Getting Defensive: A Search for Colorado's Best Defender

Yesterday, Pacific Takes posted an article looking at the top contenders for the Pac-12 Defensive Player of the Year award. They listed six candidates: Gary Payton II (Oregon State), Kadeem Allen (Arizona), Ivan Rabb (Cal), Chris Boucher (Oregon), Jakob Poeltl (Utah) and the one and only Josh Scott.

Personally, I'd vote for Payton or Boucher, but I am not here to discuss who should win the DPOY. I am here to answer a question inspired by the aforementioned list: Who is the better defender between Josh Scott and Wesley Gordon?

To answer that question, I looked at both player's defensive stats on Basketball-Reference. Before the big reveal, some of these stats need some explaining. First, defensive rating is how many points per 100 possessions a player allows. 105 is about average for a player and anything below 100 is considered excellent.

Second, percentage stats are based on a single possession. For instance, a player with a block rate of 4.2 means that he blocked 4.2% of all shots when he was playing defense. Likewise, a player with a 14.6 rebound rate means they rebounded 14.6% of all available rebounds.

Third, box plus/minus stats look at the team performance when a given player is on the court. A player with a positive score means that they were a net positive for their team when they were on the court. Conversely, a player with a negative score means that their presence on the court was detrimental for his team.

Now for the data.

Perhaps the biggest takeaway is that Josh Scott leads almost every category, but just slightly. With Scott's leads negligible -- which could be a result of simple variance -- it's impossible to conclude that he's the better defender. We need to determine which stats to ignore and which to pay attention to.

Gordon and Scott have nearly identical defensive ratings and defensive win shares, and given the amount of shared playing time, it's a testament to how well their defensive games mesh. Still, there's nothing here to tell which defender is better.

Scott leads in defensive rebounding by a bit, but he may not be the better rebounder. Gordon isn't always in rebounding position because he's altering the opposition's shots. If Scott contested shots more than Gordon, these numbers could easily be flipped. After all, we've seen on the offensive boards just how gifted a rebounder Gordon is. On that note -- and please excuse this tangent -- Gordon is first in the Pac-12 with 3.8 offensive rebounds per game with a huge lead on second place Thomas Welsh (3.1). His offensive rebound rate is also first (14.4%), a full percentage point ahead of second place ... Tory Miller. I wonder why Colorado is far and away the best rebounding team in the conference.

Back on task. Since neither Scott nor Gordon have a particularly impressive steal rate and because they're identical figures, that's a stat you can quite obviously ignore.

That leaves us with defensive box plus/minus, block rate and fouls per 40 minutes played, three stats Gordon leads, for better or for worse.

First off, I'd like to denounce the myth that blocked shots equate to good defense. The main problem with this idea is that many players hunt for blocks (looking at you, Hassan Whiteside) for whatever reason. Not only does this aggression leave a player susceptible to foul trouble, but it cedes valuable position down low. Once the player is in the air he has little control of where he ends up. A heady offensive player can recognize this and either pump fake or dump the ball off to a teammate, who can take advantage of the temporary 4-on-5 opportunity the would-be shot blocker so eagerly gifted the offense. Furthermore, weakside shot blockers lose their valuable rebounding position (I mentioned this in discussing Gordon and Scott's rebounding rates) as the player they just abandoned to block a shot -- who is usually on the lower block -- remains in prime rebounding position without much contention. As such, some defenders rack up bazillions of blocked shots, but because of the downsides to their hunting, they're actually a net negative for their team's defense.

Wesley Gordon is a terrific and prolific shot blocker, but he does so without being overly aggressive. Gordon's service as a rim-protector give the Colorado defense much needed insurance, especially against dribble-drive oriented teams like Oregon State.

Speaking of Oregon State, Gordon is capable of annihilating an opponent's chances of scoring at the rim when he's on. Against the Beavers, Gordon had six blocks to one foul, and he was in good enough position to snag seven defensive rebounds. He had a career-best defensive rating of 72 that game. Any way you look at it, Gordon had a major hand in the Beavers' sub-30% shooting performance with his dominant interior defense.Gordon posted similar numbers against Colorado State and the Washington schools at home (more on that in a bit).

