| View previous topic :: View next topic |
| Author |
Message |
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
 Posted: Tue Mar 11, 2008 8:56 pm Post subject: The PowerGriz numbers. Last update: BOS@LAL, GAME 6 (06-15) |
 |
|
=============================================================
The following is an analysis of the game using a new mathematical model that I have developed based on an artificial neural network. The neural network is very accurate in predicting the probability of victory from the boxscore stats. It has been trained with data from all NBA games during the 2001-2007 seasons and is capable of predicting the number of victories for each team during a season with an error of less than 1 game. Therefore, it is possible to explore how the outcome of a game would change if the boxscore stats are changed. Shooting percentages, rebounds, steals, blocks, assists, defensive efficiency, etc are all factored in. As far as I am aware, this model is the most sophisticated method to gauge individual and collective performances. However, it is possible that I may still have errors in the code that implements the analysis so if you see any obvious problems in these numbers, please let me know.
Related threads:
Brief description and some intitial discussion
A more detailed description of the method including a FAQ
Source code
The number next to each player indicates how much the probability of winning the game would change if we replace that player with the average NBA player at his same position.
=================================================================
The PowerGriz numbers: TOR@LAL (2008-03-11)
Visiting team:raptors. Chance of success: 0.188856
Anthony Parker 0.117321
Rasho Nesterovic 0.0986689
T.J. Ford 0.0864977
Andrea Bargnani 0.0473242
Jamario Moon 0.0319616
Primoz Brezec -0.00306596
Jason Kapono -0.00547166
Joey Graham -0.0113910
Kris Humphries -0.0559231
Jose Calderon -0.0574500
Carlos Delfino -0.0772180
Home team:lakers. Chance of success: 0.811144
Kobe Bryant 0.335724
Lamar Odom 0.280343
Derek Fisher 0.126525
Luke Walton 0.0534647
Coby Karl 0.0489917
Jordan Farmar 0.0282038
Ronny Turiaf 0.0127247
Pau Gasol 0.00789177
Sasha Vujacic -0.0271800
Vladimir Radmanovic -0.0934605
Last edited by PowerGriz on Wed Jun 18, 2008 11:49 am; edited 32 times in total |
|
| Back to top |
|
 |
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Tue Mar 11, 2008 9:01 pm Post subject: |
|
|
A great game by Kobe and Lamar. Fisher had the hot hand and sank a few important treys, although I still think he's being too selfish. Luke and Coby Karl provided a much needed spark of hustle off of the bench. Gasol was not a big factor tonight. I have the feeling that he's been too tentative in the last few games.
As for the Raptors, you gotta love Parker. He's such a pure shooter and doesn't care if he has a hand in his face. Nesterovic played really well although rather quiet, as usual in him. I hadn't noticed that he had 18 points with 9/15 shooting and no less than 5 offensive rebounds(!!!) until I saw the boxscore. TJ Ford was unstoppable late in the game. Fortunately, the Lakers already had the game under control by then. I was disappointed that Calderon had probably his worst game of the season, he has been outstanding for the Raptors thus far but tonight he was one of the worst players on the floor. |
|
| Back to top |
|
|
Xepa
Joined: 11 Jul 2004
Posts: 6677
Location: somewhere in the world
|
| Posted: Tue Mar 11, 2008 9:02 pm Post subject: |
|
|
Interesting statistics.
Thank you for your contribution; it means a lot for our community.
Keep posting strong! And lets find out how accurate your system
ends up over the long haul.
Just my .02
_________________
Team FARMAR.
alleon86, The Bynum Supremacy, Barnstable, Massacre, nicehair911, melo061, Chicano, Kwame's Cake, lizlakers31, KobeBryant583, popo, C-mac 4.35, dj vitus, M4mb4 24 |
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Tue Mar 11, 2008 9:08 pm Post subject: |
|
|
Thanks, Xepa
I've been posting these stats regularly in the Grizzlies forum. I even had a section for the Lakers after the Gasol trade. But I think interest in Gasol is slowly fading away over there so I thought it would make more sense to post the LAL numbers here :) |
|
| Back to top |
|
|
SublimeAnarky
Joined: 27 Sep 2005
Posts: 131
Location: India
|
| Posted: Wed Mar 12, 2008 12:28 am Post subject: |
|
|
Props to ya PowerGrizz.. 
