Robert Covington: top 30 NBA player?

What normally happens is that fans of a particular team overrate the players on their own squad, while fans of other teams rate those players properly or underrate them a bit. But in the case of Robert Covington, fans around the league massively underrate him, and it sometimes seems as though Sixers fans underrate him even more! Even the excellent article earlier this year by Fearless Leader KFL has apparently left some fans unconvinced of the excellence of Lord Covington.

In this essay I argue that Covington is not best described as a “decent role player,” or a “solid starter,” or a “quality professional,” but rather as a hidden star, a top-30 player in the NBA with a strong case at top 20. Given the enormous gap between the general perception of Cov and what I see as the reality, it’s going to take more than a few words for me to make the case, and there will be some digressions along the way into such subjects as Gilligan’s Island, card game tournaments, economics, philosophy, and statistics. The journey will be very enjoyable if, like me, you find that sort of thing interesting. If not, well, just skip the article and take my word for it that Cov really is that good!

1. Bridge to somewhere

When I was a kid, the card game Bridge was extremely popular. They ran bridge columns in the newspaper, and people talked about them, got together for games all the time, that sort of thing. My sense is it’s not nearly as big a deal these days, though I’m sure there are bridge fanatics now just as there were then. Anyway, do you know about “duplicate bridge”? It’s how they play bridge tournaments, how they determine who the best of the serious players is. As Wikipedia puts it:

“Duplicate bridge is the most widely used variation of contract bridge in club and tournament play. It is called duplicate because the same bridge deal (i.e. the specific arrangement of the 52 cards into the four hands) is played at each table and scoring is based on relative performance. In this way, every hand, whether strong or weak, is played in competition with others playing identical cards, and the element of skill is heightened while that of chance is reduced. Duplicate bridge stands in contrast to rubber bridge where each hand is freshly dealt and where scores may be more affected by chance in the short run.”

You can see what a great idea that is, right? How it makes it far more likely that the champions will be a team that played fantastically well, rather than just being dealt strong cards.

Well, there’s no such thing as “Duplicate Basketball.” We can’t have a game where Kevin Durant and four basketball-playing robots, Models PG, SG, PF and Model C (each representing a prototypical NBA player at that position and possessing the skills of such a player) take on a team made up of LeBron James and four robots who are identical in every way to the four models on Team KD. There are no basketball-playing robots... yet! This saddens me as the scene in which the Harlem Globetrotters take on a team of basketball-playing robots known as The New Invincibles in “The Harlem Globetrotters On Gilligan’s Island” was a highlight of my childhood.

https://en.wikipedia.org/wiki/The_Harlem_Globetrotters_on_Gilligan%27s_Island

2. Derek Parfit, science fiction, and the trolley-car problem

Switching gears for a moment: the greatest philosopher of modern times was the recently-deceased Derek Parfit, who pushed the field forward by creating science-fiction-like thought experiments. For example, in his classic Reasons and Persons he explores variants of the famed Trolley-Car Problem. The problem imagines that five people are unconscious on trolley tracks and will be killed by the onrushing trolley if you don’t throw the track switch. But if you do throw it, you will kill the one person stuck on the other track. Most people, after varying amounts of anguish, say they would throw the switch. Understandable! But then Parfit asks their feelings on the Fat Man Variant, in which there is one track with five potential victims on it, and a fat man overlooking the track, and no switch. Your choices are to let five people die or to push the fat man onto the track, killing him and saving the others. Is your answer the same? If not, why not?! If you can drop him onto the track and save the five lives by flipping a switch that opens a trapdoor, instead of having to push him, does that change your mind? What if he isn’t really all that fat, is that a difference-maker? And so forth.

Let’s honor Parfit’s memory by trying a thought experiment of our own. I swear this will eventually lead us to Rock Covington! Suppose the robot models above existed, and we had a pair of each, and we arranged a bunch of games between KD + robots and LBJ + (identical) robots. How would we use the results of the contests to determine which of Durant and James was the superior player?

