Pregame: Illinois at Texas Tech, Tuesday, November 10th, 7:00pm CT, Peacock

Status
Not open for further replies.
#126      
hangdog
wc_illini said:
EvanMiya uses a very interesting approach, and I think in large sample sizes it could be very powerful (it promises to naturally account for things like lower competition in garbage time, or when key players are hurt, etc). But the player ratings should not be used with small sample sizes, and that probably goes for the team rankings as well.

Edit: Torvik has Tomi 8th best on our team
Click to expand...
I am extremely confident that Tomi is somewhere between the 4th best player in the country and the 8th best player on his team.
 
#127      
wc_illini
NarrowJ said:
You asked what the difference was, and then kind of explained it yourself.

1. Value vs volume, mainly: "Best Player" rating would heavily favor players who score 20+ per game or play heavy minutes. BPR cares about how much better the team functions when a player is on the floor, regardless of having a ginormous usage rating or playing full minutes.

2. BPR uses play-by-play data to isolate the player. It accounts for the strength of the other 9 players on the court. A simple "Best Player" rating probably doesn't consider the individual within the context of the team/game environment.

3. Projection vs Accomplishment: A "Best Player" rating might tell you what a player has accomplished. BPR is a predictive metric design to estimate how valuable a player is going forward. NBA front offices and draft analysts rely on tools like BPR alongside scouting reports. It doesn't care he only played 36 minutes in 2 games and it knows his poor shooting percentages are likely to experience positive regression.
Click to expand...
Yes, it's different from other definitions of "best", but that doesn't mean it isn't attempting to be one.

In fact, he says:
"Bayesian Performance Rating is the best all-in-one player metric publicly available in college basketball."
"Typically, the best player in the country each year finishes with a BPR above +10.0, a top 50 player in the country would have a BPR of about +7.0, an average high-major starter would have a BPR of about +4.0, and an average D1 player would have a rating of precisely 0.0."

Even if his second quote means the best player isn't necessarily the top BPR, he's at least claiming it's highly correlated and puts players in predictable ranges.
 
#128      
NarrowJ
wc_illini said:
Yes, it's different from other definitions of "best", but that doesn't mean it isn't attempting to be one.

In fact, he says:
"Bayesian Performance Rating is the best all-in-one player metric publicly available in college basketball."
"Typically, the best player in the country each year finishes with a BPR above +10.0, a top 50 player in the country would have a BPR of about +7.0, an average high-major starter would have a BPR of about +4.0, and an average D1 player would have a rating of precisely 0.0."

Even if his second quote means the best player isn't necessarily the top BPR, he's at least claiming it's highly correlated and puts players in predictable ranges.
Click to expand...

Yeah, very fair point. It does seem the description talks out both sides of it's mouth, so to speak. Though, the first quote doesn't necessarily call it a "best player" metric (just 'best all-in-one player metric'... so 'best metric' not 'best player'). The second quote means he absolutely intends it to identify the "best" players. I think the nuance is instead of defining "best" players by flashy box-score volume, BPR proves it out mathematically by measuring how a player's presence actually drives team efficiency and winning basketball.
 
#129      
wc_illini
NarrowJ said:
Yeah, very fair point. It does seem the description talks out both sides of it's mouth, so to speak. Though, the first quote doesn't necessarily call it a "best player" metric (just 'best all-in-one player metric'... so 'best metric' not 'best player'). The second quote means he absolutely intends it to identify the "best" players. I think the nuance is instead of defining "best" players by flashy box-score volume, BPR proves it out mathematically by measuring how a player's presence actually drives team efficiency and winning basketball.
Click to expand...
I think it's fair to argue it shouldn't be used to identify the best player (I already said I think it's very inaccurate in small sample sizes- in fact, as a predictive metric, I'm surprised it moved so much for Tomi in just 2 games; I'd think his pre-season rating would have a much higher weight). I just wouldn't say the rating doesn't claim to be such.
 
#131      
lstewart53x3
wc_illini said:
Yes, it's different from other definitions of "best", but that doesn't mean it isn't attempting to be one.

