ClutchFans Game Thread: Rockets @ Celtics 11/1/2025

Good hierarchy! Although, I don't think DFS will need much time to prove he belongs in the "Dominant Role Players" category once he gets on the floor.

 

You forgot Michael Finley

 

Watching the actual events seems to help too

 

I guess we need to put asterisks on OKCs championship and Shais MVP

 

He projects authenticity, unlike a blabbering kissup like Ryan Hollins.

 

It can help some, but not all. It depends on if you know enough to know what you are seeing and how strong your bias is

 

I guess the Suns play better without Dillon and Jalen, what a shocker.

 

I would. Let’s see who has the more distinguished career 15 years from now.

 

Wemby would still bring a lot other than basketball activity, merchandising and fans.....

I see a lot of smaller markets just going for that.

 

The two stats you are comparing have nothing to do with each other, other than having the same name.

nba.com’s off and def ratings are how many points does a team score and give per 100 possessions when the player is on the team. It doesn’t measure personal contribution and a lot of people make that jump because it is called ‘player’s off and def ratings’ but it is still a team stat. A player sharing the court with jokic at exactly the same minutes will have the same ratings with jokic. The stat will not distinguish between the two.

because this stat doesn’t measure individual contribution, many others are trying to measure individual contribution. Bbref is one of them. But individual contribution is extremely hard to measure. It has a complicated formula which you can find on bbref website. It gives out a number that is supposed to be a measure of player contribution. The problem is there is no way you can verify whether the stat actually measures what it claims to measure. Typically these stats perform well on the 'best' players. Because they are good in measurables. But as you go down the ladder, the contributions are harder to measure and they don't perform well any more. For an average player, moving correctly on the court is very important defensively, which is not measured. For example bbref had j green as the most 'impactful' player in its on/off stat, despite he played horribly.

So my take is neither of the two numbers show Reed's defensive contributions. The first one because it doesn't even intend to, the second one because it doesn't measure well. And it is meaningless to compare the two numbers.

 

Spoiler

Yes, I know. I’ve said that previously in other game threads.

NBA is based on pbp, bbref is a fancy formula to use only boxscore stats (for historical reasons).

what I’m doing is providing a game-by-game result in each game thread so we can see how different the results can be. I might get bored doing this, but for now I think it is helpful to see the wild differences for those who like to quote individual DRtg. Even the full season results are significantly different wrt team rankings.

 

NBA is not pbp(what do you mean by pbp). It is just what the team scores/concedes when the player is on and off court. There is no component that comes from each play or what the player does on that play. So it is normal that the numbers are wildly different. Because they measure different things.

You are right though, you can’t compare player by their nba.com ratings. You could compare them by their bbref ratings if you believed that it does what it intends to do, personally I don’t.

 

Typically in large samples, the offensive and defensive ratings can correlate at the extremes with player value. In the past, I think it was very easy to see who the weak link was in the rotation with just raw defensive ratings. I think that is true for this season as well. It is the middle area that gets tough to see who is contributing. For those, one really needs to look at either on/off or even better a stat like EPM that uses a bunch of individual stats to help correlate +/- to a player.

 

It is PBP in that it assigns team scoring to a player using PBP data while a player is in the game.

 

What is pbp data? Can you give an example? Do you say if two players are on the court always together, their ratings can be different?

From nba.com:

Definition
The number of points allowed per 100 possessions by a team. For a player, it is the number of points per 100 possessions that the team allows while that individual player is on the court
Formula
100*((Opp Points)/(Opp POSS))
Type
Advanced

 

yes, typically the best and worst ones correlate better. Particularly not helpful with players who get limited minutes from the bench.

I mentioned in my other reply. The starters get the same contribution to their rating when they are playing together. The differences in their rating comes from the times when they are not playing with the starting 5, not from their contributions within the starting 5. That’s exactly why it cannot be used as personal defense level.

 

I'd generally agree with what you say here. There are some things that these numbers might be helpful.

For example, if you want to see whether the double-big lineup is good or bad for defense, you can look at the conbined Sengun-Adams DefRtg compared it to the whole team's DefRtg. Even that has a bit of noise because it does not account for the strength of opponents while the two guys were both on the court. Any metric based on +/- has this defect.

 

when you are comparing a 2 man rating to a team, you are essentially comparing team ratings, to conclude something about the 5 man line up, which is okay, with its limitations such as different opponents etc. Ideally you would compare to line ups where only Adam’s is replaced, because these are the main two options.

the problem starts when you use Sengun’s and Adam’s ratings to compare one to another, to judge whose individual defense is better which is very common in the forum.

 

No one said anything about a “component that comes from each play”

Play-by-Play is just what we call the data that NBA.com provides for everything that happens in a game. PBP for short.

So, from the official PBP data, we can count the factual amount of possessions in a game. Thus, determining TEAM ORtg and DRtg is simply Total Points / Total Possessions x 100, as you said.

We agree on the formula.

So, it is common for ppl to say the NBA official ORtg and DRtg stats are PBP-based. We say this to contrast from non-PBP based stats. Those are what ppl refer to as Boxscore-based stats.

bbref has decided they want to provide stats that ”work” beyond the years that pbp was available. So they have to restrict themselves to boxscore-based formulas.

Bottomline: we seem to agree. I’m just clarifying what the term PBP-based stats mean, in general.​