Well, you could look to see which of the two humans scored the most, or had the most points + assists + rebounds, or blocked the most shots, or had the fanciest dunks. You could use more complicated combo stats like PER or BPM to see who’s stats were superior. But I dare say those would all be silly choices. All we’d want to do is to see whose team did the best, who won the games! After all, we have controlled for every player on the court except our two stars. If LeBron plus typical NBA players (which the robots are coded to be) consistently beats KD and identical players, that is strong evidence LeBron is the superior player. Not incontrovertible evidence; maybe robots with different skill sets would favor KD; you’d really want a whole season of games in which they played with different kinds of players/robots, but, following the principle of Duplicate Bridge, making sure that they each got to play with and against equally good players across the many games. By controlling for both teammates and opponents, we could conclude with reasonable confidence at the end that the guy who won more is indeed the better player.

3. Physics envy

OK, one more seemingly-unrelated topic and then I’ll pull all the threads together. Economists are often said to have “physics envy.” This can actually refer to one of two things. Economists envy physicists their incredible brilliance and mathematical abilities. But economists also envy the fact that, while Galileo could drop two balls of different weights off the tower and see that they ker-thunked at the same time, economists cannot, say, observe the economy with one interest-rate policy and then turn back time and run the economy again with higher rates and see how growth is affected. Economists try to address this challenge by working with statisticians to create special tools (“econometrics”; often variants on linear regression) that attempt to capture as much as possible of the benefits of the controlled experiments economists are so envious of. So, for example, an economist might look at 50 countries over 20 time periods each, creating 1,000 separate periods, and then look at interest rates and economic growth in each of those thousand country-periods, and attempt to discern a pattern. Of course — just as one example of the frustration that comes from not running controlled experiments — one wouldn’t want to reach conclusions about the economy of a wealthy country like the USA by using, without adjustment, data primarily from much less developed countries, so care must be taken to control for differences among nations. That’s where econometrics comes in.

4. RAPM

The upshot is, we can’t create a “duplicate basketball” season in which KD and LBJ play games with matched sets of teammates. But we can take the actual seasons of games played by hundreds of players, and use that data to estimate how each player would do if they were neither benefitted by, nor harmed by, their context. We can estimate the results of Duplicate Basketball. Methods that do this are called “adjusted plus-minus” methods. Of the different variants of adjusted plus minus approaches, the most commonly used is RAPM, “Regularized Adjusted Plus Minus.” It’s the best player-evaluation statistic I know of. RAPM tries hard to do what we would do if we had duplicate basketball. It doesn’t rely AT ALL on counts of rebounds or blocks or three-point percentage. After all, that stuff isn’t what really matters, all that really matters is: did the team score more or fewer points because you played; and what about points allowed, how were they affected by your presence?

RAPM starts with simple on-off figures — how did the team do when you played and when you didn’t? It then adjusts for who else you played with and against, so that if you got to play with Steph and KD and Klay, which will mean your team scored a lot and allowed relatively little when you were out there, the credit for that good performance will be distributed among all the participants, with the best players getting the most credit, so that you don’t look artificially good just because of your teammates. But then it goes a step farther and takes into account who you played AGAINST, so that if you are a defensive savant who gets consistently matched up against other teams’ stars, your RAPM doesn’t suffer for that. In the end, RAPM delivers a single number which amounts to the answer to “how many more points per 48 minutes does your team score, and how many fewer does it allow, with you on the floor than they would with an average player in that spot?” Which is as close as we can reasonably get to a measure of what it means to be a valuable NBA player.

What’s great about RAPM is that it accounts for everything you do that helps or hurts your team. Points and assists and rebounds and steals and blocks and deflections of course, but also uncountable things like the extent to which you spread the floor, and other stuff too, like boxing out, man-to-man defense, help defense, diving for loose balls, and all the bad stuff too; if a guy blows by you to score, if you miss a free throw, if you get in your teammate’s way on defense, you name it, anything that affects the score, affects your RAPM.

