z-points @ work
To give you a flavor of how z-points work, I’ll show here a couple of totally different examples: the buzzer beater and the blow out game.
The buzzer beater
This is the box score of game 2 of the NBA 2009 play-off series between Orlando and Cleveland, in excel format. This was the game where Lebron James won it for the Cavs with an unbelievable 3 at the buzzer.
There are two different parts in the score box. The grey-colored part on the left is the regular box score. It shows points, assists and field goal percentages for each player. Note that it also includes 2pt percentages!
The orange part, on the other hand, corresponds to z-points. The column named zpoints1 shows the real z-points. They range from 0 to anything, but 0.5 is already a good performance and scores over 2.0 are extremely rare. The column named zpoints is just an adjustment to ‘convert’ the real zpoints1 to their equivalent in normal points. It works like this:
LeBron James scored 1.36 zpoints1 in the game. Out of total zpoints1 in this game, Lebron scored 23.14% (1.36/5.86; 5.86 is the sum of column AD). So we calculate the equivalent of zpoints1 as 0.2314*(95+96), where 95 + 96 would be the total combined score. That is, adjusted zpoints credits Lebron with the same % out of total points as in real zpoints, because the latter take into account the competitive context in which Lebron scored.
Now, back to the box score we can see how LeBron stands out. In fact, just by looking at it we know that Lebron won it from downtown, without ever needing to watch the game. How’s that? First, look at his points vs. adjusted zpoints (35 vs. 44.2). Then look at the zpoints1 column (1.36). That’s a pretty high number, showing that he scored when it mattered. Adjusting that results in 44.2 adjusted zpoints, which is higher than the 35 ‘face value’ points that he actually scored.
Most of his zpoints1 came from 3 point shots (0.69), and his z-3pt% is much higher that his real 3pt% (0.836 vs 0.333). That’s telling us that he delivered in crunch time from the 3-point line, and since his teammates have only so-so zpoint stats, we know it was James who won it with a trey.
The notation we use to show his performance would be like this:
LeBron James (CLE)
35 pt, 0.55 2pt%, 0.33 3pt%, 0.83 FT%
44.2 zpt (1.36), 0.43 z2pt%, 0.84 z3pt%, 0.83 FT%
But look now at Hedo Turkoglu’s stat line. Even though he didn’t score as many points as James (21 vs. 35), he’s got almost as many z-points! (1.34 vs 1.36)! Considering that Turkoglu’s didn’t score a buzzer beater, which is giving LeBron a ton of z-points, he must have done a hell lot of good things at the end of the game to keep up with him.
Let’s take a look at it:
0:48 Hedo Turkoglu makes 25-foot three point jumper (Rashard Lewis assists) 93-93
0:01 Hedo Turkoglu makes jumper 95-93
So the reason why he had so many z-points is that, in fact, he almost won the game by beating the buzzer. Only that he didn’t, and it was LeBron who ended up beating the buzzer in the little time left in the game.
Turkoglu’s stat line would read like this:
Hedo Turkoglu (ORL)
21 pt, 0.54 2pt%, 0.5 3pt%, 0.5 FT%
43.5 zpt (1.34), 0.79 z2pt%, 0.81 z3pt%, 0.5 FT%
Note how the difference between shooting percentages reflect that he scored big from 2pt and 3pt range.
LeBron on his way to some big z-points
The blow out game
On October 28, 2009 the Celtics crushed the Bobcats in their second game of the season (CHARLOTTE @ BOSTON 59-92).
Let’s focus on Ray Allen’s line. He had 18 points vs 26.23 z-points, which initially would make us believe that he was great in the clutch. But we know (by the score) that this game wasn’t even close, and that shows also in Allen’s zpoints1 column: see how he’s got only 0.28 z-points, much lower than what LeBron or Turkoglu had in our example above. This is telling us that Ray Allen scored when the game was still a game. That’s why it’s important that we indicate the real zpoints in a player’s stat line. Allen’s adjusted zpoints are high compared to his real points (26.23 vs 18) because COMPARED TO THE REST OF THE PLAYERS IN THE GAME, he scored in more competitive situations. In fact most of his z-points came from 3’s that he made as the Celtics were breaking the Bobcats in the first and second quarter of the game.
Allen’s stat line would read as follows:
Ray Allen (BOS)
18 pt, 0.29 2pt%, 0.4 3pt%, 0.67 FT%
26.2 zpt (0.28), 0.23 z2pt%, 0.54 z3pt%, 0.84 FT%
The other thing we can tell is the guys making up their stats playing garbage time. See Derrick Brown (9 points vs 0.71 z-points) and Gerald Henderson (8 points vs 1.03 z-points) in Charlotte, or Shelden Williams (12 points vs 0.71 z-points) and Eddie House (7 points vs 0.15 z-points) in Boston.
In this case, Shelden William’s stat line would read:
Shelden Williams (BOS)
12 pt, 0.67 2pt%, 0.73 FT%
0.71 zpt (0.01), 0.07 z2pt%, 0.54 FT%
One last thing to keep in mind: the real z-points (the figure in parenthesis, zpoints1 in the excel spredsheet) can be used to compare players in different games, because it’s an absolute measure. On the other hand, the adjusted zpoints (0.71 for Shelden Williams, or column zpoints in the spreadsheet) can be used only to compare players within a particular game. We use them because they give us an intuitive measure of the real z-points that can be easily interpreted, by comparing it with real points scored.
Hope it’s all clear, but feel free to leave any comments with your questions!
