Tempo-Free Stats

#1
This season, I plan to track tempo-free stats (see disclaimer below!) for all players in the Big Ten to see how Illinois compares. It's obviously a bit early to look at the results because of small sample size and variance in level of competition. But I'll post some anyway. ;)



The green (best) to red (worst) shading indicates the stat's rank among all players in the Big Ten (regardless of minutes played). Yes, Mike LaTulip is currently leading the Big Ten in points-per-normalized-possession. :thumb:

The numbers don't indicate what one should expect a player to achieve in a game, but larger numbers are still better (or worse, in the case of fouls and turnovers). The column at the end is Hollinger's Game Score, an attempt at an overall rating (which of course doesn't take defense into account except blocks and steals; it's biased toward players that put up good stats across the board).

Based on Game Score, here are the "top" 25 players in the Big Ten, through November 23, that have averaged 10 or more minutes. Unlike the table above, the green to red shading here indicates a stat's rank among players in the Big Ten that have averaged 10 or more minutes.



As an example of small sample size and level of competition, we'll see if Shannon Scott can continue to hand out assists at a rate twice that of the next best player. :D


DISCLAIMER!

I have no idea how others might track tempo-free stats, or if there is some standard way of doing so. This is the formula I'm using for it:

(NP * (STAT / ((ATP / (TTM / (TGP * 5))) * PTM)))

NP = Normalized possessions. Stats per possession are multiplied by this to give the numbers seen in the table. It's currently set at 66. So the numbers indicate how a player's per-possession stats would translate to a 66 possession game in which they played 40 minutes.
STAT = The stat in question.
ATP = Average team possessions.
TTM = Total team minutes.
TGP = Team games played.
PTM = Player total minutes.

In short: I'm determining the average number of possessions per minute for a particular team, calculating the stats per possession for each player on that team, then multiplying the per possession stats of all players on all teams by 66 to normalize.

This is obviously not perfect. It's not tracking how each player performed on each actual possession they participated in. It's tracking each player's cumulative performance over the team's average possessions.

All data is from bigten.org except average team possessions, which is from teamrankings.com. As a result of the former, these stats include non-D1 games.
 
#3
That's pretty interesting to me. Thanks for posting.
It will be a lot more significant in about 2 weeks, when everyone has played someone of consequence. Thanks again
 
#5
i love stuff like this. thanks for taking the time to create and post
 
#7
CU Expat
Chicago
This is awesome Urbanite. I know we did something similar last season for the Illini squad, but having a ranking of big ten players is great.
 
#8
The Chief Lives
Chicago, IL
This is fantastic! Definitely look forward to seeing this further into the season.

Do you think you could add a column for assist/turnover ratio? I realize this is easy to calculate just from looking at it, but it would be much easier to compare across all players if you there was a column with your color coding.

By the way, I like that you went with slightly more subtle colors rather than the in-your-face bright colors that people tend to use for things like this. :thumb:
 
#10
Do you think you could add a column for assist/turnover ratio?
Sure. I've added it and will include it going forward. For now, here's an updated table for Illinois:



I won't update the other table, but here are the top 10 (11) for A:T in the Big Ten:
9.0 - Starks (ILL)
6.7 - Trice (MSU)
6.5 - Tate (ILL)
6.0 - Albrecht (MICH)
6.0 - LeVert (MICH)
5.9 - Scott (OSU)
5.0 - Robinson (IND)
4.5 - Koenig (WIS)
4.5 - Pack (MD)
4.0 - Hartman (IND)
4.0 - Thompson (PUR)

I hope you update this throughout the year.
I was planning to update it every Monday. Maybe that's too often, though, I don't know.
 
#12
Sure. I've added it and will include it going forward. For now, here's an updated table for Illinois:



I won't update the other table, but here are the top 10 (11) for A:T in the Big Ten:
9.0 - Starks (ILL)
6.7 - Trice (MSU)
6.5 - Tate (ILL)
6.0 - Albrecht (MICH)
6.0 - LeVert (MICH)
5.9 - Scott (OSU)
5.0 - Robinson (IND)
4.5 - Koenig (WIS)
4.5 - Pack (MD)
4.0 - Hartman (IND)
4.0 - Thompson (PUR)


I was planning to update it every Monday. Maybe that's too often, though, I don't know.

haha, it might be too often, but I would enjoy it, so if you don't mind . . .
 
#14
Just doing a quick scan of the table above and I must say, the color coding is very helpful. You can almost instantly see that we are very good at scoring (lot of green in the first few columns) and very weak at rebounding and blocking shots. Even if you don't get into the details, the ability to figure that out in 2-3 seconds is helpful.

Sort of explains why USA Today uses charts and pictures, not words, whenever they can.
 
#15
Orlando, FL
Would love to see it weekly on Monday. I will look forward to it.
 
#16
Tampa, FL
Sure. I've added it and will include it going forward. For now, here's an updated table for Illinois:

Points per shot is also a good stat for scoring efficiency.

Good job! :thumb:
 
#18
Typically when people use tempo-free stats they change the denominator based on the particular stat they are using. Points (and points against) per possession is fantastic on the team level, for example, but not as useful for individuals. As mentioned above, points per shot is a pretty effective measure of scoring efficiency.

Rebounds are usually given as a percentage of opportunities. OR% is rebounds per missed field goals on offense, DR% per missed field goals by the other team. Blocks are per opponent FGA, assists per own FGM, etc...

Most of the quantities in the denominator will be estimates unless you use play-by-play data, but they will get better as the season goes along.
 
#21
Captain 'Paign
Phoenix, AZ
Ray is at least average or above in every single statistical category (no red is any one). The only other Big Ten player who can say that is Kaminsky. That's pretty good company.
 
#24
Interesting that Wisconsin has 4 players in top 25...OSU and IU both have 3. It will be interesting to see if there is any correlation on # players in top 25 on this list and corresponding B10 team success.