The only problem of any consequence with RAPM is the one shared by all statistics: it is affected by randomness. If you’re on the floor when an opposing player hits a lucky half-court shot, it hurts your RAPM. If your teammates go cold from the free-throw line while you’re playing and get hot when you’re on the bench, you’ll look unfairly bad. That sort of randomness washes out over time, but slowly — over one game, luck plays a huge role, and even over a full season, some players will be meaningfully overrated or underrated by this measure. Of course that’s true of all other measures too, sometimes a great scorer gets fewer points than his ability warrants because of bad teammates, the defense focusing on him, bad luck, etc., so traditional measures like “scoring average” are also deeply flawed at measuring what they attempt to.

With RAPM there are a couple things you can do to try to alleviate the problem. The simplest is to just get more data — if a guy has a great RAPM one year, he’s probably great, but if he has a great RAPM over 5 years, he’s pretty much DEFINITELY great. The other way you can deal with the problem is to supplement RAPM with other information that helps reduce the noise; for example you could use various methods to mix in counting stats like points and rebounds. ESPN has a stat called “RPM” which they don’t reveal the methodology behind but which probably uses the second approach (and may also use some of the first). RPM is a good shortcut to use as it’s easily accessible on ESPN’s website and it makes a sincere, though of course imperfect, effort to reduce the noise in RAPM. I will reference both RAPM and RPM in my writing and will try hard not to cherry-pick from them by choosing the one that best supports my argument.

5. Who has the highest RAPM?

Do you know who the number one player in the entire NBA was in RAPM last season? Go ahead, guess! Yes, that’s right, it was Robert Covington. No player in the entire league added more value per minute by this measure than our Cov, who was at almost +6 per 48 minutes. Steph Curry was second. Joel Embiid was fourth in the league, Ben Simmons was around 25-30 among starters, there were also a half-dozen or so part-time players who outranked Ben, as RAPM is per-minute and it’s easier for short-time guys to get a lucky RAPM draw. Remember, Ben’s numbers include his mid-season slump, the Ben we’ve seen over the past two months would rank a lot higher.

Now, does that mean Covington is the best player in the NBA? Of course not, Robert Covington’s own mother probably doesn’t think her son is the best basketball player in the world! LeBron James is probably a +10 when he gives his all, but wisely he doesn’t make that sort of effort over every minute of the 82 games. But isn’t even half-speed LeBron, regular-season LeBron, more valuable than Cov? In other words, granting that Cov is not AS GOOD as the very best, not capable of their heights, did he nevertheless contribute as much as the best players this season, a result of him giving 100% and them saving something for the postseason?

Maybe, but probably not. Probably Cov was a beneficiary of some luck. As I mentioned above, RPM tries to remove, or at least reduce, the luck factor, and RPM ranks Cov as the 7th-most-valuable player this season. Last season RPM had him 25th. The average of those is 16 and that feels about right to me; I’d say Cov is a top-20 player in the NBA but not top 10. If you think Cov is 22nd or even 32nd, not 16th, well, maybe you’re right, that’s a modest difference, and our tools are blunt. But if you think that 16th is way off, that Cov is, say, 43rd or 65th or 90th, go ahead and list your top players. My prediction is, before you get to 40 you’re going to list a guy like DeMar DeRozan or Kristaps Porzingis. And, honestly, it just couldn’t be clearer in the data that Cov does far more than those guys to help his team win. To think otherwise you really have to believe defense has almost exactly zero importance. Now, I know the typical basketball pundit isn’t convinced; to them DeRozan is a perennial All-Star; after all, the pundits are among the ones who vote him onto the team! Those pundits probably think adjusted plus-minus is stupid, just some meaningless numbers nerds like. And to that I’d say this: it’s exactly the opposite! The thing a nerd who didn’t understand basketball would do is count things like “points scored by an individual” and “rebounds taken by an individual.” I.e. traditional metrics are OCD nerd material. Whereas RAPM is the opposite: ALL IT CARES ABOUT is whether your TEAM did well or badly by putting you in the mix. In the past people made up “advanced” statistics like PER that simply added up weighted sums of points and rebounds and blocks and such. That was a reasonable thing to do when those numbers were all we had. But now we have something much better — we actually have data on which players cause the team to play winning basketball. Usually it’s the same guys who get lots of points and rebounds and blocks. Most of the top RAPM guys are famous superstars, and that’s especially true if you average multiple years to wash out noise. Over the period 2001-2014, the top five players in RAPM were:

