Good thread @durvasa. I'm with @A_3PO and the others. Gotta watch them play. However I do look for statistical clues when I'm mining data and scouting players. Steals and steal percentage, defensive fg% and defensive .efg and defensive rebound rate (if looking at bigs, defensive rebound rate sometimes gives you a clue about a players positioning instincts) However everything has to pass the eye test. Gotta see them play defense. Because just mining data will cause you to miss good defensive players on bad defensive teams. And not watching them play will cause you to think a guy is a really good defender when he isn't fundamentally sound, he plays rogue in a system and gets away with it because of the discipline of the other defenders. IMO, defense is such a team thing. Scheme and effort are what makes TEAMS great defensively a long as you have a couple guys who are really good defenders in your rotation. If you get two very good defenders in the top 7 rotation (one guy at the rim and one guy defending the point) and the other guys are just average but they work hard and put in the effort defensively and they have a great scheme then you can have a great defense.
No I agree. Just as stats can be misleading so can biases and reputation. It’s a reason Kobe made all defense teams later in his career. So, while nothing is perfect, I still trust an advanced scouts eye test than a collection of advanced stats and data. But in regards to laymen and fans(who likely don’t know what to look for) trying to look round the league and find good defenders who may not watch a lot of games outside their home team; I think collecting as much data as possible is the easier and more efficient route. But any one stat regarding defense just isn’t available and even a good collection of data doesn’t paint the whole story. I would love for a stat to come out but it hasn’t. Maybe Morey has something but he is quoted saying their isn’t a good defensive stat available
This is a very good point and it’s the emphasis behind analytics in the first place. #s and data are unforgiving and aren’t biased except for the parameters you program it to decipher. Offensively, it has done miracles to the advancement of the modern game and specifically the evolution of offenses and the 3 ball. But defensively, while extremely difficult to decipher, is in large part a big failure in regards to the nba analytics scene. It gets a pass becaus of the extreme difficulty to put a number or value of a players defense on a team concept and team design.