In fact, he says:
"Bayesian Performance Rating is the best all-in-one player metric publicly available in college basketball."
"Typically, the best player in the country each year finishes with a BPR above +10.0, a top 50 player in the country would have a BPR of about +7.0, an average high-major starter would have a BPR of about +4.0, and an average D1 player would have a rating of precisely 0.0."

Even if his second quote means the best player isn't necessarily the top BPR, he's at least claiming it's highly correlated and puts players in predictable ranges.
Click to expand...
NarrowJ said:
Yeah, very fair point. It does seem the description talks out both sides of it's mouth, so to speak. Though, the first quote doesn't necessarily call it a "best player" metric (just 'best all-in-one player metric'... so 'best metric' not 'best player'). The second quote means he absolutely intends it to identify the "best" players. I think the nuance is instead of defining "best" players by flashy box-score volume, BPR proves it out mathematically by measuring how a player's presence actually drives team efficiency and winning basketball.
Click to expand...
I look at the stat to mean: usage aside, who are the best players?

But obviously, in real life usage shouldn't be set aside. So the true best players are the ones who have high BPR and high usage.
 
#133      
NarrowJ
lstewart53x3 said:
I look at the stat to mean: usage aside, who are the best players?

But obviously, in real life usage shouldn't be set aside. So the true best players are the ones who have high BPR and high usage.
Click to expand...

So maybe “highest impact on team efficiency” and not “best players” as the tweet claimed.
 
#134      
wc_illini
lstewart53x3 said:
I look at the stat to mean: usage aside, who are the best players?

But obviously, in real life usage shouldn't be set aside. So the true best players are the ones who have high BPR and high usage.
Click to expand...
I wondered that too.

What's strange is that he claims: "One primary choice I made was to include some measurements of playing time and usage in the model to reward players who play a larger percentage of their team’s minutes and carry bigger offensive loads."

I'd have to guess he doesn't do it as much as Torvik (Miya had Z as our 3rd best player last year just behind Mirk, while Torvik had him 8th)
 
#135      
lstewart53x3
wc_illini said:
I wondered that too.

What's strange is that he claims: "One primary choice I made was to include some measurements of playing time and usage in the model to reward players who play a larger percentage of their team’s minutes and carry bigger offensive loads."

I'd have to guess he doesn't do it as much as Torvik (Miya had Z as our 3rd best player last year just behind Mirk, while Torvik had him 8th)
Click to expand...
Yeah, there's just no way BPR does a good job at accounting for usage. He's got big Z at #27.
 
#138      
RollChief
Here's a fun one for you. TTU's average height through two games is 76.6". The last time that Illinois lost to a team this short or shorter was on December 9th, 2023 away against Tennessee.
 
#141      
audie
RollChief said:
Here's a fun one for you. TTU's average height through two games is 76.6". The last time that Illinois lost to a team this short or shorter was on December 9th, 2023 away against Tennessee.
Click to expand...
My personal average height per 24 hours is on an ever-so-slow decline, however, I've never ever seen this stat before.
It's awesome though. Thanks, Chief.
 
#146      
Ga65
IlliniCC16 said:
I really believe this is a 15-20 point win that never feels that close. Source: Gut.
Click to expand...
This is the game we tell if this offense is for real. Actually trust the defense but need to know they can get away with all those cross court passes against a really good team. If turnovers are under 10 we will win by 20.
 
#149      
Heisenbobble
I don't have a good feeling on this game. I love our freshmen, but this will be a tough one for them. A true road game against a top-20 opponent. Seems like one of those games you wish you had more veterans available to play (Mirk & Jake). We'll see. I predict Tech wins this one, 84-73.
 
#150      
Illini4Christ
illini0440 said:
If I know Dravyn like I think I know him, he’s going to try and prove a point and it feels like he’s gonna have the 25 points on 18 shot attempts kind of night. He will try and force things
Click to expand...
Oh yeah. I guarantee you are dead on.
 
Status
Not open for further replies.
Back