re: LSU Basketball: The SoS Rating (Score on Stats)

quote:

---

SoS, this is my first chance to have a look at your formula. One question I have is how you control for what I call the Iverson effect.

Iverson effect is the relationship between points scored and shots taken to achieve that point total. Thirty points are great, but if you are taking 30 shots to get there, not so much.

Is their something in the formula that I am missing that accounts for scoring efficiency, or is that not what you are going for?

Thanks.

---

I really appreciate everyone's input. People's remarks have pushed me to think even more about this and I have made some major adjustments. I made a correction for the "Iverson effect" which will be shown below as well as an adjustment to rebounds to use both offensive and defensive rebounds.

Here are the formula's:






Change in variables:


the variable reb was changed into two separate variables; Dreb and Oreb. Oreb receives a 2.3 value (a +.5 from reb) and Dreb receives a 1.3 value (a -.5 from reb). This will give those who get offensive rebounds a better SoS score due to an offensive rebounds ability to help a team score.

The shooting change looks much more significant. I did not want to penalize people for taking more shots because some teams require their star players to shoot more often; however, I did want to penalize players from shooting a poor percentage.

The numbers inside the () is the average of the last 5 years average in each of the categories:

2 point shots
3 point shots
Free throws

Oshot is a variable equal to the number of 3 point shots taken

Ishot is a variable equal to the number of 2 point shots taken

Fshot is a variable equal to the number of Free throws taken

out is a variable equal to the percentage a player shoots 3 pointers

in is a variable equal to the percentage a player shoots 2 pointers

ft is a variable equal to the percentage a player shoots free throws

So, for each 3 point shots, 2 point shots, and free throws the number of shots taken (in that category) is multiplied by the difference in % made and NCAA average % and then multiplied by the worth of the shot (3, 2, or 1).

This hasn't changed anyone very much for LSU, probably because noone has had a game where they use the "Iverson effect" to increase their SoS.

The new SoS ratings are on the left while the right graph represents the difference between the new SoS rating and the old rating.

The biggest changes are in Shavon's 2nd and third game where he shot over 60% from 2 point and 67% from 3 point. It gave him a boost of over 5 SoS.

Anthony Hickey, however, saw a decrease in SoS rating due to his poor shooting numbers to start the season.



The change for rebounds hasn't had as much of an effect on LSU due to the high percentage of offensive rebounds we have gotten (3rd in nation). But it will have an effect on players who do not get many offensive rebounds.

Any other input is gladly appreciated. With the changes in the formula, it does make it a bit harder to track the SoS of other teams, but I will work on finding a source that has all the information required. I hope people don't forget about the basketball squad during this 11 day break.