Looking forward to seeing more of your model and analysis. |
|
| Back to top |
|
|
Critical Beatdown
Joined: 26 Oct 2006
Posts: 11101
|
| Posted: Wed Mar 12, 2008 1:58 am Post subject: |
|
|
Interesting, but I can't see any possible explanation for Fisher having a better score than Parker.
_________________
...as I enter your mind, decorate and paint my sign...LA Lakers will be 2008-09 NBA Champions! |
|
| Back to top |
|
|
Critical Beatdown
Joined: 26 Oct 2006
Posts: 11101
|
| Posted: Wed Mar 12, 2008 2:06 am Post subject: |
|
|
This method for calculating defensive efficiency is very questionable:
| Quote: |
Does your method consider defense?
Yes, it does in some average sense. To evaluate the defensive abilities of player A it goes through each game in the season and looks at the stats of opponents playing in the same position as player A (defined only as perimeter or interior; in the case of forward both are considered). I then compute the difference between opponent stats and their season averages. Player A is credited (or charged) with the proportional fraction of minutes that he was in the game. This method can sometimes be a bit unfair if player A is a very good defender but all of his teammates are bad defenders (or viceversa), because the blame (or credit) is spread equally over all perimeter/inside defenders. Besides this only works if you have a sufficiently large statistical sample. I wouldn't trust it for fewer than 30 games or so. |
PowerGriz FAQ
_________________
...as I enter your mind, decorate and paint my sign...LA Lakers will be 2008-09 NBA Champions! |
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Wed Mar 12, 2008 5:44 am Post subject: |
|
|
| Critical Beatdown wrote: |
| Interesting, but I can't see any possible explanation for Fisher having a better score than Parker. |
They play in different teams so they are not directly comparable. The ranking refers to the importance of each player to his team.
From the FAQ:
| Quote: |
Your method tells us who are the best and the worst players, right?
Nope. The method tells you how much a given player contributes to increasing the probability of wining a game.
|
| Quote: |
If player A of team X has a very similar statline to player B of team Y, shouldn't they have a similar ranking?
Not at all. Remember how the method works. Take the total statline for team X with and without player A. Then take the total statline for team Y with and without player B. Now compare the winning chances for both teams. Only if team X and team Y had the same total stats then the ranking for player A and player B would have to be similar.
|
|
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Wed Mar 12, 2008 5:46 am Post subject: |
|
|
| Critical Beatdown wrote: |
This method for calculating defensive efficiency is very questionable:
|
If you have a better idea I'll be delighted to hear it |
|
| Back to top |
|
|
trodgers
Clublakers Moderator
Joined: 22 Oct 2004
Posts: 22092
Location: Tallahassee
|
| Posted: Wed Mar 12, 2008 2:33 pm Post subject: |
|
|
| PowerGriz wrote: |
| Critical Beatdown wrote: |
This method for calculating defensive efficiency is very questionable:
|
If you have a better idea I'll be delighted to hear it |
Look, if you're just going by the boxscore, your system looks like it works fine. You're not purporting to show that it's the single greatest stat on the planet at determining how good a player is or how much a defender he is. As I see it, it's not your problem.
_________________
PM me to join!
Or play some football:
http://goallineblitz.com/game/signup.pl?ref=324001 |
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Wed Mar 12, 2008 9:44 pm Post subject: |
|
|
| Thanks |
|
| Back to top |
|
|
Critical Beatdown
Joined: 26 Oct 2006
Posts: 11101
|
| Posted: Wed Mar 12, 2008 10:40 pm Post subject: |
|
|
Constructive criticism is what makes theories great instead of dusty, noone needs to get sensitive about it, especially outside actors.
_________________
...as I enter your mind, decorate and paint my sign...LA Lakers will be 2008-09 NBA Champions! |
|
| Back to top |
|
|
Critical Beatdown
Joined: 26 Oct 2006
Posts: 11101
|
| Posted: Wed Mar 12, 2008 10:46 pm Post subject: |
|
|
| PowerGriz wrote: |
| Critical Beatdown wrote: |
| Interesting, but I can't see any possible explanation for Fisher having a better score than Parker. |
They play in different teams so they are not directly comparable. The ranking refers to the importance of each player to his team.
From the FAQ:
| Quote: |
Your method tells us who are the best and the worst players, right?
Nope. The method tells you how much a given player contributes to increasing the probability of wining a game.
|
| Quote: |
If player A of team X has a very similar statline to player B of team Y, shouldn't they have a similar ranking?
