There has recently been a lot - a very lot - of speculation on what effect Kevin Martin's return has had on the Kings. Clearly the Kings have not had a great record since he has returned (2-12) and Kevin's performance (in terms of scoring, in terms of efficiency, etc.) has been sub-par given what we expect from him.
These facts have led many readers here to conclude that Kevin is not a good fit for this team, and that he is bringing the team down.
Other readers, however, have pointed out some possible mitigating factors for the Kings W-L record following Martin's return. For instance, of the 14 games since Kevin has come back, 10 have been on the road. Many of these games have also been against strong teams. Also mentioned has been a suggestion that the decline in the Kings performance (down from the near-euphoric heights we saw earlier in the season) began before Martin's return.
This latter suggestion has in some cases not gone over particularly well, and those who suggest that the Kings' slide began before Martin's return have been accused of intentionally and falsely denigrating the rest of the team (especially its performances forcing overtime against the Cavaliers and the Lakers) in order to support Kevin Martin:
This is a pretty serious accusation to throw around, and when I saw it I felt that those accused should at least have the opportunity of a factual defense - if the facts support their point of view, it's not a fair accusation. So I decided to see if the facts bore them out.
Of course, it's not the easiest thing in the world to determine how well a team is doing - simply looking at the win-loss record won't suffice for several reasons. First off, there's a big difference between a close win and a blowout, and a big difference between a close loss and being blown out. In addition, the quality of the opponent, and whether a game is at home or on the road will be expected to have an effect on how the team performs. Losing to a playoff team by 2 points on the road suggests a good performance, while barely beating a bad team at home suggests a poor performance.
In order to quantify this, I reasoned as follows: the expected final margin of a game can be estimated by combining three factors - the Kings' average margin ("K", -4.0 points this year), the opponent's average margin ("O", varies), and the margin associated with home court advantage ("HC", +3.3 points, calculated over the entire 2008-2009 season). Thus, the expected margin would be K - O + HC for a game at Arco, and K - O - HC for a road game. For example, the expected margin in a road game against Utah would be -4.0 - 5.0 - 3.3 = -12.3 points, we would expect to lose by about 12. The expected margin for a home game against the Warriors would be -4 -(-4.6) + 3.3 = 3.9 points, we would expect to win by about 4.
Following that procedure, I calculated the expected final margin for each of the 51 games this season, and subtracted it from the actual final margin (after four quarters) to get a measure of Margin Relative To Expected Margin ("RM", or relative margin). Here, positive RM numbers indicate a Kings performance better than expected, and negative RM numbers indicate a Kings performance worse than expected. (I should note that for the four overtime games this season, I set the final margin to 0 - the result after four quarters - so as not to "denigrate" the Kings' excellent performances in the hot-button Cavs and Lakers games, which were ultimately lost by 13 and 9 points.)
Below I present a plot of RM for each game of the season thus far. Red dots are games in which Kevin Martin played, blue dots are games in which Kevin did not play.
Since the data are so noisy (and expectedly so), I calculated a running average, which I plotted as a line. This running average takes the mean of the surrounding 7 values (the value at that point, and three games on either side). Where the line is red, every game that went into the average had Martin playing. Where the line is blue, every game that went into the average had Martin out. The line is green during transitions - some games in the average included Kevin and some games did not.
As you can see from the plot, the Kings' Relative Margin plateaued about 10 games into the season between about 2-4 points above expected, then rose to a peak of about 6 points above expected about 30 games in (not coincidentally, the Cavs and Lakers games were games 28 and 29). However, following that peak, there was a stretch of about 7 games where the Kings' RM plummeted before Martin's return, and it continued to drop to about 5 points below expected as the running average transitioned across the No-Kevin/Kevin border. Following Kevin's return, the RM has actually improved back to even.
These data paint a picture that is not very compatible with the hypothesis that Kevin Martin's return is hurting the team. It appears that the Kings were in fact declining before Kevin Martin's return (those "fans" were right!), and that, somewhat paradoxically, that decline stopped - and reversed - shortly after Martin's return despite the fact that Martin has been playing poorly. Why? I certainly don't know, but I would guess that "rookie walls" and random chance have a lot more to do with it than Kevin Martin.
