ISU Individual Player OFF vs. DEF ratings

[EarthIsMan]

Offensive and defensive ratings (points per possession x 100) for each individual player as of 1/15/2019. For reference the farther right on the x-axis and lower on the y-axis, the better.

These are using estimated point per possession for each player while they are on the court. The size of the dot is proportional to their % usage each possession. This is a decent way for determining their defensive and offensive contribution or efficiency while they are on the court.

I will just point out that Haliburton has one of highest offensive ratings in the country right now. The other thing I found interesting is Wigginton's rating, but in the context of returning from an injury- it is too early to draw sweeping conclusions.

Lastly, using this to compare players has its flaws so do not interpret this as simplistically as X player should be playing over Y player. A player's usage and how much they play within different lineups is important context when interpreting these values. Also realize, for some players the sample size is quite small.

p.s.
Offensive ratings: you want the largest rating
Defensive ratings: you want the smallest rating

 

[CloneFan4]

This helps solidify what I've thought since December: Zoran needs more minutes.

Don't @ me

 

[Tre4ISU]

Yeah, Lewis needs to be on the floor more.

 

[EarthIsMan]

Yeah, Lewis needs to be on the floor more.

Defensively, Lewis has the worst defensive rating on the team- which is likely why he is not seeing the floor.

 

[CascadeClone]

My first thought is that THT and Wigg are roughly the same player. Volume shooters, not always efficient, but we need someone to be super aggressive offensively to open stuff up. And if either could get on track it would be a HUGE positive for this team on offense when they get stagnant. More than once I've thought that you only want one of those two on the court at a time, but that you want one of them on the court at ALL times.

Second thought is Halliburton needs more involvement on offense, but having the ball in his hands more would likely bring that rating down a bit too.

 

[EarthIsMan]

My first thought is that THT and Wigg are roughly the same player. Volume shooters, not always efficient, but we need someone to be super aggressive offensively to open stuff up. And if either could get on track it would be a HUGE positive for this team on offense when they get stagnant. More than once I've thought that you only want one of those two on the court at a time, but that you want one of them on the court at ALL times.

Second thought is Halliburton needs more involvement on offense, but having the ball in his hands more would likely bring that rating down a bit too.

Agree, generally as usage increases, offensive efficiency decreases. Halliburton is not there yet scoring in isolation, but a lot of great things happen he has the ball in his hands.

 

[Tre4ISU]

Whoops. Didn't pay attention to how the graph was set up. I still hold the same opinion.

 

[EarthIsMan]

@JStanz51 Any Synergy nuggets you can provide?

 

[CascadeClone]

One other thought. I'd get Conditt on the floor more (likely anyway with Solo done) . Best defender, rim protector. And with all the other guys who can create he can be around to rebound or take out the trash. He seems to be more natural with that than Lard.

 

[Cyclonepride]

This helps solidify what I've thought since December: Zoran needs more minutes.

Don't @ me

@

 

[FinalFourCy]

Defensively, Lewis has the worst defensive rating on the team- which is likely why he is not seeing the floor.

Whoops. Didn't pay attention to how the graph was set up. I still hold the same opinion.

It’s distorted for guys like Talley and Lewis that get sporadic and limited quality minutes (which in itself doesn’t lend to performing well), combined with a larger sample of scrub time pickup ball minutes. If you’re getting steals or blocks like Conditt in walkon time, it’s great though. Lewis isn’t a great defender, but he’s closer to others than that rating.

It’s better than nothing, but individual defensive ratings are limited. It rewards making box score plays more than doing your job.

 

[DurangoCy]

Initial reaction - No way is Terrence Lewis our best defender. I just wanted him play 2 minutes vs. KSU and completely forget to watch the guy he was guarding, who simply ran to the other corner and nailed a wide open three.

Read the explanation and Secondary Reaction - Yep, that makes sense, TL is by far our worst defender.

Third Reation - My GAWD I love Haliburton.

Fourth Reaction - Wigginton and THT have been pretty underwhelming thus far, which makes me sad.

