 |
APBRmetrics
The statistical revolution will not be televised.
|
| View previous topic :: View next topic |
| Author |
Message |
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
|
 Posted: Sat Aug 20, 2005 10:44 am Post subject: Question about the 82games.com project |
 |
|
Hey, everyone. This is my first post on this board. I recently signed up to work for the 82games.com charting project next season, and I have what I think are some good ideas about things to study and record. I looked around at 82games.com, and I couldn't find out if they were planning on doing some of these things. So I thought I would throw these ideas around on this board, and see if they are possible.
When charting the games, one thing the charters should do is record not only who forces misses on defense but who took that shot. For example, if Kobe misses a contested jumper with Bruce Bowen in his face, Bowen would get credit not just for forcing a miss but for forcing a good player to miss. Then, you could compare the offensive ratings of the players that took shots when Bowen was guarding them to the offensive ratings of those players when the rest of the league was guarding them. This way you account for good defenders always being matched up with the other team's best scorer. Here's a situation....
The Spurs just played the Pistons, who were without injured Chauncey Billups. While being guarded by Tony Parker, Lindsey Hunter scored 6 points on 5 possessions. Carlos Arroyo scored 3 points on 5 possessions. Tayshaun Prince, while being guarded by Bruce Bowen, scored 12 points on 10 possessions. Rip Hamilton, guarded by Bruce Bowen, scored 8 points on 10 possessions.
For the game, Tony Parker had a defensive rating of 90 (9 points/10 posessions times 10). Bruce Bowen had a defensive rating of 100 (20 points/20 pessessions times 5). According to these numbers, it seems Parker is the better defender.
However, the average offensive rating of the players Parker guarded was 95.5 (5 possessions for Hunter at a 93 rating and 5 for Arroyo at a 98 rating). The average offensive rating of the playerrs Bowen guarded was 113.5 (10 possessions for Hamilton at a 106 rating and 10 for Prince at a 121 rating). Comparing the players, Parker reduced the ratings of his opponents by 5.5. Bowen reduced the ratings of his opponents by 13.5. When looking at this stat, it's clear that Bowen is the superior defender.
I have two questions. Is it realistic to expect to be able to record this data for a whole season? And would this data be a good indicator of defensive ability? |
|
| Back to top |
|
 |
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 978
Location: Seattle
|
| Posted: Sat Aug 20, 2005 11:50 am Post subject: Re: Question about the 82games.com project |
|
|
| JNichols42887 wrote: | | I have two questions. Is it realistic to expect to be able to record this data for a whole season? And would this data be a good indicator of defensive ability? |
It would be more challenging, but I suspect almost certainly doable with a good TiVo (which I'm assuming is how the project is going to be done).
Would it be valuable? Yeah, I think that logic is accurate. The one thing I'd point out is you're still somewhat undervaluing elite defenders because, to some extent or another, you're comparing them to other elite defenders (who would be the ones guarding elite offensive players), but not accounting for defensive assignment was the biggest problem I found with Project Defensive Scoresheet. |
|
| Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004
Posts: 785
Location: Toronto
|
| Posted: Sat Aug 20, 2005 12:10 pm Post subject: Re: Question about the 82games.com project |
|
|
| JNichols42887 wrote: | | Is it realistic to expect to be able to record this data for a whole season? |
The more data recorded, the better. However, if you're new to game scoring, I strongly suggest you do at least a couple of practise runs, just to see what you're getting yourself into, and to see where you're limits are. It is a lot more work than it seems—you have to prepare yourself for it, both mentally and physically. Mentally, you have to understand that at minimum each game will occupy 3 hours of your time, more if you score more data. Physically, you have to assemble whatever tools you're going to need to record the data: some put everything straight onto a spreadsheet, some write a program for the purpose, and some go old-school with a notebook. The latter requires more time, because you have to transfer the data to the computer, but I find it the easiest way. You should try different methods on your own, see which works best for you.
Because it will be a lot of work. You have to prepare yourself.
_________________
ed |
|
| Back to top |
|
|
Roland_Beech
Joined: 14 Jan 2005
Posts: 43
|
| Posted: Sat Aug 20, 2005 1:45 pm Post subject: |
|
|
hello to JNichols and anyone else wondering about the extreme charting venture.
We will be undertaking this massive project this season, and I imagine we will start contacting folks who have expressed interest in being involved sometime in September. We will then do some preseason work in Oct. and get ready to go for the real season.
Certainly part of the defensive charting will allow us to 'cross index' players so you will know the results when player X is guarded by player Y and so on.
There will also be opportunities for people to contribute content to the site (articles featuring stats analysis, etc) and we will also be trying to give more exposure to the individuals seeking to establish their basketball credentials.
