:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/31291487/20140331_krj_ai4_0496.0.jpg)
In terms of better understanding basketball we have to look at both the game and the data from the game. The data from the game merely quantifies the behaviors that we see and value. And the data, be it scouting information, or film, or just numbers are serve one purpose: helping to identify good behaviors and poor behaviors. Sorry for bringing this back down to B.F. Skinner levels but that's really what it is all about. You see good things on the court with your eyes, and you confirm that with data.
This is what Utah Jazz head coach Tyrone Corbin would be likely to call "that's just your basketball game."
Indeed, good sir. Indeed.
So we don't need more stats, right? We already know a few things about our team with the stats we have. Mainly:
- Utah is the worst team in the conference
- They have one of the worst offenses in the league, by production
- They have one of the worst defenses in the league, by proportion
- They play at one of the slowest paces in the league, by value
- They do not force a lot of turn overs on defense
- and they seem to make a lot of pointless passes on offense
Furthermore, we know that the majority of our team is below a league standard (lower than 15.0 PER), and some of our starters are one-dimensional. We know our team. We don't need new numbers, right? We even know who is the best player on the team. It's Gorrack Favward. (Who is also my next Dungeons & Dragons character.)
Well, what does ESPN's RPM have to say about it? (Btw, did you know that RPM is waay old and used to be called something else?)
| ESPN Ranks | |||||||||
| Player | RPM | NBA | PG | SG | SF | PF | C | Rk | |
| 1 | Gordon Hayward | 1.47 | 95 | 15 | 1 | ||||
| 2 | Alec Burks | -0.08 | 158 | 33 | 2 | ||||
| 3 | Marvin Williams | -0.35 | 174 | 43 | 3 | ||||
| 4 | Derrick Favors | -0.99 | 200 | 40 | 4 | ||||
| 5 | Trey Burke | -1.66 | 254 | 49 | 5 | ||||
| 6 | Malcolm Thomas | -1.73 | 264 | 59 | 6 | ||||
| 7 | Ian Clark | -2.03 | 289 | 59 | 7 | ||||
| 8 | Rudy Gobert | -2.13 | 294 | 56 | 8 | ||||
| 9 | Jeremy Evans | -2.26 | 303 | 60 | 9 | ||||
| 10 | Andris Biedrins | -2.55 | 322 | 62 | 10 | ||||
| 11 | Brandon Rush | -2.87 | 339 | 72 | 11 | ||||
| 12 | Richard Jefferson | -2.89 | 341 | 66 | 12 | ||||
| 13 | Diante Garrett | -3.26 | 365 | 68 | 13 | ||||
| 14 | John Lucas III | -3.87 | 385 | 74 | 14 | ||||
| 15 | Enes Kanter | -4.75 | 410 | 75 | 15 | ||||
Yeah, ESPN lists Evans as a SF, and Alec as PG. But still, so much of this seems to track with what we already know and feel. Marvin gets a huge bump here -- I guess part of that has to deal with how he's always pulled from the lineup when the going gets rough.
But this is just one metric, and there are plenty to choose from. I'm not really sold on RPM (and not just because our best player is the 94th best player in the NBA at it), there's also PER, Win Shares, WAR (wins above replacement), Net RTG (Off RTG - Def RTG), ON/Off Stats, the +/- per 100 possessions, and something called PIE. Hmmmm. Pie.
These things all show different aspects of the game, but the general idea here is that they all show things that can be classified as good behavior. (Going back to the whole behaviorism psychology aspect of why we even look at basketball data....) The idea, then, is that if you are good at a lot of these things you are probably good at basketball.
So in order to try this out the first thing to do was find all the values for our players for these eight categories. Then I gave each Jazz player a rank compared to their Jazz teammates. So in the above case Kanter would get a Rank of 15 in RPM. Then I did the next logical thing and average out the ranks for each player.
Surely this would give us useful data and help us identify who our best player is. After all, we already know who it is, it's Derrdan Hayors. (Also a good Dungeons & Dragons name). So in this case we know what the answer should be so we can also test how good a method this is at confirming our own biases.
