baskets maybe?  ....
these answers are not meant to make light of the above questions, but simply to show that imho these questions can be answered via stats and stats analysis, and the methodology will only get better with time...
Stats can't answer all of those questions
well that depends on just exactly what all those questions are, but imho they can answer those above questions now, and answer even more questions later with the acquisition of even more data - via efforts like 82games.com and whomever else decides to bring innovative stats gathering projects to the forefront (like the WNBA boxscore project)...
but they can rule out some wrong answers that have intuitive appeal and focus attention on
absolutely...
I am a firm believer that basketball is a science that can be accurately characterized, and the statistical Holy Grail... is within reach.
agreed. in this day and age many many processes - from space exploration to medical technology to whatever - can be mapped out, analyzed, projected, and simulated. basketball is no different...
...the value of statistical analysis in basketball can be huge...
yes, to those willing to have an open mind and especially to those that are looking for an edge over their competition...
I've noticed a stubborness to accept stats unless there is a clear preponderance of evidence.
the evidence is there, and has been there for some time now. all it takes is somebody who really wants to see it... |
|
| Back to top |
|
|
jambalaya
Joined: 30 Jan 2005
Posts: 282
|
| Posted: Thu Jul 28, 2005 11:08 am Post subject: the value of watching the game directly |
|
|
| NickS wrote: | The reason I'm a stat guy is because I don't have time to watch hundreds of hours of basketball. It's a way for me to follow the game and the players. So I would find stats interesting regardless of whether they provide anything that can't be gotten by watching the game. But I believe that they do.
The reason to use stats in any field is because humans are poor at evaluating probability. We tend to see patterns where there aren't, overestimate the probability of low frequency events and, most importantly, have a tendency towards confimation bias -- looking for evidence that confirm our preexisting beliefs.
One of the things that's said in defense of stats in baseball is that you can't tell the difference between a .260 hitter and a .280 hitter by watching one game or one series. The difference amounts to one extra hit every 2 weeks. Similarly is there any way to tell just by watching whether Eddy Curry is more or less prone to turnovers than Yao Ming?
Similarly I think that one of the best uses of stats is to provoke questions and try to map out ways in which questions can be answered. How can we tell if a team is shooting "too many" or "too few" three-pointers? Do shot-blockers have an "intimidation" effect? How valuable are "scoring" point guards compared to "traditional" point guards? Are specialists more or less valuable than generalists? How valuable is it to have guards who can rebound or big men who can pass? What separates a good shooter from a great shooter? Stats can't answer all of those questions but they can rule out some wrong answers that have intuitive appeal and focus attention on possibilities that are more likely to be correct.
Finally I would link to DeanO's old post on the power of context |
i agree with the danger of confirmation bias, but... to state the other side, which also has value, value that is obvious and of course is not necessarily really being denied by anyone, i would also say that watching the game directly will tell you many things stats never will.
stats will tell you the ball went in the hoop or not. it won't tell you if the player created the best shot possible or overlooked a open teammate. and how often it was one or the other. watching a ray allen move you can see quality team-oriented decisionmaking as well as an aggressive shooter with an excellent stroke.
you can also learn a lot about players by observing the technique and decisonmaking of rebounders and passers. stats can't and won't replace observation. it is the synergy of the two that makes advances. |
|
| Back to top |
|
|
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
|
| Posted: Thu Jul 28, 2005 11:38 am Post subject: Re: the value of watching the game directly |
|
|
| jambalaya wrote: | | NickS wrote: | The reason I'm a stat guy is because I don't have time to watch hundreds of hours of basketball. It's a way for me to follow the game and the players. So I would find stats interesting regardless of whether they provide anything that can't be gotten by watching the game. But I believe that they do.
The reason to use stats in any field is because humans are poor at evaluating probability. We tend to see patterns where there aren't, overestimate the probability of low frequency events and, most importantly, have a tendency towards confimation bias -- looking for evidence that confirm our preexisting beliefs.
One of the things that's said in defense of stats in baseball is that you can't tell the difference between a .260 hitter and a .280 hitter by watching one game or one series. The difference amounts to one extra hit every 2 weeks. Similarly is there any way to tell just by watching whether Eddy Curry is more or less prone to turnovers than Yao Ming?
Similarly I think that one of the best uses of stats is to provoke questions and try to map out ways in which questions can be answered. How can we tell if a team is shooting "too many" or "too few" three-pointers? Do shot-blockers have an "intimidation" effect? How valuable are "scoring" point guards compared to "traditional" point guards? Are specialists more or less valuable than generalists? How valuable is it to have guards who can rebound or big men who can pass? What separates a good shooter from a great shooter? Stats can't answer all of those questions but they can rule out some wrong answers that have intuitive appeal and focus attention on possibilities that are more likely to be correct.