Thoughts?
These facts have led many readers here to conclude that Kevin is not a good fit for this team, and that he is bringing the team down.
Other readers, however, have pointed out some possible mitigating factors for the Kings W-L record following Martin's return. For instance, of the 14 games since Kevin has come back, 10 have been on the road. Many of these games have also been against strong teams. Also mentioned has been a suggestion that the decline in the Kings performance (down from the near-euphoric heights we saw earlier in the season) began before Martin's return.
This latter suggestion has in some cases not gone over particularly well, and those who suggest that the Kings' slide began before Martin's return have been accused of intentionally and falsely denigrating the rest of the team (especially its performances forcing overtime against the Cavaliers and the Lakers) in order to support Kevin Martin:
This is a pretty serious accusation to throw around, and when I saw it I felt that those accused should at least have the opportunity of a factual defense - if the facts support their point of view, it's not a fair accusation. So I decided to see if the facts bore them out.
Of course, it's not the easiest thing in the world to determine how well a team is doing - simply looking at the win-loss record won't suffice for several reasons. First off, there's a big difference between a close win and a blowout, and a big difference between a close loss and being blown out. In addition, the quality of the opponent, and whether a game is at home or on the road will be expected to have an effect on how the team performs. Losing to a playoff team by 2 points on the road suggests a good performance, while barely beating a bad team at home suggests a poor performance.
In order to quantify this, I reasoned as follows: the expected final margin of a game can be estimated by combining three factors - the Kings' average margin ("K", -4.0 points this year), the opponent's average margin ("O", varies), and the margin associated with home court advantage ("HC", +3.3 points, calculated over the entire 2008-2009 season). Thus, the expected margin would be K - O + HC for a game at Arco, and K - O - HC for a road game. For example, the expected margin in a road game against Utah would be -4.0 - 5.0 - 3.3 = -12.3 points, we would expect to lose by about 12. The expected margin for a home game against the Warriors would be -4 -(-4.6) + 3.3 = 3.9 points, we would expect to win by about 4.
Following that procedure, I calculated the expected final margin for each of the 51 games this season, and subtracted it from the actual final margin (after four quarters) to get a measure of Margin Relative To Expected Margin ("RM", or relative margin). Here, positive RM numbers indicate a Kings performance better than expected, and negative RM numbers indicate a Kings performance worse than expected. (I should note that for the four overtime games this season, I set the final margin to 0 - the result after four quarters - so as not to "denigrate" the Kings' excellent performances in the hot-button Cavs and Lakers games, which were ultimately lost by 13 and 9 points.)
Below I present a plot of RM for each game of the season thus far. Red dots are games in which Kevin Martin played, blue dots are games in which Kevin did not play.
Since the data are so noisy (and expectedly so), I calculated a running average, which I plotted as a line. This running average takes the mean of the surrounding 7 values (the value at that point, and three games on either side). Where the line is red, every game that went into the average had Martin playing. Where the line is blue, every game that went into the average had Martin out. The line is green during transitions - some games in the average included Kevin and some games did not.
As you can see from the plot, the Kings' Relative Margin plateaued about 10 games into the season between about 2-4 points above expected, then rose to a peak of about 6 points above expected about 30 games in (not coincidentally, the Cavs and Lakers games were games 28 and 29). However, following that peak, there was a stretch of about 7 games where the Kings' RM plummeted before Martin's return, and it continued to drop to about 5 points below expected as the running average transitioned across the No-Kevin/Kevin border. Following Kevin's return, the RM has actually improved back to even.
These data paint a picture that is not very compatible with the hypothesis that Kevin Martin's return is hurting the team. It appears that the Kings were in fact declining before Kevin Martin's return (those "fans" were right!), and that, somewhat paradoxically, that decline stopped - and reversed - shortly after Martin's return despite the fact that Martin has been playing poorly. Why? I certainly don't know, but I would guess that "rookie walls" and random chance have a lot more to do with it than Kevin Martin.
Thoughts?
thats so true... but other factors do go into it... 