 

[Deleted member 8507]

The three freshman (Condit, THT and Halliburton) have the best defensive ratings. Interesting.

 

[FinalFourCy]

“The core of the Defensive Rating calculation is the concept of the individual Defensive Stop. Stops take into account the instances of a player ending an opposing possession that are tracked in the boxscore (blocks, steals, and defensive rebounds), in addition to an estimate for the number of forced turnovers and forced misses by the player which aren't captured by steals and blocks.”

The formula for Stops is:

• Stops = Stops1 + Stops2

where:

• Stops1 = STL + BLK * FMwt * (1 - 1.07 * DOR%) + DRB * (1 - FMwt)
• FMwt = (DFG% * (1 - DOR%)) / (DFG% * (1 - DOR%) + (1 - DFG%) * DOR%)
• DOR% = Opponent_ORB / (Opponent_ORB + Team_DRB)
• DFG% = Opponent_FGM / Opponent_FGA
• Stops2 = (((Opponent_FGA - Opponent_FGM - Team_BLK) / Team_MP) * FMwt * (1 - 1.07 * DOR%) + ((Opponent_TOV - Team_STL) / Team_MP)) * MP + (PF / Team_PF) * 0.4 * Opponent_FTA * (1 - (Opponent_FTM / Opponent_FTA))^2

 

[Sigmapolis]

View attachment 61635
Offensive and defensive ratings (points per possession x 100) for each individual player as of 1/15/2019. For reference the farther right on the x-axis and lower on the y-axis, the better.

These are using estimated point per possession for each player while they are on the court. The size of the dot is proportional to their % usage each possession. This is a decent way for determining their defensive and offensive contribution or efficiency while they are on the court.

I will just point out that Haliburton has one of highest offensive ratings in the country right now. The other thing I found interesting is Wigginton's rating, but in the context of returning from an injury- it is too early to draw sweeping conclusions.

Lastly, using this to compare players has its flaws so do not interpret this as simplistically as X player should be playing over Y player. A player's usage and how much they play within different lineups is important context when interpreting these values. Also realize, for some players the sample size is quite small.

p.s.
Offensive ratings: you want the largest rating
Defensive ratings: you want the smallest rating

If I may make a few small requests (with data that you should mostly already have from the above) that would help in conceptualized this for most people...

Graph of...

Net rating per 100 (which is just offensive rating minus defensive rating)...?

Graph of either possessions played and/or minutes played?

This would help one see who is good at both (or at least nets you out the highest) and who has a sample size to give you a reasonable view of their capabilities.

I do not think these statistics adjust for your opponent, however, which is troublesome, given that Lewis mostly played against cupcakes and Wigginton has mostly played against the Big 12, for instance, skewing what the numbers would really tell you.

 

[mwwbbfan]

It’s distorted for guys like Talley and Lewis that get sporadic and limited quality minutes (which in itself doesn’t lend to performing well), combined with a larger sample of scrub time pickup ball minutes. If you’re getting steals or blocks like Conditt in walkon time, it’s great though. Lewis isn’t a great defender, but he’s closer to others than that rating.

It’s better than nothing, but individual defensive ratings are limited. It rewards making box score plays more than doing your job.

I would be interested to see Conditt/Lewis vs Steyer's stats because we know that all of his time is all end of game possessions, that may give an indication of where they are at in competitive situations.

 

[CloneFan4]

@

Can we please get a sad react button?

 

[EarthIsMan]

If I may make a few small requests (with data that you should mostly already have from the above) that would help in conceptualized this for most people...

Graph of...

Net rating per 100 (which is just offensive rating minus defensive rating)...?

Graph of either possessions played and/or minutes played?

This would help one see who is good at both (or at least nets you out the highest) and who has a sample size to give you a reasonable view of their capabilities.

I do not think these statistics adjust for your opponent, however, which is troublesome, given that Lewis mostly played against cupcakes and Wigginton has mostly played against the Big 12, for instance, skewing what the numbers would really tell you.