Kevin Garnett

LeBron James

Tim Duncan

Chris Paul

Dirk Nowitzki

Those might not be exactly your personal top 5 of that era, but you can see it’s not a crazy list. And remember, this is based on nothing but TEAM accomplishments when players were on the floor, it didn’t use any information at all about these players’ points or rebounds or steals. So that should provide some comfort that the method is credible. And when I tell you that Manu Ginobili was sixth and Tracy McGrady 12th,, well, that certainly doesn’t prove Manu was superior, but if someone tells you that Tracy was a legitimate superstar bucket-getter and Manu a mere role-playing sixth man, this is real reason not to take that way of seeing things seriously. And similarly, when you see a one-year RAPM list, and it has Curry and Paul and James and Butler on it, and then you see a Robert Covington or a Jrue Holiday in the top 10, as both were this season, well, that’s telling you something. Not that Cov and Jrue are better than CP3 or Butler, but that they are a lot better than people previously thought, that they are in the top-20 conversation, or should be.

6. Cov and the Sixers

As I say, my guess is that right at this moment Cov is around the 15th or 20th-best player in the NBA, around a +4 player. He’s 27, so he’ll probably be similarly good next year and start declining sometime between two and four years from now. Of course this is all just educated guesswork. But the biggest mistake people make in mis-rating the Sixers is thinking of Covington as “just a role player” or words to that effect. You can argue whether he was our best player, or our second best after Joel, for the season as a whole. Since Ben has improved so much late in the year, you can make a case that right now Cov is only our third-best player. But that’s not because Cov is “just OK.” It’s because we have a team with extraordinary talent — probably 3 of the league’s top 20; amazing!

I’ll just add here that I know many, many Sixer fans cannot accept that Cov is any good at all, let alone one of the best players in the NBA. Since I am blind I cannot see the games. It’s still painful to me when I hear Tom McGinis moan after Cov misses a layup or makes a bad pass, but, no doubt, not as painful as it is for those who are watching it. And all I can say is, sure, that stuff is annoying, and if Cov didn’t do those things, he’d rate even higher; just as Joel would rate higher if he cut his turnovers and Ben would rank higher if he hit more of his free throws.. But Cov is still an extremely effective offensive player; he not only shoots around 37% from three on heavy volume, he does it despite taking many, many shots with a defender nearby. Cov gets guarded much closer out there than, say, Dario Saric, and what that means is, he spreads the floor very effectively, opening space for others to score. He has the occasional frustrating turnover, but frustrating turnovers don’t count more on the scoreboard than less-frustrating turnovers, and overall Cov’s turnovers are quite low (plus his TO/36 have plummetted in each of the past two seasons, so the trend is excellent). This season his offensive RAPM was almost +2; that was probably partly luck, but I’d be quite surprised if he isn’t above +1 offensively over the next few years. +1 is really impressive; the average starter is +1 OVERALL, i.e. something like +.5 on offense and the same on D, so Cov is, as best I can tell, an above average starting SF at the offensive end.

And defensively, well, he’s just a one-man wrecking crew out there. People see him get blown by from time to time and think he’s letting down the team, but it’s just very difficult to judge a player’s overall defensive contribution by eye. We see all the steals and blocks and deflections, though even seeing them I suspect we underestimate their importance. But we don’t see the pass that doesn’t get made because of fear of the long arm of Robert Covington. We don’t see the easy dunk that would have happened with even a good defender at SF, but that doesn’t happen because ridiculously-good Cov is playing.

What we don’t see, but are able to measure later, is how many points the Sixers let up when Cov is playing and when he isn’t, with Cov and Amir vs. with other SFs and Amir, with Cov, Joel and JJ vs. with Anderson, Joel, and JJ, and all the other combinations. And what that data shows us is that teams score 3 or 4 fewer points per 48 when it’s Cov on patrol, compared to an average defender. If you don’t like the fancy regression stats, just look at the simple on-off, you’ll see he does as much for our defense as even the amazing Joel Embiid. Look at the lineups that work and you’ll see Cov’s name in almost all of them. He’s an amazing player, and we’re lucky to have him, and we’re super-lucky to have him for a fraction of the max salary.