Not at all. Remember how the method works. Take the total statline for team X with and without player A. Then take the total statline for team Y with and without player B. Now compare the winning chances for both teams. Only if team X and team Y had the same total stats then the ranking for player A and player B would have to be similar.
|
|
Parker and Ford were their most important players as far as winning the game, and the outcome was in doubt until close to the end. So how is Fisher, the 3rd best performer according to your system, more important than Parker, the 1st more important player on his team, a team that had a legit chance of winning this game?
Methinks it has to do with your defensive schema, which is seriously flawed, and I've already suggested an alternative in the other thread, if you were paying attention.
Any system must pass the eyeball test from seasoned basketball observers, and just about any scout would have not a single conflict in telling you that Parker and Ford were more impactful players in this game than Fisher.
_________________
...as I enter your mind, decorate and paint my sign...LA Lakers will be 2008-09 NBA Champions! |
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Wed Mar 12, 2008 10:56 pm Post subject: |
|
|
| Critical Beatdown wrote: |
| just about any scout would have not a single conflict in telling you that Parker and Ford were more impactful players in this game than Fisher. |
Read back your post and you'll find your answer right there. "more impactful in this game". I'm repeating myself here for the 10th time but here it goes again. The numbers are not an absolute measure, they do not reflect the impact on the game. They reflect how the chance of winning is affected by the presence of that player.
Let's go back and look at the numbers. The Raptors chance of success given by the model was 0.188. If you replace Parker with John Doe that chance goes down by 0.11. That's what the numbers actually mean. You cannot have Parker get a 0.30 ranking because that would imply that the chance of success without him would have been negative. That would be kind of strange. If you want you could normalize the individual rankings to the team chance of winning. I don't want to do that because then the strict meaning of the numbers would be obscured. If you want to do it, here is what you'd get (the last column, but again I'm not sure about the meaning):
| Code: |
Visiting team:raptors. Chance of success: 0.188856
Anthony Parker 0.116315 0.615894
Rasho Nesterovic 0.0991530 0.525018
T.J. Ford 0.0854602 0.452515
Andrea Bargnani 0.0473261 0.250593
Jamario Moon 0.0300798 0.159273
Primoz Brezec -0.00274152 -0.0145164
Jason Kapono -0.00588789 -0.0311766
Joey Graham -0.0111235 -0.0588994
Kris Humphries -0.0569271 -0.301431
Jose Calderon -0.0586222 -0.310407
Carlos Delfino -0.0777810 -0.411853
Home team:lakers. Chance of success: 0.811144
Kobe Bryant 0.334122 0.411915
Lamar Odom 0.283575 0.349599
Derek Fisher 0.125592 0.154833
Luke Walton 0.0545035 0.0671934
Coby Karl 0.0487945 0.0601552
Jordan Farmar 0.0279620 0.0344723
Ronny Turiaf 0.0137018 0.0168920
Pau Gasol 0.0106889 0.0131775
Sasha Vujacic -0.0273361 -0.0337007
Vladimir Radmanovic -0.0925218 -0.114063
|
Again. The docs are there, the FAQ is there, heck, even the source code is there. Please go read some of that because, quite frankly, I'm getting tired of repeating the same things over and over. |
|
| Back to top |
|
|
Critical Beatdown
Joined: 26 Oct 2006
Posts: 11101
|
| Posted: Wed Mar 12, 2008 11:19 pm Post subject: |
|
|
Man, I've read your documents and FAQ, I wish you'd put that feather down dude, because it ain't working, not relevant, no weight.
Just so I'm getting this straight though, you're trying to tell me that if you replace Fisher with an average player, aka John Doe, that the Lakers still come out better in this game than if we replace Parker or Ford with an average player, aka John Doe?
No, I'm sorry dude, I'm not buying it, and neural algorithms can be misused just as much as any other method, especially since you're obviously keeping a tight leash on your "training" method, could very well be a newbie to the neural game as far as I'm concerned, so it sounds great and magical, but as I said you always have to pass the eyeball test, this isn't a statistical seminar, it's a hardcore basketball board.
Now, if you put down the attitude, and stop accusing me of "misunderstanding" all the time, especially since you are not a master of words and this is not a stat geek board, than we might have a more peer-like discussion, because I have shown you no disrespect, but I'm starting to pick up something I don't like from you.
_________________
...as I enter your mind, decorate and paint my sign...LA Lakers will be 2008-09 NBA Champions! |
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Thu Mar 13, 2008 1:06 pm Post subject: |
|
|
Ok, sorry if I've been rude. It's not usually my style. But you are being pretty harsh yourself, there are ways to criticize and still be tactful. You know, treat others the way you'd like to be treated.
