while we are at it- what is the +/_ in a box score?

[CTrefugee]

I often see a column where a player might be something like +2 or -4. How is that computed?
Thanks

 

[DavidinNaples]

In a basketball box score, the plus/minus (+/-) statistic measures a player's point differential while they are on the court during a game. A positive number means their team outscored opponents while they were playing, while a negative number means their team was outscored. This stat shows a player's overall impact on the team's score, regardless of whether the points were scored by them or their team mates.

 

[ucmiami2]

To me it is one of the least useful stats in basketball because who the other 9 players on the floor are when a player is in the game usually has much more impact on point differential than any single player. Quite often when looking at a box score the losing team's bench players will all have much better +/- than their starters, especially in blow outs - because they played most of their minutes when the opposing teams coach had called off the dogs and against the back end of the roster.

 

[Bone Dog]

To me it is one of the least useful stats in basketball because it who the other 9 players on the floor are when a player is in the game usually have much more impact on point differential than any single player. Quite often when looking at a box score the losing team's bench players will all have much better +/- than their starters, especially in blow outs - because they played most of their minutes when the opposing teams coach had called off the dogs and against the back end of the roster.

I tend to agree. It's a very rough-cut stat without a lot of nuance. I suppose if the data set were large enough, it could become quite meaningful, at least comparatively. But I'm not sure a single season is large enough.

 

[stamfordhusky]

To me it is one of the least useful stats in basketball because it who the other 9 players on the floor are when a player is in the game usually have much more impact on point differential than any single player. Quite often when looking at a box score the losing team's bench players will all have much better +/- than their starters, especially in blow outs - because they played most of their minutes when the opposing teams coach had called off the dogs and against the back end of the roster.

Yes, rarely useful. It started out many decades ago as a hockey statistic, but was only marginally more useful there. One classic example is defenseman Dallas Smith of the Bruins, who was in the top three in NHL history in plus/minus when he retired. But he was regarded only as a good but not great player. The reason for his extraordinarily good +/- is that he was normally paired with Bobby Orr for his entire career. That guaranteed a spectacular +/- if he only played average hockey.

 

[oldhusky]

Agree with those that say it depends greatly on who you are playing with, and the opponents lineup at the time. On the other hand some sites compute on and off plus minus statistics which have considerably more value but still some flaws. They look at the plus minus stats while you are in the game and compare it to when you are not, and then take the difference.

Taken in context with many other stats I think it has some value, but more based on a season than an individual game. I think the best overall comprehensive stat is the plus/minus at Sports-Reference.com. PER is very good too, but is measured against the competition you play with 15 being average. The problem is comparing players that have faced very different competition. A player in one of the weakest conferences might have a PER of 25 against their competition, but maybe it only would have been 15 if they played in a power 4 conference for example. The other issue is no adjusting for pace. Fast paced teams put up bigger numbers across the board in most categories.

Sports-Reference.com for their college basketball adjusts for both of those factors. In their case, the data they use is the totals from all the division 1 schools in all categories. Then the player's stats are adjusted for the pace of the team. Increasing them if they played on a slow pace team, and decreasing them on a real go go team so they are on a level playing field. Then a similar adjustment is made for the level of competition, increasing them if obtained against top competition, and decreasing them if they played in one of the weakest conferences.

In the end, as I understand it, their plus/minus is an estimate of how many extra points a player produces when compared to the average division 1 player, playing average competition, per 100 possessions. Now the average number of possessions per game I think is closer to 70, and a starter might be in for say close to 70% or more of a teams possessions, so I consider half of that site's plus minus to be roughly how much a starter would be worth vs. an average division 1 player per game as a ballpark.

 

[Jaydumo20]

I tend to agree. It's a very rough-cut stat without a lot of nuance. I suppose if the data set were large enough, it could become quite meaningful, at least comparatively. But I'm not sure a single season is large enough.

Well said. Also, our nuisance variable is the fact that we tend to blow most teams out by 50. This distribution skews waaaaay right