There’s a Fanpost poll up now asking if we should trade Covington and the Lakers pick for Andrew Wiggins. About half the respondents voted “yes.” This is pure insanity! Andrew Wiggins is a promising young player and may someday be terrific. But Wiggins, together with his contract, has negative value to the Sixers. I honestly wouldn’t trade even “cash considerations” for him, and it’s not my money! I’d rather have the cap space to sign a real star, like LeBron this summer or Butler next summer. Whereas Cov, with his conttract, is almost certainly among the ten most valuable assets in the NBA. If Kawhi is physically and mentally healthy, well, that guy is so unbelievably good at both ends that we should try to get him, and even consider including Cov in the deal if we absolutely, positively have to. But I would not trade Cov straight up for Paul George. i wouldn’t trade him for Klay. He’s as good as those guys, and quite possibly better, plus is cheaper and signed for longer. And he adds all that value without using up many touches, and with the talent on this team looking ahead, finding enough touches to go around will be difficult; people who can help the team while giving Ben and Joel and Dario and maybe Markelle and LeBron the shots they deserve are going to be the kind of players we most desperately need.

7. How could Cov possibly be that good when he doesn’t look all that impressive?

I can’t see the games, but I used to be able to and so I understand what people mean when they say that a guy like DeRozan looks dramatically more impressive than Cov. How could Cov be better? I suggest an analogy to baseball. When we finally got really good at evaluating defensive play in baseball, one of the big shocks was who turned out to be the good defensive shortstops. Some of the guys, like Ozzie Smith, who the eye test told us were glove wizards were indeed among the best ever. But in other cases, our eyes misled us. The players that impressed with the leather were those that ran quickly to the ball and made athletic, diving stops. But it often turned out that the most effective shortstops were those who knew where to play the hitters and positioned themselves so that they didn’t need to run and dive; the ball was hit three feet to their left and they simply took one step and speared it easily. Moreover it turned out that some slow shortstops with strong arms — often players like Cal Ripken who were thought of as ordinary fielders who earned their playing time with their bats — were devastatingly effective in the field. The key was that they could play the hitters several steps deeper than other shortstops, increasing hugely the “cone” of the field they were able to cover as the ball took longer to get to them. They made up for that extra fraction of a second lost with their cannon arms, gunning down runners at first from the edge of left field.

My point is, we tend to overrate fast, athletic players who move impressively and underrate players who are where they’re supposed to be, when they’re supposed to be there. Famously a key to the Sixers’ last title was a trade of the high-scoring athlete George McGinnis for the right-place, right-time Bobby Jones. Robert Covington is this team’s Bobby Jones, except he’s even better than Bobby because he is so skilled at today’s most important offensive weapon, the three-point shot. Before the season I spent a lot of pixels here at LB singing the praises of Amir Johnson. Many, many fans were puzzled — how could Amir be as good as, or better than, an athlete like Richaun Holmes. But here’s the thing: it’s not just that Amir is a better player than Richaun. It’s that even at the specific thing you’d think an athlete like Richaun would be great at and an earthbound old man like Amir would fail miserably at, Amir is just as good. Amir’s blocked shots per minute were just as high as Holmes’ because Amir was where he needed to be. The extra two feet off the ground Richaun can jump was not enough to overcome Amir’s edge in positioning and savvy. Cov has the Amir virtues without the limitations that Amir’s ankle injuries have created.

8. What a great team needs

So that’s part of it, great athletes like Cov are underrated when compared to mind-bogglingly, absurdly-great athletes like Andrew Wiggins. Just at the offensive end, Cov is the better player right now, but it doesn’t look that way because Cov doesn’t run as fast or jump as high.

But there’s another element of Cov’s underappreciation, which is that people notice offense more than defense, and appreciate it more. After all, when you watch a game, it’s just normal to watch the ball, and to watch baskets. A player who scores lots of points is going to be noticed and appreciated. A player who is shutting down his defensive assignment, denying him the ball... well, that’s not all that noticeable, as the ball never even goes there!