You're right that I'm not a "master of words". English is not my first language as have probably guessed already. Which, by the way, means extra effort in writing all that documentation.
So, going back to the point. Yes, if the Lakers had John Doe instead of Derek last night, the winning chances would have been LAL:0.69-TOR:0.31 instead of what it was LAL:0.81-TOR:0.19. Still the Lakers are clearly on top. Without Parker, the chances would have been: LAL:0.94-TOR:0.06. That's pretty much a blowout.
In other words, with or without Fisher, with or without Parker, Lakers win. Without Kobe or Lamar, on the other hand, Lakers probably loose.
As for "keeping a tight leash on the training code", it's true that I haven't released it. But that doesn't mean the network can't be tested. There is a network that I made public. It works pretty much the same as mine except that it's a little less accurate. And you don't need to see the training to test the network. You can (in fact I encourage you to) pick a game, take the boxscore stats and see if the network is capable of telling who won (and with how much certainty). That's all it does really, and it's all that's needed in my method. So if you can verify that it's doing that correctly, then it is working as it's supposed. to. And by the way, there's nothing magic or trascendent about neural networks other than the name. All they are is a glorified fitting method. You take a bunch of data and the network provides a non-linear multi-dimensional function that fits those data. It's pretty much like fitting a 2D surface, except that it's not 2D (in this case it has about 20 dimensions) and the function is numeric, not analytic. |
|
| Back to top |
|
|
kenside34
Joined: 29 Dec 2006
Posts: 748
Location: LakerLand
|
| Posted: Thu Mar 13, 2008 1:58 pm Post subject: Re: The PowerGriz numbers. Last update: TOR@LAL (2008-03-11) |
|
|
Powergriz can your systems be used for EVIL. Say take bodog out of business?
_________________
 |
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Thu Mar 13, 2008 2:34 pm Post subject: Re: The PowerGriz numbers. Last update: TOR@LAL (2008-03-11) |
|
|
| kenside34 wrote: |
| Powergriz can your systems be used for EVIL. Say take bodog out of business? |
You mean, like this? ;)
I should warn you, though. I'm investigating a potential bug right now (that's why I haven't updated the thread since yesterday). That could take a few days, maybe more because I'm terribly busy at work and traveling too much.
Also, I explicitly discourage people from using those numbers for any gambling purposes. For one thing, I don't think they're accurate enough. |
|
| Back to top |
|
|
epak
Joined: 29 Jun 2005
Posts: 1857
|
| Posted: Thu Mar 13, 2008 2:39 pm Post subject: Re: The PowerGriz numbers. Last update: TOR@LAL (2008-03-11) |
|
|
| kenside34 wrote: |
| Powergriz can your systems be used for EVIL. Say take bodog out of business? |
LOL.
Or can your system bring down Hollinger's system?
In the end, I'm just glad that life can't be predicted. :)
_________________
Jerry West:
"You give the same amount of shots to everybody else and they're not making that many, I know it. Players are jealous of greatness. Kobe is a unique talent and a unique person." (after Kobe's 81 points) |
|
| Back to top |
|
|
knm131
Joined: 07 Jul 2007
Posts: 4252
|
| Posted: Thu Mar 13, 2008 3:05 pm Post subject: |
|
|
| PowerGriz wrote: |
| So, going back to the point. Yes, if the Lakers had John Doe instead of Derek last night, the winning chances would have been LAL:0.69-TOR:0.31 instead of what it was LAL:0.81-TOR:0.19. Still the Lakers are clearly on top. Without Parker, the chances would have been: LAL:0.94-TOR:0.06. That's pretty much a blowout. |
Wait, that doesn't make sense. Those ratios determine likelihood of victory, they have no bearing on margin of victory.
The only thing your conclusions can possibly conclude (assuming the math is correct) is the likelihood of winning by 1 or more points. |
|
| Back to top |
|
|
knm131
Joined: 07 Jul 2007
Posts: 4252
|
| Posted: Thu Mar 13, 2008 3:22 pm Post subject: |
|
|
| Critical Beatdown wrote: |
Man, I've read your documents and FAQ, I wish you'd put that feather down dude, because it ain't working, not relevant, no weight.
