- My Forums
- Tiger Rant
- LSU Recruiting
- SEC Rant
- Saints Talk
- Pelicans Talk
- More Sports Board
- Fantasy Sports
- Golf Board
- Soccer Board
- O-T Lounge
- Tech Board
- Home/Garden Board
- Outdoor Board
- Health/Fitness Board
- Movie/TV Board
- Book Board
- Music Board
- Political Talk
- Money Talk
- Fark Board
- Gaming Board
- Travel Board
- Food/Drink Board
- Ticket Exchange
- TD Help Board
Customize My Forums- View All Forums
- Show Left Links
- Topic Sort Options
- Trending Topics
- Recent Topics
- Active Topics
Started By
Message
re: LSU Basketball: The SoS Rating (Score on Stats)
Posted on 11/28/12 at 9:44 am to Dooshay
Posted on 11/28/12 at 9:44 am to Dooshay
quote:
Some of that can be attributed to luck and some of it can be attributed to offensive ability (which is hard to distinguish statistically) whereas the other statistics are primarily off/def ability, hence my suggestion.
But i guess if you subtract it out like you said it normalizes it
I certainly appreciate the quesitons. I am not too proud to admit I did something wrong if the that is the case.
I want this to be the best it can. I am thinking of breaking out defensive and offensive rebounds because offensive rebounds lead to more points on average than defensive. I just would have to put in some extra work.
I am also thinking about adding fouls in as a negative variable.
Posted on 11/28/12 at 10:01 am to SouthOfSouth
Yeah I'm not a basketball person, but I was just trying to give some positive critiques mathematically.
It seems that having the two coefficients for assists and pts with a wide margin would lower the regression (accuracy) since one is dependent on the other (it's not an assist unless a point is scored, so that person making the basket is pretty important). It's a differential equation, not linear. So taking it out is easier than breaking it down. If the goal is "should be rewarded with more playing time" then the scoreboard doesn't matter and points being scored are included in the fact that an assist happened. Who those points (statistic, not literal) are assigned to is probably attributed more to position than anything else. That's kind of taken care of since there is a finite number of each.
Whether points made it on the board or not depends on if the player made the basket (a whole other statistic). Free throws and break-a-ways could be assumed that "a player with above average ability" (ie any college basketball player warranted playing time) would make more often than not.
I'm not saying change anything up as it appears to be a good reflection as it, I was just giving some mathematical criticism (hopefully positive) on it.
It seems that having the two coefficients for assists and pts with a wide margin would lower the regression (accuracy) since one is dependent on the other (it's not an assist unless a point is scored, so that person making the basket is pretty important). It's a differential equation, not linear. So taking it out is easier than breaking it down. If the goal is "should be rewarded with more playing time" then the scoreboard doesn't matter and points being scored are included in the fact that an assist happened. Who those points (statistic, not literal) are assigned to is probably attributed more to position than anything else. That's kind of taken care of since there is a finite number of each.
Whether points made it on the board or not depends on if the player made the basket (a whole other statistic). Free throws and break-a-ways could be assumed that "a player with above average ability" (ie any college basketball player warranted playing time) would make more often than not.
I'm not saying change anything up as it appears to be a good reflection as it, I was just giving some mathematical criticism (hopefully positive) on it.
This post was edited on 11/28/12 at 10:07 am
Posted on 11/28/12 at 10:05 am to SouthOfSouth
quote:
I want this to be the best it can. I am thinking of breaking out defensive and offensive rebounds because offensive rebounds lead to more points on average than defensive. I just would have to put in some extra work.
I am also thinking about adding fouls in as a negative variable.
That would certainly make it more accurate, but each variable adds an increasing level of complexity to it.
If you made two equations with positives (points, assists, blocks, etc) and and another with negatives (tos, fouls, missed baskets, etc) and added them together, it would also be a good reflections of the player's overall contribution to the team, rather than just who is out there making an impact, but those are really two different things to look at.
Popular
Back to top
Follow TigerDroppings for LSU Football News