View previous topic :: View next topic |
Author |
Message |
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Sat Apr 22, 2006 8:27 pm Post subject: Playoff Win Probabilities |
|
|
Hopefully, I'll have the time to post one of these for every game. Here's game 1 of the WAS/CLE series:
_________________ ed |
|
Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Sat Apr 22, 2006 8:28 pm Post subject: |
|
|
Game 1 SAC/SAN:
_________________ ed |
|
Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Sat Apr 22, 2006 9:53 pm Post subject: |
|
|
_________________ ed |
|
Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Sun Apr 23, 2006 12:39 am Post subject: |
|
|
_________________ ed |
|
Back to top |
|
|
Eli W
Joined: 01 Feb 2005 Posts: 402
|
Posted: Sun Apr 23, 2006 1:03 am Post subject: |
|
|
Cool pictures, and I read the thread you had on these before, but I still don't really get what these are for. What can we learn from after-the-fact probabilities? I'm not trying to be rude but I'm just not really sure I understand what you're trying to show. |
|
Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Sun Apr 23, 2006 1:45 am Post subject: |
|
|
John Quincy wrote: | Cool pictures, and I read the thread you had on these before, but I still don't really get what these are for. What can we learn from after-the-fact probabilities? |
Well, nothing, really. It just combines two numbers into one, making the story of the game easier to grasp. For example, the charts show that even though SAN won their game by 34 and CLE won thiers by 11, both games were essentially decided by the second quarter. And they also show that even though LAC won by a single basket, the 5-point MIA win had the scarier final quarter for the home team. _________________ ed |
|
Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Sun Apr 23, 2006 3:27 pm Post subject: |
|
|
Tight finish to this game. Here's a detail of the final 5 minutes, along with the running scoreline from the Pacers' perspective:
_________________ ed |
|
Back to top |
|
|
Kevin Pelton Site Admin
Joined: 30 Dec 2004 Posts: 979 Location: Seattle
|
Posted: Sun Apr 23, 2006 4:58 pm Post subject: |
|
|
Ed, do you have the ability to tie this to specific players and their effect (offensively, at least) on the probability of winning?
You see that regularly in baseball, but I don't know if I've ever seen it in the NBA. |
|
Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Sun Apr 23, 2006 5:24 pm Post subject: |
|
|
This thread is getting cluttered. After todays games, I'll just provide links to the images. _________________ ed |
|
Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Sun Apr 23, 2006 5:31 pm Post subject: |
|
|
admin wrote: | Ed, do you have the ability to tie this to specific players and their effect (offensively, at least) on the probability of winning?
You see that regularly in baseball, but I don't know if I've ever seen it in the NBA. |
Two ways of doing this:
1. Assign credit to players for each change in win probability. How to do this? Your guess is as good as mine. Three point attempt coverted -- some credit has to go to the shooter of course, but what about the other players on the floor? Do they get any?
2. Assign all players equal credit. What this would boil down to is a DanVal system, with win probability as the response variable instead of plus/minus. From previous experiments on last season's finals, I know that the results from only playoff games would be nowhere near statistically significant.
While some system based on changing win probability would seem to be the element missing from a holy grail rating system, I prefer to use it at a smaller scale: coaching decisions, shooting decisions, the effects of a single missed shot or offensive rebound, the effects of foul trouble. _________________ ed |
|
Back to top |
|
|
Kevin Pelton Site Admin
Joined: 30 Dec 2004 Posts: 979 Location: Seattle
|
Posted: Sun Apr 23, 2006 6:40 pm Post subject: |
|
|
Ed Küpfer wrote: | 1. Assign credit to players for each change in win probability. How to do this? Your guess is as good as mine. Three point attempt coverted -- some credit has to go to the shooter of course, but what about the other players on the floor? Do they get any? |
I was thinking of a very simplistic method: Credit or blame only to the shooter (and maybe the assister). Rebounds change the win probability by changing possession, correct? So credit there.
I don't see this as anything with any real meaning, at least at this smaple size, but I think it would be interesting to see nonetheless. |
|
Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Sun Apr 23, 2006 8:55 pm Post subject: |
|
|
_________________ ed |
|
Back to top |
|
|
Ben
Joined: 13 Jan 2005 Posts: 264 Location: Iowa City
|
Posted: Sun Apr 23, 2006 10:32 pm Post subject: |
|
|
admin wrote: | Ed Küpfer wrote: | 1. Assign credit to players for each change in win probability. How to do this? Your guess is as good as mine. Three point attempt coverted -- some credit has to go to the shooter of course, but what about the other players on the floor? Do they get any? |
I was thinking of a very simplistic method: Credit or blame only to the shooter (and maybe the assister). Rebounds change the win probability by changing possession, correct? So credit there.
I don't see this as anything with any real meaning, at least at this smaple size, but I think it would be interesting to see nonetheless. |
I assume this is the sort of thing you had in mind with respect to baseball?
http://www.amstat.org/sections/sis/pgp/
I've only seen it done world series time. Have you seen it done more extensively? |
|
Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Sun Apr 23, 2006 10:37 pm Post subject: |
|
|
Ben wrote: | I've only seen it done world series time. Have you seen it done more extensively? |
Yeah, it generally goes under the name Win Probability Added. This site has a million of them. _________________ ed |
|
Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Sun Apr 23, 2006 11:32 pm Post subject: |
|
|
_________________ ed |
|
Back to top |
|
|
|