Finally I would link to DeanO's old post on the power of context |
i agree with the danger of confirmation bias, but... to state the other side, which also has value, value that is obvious and of course is not necessarily really being denied by anyone, i would also say that watching the game directly will tell you many things stats never will.
stats will tell you the ball went in the hoop or not. it won't tell you if the player created the best shot possible or overlooked a open teammate. and how often it was one or the other. watching a ray allen move you can see quality team-oriented decisionmaking as well as an aggressive shooter with an excellent stroke.
you can also learn a lot about players by observing the technique and decisonmaking of rebounders and passers. stats can't and won't replace observation. it is the synergy of the two that makes advances. |
I don't think many (if any) people in here would disagree that there is value in watching and listening to the game, telling you things that stats aren't telling you. If it sometimes seems like we don't believe that, well, at least you don't have to bind us up and force us to watch games in order to convince us. From the other perspective, there are people who will never accept that stats can tell you something more clearly than watching the game, even if you do a Clockwork Orange type torture to them.
A friend of mine has drawn 3 circles, with "eyes", "ears", and "numbers" in them. Where the 3 circles intersect is a good decision. Most decisions are made on the overlap of eyes and ears, which means that roughly 2/3rds of those decisions not coinciding with numbers and probably not optimal (2/3rds being an artifact of drawing things and not meant to be exact in any way).
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers. |
|
| Back to top |
|
|
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
|
| Posted: Tue Aug 02, 2005 5:08 pm Post subject: Re: the value of watching the game directly |
|
|
| HoopStudies wrote: |
I don't think many (if any) people in here would disagree that there is value in watching and listening to the game, telling you things that stats aren't telling you. If it sometimes seems like we don't believe that, well, at least you don't have to bind us up and force us to watch games in order to convince us. From the other perspective, there are people who will never accept that stats can tell you something more clearly than watching the game, even if you do a Clockwork Orange type torture to them.
A friend of mine has drawn 3 circles, with "eyes", "ears", and "numbers" in them. Where the 3 circles intersect is a good decision. Most decisions are made on the overlap of eyes and ears, which means that roughly 2/3rds of those decisions not coinciding with numbers and probably not optimal (2/3rds being an artifact of drawing things and not meant to be exact in any way). |
so then do you watch the game to confirm what you've observed through statistics, or do you try to use statistics to confirm what you've observed while watching the game? |
|
| Back to top |
|
|
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
|
| Posted: Tue Aug 02, 2005 6:04 pm Post subject: Re: the value of watching the game directly |
|
|
| gabefarkas wrote: | | HoopStudies wrote: |
I don't think many (if any) people in here would disagree that there is value in watching and listening to the game, telling you things that stats aren't telling you. If it sometimes seems like we don't believe that, well, at least you don't have to bind us up and force us to watch games in order to convince us. From the other perspective, there are people who will never accept that stats can tell you something more clearly than watching the game, even if you do a Clockwork Orange type torture to them.
A friend of mine has drawn 3 circles, with "eyes", "ears", and "numbers" in them. Where the 3 circles intersect is a good decision. Most decisions are made on the overlap of eyes and ears, which means that roughly 2/3rds of those decisions not coinciding with numbers and probably not optimal (2/3rds being an artifact of drawing things and not meant to be exact in any way). |
so then do you watch the game to confirm what you've observed through statistics, or do you try to use statistics to confirm what you've observed while watching the game? |
Depends. It goes both ways, of course. Sometimes I find things in numbers I never saw. I then watch the game to see if it is there. Other times I see things in the game and then look in the numbers to see if it is reflected. If someone comes to me with a suggestion of something that may be right, I approach it both ways. If I don't find something in numbers, that may suggest rephrasing the study. If I don't find it in the game, it could be a game where things don't work that way. You gotta watch a lot of games sometimes to see big issues. Numbers can help that way.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers. |
|
| Back to top |
|
|
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
|
| Posted: Wed Aug 03, 2005 10:56 am Post subject: Re: the value of watching the game directly |
|
|
| HoopStudies wrote: |
Depends. It goes both ways, of course. Sometimes I find things in numbers I never saw. I then watch the game to see if it is there. Other times I see things in the game and then look in the numbers to see if it is reflected. |
not to nitpick, since I agree with you that it goes both ways, but wouldn't either case be a prime example of confirmation bias? |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
