STATS: AP Top 25 Team Efficiency by Role

[maccca3232]

Here is breakdown by team versus opponents in Level 1 and 2 of RPI sorted by Efficiency For

Level	Home	Neutral	Away
1​	30​	50​	75​
2​	75​	100​	135​
3​	160​	200​	240​
4​	351​	351​	351​

Where:
Starting Backcourt (Guard in Blue). Lead by Arizona at 49
Starting Frontcout (Forward in Red). Lead by UConn at 53.6
Bench (Reserves in Green). Lead by Stanford at 32.4
Misc (Team in Purple).
Overall Lead by UConn at 107.5

The data based on games until end of Week 7 (i.e. 22 Dec 19) and RPI as at 21 Dec 2019

 

[maccca3232]

Here is breakdown by team versus opponents in Level 1 and 2 of RPI sorted by Efficiency Differential for AAC.

Note: I edited post to replace the Efficiency For data with Efficiency Differential data

 

[WestCoast]

Does this mean that a UConn vs. Oregon State game would be a toss up? More fuel for further discussion for supporters for both teams.

 

[maccca3232]

Here is breakdown by team versus opponents in Level 1 and 2 of RPI sorted by Efficiency Against

Level	Home	Neutral	Away
1	30	50	75
2	75	100	135
3	160	200	240
4	351	351	351

Where:
Starting Backcourt (Guard in Blue). Lead by Arizona at 3.0
Starting Frontcout (Forward in Red). Lead by Stanford at 17.9
Bench (Reserves in Green). Lead by NC State at 5.0
Misc (Team in Purple).
Overall Lead by Gonzaga at 48.0

 

[eebmg]

@maccca3232

Very interesting.

To be sure what you are doing;

I am assuming You are summing the efficiency scores based on

(PTS + REB + AST + STL + BLK − Missed FG − Missed FT - TO) / GP

of each player in the positions you are focusing on

Question

Is your procedure taking into account that starters from game to game may change?

 

[maccca3232]

Here is breakdown by team versus opponents in Level 1 and 2 of RPI sorted by Efficiency Differential

Level	Home	Neutral	Away
1	30	50	75
2	75	100	135
3	160	200	240
4	351	351	351

Where:
Starting Backcourt (Guard in Blue). Lead by Arizona at 46.0
Starting Frontcout (Forward in Red). Lead by UConn at 33.7
Bench (Reserves in Green). Lead by Florida St. at 16.2
Misc (Team in Purple).
Overall Lead by UConn at 56.7

 

[maccca3232]

@maccca3232
Question

Is your procedure taking into account that starters from game to game may change?

Yes, stats are based on actual starting lineups.

If a player is out of the lineup the stats include their replacement (e.g. Bent when Dangerfield was out for 2 games).

 

[eebmg]

Yes, stats are based on actual starting lineups.

If a player is out of the lineup (e.g. Dangerfield for 2 games) the stats include the replacement (i.e. Bent) for those games.

Fantastic analysis. Watch out 538.

 

[eebmg]

Yes, stats are based on actual starting lineups.

If a player is out of the lineup the stats include their replacement (e.g. Bent when Dangerfield was out for 2 games).

Suggestion

There are probably some teams on the top 25 list that do not have that many Level 1 and Level 2 games. Maybe another column identifying the number of suitable games each team played.

 

[maccca3232]

Suggestion

There are probably some teams on the top 25 list that do not have that many Level 1 and Level 2 games. Maybe another column identifying the number of suitable games each team played.

Agreed.

I will edit the posts tomorrow along with adding Misc column (i.e. non player stats e.g. team rebounds). This is why the team totals don't currently match the sum of G+F+B in Against and Differential graphs.

 

[maccca3232]

Does this mean that a UConn vs. Oregon State game would be a toss up? More fuel for further discussion for supporters for both teams.

After i posted the ACC graph (where UCF was last) I realized much like scoring I needed to not only show For (team vs opposition) but Against (opposition vs team) and finally Differential (team vs opposition less opposition vs team).

Example: DePaul
For 89.8 (6th)
Against 92.1 (25th i.e. last)
Differential -2.3 (21st)

Oregon St. is still 2nd behind UConn in Differential, even though it isn't as close, no doubt it would be a great match-up.

 

[maccca3232]

Posts have been edited.

For
Against
Differential

I have added:
GP column
Team column (those stats not allocated to a player e.g. team rebounds)

 

[maccca3232]

I have made available a google spreadsheet where you can changes the options yourself (via drop-down for those cells in yellow)

Level:

• 1
• 2
• 3
• 4
• 1-2
• 1-3
• T (a.k.a. Total i.e. 1-4 or All)

Efficiency:

• F (For)
• A (Against)
• D (Differential)

There is also a 'Conference' tab where you in addition to the above options you can specify a 'Conference'.

 

[willtalk]

Does this mean that a UConn vs. Oregon State game would be a toss up? More fuel for further discussion for supporters for both teams.

Actually this sort of validates my response to those that claimed it was out of line to even consider any other team at this time deserving the #1 rating. I responded that either Oregon St or Uconn, could reasonably be rated #1 at this point in time. It all would depend on how much value was placed on the variable and different metrics.