Just so I'm getting this straight though, you're trying to tell me that if you replace Fisher with an average player, aka John Doe, that the Lakers still come out better in this game than if we replace Parker or Ford with an average player, aka John Doe?
No, I'm sorry dude, I'm not buying it, and neural algorithms can be misused just as much as any other method, especially since you're obviously keeping a tight leash on your "training" method, could very well be a newbie to the neural game as far as I'm concerned, so it sounds great and magical, but as I said you always have to pass the eyeball test, this isn't a statistical seminar, it's a hardcore basketball board.
Now, if you put down the attitude, and stop accusing me of "misunderstanding" all the time, especially since you are not a master of words and this is not a stat geek board, than we might have a more peer-like discussion, because I have shown you no disrespect, but I'm starting to pick up something I don't like from you. |
Critical, I would think PowerGrizz is right here.
If you replace Derek Fisher with Jerry Stackhouse, the Lakers would still come out better in this game than if you replace Parker/Ford with Stackhouse.
In this particular example, it passes the eyeball test.
What I would like to see, to legitimize the product, is some predictions and a W/L record. |
|
| Back to top |
|
|
kenside34
Joined: 29 Dec 2006
Posts: 748
Location: LakerLand
|
| Posted: Thu Mar 13, 2008 3:35 pm Post subject: Re: The PowerGriz numbers. Last update: TOR@LAL (2008-03-11) |
|
|
| PowerGriz wrote: |
| kenside34 wrote: |
| Powergriz can your systems be used for EVIL. Say take bodog out of business? |
You mean, like this? ;)
I should warn you, though. I'm investigating a potential bug right now (that's why I haven't updated the thread since yesterday). That could take a few days, maybe more because I'm terribly busy at work and traveling too much.
Also, I explicitly discourage people from using those numbers for any gambling purposes. For one thing, I don't think they're accurate enough. |
They say the best handicapping % is under 65% . If you hit 70% of your picks you're considered a GOD. We're not looking for a 100 sure win system. Just above 50% ahha
Nice system you have created. Glad you're apart of CL. Enjoy your stay and hook a brother up with the pick of the night . ahhaha
_________________
|
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Thu Mar 13, 2008 5:46 pm Post subject: |
|
|
| knm131 wrote: |
Wait, that doesn't make sense. Those ratios determine likelihood of victory, they have no bearing on margin of victory.
|
Actually they do, a little bit, even though you are correct in that they were not originally meant to. So, when you get .90/.10 that's almost always a blowout by 20 or 30 points. A .55/.45 is a very close game. |
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Thu Mar 13, 2008 5:50 pm Post subject: |
|
|
| knm131 wrote: |
What I would like to see, to legitimize the product, is some predictions and a W/L record. |
I'm afraid it doesn't work that way. I do make predictions but that is a different method, it has nothing to do with the analysis I present here. Look at it this way. I do the analysis by looking at the player stats after the game has been played. When I make predictions I have to come up with a guess of what those stats will be. Usually I just throw in some average stats for each team (corrected with a factor that says how much the other team changes its opponent's stats on average). I think I'm making it sound more confusing than it actually is but the point is that the predictions and the analysis are two different things because the predictions are based on a lot of guesswork that the analysis doesn't require. |
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Thu Mar 13, 2008 5:54 pm Post subject: Re: The PowerGriz numbers. Last update: TOR@LAL (2008-03-11) |
|
|
| kenside34 wrote: |
They say the best handicapping % is under 65% . If you hit 70% of your picks you're considered a GOD. We're not looking for a 100 sure win system. Just above 50% ahha
|
Really??? That seems too easy. I can get 50% by just flipping a coin. My predictions get me to around 80%. You can check them in the link that I posted in my message above. Lately I'm closer to 70% and I think that's related to a pesky bug that I'm trying to squash.