In the meantime there is some work going on for a redesign of the site which will change the page layouts and look/feel. It will also include some long asked for features like sortable tables showing all players/teams for given stats. |
|
| Back to top |
|
|
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 978
Location: Seattle
|
| Posted: Sat Aug 20, 2005 1:50 pm Post subject: |
|
|
I found a Sonics playoff game left on my TiVo and give it a shot. It doesn't actually seem that challenging in terms of scoring (I'm old school, since I've mostly charted in person).
Instead of using tally marks for each defensive category, I used the offensive player's number. It's a bit challenging to sort out 7 7 - two Lewis shots from 77 - Radmanovic shot and dividing half credit becomes a bit trickier, but that part wasn't any more time-consuming.
Getting it to the spreadsheet, I imagine, would be a different story. |
|
| Back to top |
|
|
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
|
| Posted: Sat Aug 20, 2005 3:10 pm Post subject: |
|
|
| Roland_Beech wrote: | hello to JNichols and anyone else wondering about the extreme charting venture.
We will be undertaking this massive project this season, and I imagine we will start contacting folks who have expressed interest in being involved sometime in September. We will then do some preseason work in Oct. and get ready to go for the real season.
Certainly part of the defensive charting will allow us to 'cross index' players so you will know the results when player X is guarded by player Y and so on.
There will also be opportunities for people to contribute content to the site (articles featuring stats analysis, etc) and we will also be trying to give more exposure to the individuals seeking to establish their basketball credentials.
In the meantime there is some work going on for a redesign of the site which will change the page layouts and look/feel. It will also include some long asked for features like sortable tables showing all players/teams for given stats. |
Also, I think that once a good amount of data is collected, you can expand the study even more. You could analyze players to determine how well they play defense against players who shoot better than 40% on three pointers, or players who take more than 50% of their shots inside. If accurate, that kind of data would be enormously valuable for both coaches and GMs. |
|
| Back to top |
|
|
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
|
| Posted: Sat Aug 20, 2005 3:12 pm Post subject: |
|
|
| admin wrote: | I found a Sonics playoff game left on my TiVo and give it a shot. It doesn't actually seem that challenging in terms of scoring (I'm old school, since I've mostly charted in person).
Instead of using tally marks for each defensive category, I used the offensive player's number. It's a bit challenging to sort out 7 7 - two Lewis shots from 77 - Radmanovic shot and dividing half credit becomes a bit trickier, but that part wasn't any more time-consuming.
Getting it to the spreadsheet, I imagine, would be a different story. |
Dividing the credit is interesting. It's necessary, but how is it done? Do you just split the credit based on what seems right to you personally? Or is there a more scientific way of doing it? |
|
| Back to top |
|
|
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 978
Location: Seattle
|
| Posted: Sat Aug 20, 2005 3:50 pm Post subject: |
|
|
| There are general rules that DeanO laid out, but it's primarily scorer's decision. |
|
| Back to top |
|
|
kjb
Joined: 03 Jan 2005
Posts: 864
Location: Washington, DC
|
| Posted: Sun Aug 21, 2005 7:06 am Post subject: |
|
|
| admin wrote: | I found a Sonics playoff game left on my TiVo and give it a shot. It doesn't actually seem that challenging in terms of scoring (I'm old school, since I've mostly charted in person).
Instead of using tally marks for each defensive category, I used the offensive player's number. It's a bit challenging to sort out 7 7 - two Lewis shots from 77 - Radmanovic shot and dividing half credit becomes a bit trickier, but that part wasn't any more time-consuming.
Getting it to the spreadsheet, I imagine, would be a different story. |
What really boosted the accuracy and efficiency of my charting last season was to print out a play-by-play from NBA.com and make the notations directly on the margins. |
|
| Back to top |
|
|
mtamada
Joined: 28 Jan 2005
Posts: 376
|
| Posted: Tue Aug 23, 2005 5:24 am Post subject: |
|
|
| WizardsKev wrote: | | What really boosted the accuracy and efficiency of my charting last season was to print out a play-by-play from NBA.com and make the notations directly on the margins. |
Yes, as one of the stats recorders for DeanO's Project Defensive Scoresheet for theWNBA, it was enormously valuable for me to have a play-by-play (and box scores for each half) to double-check my work.
On the other hand, I was simply recording the data live (we had only a small number of stats to record, specifically, only the defensive ones). The 82Games.com stats recorders I believe will all be viewing video, presumably with re-wind and slo-mo capability, which is another way of enormously increasing accuracy. |
|
| Back to top |
|
|
mtamada
Joined: 28 Jan 2005
Posts: 376
|
| Posted: Tue Aug 23, 2005 5:30 am Post subject: |
|
|
Oh, in regard to the multiple credits question: I don't know how RolandB is planning to handle it, but DeanO, properly in my view, told his stats recorders to give multiple (but fractional) credit for missed FGAs and turnovers when appropriate. E.g. Tamecka Dixon deflects a pass and Mawadi Mabika picks it up: 1/2 of a TO caused by each of them. Marie Ferdinand blows past Dixon, Lisa Leslie slides over to help but Ferdinand scoops in the layin under Leslie's arm: 1/2 of a FGM charged to each of them.
