scott
02-19-2025, 01:29 AM
Hi all,
Based on a conversation I was having with a Vassell stan on the Spurs subReddit, I ended up doing a little bit of a Tuesday night stat session. I'll eventually expand this to look at some other Spurs, but essentially I made a comment that Vassell is inconsistent and someone challenged me on that. Seems like people here appreciate some stat drops, so I figured I'd share it here. Below is a copy/paste from what I posted on Reddit, with some additional references.
I'll look to run some similar stats for our other players, and their levels of consistency. I can also go back and do this for historical players if you are curious, so long as their GameLogs are available in BBRef (it at least goes back to Tim Duncan's rookie season). One thing I won't do is go calculate every player in the NBA, but I'm happy to pull the statistics for specifically requested players. Just tell me the player and the year you are curious in and I'll get to it when I can (you have to tell me both. I won't go through the effort to look at multi-year periods)
The Reddit post is below.
First I wanted to establish a baseline of what constitutes consistency. To that I just googled "most consistent NBA players" and thankfully it brought me to one of my favorite analytic sites, CraftedNBA: https://craftednba.com/player-traits/consistency
I like that they use GameScore to measure consistency, because I've done the same in other data projects I've done to where I was looking at how whether or not certain players having a good game or a bad game was correlated with winning and losing (spoiler alert: when Wemby has a good game, we are more likely to win... when he has a bad game, we are far more likely to lose... you probably could have guessed this). NOTE NOT FROM THE REDDIT POST: here is my post on how GameScore correlates to wins/losses. https://www.spurstalk.com/forums/showthread.php?t=304748. I like the way CraftedNBA has scaled GameScore to each individual player in their methodology and I think I'll go back to this post and update it with these breakouts rather than the hard and fast cut offs. That will come at a later date.
I randomly picked four guys, kind of scattered throughout the list. I ended up with KD, Fox, Desmond Bane and Jalen Williams. You can go to the link I provided and check out everyone else, I just wanted to include 4 random guys in a summary table for comparison purposes. There is nothing special about the 4 guys I randomly chose.
I calculated the same consistency metric for Devin as the site uses (which I like, especially in that it is scaled for each individual player. A great game for Vassell is not the same as a great game for KD. You can look at the site, and it explains what is a great game (basically, greater than 125% of the player's average GameScore, a Good Game is +/- 25% of the players Average, a Fair game is between 50 and 75% of the player's average, and a Poor Game is below 50% of the Player's average Game Score. I've included a note on GameScore at the end of this post)).
Here is what I found:
https://i.imgur.com/fXzZspV.png
As expected, 1) Devin isn't nearly as consistent as these guys and 2) as expected his consistency is significantly lower this year.
One thing that stands out is the percentage of great games Devin had in the previous two seasons. On the surface level, this seems like a good thing - but it actually feeds into his relative inconsistency. You'll see in the link I provided, Jokic (rated the most consistent) has the lowest % of "Great Games" because all of his games are great compared to other players and to hit 125% of that would be rather extraordinary. The "most inconsistent" version of this statistic would be if 50% of your games were "Great" and the other 50% were "Poor". To be consistent, you want the bulk of your games to fall into the "Good" bucket.
To further illustrate this, I calculated the standard deviation of Devin's Game Score in each of the last 3 years, and also the standard deviation of the game score for the 4 guys I randomly selected. I then took that Standard Deviation as a percentage of the average, which tells us that 68% of all the players Games fall within X% of their average. The X being the number I have reported below:
Dev 22-23: 48.7%
Dev 23-24: 51.7%
Dev 24-25: 56.8%
KD: 28.7%
Fox: 42.3%
Bane: 43.1%
JWill: 43.2%
A smaller number here is better. Basically, you can read this as "In 2024-25, 68% of Devin's games result in a GameScore within 56.8% of his Average GameScore". That sounds kind of convoluted, but basically tells us how tight of a range his game outcomes are.
So... at the end of the day, Devin is definitely becoming more inconsistent. I also stand by my assertation that he is quite inconsistent.
A note on GameScore: it's a composite metric of all of a players box score contributions in a game. It's not perfect, but it's the best that we really have to measure all of the box score contributions on a game by game basis. It's kind of like PER, but on a game by game basis. Not great at measuring true game impact, but a good measure of counting stats.