| Quote: |
Nice system you have created. Glad you're apart of CL. Enjoy your stay and hook a brother up with the pick of the night . ahhaha |
Thanks, bro. If you are interested in the predictions I'll resume posting them daily after I fix that bug. Usually I post them between 10am and 1pm (CO time). |
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Fri Mar 14, 2008 10:39 pm Post subject: |
|
|
I haven't had a chance to watch the game yet. It's on my TiVo so hopefully I'll get to watch it tomorrow some time. But I didn't want to keep you guys waiting for the numbers so here they are:
The PowerGriz numbers: LAL@NOH (2008-03-14)
Visiting team:lakers. Chance of success: 0.195541
Kobe Bryant 0.136575
Vladimir Radmanovic 0.102143
Lamar Odom 0.0844374
Luke Walton 0.0670710
Jordan Farmar 0.0407300
Pau Gasol -0.0113459
Sasha Vujacic -0.0348809
Ronny Turiaf -0.0901200
Derek Fisher -0.109579
Home team:hornets. Chance of success: 0.804459
Tyson Chandler 0.240901
Chris Paul 0.168050
Melvin Ely 0.134093
Jannero Pargo 0.0543541
Bonzi Wells 0.0246061
Ryan Bowen 0.00880050
Julian Wright 0.00593712
David West -0.00288028
Morris Peterson -0.0178488
Peja Stojakovic -0.0513593
It always makes me feel a bit insecure when I post the rankings without watching the game but that's of course rather silly. |
|
| Back to top |
|
|
L4L
Clublakers Analyst
Joined: 18 Jun 2007
Posts: 8008
Location: Basking in Glory
|
| Posted: Fri Mar 14, 2008 10:44 pm Post subject: |
|
|
Not a great one to watch... Gasol's injury in the first 3 minutes pretty much sealed the deal though it was close all the way up until the late 3rd.
_________________
Help me name my Puppy! |
|
| Back to top |
|
|
Critical Beatdown
Joined: 26 Oct 2006
Posts: 11101
|
| Posted: Sat Mar 15, 2008 9:12 am Post subject: |
|
|
| knm131 wrote: |
Critical, I would think PowerGrizz is right here.
If you replace Derek Fisher with Jerry Stackhouse, the Lakers would still come out better in this game than if you replace Parker/Ford with Stackhouse.
In this particular example, it passes the eyeball test. |
Um, no it doesn't, and that doesn't make any sense, if you really believe that you don't know basketball bro, and I like you so I"m not saying that with malice.
_________________
...as I enter your mind, decorate and paint my sign...LA Lakers will be 2008-09 NBA Champions! |
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Sun Mar 16, 2008 8:00 pm Post subject: |
|
|
I'm on travel for two weeks with no chance of catching the games. This one was particularly tough to miss so again I'm posting here the numbers blindly:
The PowerGriz numbers: LAL@HOU (2008-03-16)
Visiting team:lakers. Chance of success: 0.177993
Lamar Odom 0.114598
Ronny Turiaf 0.0949383
Sasha Vujacic 0.0510936
Vladimir Radmanovic 0.0408724
Luke Walton 0.0368148
Derek Fisher 0.0184303
Jordan Farmar -0.0188345
Didier Ilunga-Mbenga -0.0661832
Kobe Bryant -0.356740
Home team:rockets. Chance of success: 0.822007
Bobby Jackson 0.327923
Luis Scola 0.131286
Shane Battier 0.106800
Rafer Alston 0.103369
Mike Harris 0.0920030
Luther Head 0.0174899
Dikembe Mutombo 0.0111650
Chuck Hayes -0.0438338
Tracy McGrady -0.151297
Last edited by PowerGriz on Sun Mar 16, 2008 8:12 pm; edited 2 times in total |
|
| Back to top |
|
|
PowerGriz
Joined: 09 Mar 2008
Posts: 181
|
| Posted: Sun Mar 16, 2008 8:02 pm Post subject: |
|
|
What the heck happened?!?!?
Bobby Jackson?!? Are you f.. kidding me??? We had him in the Grizzlies and I was sooo happy to see that ballhog go. Now it turns out he was one of the best players in a championship caliber game?
And T-Mac and Kobe both having a negative impact?  Is that even possible???
By the way, I've noticed several times with different teams and different players that only the superstars can have very large negative impact (say below -0.20 or so). This is what I call the "superstar off-night effect". Basically, when any human player is having a bad game (and every player, be human or super-human has one every now and then) his negative impact is mitigated by his own judgment (ok, let me not force this shot because clearly tonight is not my night; if I can't get a good look I'll find a teammate with a better one) and the coach's (ok, I'm going to sit this guy for a few minutes to see if he gets his game back). But that doesn't happen with superstars. They will still take their 30 or 40 shots a game because that's their job. And of course, no coach is going to bench the LeBrons, Kobes, Tmacs of the league because they've missed a few shots. Anyway, the explanation is just a theory but I'm pretty sure that basically only superstars can reach very negative numbers. |
|
| 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
|
|
Basketball 
Football 
Baseball
FAQ
Search
Memberlist
Register
Profile
Log in to check your private messages
Log in
