Avery Bradley shocking defensive +/-

[Donoghus]

Wouldn't you need to take a look at the 5 man units that Bradley is on the floor with?  Doesn't that help drawing a better conclusion to see where he's at (even though, I feel like these defensive stats are pretty noisy to begin with), by looking at who is is out on the floor with and each of their defensive capabilities?

[loco_91]

Yeah, this has been noted at various points over the last coupe years. It's a combination of two things:

1. Defensive statistics are inaccurate. 90% of defensive plays do not result in a steal or block, and team-based metrics (plus-minus) are heavily affected by lineups - for example, the +/- might just be telling you that Smart+ET are better on defense than AB+IT, which is very different from saying that AB is worse than average. These types of biases don't necessarily improve when you have a large sample size.

2. The eye test is also inaccurate.  As others have noted, AB's on-ball defense is excellent, and that's mostly what casual fans see when they watch people playing defense; people are less good at appreciating good team defense, which is something that AB is not nearly as good at.

Overall, I think AB is a very good defender, but I think that Crowder is at least as good as him and Smart is significantly better.

[Fafnir]

Honestly given all the things calculated into this box score derived metric its probably his poor rebounding. You can find the formula and derivation on basketball reference and while interesting I don't think its a useful tool for evaluating defense in the way you conventionally think of it.

[td450]

Wouldn't you also need to compare this parallel to each respective minutes True Shooting % to give a blanket statement like that some actual credibility?

Maybe guys are just hitting a higher percentage of their shots, by coincidence, during the times Bradley is on the court. Not necessarily who he's guarding, but other players on the opposing team.

Wouldn't you have to run each opposing teams/players stats to see if there was a dip or spike in FG% when Bradley is on the court. Plus you also have to factor Free Throws, which is why true shooting percentage came to mind. Maybe teams are getting to the free throw line more often when he is in the game.....and by more often, I mean by circumstance, not a dip in Celtics team-defensive performance.

When it comes to measuring D, we have raw, factual stats like BPG, SPG, DEF REB, but the rest is kind of based on "eye test", as mentioned before. I refuse to think that Avery Bradley is solely responsible for dips in overall team defense, when he happens to be on the court.

If players are hitting a higher % of shots when Avery is on the floor compared to off the floor over a period of 10,000 possessions, the probability of that being "by coincidence" like you suggest is incredibly incredibly small.

As for teams shooting more free throws when he's in the game (like in end-of-game scenarios), that's an interesting idea, but it would affect the ratings of all 5 celtics players on the floor, and Bradley (our lone all-defensive teamer) has the lowest rating out of every single one of our rotation players except IT, so I don't think that can be it.  Also, we're more likely to send them to the foul line in end of game scenario if we're a losing team, not a 48 win team.

We play Bradley against starters with IT being the worst defender on the team out of anyone that gets serious minutes, and Sullinger, who is second worst. Smart gets to play more with Turner and Jerebko or Olynyk, who are far superior defensive teammates, and they play more against reserves. It's not hard to see why Smart has better numbers.

[Androslav]

Was digging through some stats today and I found something interesting.  Defensive box plus/minus is defined as "A box score estimate of the defensive points per 100 possessions a player contributed above a league-average player, translated to an average team.  Avery Bradley has posted a negative defensive plus/minus compared to average players on average teams over the past three consecutive seasons (-1.4, -0.9, -0.3).  Further, let's look at Avery compared to the Celtics team defensive rating (points per 100 possessions) over the past three seasons:

2015-16:  Celtics - 103.6 pp100;  AB - 106 pp100
2014-15:  Celtics - 104.5 pp100;  AB - 107 pp100
2013-14:  Celtics - 107.7 pp100;  AB - 109 pp100

Our team's defense has been worse when he's on the floor for three consecutive seasons now.  This doesn't feel like it passes the eye test, but this is no small sample, we're talking about 6,819 total minutes of game time AB has played which I'd guess is somewhere around 10,000 or more possessions.

I always thought AB was a truly elite defender, a player that I put very high on my do not trade list for that very reason, but what gives here?  Is Avery Bradley an overrated defender?  Or does anyone have any alternative theories?

Avery is a terrific on ball defender, but his overall defensive impact I have always felt was far less than Marcus Smarts. However thats a high bar. RPM is a tricky stat. AB plays with IT and is backed up by Smart so it could be misleading.

That said with winning 1st team all defense I think his D is overrated I would love to sell high on him except Im not sure we can survive without his shooting.

TP action,
He is a small guard and that would suggest he will have a shorter prime.
His value is probably at its peak atm.
Rozier looks ready too.
One big selling point with ABs is that he can "guard", slow down, the Currys and the Lillards of the world.

 

[kozlodoev]

Overall, I think AB is a very good defender, but I think that Crowder is at least as good as him and Smart is significantly better.

The unfortunate part is that even so, Avery Bradley is still the better basketball player.

[mctyson]

Him and Sullinger had the lowest points per possession defending the pick and roll in the entire NBA.

He is really that good.

[Scintan]

Here we have yet another example of a stat being used without proper context, and then producing misleading results.

[PhoSita]

His backup was Marcus Smart.

[PhoSita]

Honestly given all the things calculated into this box score derived metric its probably his poor rebounding. You can find the formula and derivation on basketball reference and while interesting I don't think its a useful tool for evaluating defense in the way you conventionally think of it.

Avery's lack of rebounds, assists, free throws, and shots inside of fifteen feet have always torpedoed his efficiency / advanced box score stats.

[snively]

Honestly given all the things calculated into this box score derived metric its probably his poor rebounding. You can find the formula and derivation on basketball reference and while interesting I don't think its a useful tool for evaluating defense in the way you conventionally think of it.

Avery's lack of rebounds, assists, free throws, and shots inside of fifteen feet have always torpedoed his efficiency / advanced box score stats.

This is plus-minus - not affected by his own stats.

[byennie]

Any single stat measure is going to be highly questionable, but I do think it lends *some* insight.

Bradley is a great defensive weapon, because against certain point guards he completely disrupts the opposing offense. Offensively, we like to talk about "clutch" ability. I think on defense there is some equivalent where Bradley excels. For short stretches and/or against certain players he can suffocate the opponent. Those stretches don't get any extra weight in +/-, but they are important.

Others have pointed out the difficulty with +/- on a given team, and I think those are valid as well. The Cs have traditionally had a bench with good defenders, who then play against the other team's backups. To have a good +/-, Bradley has to outperform some of those (pretty good) bench players while often guarding the other team's 1st or 2nd option.

His value may have piqued, but he's not secretly and average defender. He's well above average for a starter, at least.