Based on a conversation I was having with a Vassell stan on the Spurs subReddit, I ended up doing a little bit of a Tuesday night stat session. I'll eventually expand this to look at some other Spurs, but essentially I made a comment that Vassell is inconsistent and someone challenged me on that. Seems like people here appreciate some stat drops, so I figured I'd share it here. Below is a copy/paste from what I posted on Reddit, with some additional references.
I'll look to run some similar stats for our other players, and their levels of consistency. I can also go back and do this for historical players if you are curious, so long as their GameLogs are available in BBRef (it at least goes back to Tim Duncan's rookie season). One thing I won't do is go calculate every player in the NBA, but I'm happy to pull the statistics for specifically requested players. Just tell me the player and the year you are curious in and I'll get to it when I can (you have to tell me both. I won't go through the effort to look at multi-year periods)
The Reddit post is below.
First I wanted to establish a baseline of what constitutes consistency. To that I just googled "most consistent NBA players" and thankfully it brought me to one of my favorite analytic sites, CraftedNBA: https://craftednba.com/player-traits/consistency
I like that they use GameScore to measure consistency, because I've done the same in other data projects I've done to where I was looking at how whether or not certain players having a good game or a bad game was correlated with winning and losing (spoiler alert: when Wemby has a good game, we are more likely to win... when he has a bad game, we are far more likely to lose... you probably could have guessed this). NOTE NOT FROM THE REDDIT POST: here is my post on how GameScore correlates to wins/losses. https://www.spurstalk.com/forums/showthread.php?t=304748. I like the way CraftedNBA has scaled GameScore to each individual player in their methodology and I think I'll go back to this post and update it with these breakouts rather than the hard and fast cut offs. That will come at a later date.
I randomly picked four guys, kind of scattered throughout the list. I ended up with KD, Fox, Desmond Bane and Jalen Williams. You can go to the link I provided and check out everyone else, I just wanted to include 4 random guys in a summary table for comparison purposes. There is nothing special about the 4 guys I randomly chose.
I calculated the same consistency metric for Devin as the site uses (which I like, especially in that it is scaled for each individual player. A great game for Vassell is not the same as a great game for KD. You can look at the site, and it explains what is a great game (basically, greater than 125% of the player's average GameScore, a Good Game is +/- 25% of the players Average, a Fair game is between 50 and 75% of the player's average, and a Poor Game is below 50% of the Player's average Game Score. I've included a note on GameScore at the end of this post)).
Here is what I found:
https://i.imgur.com/fXzZspV.png
As expected, 1) Devin isn't nearly as consistent as these guys and 2) as expected his consistency is significantly lower this year.
One thing that stands out is the percentage of great games Devin had in the previous two seasons. On the surface level, this seems like a good thing - but it actually feeds into his relative inconsistency. You'll see in the link I provided, Jokic (rated the most consistent) has the lowest % of "Great Games" because all of his games are great compared to other players and to hit 125% of that would be rather extraordinary. The "most inconsistent" version of this statistic would be if 50% of your games were "Great" and the other 50% were "Poor". To be consistent, you want the bulk of your games to fall into the "Good" bucket.
To further illustrate this, I calculated the standard deviation of Devin's Game Score in each of the last 3 years, and also the standard deviation of the game score for the 4 guys I randomly selected. I then took that Standard Deviation as a percentage of the average, which tells us that 68% of all the players Games fall within X% of their average. The X being the number I have reported below:
Dev 22-23: 48.7%
Dev 23-24: 51.7%
Dev 24-25: 56.8%
KD: 28.7%
Fox: 42.3%
Bane: 43.1%
JWill: 43.2%
A smaller number here is better. Basically, you can read this as "In 2024-25, 68% of Devin's games result in a GameScore within 56.8% of his Average GameScore". That sounds kind of convoluted, but basically tells us how tight of a range his game outcomes are.
So... at the end of the day, Devin is definitely becoming more inconsistent. I also stand by my assertation that he is quite inconsistent.
A note on GameScore: it's a composite metric of all of a players box score contributions in a game. It's not perfect, but it's the best that we really have to measure all of the box score contributions on a game by game basis. It's kind of like PER, but on a game by game basis. Not great at measuring true game impact, but a good measure of counting stats.