What will we analyze?
Olympiacos has been under the spotlight for the past five years and rightfully so since the team has reached the Final Four three times and has been crowned Euroleague champion two of those. The core of the team has remained the same, so how about having a closer look at some of the numbers of this success? Note that all data used for the following analysis are based on the first 21 games of the 2016-2017 Euroleague regular season.
Does Spanoulis take up too many of Olympiacos’ offenses? Should Mantzaris be shooting more spot-up threes? Is Papapetrou actually that good in defense? Let’s dig in and find out!
There are quite a few metrics that are used to assess offensive performance, Points Per Game, FG% and PIR to name a few but we will not be using any of those as they carry bias and don’t adjust for pace or playing time. A metric that overcomes the shortfalls mentioned before is Offensive Rating(Ortg). It is an advanced statistic widely used in the NBA that estimates the total points produced by a player over the total possessions that the player was involved in (detailed formula can be found here).
Players with higher Ortg, should take up more offenses and if this is not the case then a player’s full potential is not materialized. To spot if this inefficiency exists in Olympiacos, we can compare Ortg with Usage % (Usg), a metric that estimates the percentage of team plays used by a player while he was on the floor.
The graph below illustrates what was described previously. If a player is on the right side of the bold line, we can infer that the coach undervalues him whereas the left side of the line portrays the opposite relationship.
One thing that stands out from the plot, is how overvalued Spanoulis is. He is used in, by far, the most plays in the team whereas his Ortg is below average. Players of the caliber of Spanoulis are hard to manage and I will not debate against his accomplishments, however, analytics suggest that Olympiacos could improve offensively if Spanoulis used the ball less. It could be argued that Spanoulis has the best defenders guarding him and he is the one that takes the shots when the game is on the line so factors like this might attribute to a lower Ortg, however, the difference between him and his teammates still remains significant.
It appears that Lojeski could take the load off Spanoulis occasionally, as there is plenty of untapped potential in that player. Also, Birch could get involved in some extra pick-n’-roll plays and Mantzaris could hit some more of his signature spot-up threes.
According to the numbers (quantitative analysis), all these adjustments would improve the offense of Olympiacos, however, coach Sfairopoulos has (qualitative) information on his players so he might view these facts from a different perspective. In any case, these are suggestions based on the data that has been produced up to today.
Similar to the offensive evaluation, in order to assess performance, we need to find a valid measure. A metric that is widely used in the NBA is Defensive Rating (Drtg) which estimates how many points a player allowed while on the court (detailed formula here). It follows that a lower Drtg is a better one since that means that the defensive player allowed less points scored while he was on the court.
Physical characteristics, such as age, height, and athleticism, can be significant in a player’s effort to play good defense. Diamandidis, for example, arguably one of the best defenders in Euroleague history, was big for his position and was relatively athletic. To measure each player’s physical characteristics we developed a formula that takes into account age, height and ethnicity and rates players on a scale of -20 to +20 (detailed formula at the end of the post). Spanoulis, for example, is older than the average Euroleague age and relatively short for his natural position of a shooting guard so he will have a low rating. On the other hand, Papapetrou is young and very big for a small forward so he gets the highest rating.
In the graph below, the two variables that were previously described are compared. Note that a lower Drtg indicates better defense. Also, think of the bold line as the defensive expectations based on a player’s physical attributes.
Let’s focus initially on the x-axis and the Drtg scores.
As expected, there is a high correlation between Physical Characteristics and Drtg. Players that are tall and in the early stages of their careers, such as Milutinov and Agravanis have the best Drtg, whereas aged and small players such as Spanoulis and Waters have the lowest.
The important analysis of this graph though stems from the comparison of Drtg with the Physical Characteristics.
Birch and Young are true fighters in defense and over-exceed expectations by a significant amount. Papanikolaou and Lojeski closely follow in that category.
On the other hand, Papapetrou does not perform as well as his physical characteristics would allow him to. He is the most gifted player, however, his Drtg is only average. Truth is that numbers might be deceiving in regards of Papapetrou as he is the primary defender of talented perimeter scorers which would bring his Drtg down. In addition, he might often guard a power forward in which case his physical attributes should be rated a lot less as he is relatively small for that position. It appears that Papapetrou’s rare versatility is working against him in terms of the number he produces.
Mantzaris and Green also fail expectations, however, a small bias might exist due to the fact that they tend to defend high-caliber guards. Even so, there is a reason why coach Sfairopoulos has assigned them this role and evidence suggest that they have not performed as expected.
It looks like the answers to the original questions posed are that Spanoulis does take up way too much of Olympiacos’ offense, Mantzaris should be hitting more spot-up threes and there is still plenty of room for improvement in Papapetrou’s defense. However, as mentioned in the Welcome post, Euroleageeks only focuses on the quantitative analysis of the game. The qualitative one is a very different, equally significant component of basketball and it is very likely that Spanoulis would be viewed much more favorably in that analysis.
I hope that this first blog post sparked your interest in terms of Euroleague Basketball Analytics. If you disagree with any of the arguments posed in this post, I would love to hear your opinion, so just drop a comment. If you have any suggestions for analysis of a subject let us know, but in the meantime, how about have a read of our next post?
*Physical attributes formula=relative age+relative height-0.25*relative race
relative age= average Euroleague age – player’s age
relative height=player’s height – player’s natural position average Euroleague height
relative race= 1 (if the player is European/White), 0 (if the player is of African descent)