But the fractions were all 1/2, we didn't (or at least I didn't) try to do .25/.75 or .4/.6 credit divisions. |
|
| Back to top |
|
|
kjb
Joined: 03 Jan 2005
Posts: 864
Location: Washington, DC
|
| Posted: Tue Aug 23, 2005 6:32 am Post subject: |
|
|
I also did the fractional credit thing, and I always did it as a 50/50 split. If it was more than 2 players involved, I'd pick the two players most responsible, or assign it to the team category, or in a very few cases I'd split between a single player and the team category.
The deflected pass/picked-up the ball issue is an interesting one. I went the route of giving full credit in that kind of situation to the person who got the official box score steal. Even though I sometimes disagreed with the official scorekeepers decision. Basically, I did that so that I could have as much agreement as possible with the official record. I even used the official steals and blocks stats when I found them erroneously credited. I have notations on my scoresheets indicating who was "robbed" and who got the extra credit.
I don't know Roland's thinking on this, but it'd be great if we could use the 82games project to correct some of the errors that crop up in the official stats.
I'm not sure how you guys could track that stuff live. Maybe I just don't write fast enough, or maybe it was the extras I was recording. I would have had a tough time tracking with accuracy without my TiVo. |
|
| Back to top |
|
|
mtamada
Joined: 28 Jan 2005
Posts: 376
|
| Posted: Tue Aug 23, 2005 8:34 am Post subject: |
|
|
| WizardsKev wrote: | The deflected pass/picked-up the ball issue is an interesting one. I went the route of giving full credit in that kind of situation to the person who got the official box score steal. Even though I sometimes disagreed with the official scorekeepers decision. Basically, I did that so that I could have as much agreement as possible with the official record. I even used the official steals and blocks stats when I found them erroneously credited. I have notations on my scoresheets indicating who was "robbed" and who got the extra credit.
I don't know Roland's thinking on this, but it'd be great if we could use the 82games project to correct some of the errors that crop up in the official stats. |
One of the big things that I learned from the WNBA stats project (DeanO had already been aware of this) was how many judgement calls, and also outright errors, there are in the play-by-play sheets and even the box scores. The play-by-play errors or misjudgements were the more common ones, where steals and blocked shots were especially likely to be missed (or judged idiosyncratically) by the WNBA stats keeper. But occasionally a box score would clearly be in error.
So although I always compared my numbers to what was in the box scores and play-by-play, if I was sure that what I recorded was more accurate, or gave a better representation of what happened, than what the box score or play-by-play said, I went with my data.
But my interpretation of DeanO's intentions was that the collected data would be used for analytic purposes, and less for archivable record-keeping purposes where consistency with the box scores would be more important.
| Quote: | | I'm not sure how you guys could track that stuff live. Maybe I just don't write fast enough, or maybe it was the extras I was recording. I would have had a tough time tracking with accuracy without my TiVo. |
It was probably the extras. I did not even attempt to track any offensive stats nor rebounding or foul stats ('cuz the official scorekeepers and play-by-play record those events). I also pretty much wasn't really "keeping up with the game", i.e. I was not watching the lead, the scoring streaks, the strategies, etc. I had to zero in on two things: who the players were, and who made what sort of defensive play. So it did take total concentration. (I did learn to, every 4-5 minutes, write down the score and the time, so I could use that info as a sort of bookmark to match my notes with the same time and score in the play-by-play sheet.)
Another possible factor might've been that the Sparks gave me great seats, in press row. Action at the far side of the court might get a little difficult to follow if a player's shot was contested by two defenders, whose identity might be partly screened from me. But that's where knowing who the 10 players on the court were was a big help, because when the players disentangled, I could usually see or deduce who the defenders must've been.
But you're right, TiVo or the equivalent makes these stat-keeping tricks moot and unecessary. (Although players might still get screened from the camera.) |
|
| Back to top |
|
|
kjb
Joined: 03 Jan 2005
Posts: 864
Location: Washington, DC
|
| Posted: Tue Aug 23, 2005 8:54 am Post subject: |
|
|
| Even with TiVo, there are still issues. The biggest one was the network (Comcast for most Wizards games) coming back late from commercial, lingering too long on a replay, or leaving a graphic up through play. |
|
| Back to top |
|
|
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 978
Location: Seattle
|
| Posted: Tue Aug 23, 2005 10:03 am Post subject: |
|
|
| WizardsKev wrote: | | I'm not sure how you guys could track that stuff live. Maybe I just don't write fast enough, or maybe it was the extras I was recording. I would have had a tough time tracking with accuracy without my TiVo. |
I don't know about Mike, but I'm just really intelligent.  |
|
| 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
