A Further Look at Rest and Shot Difficulty

May 17, 2015; Houston, TX, USA; Los Angeles Clippers forward Blake Griffin (32) and guard Chris Paul (3) walk on the court during the fourth quarter against the Houston Rockets in game seven of the second round of the NBA Playoffs at Toyota Center. The Rockets defeated the Clippers 113-100 to win the series 4-3. Mandatory Credit: Troy Taormina-USA TODAY Sports
May 17, 2015; Houston, TX, USA; Los Angeles Clippers forward Blake Griffin (32) and guard Chris Paul (3) walk on the court during the fourth quarter against the Houston Rockets in game seven of the second round of the NBA Playoffs at Toyota Center. The Rockets defeated the Clippers 113-100 to win the series 4-3. Mandatory Credit: Troy Taormina-USA TODAY Sports /
facebooktwitterreddit
May 17, 2015; Houston, TX, USA; Los Angeles Clippers forward Blake Griffin (32) and guard Chris Paul (3) walk on the court during the fourth quarter against the Houston Rockets in game seven of the second round of the NBA Playoffs at Toyota Center. The Rockets defeated the Clippers 113-100 to win the series 4-3. Mandatory Credit: Troy Taormina-USA TODAY Sports
May 17, 2015; Houston, TX, USA; Los Angeles Clippers forward Blake Griffin (32) and guard Chris Paul (3) walk on the court during the fourth quarter against the Houston Rockets in game seven of the second round of the NBA Playoffs at Toyota Center. The Rockets defeated the Clippers 113-100 to win the series 4-3. Mandatory Credit: Troy Taormina-USA TODAY Sports /

Given that we now know that the second game of back-to-backs has a noticeable negative impact on shot selection, I wanted to dig a little deeper into the specifics of when and to whom that drop in effectiveness happens. Though it is helpful to know that shot selection will suffer generally, not every drop off is equal – in other words, some legs might be a little more tired.

To examine this effect, I split the data from back-to-back games into bins representing the amount of time each player played in the first and second game of the pair.

aXPPS diff
aXPPS diff /

As expected, players who played more minutes ended up with more difficult shots on average. Note that this assumes that players were not rested on the front end of the back-to-back, which is a rare occasion. This suggests that there’s an expected efficiency cut off at around 30 minutes played. The following again looks at some of the elements going into the aXPPS:

aXPPS stats
aXPPS stats /

Similar to the patterns found in the previous post, greater playing time (and presumably fatigue) leads to slightly less distance between the shooter and defender, as well as more dribbling prior to shooting. These both have negative impacts on the aXPPS.

Yet even within a game, fatigue might not be equal, so I decided to look at the data quarter-by-quarter. The following has the raw numbers along with a the figure that illustrating difference from the respective mean.

aXPPSbyQ
aXPPSbyQ /
aXPPSGraphJPG
aXPPSGraphJPG /

We can see the noticable impact of playing >30 minutes and the huge effect of playing >35 minutes. These players take more difficult shots as the game goes on. The players’ shot selection rebounds in the 3rd quarter after a brief half time, but then crashes in the fourth. Fatigue is likely significantly responsible, but other factors could help exaggerate the difference. If a player is playing more significant minutes in the fourth quarter, the game is likely very close and the player is likely elite; thus, the opposing team will try to make that player’s shot selection as difficult as possible. This may even explain why role players playing less minutes have a higher aXPPS, for the defense is more focused on the star players.

To try quantify this, I looked at the average final margin for the score for those shots in the 4th quarter:

q4 final margin
q4 final margin /

As expected, those who played longer minutes are in closer games while players with minutes more associated with a back up role are playing in blow outs. So the better players are playing longer in higher leverage games. For all minute intervals, the standard deviation was about 6.9 points.

Interestingly, players who played >35 minutes are not taking more three pointers than usual in 4th quarter (27.70%). Given that an average of 27% of all shots are threes, the dip in 4th quarter aXPPS  comes from a representative shot distribution; thus, they are in fact taking shots from similar areas, but those shots are more difficult. For all other players, the 3pt rate jumps up to 30% of the shots. It would seem that these more difficult shots are coming from 2pt range.

This post seems to further the hypothesis that fatigue has a significant negative impact on shot selection. Perhaps the most important take away is the change that occurs at 30 minutes played. While there may be a hidden star effect, there is no doubting that shots get more difficult for players who play longer minutes on the back end of a back-to-back.