BuzzerBeater Forums

BB Global (English) > The Quest for the Formulas

The Quest for the Formulas (thread closed)

Set priority
Show messages by
This Post:
00
46657.110 in reply to 46657.109
Date: 3/23/2009 10:23:44 AM
Overall Posts Rated:
409409
Ok. Thanks for taking the time to answer.

This Post:
00
46657.111 in reply to 46657.110
Date: 4/4/2009 10:02:51 AM
Overall Posts Rated:
22
I was wondering exactly what you do in excel with the formulas you have been providing. Could anyone help me out? I want to calculate the salaries of my roster on my own. So what do i do with this information if i want to calculate the salary of one of my SG's?

___JS___JR___OD___HA___ DR___PA___IS___ID___RB___BL___const
SG 1,116 1,149 1,115 1,006 1,013 1,000 1,000 1,004 1,061 1,005 307

From: LA-Niko

This Post:
00
46657.112 in reply to 46657.1
Date: 4/12/2009 1:41:15 AM
Overall Posts Rated:
2121
Not sure how many people noticed but there is a new formula calculation on BBStats that I thought I better bring to people's attention.

I have been working with regression analysis techniques (a perceptron with backpropagation) to get a successful classifier for BuzzerBeater players. I have got results of my testing displayed on (http://www.bb-stats.org/weightings.php5)

You can see the weightings that each class of player has for each skill.
This could be the weightings that the game engine uses to work out how well a player plays.

Alternatively it could be only be the weightings used by the classifier that BuzzerBeater has for best position and that it all...

So keep in mind this could be very useful and it could also be useless.

;)

Creator of (http://www.buzzerbeaterstats.com) and (http://www.buzzerbeaternews.com/) -- Ex GM of Australia -- Division 1 winner of Italy Season 1 then moved team to Australia after the country was created by the BBs. Australian team manager for 2 seasons. Won various tournaments and division 1 titles in the following seasons.
From: LA-Niko

This Post:
00
46657.115 in reply to 46657.114
Date: 4/12/2009 4:50:10 AM
Overall Posts Rated:
2121
The regression I used is of the "generalised linear model" family.
More info here: (http://www.dmresearch.net/Matlab/2009/0327/124744.html)

The weightings below the regression are average weightings from each best position.
They are the mean point for each skill given the 'best position'.

Creator of (http://www.buzzerbeaterstats.com) and (http://www.buzzerbeaternews.com/) -- Ex GM of Australia -- Division 1 winner of Italy Season 1 then moved team to Australia after the country was created by the BBs. Australian team manager for 2 seasons. Won various tournaments and division 1 titles in the following seasons.
This Post:
00
46657.117 in reply to 46657.112
Date: 4/13/2009 10:21:35 AM
Overall Posts Rated:
404404
there are the ideal weigthings that a player could have for the game engine?Have you consdiered small forwards as perimketrale players,looking to their inside shot

This Post:
00
46657.118 in reply to 46657.117
Date: 4/13/2009 7:10:14 PM
Overall Posts Rated:
2121
I don't consider anything.

My results are purely based on how BuzzerBeater classifies a player into a Best Position

Creator of (http://www.buzzerbeaterstats.com) and (http://www.buzzerbeaternews.com/) -- Ex GM of Australia -- Division 1 winner of Italy Season 1 then moved team to Australia after the country was created by the BBs. Australian team manager for 2 seasons. Won various tournaments and division 1 titles in the following seasons.
This Post:
00
46657.119 in reply to 46657.116
Date: 4/18/2009 11:15:47 AM
Overall Posts Rated:
22
I was wondering exactly what you do in excel with the formulas you have been providing. Could anyone help me out? I want to calculate the salaries of my roster on my own. So what do i do with this information if i want to calculate the salary of one of my SG's?

___JS___JR___OD___HA___ DR___PA___IS___ID___RB___BL___const
SG 1,116 1,149 1,115 1,006 1,013 1,000 1,000 1,004 1,061 1,005 307

Advertisement