Offensive Pace and Defensive Pace

[nonemus]

This is a question that I had for quite a while now. If you look at the formulas for the most commonly used method for calculating the number of possessions in a game:

B-R:

0.5 * ((Tm FGA + 0.4 * Tm FTA - 1.07 * (Tm ORB / (Tm ORB + Opp DRB)) * (Tm FGA - Tm FG) + Tm TOV) + (Opp FGA + 0.4 * Opp FTA - 1.07 * (Opp ORB / (Opp ORB + Tm DRB)) * (Opp FGA - Opp FG) + Opp TOV))

There are basically two calculations, its, 0.5*(A+B), where A is supposed to represent the teams possessions (ideally) and B is supposed to represent the number of offensive possessions of the other team (or, Team A's defensive possessions). So, these two are averaged to form a number which represents the number of possessions of a team in a game. Why though? If they are separated, you have Offensive Possessions and Defensive Possessions (or Opponent's possessions), which you can use to calculate Offensive Pace and Defensive Pace. This would provide for a more accurate estimate of a team's actual O and D ratings.

Consider this, say that team A uses 100 possessions and Team B uses 80. Both teams score 90 points. Which team was more efficient on offense/defense? By the normal method, the O/D ratings of both teams are 100. But why is this? If Team A took 20 more possessions than Team B to score those 90 points, doesn't that point to the fact that they are less offensively efficient than Team B ? Team A's O/D rating would be 90/112.5, which makes sense, because although they had 20 more possessions, they scored the same number of points.

Why do your offensive possessions get factored in to your Drtg (and vice versa)? Wouldn't it be better to separate the two?

[Chicago76]

In reality offensive and defensive possessions don't deviate substantially over the course of the year and game to game. You will never have a game where one team has 100 possessions while the other has 80. Why? Because one team's possession begins when the other team's ends. You effectively take turns. The only difference is game clock issues. Theoretically, you could have 3-4 more possessions a game if you had the ball to start and end most quarters of play. That's the only difference in team possessions 99% of the time.

A decent illustation might be the 3 or so times over the course of a game where one team or the other is stuck chucking something up on a possession that began with 5 seconds on the game clock. They're not typically going to happen to both teams with equal frequency. These types of possessions could swing things +/-50 possessions over the course of a season. If by random chance your opponent has more of these possessions, then your Drtg would look artificially better if you based it only on opp pts/opp poss. Similarly, if your team is stuck with more of these possessions, your Ortg will look artifically worse if you based it only on team pts/team poss.

It's just easier to average the two to better wash out the end of clock impact.

[Bobbcats]

I'd be interested to see stats on team's average shot clock at the end of offensive possessions, and also defensive possessions. Someone good with computers should be able to do that from game logs.

[droponov]

I'd be interested to see stats on team's average shot clock at the end of offensive possessions, and also defensive possessions. Someone good with computers should be able to do that from game logs.

I'm not entirely sure if this is what you meant, but 82games.com tracks shooting efficiency per shot clock usage on both ends of the floor.

http://www.82games.com/1011/1011ATL3.HTM

For example, the Hawks attempted 20% of their shots in the first 10 seconds of the clock and had a .432 eFG% in those shots, 56% of them were assisted and they scored 13.4 ppg in that period of the clock.

[tmd_10]

Doesn't an offensive rebound count as an extra possession?

[Nivek]

Doesn't an offensive rebound count as an extra possession?

No. An offensive rebounds extends a possession, it doesn't create a new one. A basketball possession starts when one team gets the ball and lasts until the other team gets it back.

An offensive rebound creates a new "play" for the team with the ball. There can be multiple plays per possession. A team could come down, miss an FGA, get the offensive board, get fouled, miss both free throws, get THAT offensive rebound, them commit a turnover. That's 3 plays on one possession. And it sounds disturbingly familiar to a number of actual Wizards possessions this season.

[penbeast0]

Kurtz: [voiceover] The horror... the horror...

[tmd_10]

Doesn't an offensive rebound count as an extra possession?

No. An offensive rebounds extends a possession, it doesn't create a new one. A basketball possession starts when one team gets the ball and lasts until the other team gets it back.

An offensive rebound creates a new "play" for the team with the ball. There can be multiple plays per possession. A team could come down, miss an FGA, get the offensive board, get fouled, miss both free throws, get THAT offensive rebound, them commit a turnover. That's 3 plays on one possession. And it sounds disturbingly familiar to a number of actual Wizards possessions this season.

Thanks. I was thinking in terms of the shotclock---when it gets renewed it's almost like having an extra possession---but technically the way it is makes sense.

[Doctor MJ]

Thanks. I was thinking in terms of the shotclock---when it gets renewed it's almost like having an extra possession---but technically the way it is makes sense.

I'd say that the reality is that the actual official definition for what constitutes a possession was never crystal clear. From a statistical analysis perspective though, defining a possession's end based on when the other team gains control over the ball is really the only way to go.

[lorak]

Thanks. I was thinking in terms of the shotclock---when it gets renewed it's almost like having an extra possession---but technically the way it is makes sense.

I'd say that the reality is that the actual official definition for what constitutes a possession was never crystal clear. From a statistical analysis perspective though, defining a possession's end based on when the other team gains control over the ball is really the only way to go.

Yeah, but it would be interesting if offensive rebounds would be count as new possessions for attacking team.
For example team A had four 2pts shots, made two of them but also had 2 off rbs, so overall on two possessions team A scored 4 points, but needs 4 shots to do that! On the other hand team B had two shots and made them all, so also scored 4 pts on 2 possessions but needs only 2 shots to do that. Which offense was more efficient? Going by traditional possessions definition they were equal on offense, but in reality team A needs more effort, cost was bigger, more shots (2/2 FG vs 2/4 FG +2ORB)... so I'm not sure that in the game reality these situations are equal as possessions theory suggest.