And yet even that understates the issue. It’s not just that a player with +3 offense and +1 defense is going to seem more impressive than a guy who offers the reverse, +1 offense and +3 D. It’s that, for a good team, a team with stars, a team like the Sixers, the player who adds three-quarters of his value at the defensive end is not only less appreciated, he’s MORE valuable! Let’s say we sign LeBron James, now we have Ben, Joel, LeBron, Dario, and maybe JJ and Markelle in the rotation. How much benefit would we get from adding another scorer to the mix. Every shot that guy takes is a shot that LeBron isn’t taking! That Joel isn’t taking, that Dario isn’t taking. It had better be one hell of a shot to be adding ANY VALUE AT ALL. So if you could trade Cov for DeRozan in that situation, how does that help you? You’ve hugely harmed your defense, and what you get in return is replacing some LeBron possessions and Ben possessions and Joel possessions with DeRozan possessions. Of course I’m not saying DD doesn’t help your offense at all, he’s a superior offensive player to Cov and adding him as a weapon would of course help at that end. But as the saying goes, there’s only one ball, and as a consequence adding another offensive player to an excellent offensive group suffers from diminishing returns. Whereas that’s just not true on D; no matter how good your defense is, adding one more defender still helps hugely reduce points. And a great D with one weak link is not that great a D after all. So for a team like ours, with so many excellent two-way players and quite likely more to come as we develop Markelle, sign free agents and use our top-10 pick, a player like Cov is the very best thing you can add other than a two-way superstar like LeBron or Kawhi.

9. Doubting Thomas

The Apostle Thomas is known to history as “Doubting Thomas because he refused to believe that Jesus was really dead until he touched the wounds with his own hands. I myself am a doubter; when I have reason to think something that others find implausible, I question my own views, and I try to develop ways to further test my conclusions. RAPM really doesn’t have any natural “biases”; there’s no way it can artificially favor rebounders or defensive players or shooters or something. But as I’ve noted above, it is affected by luck. Is it possible Cov has just been lucky?

One way to explore this is to consider RPM, which tries to reduce the effect of randomness. As mentioned Cov was #7 in the NBA this year and #25 last year in overall RPM. What about the previous two years, when he was just learning his craft? I don’t have RAPM for those years, but RPM had him around +2 both seasons, in each year he ranked as the #11 SF. That is really impressive for an undrafted and super-inexperienced player like Cov was at that time.

Here’s another idea: we can look at similar players. If Cov is a so-so player who was just lucky, then players who play similarly to Cov but are on different teams with different teammates and different luck will likely fare poorly. Who’s the most similar NBA player to Cov? I bet that more than half of you have the same person in mind: Otto Porter, a similar 3-and-D guy. I just heard Washington fans complaining on a podcast about how overpaid he is; after all he’s “just a role player.”

Except you know what? Otto Porter is the best player on the Wizards. He was 8th in the NBA in RAPM this year, and 11th in RPM. On-off stats said he was a stud last year too; actually in RPM he was one spot ahead of Covington, at 24. Was he also lucky? This is getting to be an awful lot of luck!

Obviously we know that Sam Hinkie really valued Cov, he wrote extensively about him in his resignation letter. You know who else liked Cov? Hinkie mentor Daryl Morey, who runs what appears to be the best team in the NBA. He picked up Cov, and when he had too many good players to use a roster spot on a developing player like Cov and lost him, he nevertheless had some of the Coviest players in the NBA, like former Process Sixer Luc Richard Mbah a Moute, and Trevor Ariza. Guess how those guys have fared in RPM and RAPM lately? That’s right, they’ve been excellent — not as excellent as Cov, as they are in their 30s, plus Cov got a little lucky this year and presumably these two had a mix of good and bad luck. But excellent nevertheless. Joe Ingles is another guy in this group. Jae Crowder was in this group under Brad Stevens, his game fell off a cliff in Cleveland; the future is guaranteed to no man! OG Anunoby appears poised to join the club as his game matures over the next few years.