Usage rate vs. Net Rating (bubbles are proportional to Player Efficiency Rating)

This is to generally show (what is widely known) that usage is negatively correlated with efficiency. So if you want to see Haliburton or Talley get more usage, their efficiency will drop too (which is expected and completely fine). FWIW, Haliburton should probably get more usage during possessions.

Any player that occupies the upper right part of the graph is a statistical unicorn (e.g. Lard & Jacobson) meaning they are maintaining efficiency while having high usage.

Possessions or minutes played would probably not show much other than the sample size is quite small for quite a few players which I explained in the original post.

Totally agree about adjusting data for opponent, but unless you have a Synergy account it is difficult to impossible to get possession-level data in order to adjust for opponents and your team lineups. This website does provide Regularized Adjusted Plus Minus (RAPM) data for NCAA MBB from 2010 to 2018, but nothing for the current season.

 

[Sigmapolis]

Usage rate vs. Net Rating (bubbles are proportional to Player Efficiency Rating)

This is to generally show (what is widely known) that usage is negatively correlated with efficiency. So if you want to see Haliburton or Talley get more usage, their efficiency will drop too (which is expected and completely fine). FWIW, Haliburton should probably get more usage during possessions.

Any player that occupies the upper right part of the graph is a statistical unicorn (i.e. Lard & Jacobson) meaning they are maintaining efficiency while having high usage.
View attachment 61640
Possessions or minutes played would probably not show much other than the sample size is quite small for quite a few players which I explained in the original post.

Totally agree about adjusting data for opponent, but unless you have a Synergy account it is difficult to impossible to get possession-level data so that you adjust for opponents and your team lineups. This website does provide Regularized Adjusted Plus Minus (RAPM) data for NCAA MBB from 2010 to 2018, but nothing for the current season.

Thank you. Some observations on this...

-- Zion is scary... Texas Tech is a monster this year... it would be ironic if the Crimson Cowboys were the ones to finally dethrone Kansas considering what their program has been through some bad times, but they are really good this year

I remember some article called them the "Redneck UVA." It is pretty apt.

-- theoretically, assuming Tyrese keeps shooting 50%+ or something ridiculous from three, his efficiency can remain so high... but we all know he is due for some regression towards the mean there, especially if they start feeding him the ball more

-- even if we feel Cameron has been a shell of himself last season compared to this year, the numbers that he produces -- blocks and rebounds, especially offensive rebounds -- are inherently efficient because he generates possessions for himself

We do need him back, focused, and healthy to hit our peak. Even more than we need Wigginton to hit the heights he had late last season.

-- Nick is qualitatively different from our three high-usage, relatively low-efficiency guards in the bottom-right corner... Marial is the best of the three, surprisingly for his offense and not for his defense, though he faced the weakest schedule

-- THT and Wigginton are hard to distinguish between each other because of that same strength of schedule issue that I brought up earlier

-- not to pick on Ja Morant, but he is doing this in the Ohio Valley Conference, which goes back to my point about adjusting for quality of opponents

 

[ESciGuy]

Usage rate vs. Net Rating (bubbles are proportional to Player Efficiency Rating)

This is to generally show (what is widely known) that usage is negatively correlated with efficiency. So if you want to see Haliburton or Talley get more usage, their efficiency will drop too (which is expected and completely fine). FWIW, Haliburton should probably get more usage during possessions.

Any player that occupies the upper right part of the graph is a statistical unicorn (i.e. Lard & Jacobson) meaning they are maintaining efficiency while having high usage.
View attachment 61640
Possessions or minutes played would probably not show much other than the sample size is quite small for quite a few players which I explained in the original post.

Totally agree about adjusting data for opponent, but unless you have a Synergy account it is difficult to impossible to get possession-level data in order to adjust for opponents and your team lineups. This website does provide Regularized Adjusted Plus Minus (RAPM) data for NCAA MBB from 2010 to 2018, but nothing for the current season.

Thanks for the graphs. I like data! I think you meant e.g. instead of i.e., though. I had a colleague who would actually say "eye-ee" in a presentation, instead of "that is". Of course, whenever he did, he actually meant, e.g. (i.e., for example).