[ElGee]

Yeah but think about it this way: A possession ENDS when the other team gets the ball. There's really no way around that, is there?

So, how can the other team get the ball? A made score, a turnover, or a defensive rebound (including the "team rebound"). If there were no shot clock, it would be possible for them to never get the ball (until the quarter expired).

Given that, offensive rebounding *continues* a possession because it's part of the concept of efficiency. A team can shoot 20% but if they grab 100% of the offensive rebounds they will have the highest efficiency in league history and probably never lose. (Literally, think about it.) There isn't really an extra cost as you suggest.

[lorak]

The extra cost is team effort, fatigue and so on.
And I understand what possession is, that ORB don't end possession, but I'm not sure that points per possessions are the best way to measure offensive efficiency - because as I said in that example above: 2/2 FG in 2 possessions aren't equal to 2/4 FG in 2 possessions.

[ElGee]

Right but it's just a bias toward FG%. Who cares if it's 2/2 versus 2/4? 3 point shots count more than 2s and FTs count too. That shouldn't be dismissed. Those are, literally, the rules of the sport.

I'm having a hard time buying the argument that it's extra physical exertion to grab an offensive rebound when (a) most teams try to grab the rebound in some capacity anyway and (b) they wouldn't have to run 75 ft. down the court and play defense, which is more physically taxing than offense.

[lorak]

Right but it's just a bias toward FG%.

Bias? Really? What's more efficient: 2/2 FG (only 2pts shots) or 2/4 FG (only 2pts shots)?

I'm having a hard time buying the argument that it's extra physical exertion to grab an offensive rebound when (a) most teams try to grab the rebound in some capacity anyway and (b) they wouldn't have to run 75 ft. down the court and play defense, which is more physically taxing than offense.

It's not about extra physical exertion to grab ORB. It's about more time, so effort, fatigue (running and so one) team have to use to score 4 points (I'm still using that example). Let say when you scored 4 pts on 2/2 FG it takes only 30 seconds combined. But when scored 4 pts on 2/4 FG it takes 50 seconds. You see the difference? Sure, defense also have to work during these extra seconds, but we are talking about offensive efficiency: more efficient is to score 4 pts on 2 shots using less energy (30 seconds of "running") than to score 4 pts on 4 shots using more energy (50 seconds of effort).

[ElGee]

But that's not how efficiency is defined. It's points per possession.

Think of it this way: in American football, is a 1 play 99 yard drive more efficient than a 20 play, 14 minute 90 yard drive if they both end in a touchdown? They had the same efficiency. It's the same in basketball - two points before the other team can touch the ball is again is two points before the other team can touch the ball, regardless of whether it took 4 seconds, 20 seconds, 50, or the whole quarter.

And I'd argue energy wise, it's better to be on offense longer and force the other team to dig in on defense.

[lorak]

But that's not how efficiency is defined. It's points per possession.

Definition isn't set in stone. Long time ago efficiency = FG%, then some kind of formula (PTS+RBS-FGA and so on), now pts per possession... but in the future maybe we would have something better.

Try thinking outside of the box and answer this question:
What's more efficient: 2/2 FG (only 2pts shots) or 2/4 FG (only 2pts shots)?

And I'd argue energy wise, it's better to be on offense longer and force the other team to dig in on defense.

Maybe it's better from overall-game-point-of-view, but is it better from offense only point of view? Really more effort, more shots, more time to score (more energy used) the same amount of points = more/equal efficient offense?

[ElGee]

Yes I understand what you're saying, but as I've outlined, it's a logical way to view efficiency based on the rules of the sport. You're hung up on a stat (FGA), but that seems fairly arbitrary to me when the way to win is score more points. You're just talking about shooting efficiency from the field.

What's more efficient, a possession with 0 passes or 40 passes? I say who cares if they both result in the same number of points. Know what I mean?

[Golden State]

Basketball is played distinctly different in the final minutes when the two teams make the efforts to take over and win a game. Likewise, two distinct calcs are needed.

1. Not so easy Calc. Take over and produce the win in the final 10 minutes.

2. The easy calc: Produce a 3-1/2 quarter tie or better. (I like run their tongues out styles like SA. The occaisional blowout occurs easier and earlier in this style allowing players to log fewer minutes. Also, ...tired opposing players make more mistakes/turmnovers and throw more air balls in the final minutes.)

The winning effort is at the end of most close games. Complicating this event is coaches of varying efficiency micro-managing for advantages and higher efficiency. An experienced and knowledgeable coach may mean more to a win in those final moments than the talents and sizes of the players on the floor.

While I have never been a statistician I do see competitive teams saving their best play for the final minutes or risking a loss. It is logical that rosters be backwards engineered around this goal to win the game. On paper, you want superiority in the final minutes vs any of the 16 potential playoffs opponents.

Any remaining roster slots can go to promising prospects.

[kabstah]

@DavidStern: Efficiency measured as points per possession makes intuitive, practical sense. If we demarcate a possession as when control of the ball switches from one team to another, then in a single game both teams will have roughly the same number of possessions (like a difference of 2 possessions at most). What that means is, the team that scores more points per possession in a particular game will almost always win the game. Shooting a higher FG% or expending less energy (points per kilojoule?) has a much weaker correlation to winning, and thus they're less valuable metrics for efficiency.

@Golden State: Until I see a study that suggests otherwise, I'm unconvinced that the last two minutes of a game mean any more than any other two minute period in the game. The fundamental principal would be easy: see if a stronger correlation exists between winning the game and outscoring the opponent by X points in the last two minutes than the correlation between winning the game and outscoring the opponent by X points in any other two minute period.