10. 3 and D

It turns out that 3-and-D guys are exceptionally valuable. Probably more valuable as a group than other types; initiators, rim protectors, bucket-getters. The only group of players who help their teams a lot more than guys like Cov and Porter are the dozen or so players who can do almost everything; shoot, drive, pass, rebound, defend. That list includes LeBron and KD and Kawhi and CP3 and Jimmy Butler and a handful of others. Oh and then there’s Steph, if you can do one thing as freakishly well as Steph shoots from distance, and if that one thing is shooting from distance, and if you’re pretty decent at the other stuff, as Steph is, then you can be as good as the very best without being as well-rounded as LeBron or Kawhi. But beyond this short list of superstars, there are two types of NBA players: a dozen or two guys who are about as good as Robert Covington, and hundreds of guys who are worse than him.

I welcome attempts in comments to persuade me that i’m wrong about Cov, that in fact he’s an easily-replaceable role player who is less valuable than Dario and JJ and that my suggestion his value is in the ballpark of Joel’s and Ben’s at this point in their careers is laughable. But I’ll especially appreciate it if those attempts give me some basis for me being wrong other than that “he went 1-for-9 one night” or “he misses too many bunnies, or Dragic blew by him twice in a game last week.” Adjusted plus-minus counts all the missed threes and all the missed bunnies and all the bonehead passes and all the blow-bys. It’s all in there, and nevertheless game after game, year after year the team is just vastly better with Cov on the floor than with him off. So if you still think he’s ordinary, please explain why on good teams and bad, with superstar teammates and jokers, Cov always seems to make things better. Really there are only three possibilities:

1) Cov is an excellent, near-elite or elite player.

2) The team has been incredibly lucky whenever Cov is on the floor, consistently all season, and in past seasons too, and this is also true of the group of players whose stats are similar to Cov’s

3) There is some strange way that I’ve been unable to think of in which on-off statistics are biased toward Cov and players like him

Only the first seems likely to me, but if you think it’s one of the other two, tell me which, and what makes you think so.

11. The future

Robert Covington is 27 years old, by most accounts the age at which players peak, though it’s possible that with medical advances the peak age has crept up to 28. So this year could easily be the best year he ever has; at the minimum he sure isn’t going to finish any better in RAPM than first!Meanwhile Joel Embiid and Ben Simmons are years younger and are extraordinarily inexperienced; it’s entirely possible that their ceilings are far, far higher than even the exceptional play they’ve delivered lately. Consequently Joel and Ben are likely to be better than Cov next year, and far, far better in 2-4 years, when Cov is 30 and they are getting to around prime age. So I hope I’m not misinterpreted as suggesting Covington is as important to the Sixers’ future as Embiid or Simmons — that would be silly! I think there IS a strong case that Cov is a more important asset than Dario Saric; I love Dario and he has a very bright future but he is 24 to Cov’s 27, only three years different, and has been playing pro hoops longer than Cov, and Cov is enough better than Saric at this moment that it’s quite likely Cov will be the better player for 2-3 more years, possibly longer. Or not, maybe Dario’s game will continue to rocket upward; I sure hope so! The more important comparisons are to other players in Cov’s age range, such as the ones I mentioned above, to Klay Thompson and Paul George. Cov really is approximately as good as those guys, with a far superior contract, and so any view that we might want to sacrifice Cov as part of an effort to spend assets to obtain a player like that is, in my opinion, misguided. We should only give up Covington if we can get a superstar like Kawhi or Butler or LeBron, and ideally we would find a way to hold onto him even in adding such a player.

12. Playoff basketball

Oh, I guess before I finish I should mention playoff basketball. Everything I’ve written here is based on regular-season data, as Cov has only played five playoff games in his career as of this writing. If you think I am overrating regular-season Cov, let’s discuss that. If you think I’m right about regular-season Cov but that his game won’t translate to the playoffs or something, then please FIRST make clear that you agree with me about the regular season, and then make your case about his playoff flaws, recognizing that all the data we have is from a few games against a single opponent. My forecast is that Playoff Cov will turn out to be right around as good as normal Cov, but feel free to make your own predictions here and we’ll see what actually occurs. Hopefully Cov will be playing several more playoff rounds this season so as to provide us with plenty of data to analyze!