However, Gordon can sometimes play sometimes play himself into foul trouble -- as he's done increasingly often in conference play -- and he's required to tone it down a notch or two. In these instances, his defensive value is diminished.

Josh Scott is also a terrific blocker, but in a different sense. His 1.6 blocks per game are more indicative of his one-on-one defense than his rim-protecting. Even more impressive, he does so without fouling. As I've discussed before (skip to the bottom), Scott is able to block shots without fouling because he uses impeccable timing to block shots early in the shot attempt.

Contesting early means that the shooter has the ball lower, which mitigates the need for the defender to jump to contest. Doing this is incredibly difficult because without perfect timing, fouls are easy to accumulate. As noted earlier, jumping to contest has all kinds of downsides, which Scott seems adept at avoiding because of his unique skill.

Because Scott is a machine, his interior defense never takes a day off, but he will never take over a game with his shot blocking like Gordon can.

On the issue of shot blocking, Gordon and Scott are both excellent in their own ways. If you take Scott, you're siding with consistency over the potential to dominate. And vice versa if you take Gordon.

Lastly, there's Gordon's sizable advantage in defensive box plus/minus. And yes, a 1.6 point advantage is sizable considering how close Gordon and Scott are to each other in the other metrics.

If there is one problem with box plus/minus stats, it's that it can be too team dependent. Let's imagine something fun. There's a player -- let's call him something random like Yabe Gork -- who just kind of stands around on defense doing just enough to avoid cratering his team's defense. Meanwhile, his four surrounding teammates are all excellent defenders and work harmoniously on their switches. Because of the efforts of those four and despite the nonchalance of the other, the team puts up above average defensive numbers. Each player then gets above average box  plus/minus numbers, even the nonchalant Yabe Gork. If Yabe Gork was replaced by a better defender -- let's call him Adeem Kallen -- the team's defense would have great numbers. However, box plus/minus stats don't recognize that difference (adjusted versions of this metric are much better).

Of course, I don't mean that Gordon or Scott have their box plus/minus stats buoyed by their teammates (it's probably the opposite that's actually occurring). What I mean is that Scott and Gordon buoy each other's numbers. Thus, it's rather difficult to tell which is actually a better defender unless one of them was out for a bit of time.

Well hello there, UDub and Wazzu. Without Josh Scott playing, you two were ever so close to upsetting the Buffs at home. Without Wesley Gordon playing the way he did, you two would have walked away with a convincing win

The home game against Oregon State may have been Gordon's best defensive game of his career, but his performances against the Huskies and Cougars are without a doubt his most important performances. In those two games, Gordon had 32 points, 25 rebounds (12 offensive) and 8 blocks. I'm not here to talk about his offensive (which was rad as hell), but I can comment that his defense was absolutely critical for the Buffs, especially down the stretch. In these two games, he showed that his defensive value isn't tied directly to Scott.

Luckily for this article (not lucky for the Buffs at the time), Scott has also logged some time without Gordon playing well, if at all. Against Iowa State, most notably, Colorado was able to hang in with an elite offense by stifling them with great defense. That defense was despite Gordon fouling out after 20 minutes of mediocre play. Scott didn't play well on offense because Iowa State's Jameel McKay would make DeMarcus Cousins look like Kosta Koufos. However, Scott was excellent on defense, as he there all game to pressure the Cyclones' litany of weapons. He had a defensive rating of 82 in this game, second on the Buffs to Xavier Talton at 81. As Gordon did against the Washington schools, Scott proved his defense can be superb without his running mate.

It seems that we have yet another wash.

At the end of the day, Colorado is lucky as can be to have two defenders as deft as Scott and Gordon. The two compliment each other perfectly, especially on the boards and on the defensive block. One of the two may be a better defender, but I've been unable to decipher which of whom it is.