So . . .
| Avg | Win | NetRtg | On/Off | +/-per100 | ||||||||||||||
| Player | Rank | RPM | Rk | PER | Rk | Shares | Rk | WAR | Rk | Rk | Net | Rk | pos. | Rk | PIE | Rk | ||
| 1 | Marvin Williams | 4.6 | -0.35 | 3 | 14.2 | 6 | 3.0 | 4 | 1.98 | 3 | -5.5 | 4 | 5.7 | 3 | -5.3 | 5 | 8.9% | 6 |
| 2 | Alec Burks | 5.4 | -0.08 | 2 | 16.2 | 4 | 3.0 | 3 | 2.76 | 2 | -7.3 | 8 | 2.3 | 7 | -6.2 | 7 | 10.6% | 3 |
| 3 | Gordon Hayward | 5.4 | 1.47 | 1 | 16.2 | 3 | 3.4 | 2 | 6.01 | 1 | -8.4 | 9 | 1.7 | 8 | -7 | 9 | 11.3% | 2 |
| 4 | Derrick Favors | 6.4 | -0.99 | 4 | 18.6 | 1 | 4.6 | 1 | 1.67 | 4 | -9.3 | 10 | -2.6 | 10 | -8.2 | 10 | 11.7% | 1 |
| 5 | Ian Clark | 6.6 | -2.03 | 7 | 8.9 | 10 | -0.1 | 12 | 0.03 | 9 | 0.2 | 1 | 22.6 | 1 | 3.8 | 1 | 5.4% | 11 |
| 6 | Rudy Gobert | 6.6 | -2.13 | 8 | 12.2 | 8 | 0.3 | 9 | 0.05 | 8 | -2.4 | 2 | 5.7 | 4 | -1.3 | 3 | 8.3% | 7 |
| 7 | Trey Burke | 7.0 | -1.66 | 5 | 12.5 | 7 | 0.8 | 8 | 0.85 | 5 | -7.0 | 7 | 4.0 | 5 | -6.1 | 6 | 8.2% | 8 |
| 8 | Jeremy Evans | 7.4 | -2.26 | 9 | 16.3 | 2 | 2.6 | 5 | 0.06 | 7 | -7.0 | 6 | 1.6 | 9 | -6.6 | 8 | 10.3% | 4 |
| 9 | Andris Biedrins | 8.5 | -2.55 | 10 | 2.4 | 14 | 0.0 | 10 | 0.01 | 11 | -6.3 | 5 | 9.0 | 2 | 2.1 | 2 | 4.7% | 12 |
| 10 | Diante Garrett | 10.0 | -3.26 | 13 | 7.2 | 11 | -0.7 | 14 | -0.54 | 13 | -5.2 | 3 | 3.4 | 6 | -4.9 | 4 | 5.6% | 10 |
| 11 | Richard Jefferson | 11.0 | -2.89 | 12 | 12.1 | 9 | 2.6 | 6 | 0.68 | 6 | -10.5 | 11 | -5.4 | 12 | -9.7 | 11 | 7.5% | 9 |
| 12 | Enes Kanter | 12.6 | -4.75 | 15 | 15.1 | 5 | 1.8 | 7 | -2.88 | 15 | -13.0 | 13 | -8.7 | 14 | -11.7 | 13 | 9.7% | 5 |
| 13 | Brandon Rush | 13.5 | -2.87 | 11 | 4.1 | 13 | -0.4 | 13 | -0.13 | 12 | -12.3 | 12 | -5.0 | 11 | -11.4 | 12 | 3.1% | 13 |
| 14 | Malcolm Thomas | 14.6 | -1.73 | 6 | 0.2 | 15 | -0.1 | 11 | 0.02 | 10 | -30.9 | 15 | -16.7 | 15 | -25.3 | 15 | 2.4% | 15 |
| 15 | John Lucas III | 15.4 | -3.87 | 14 | 5.2 | 12 | -0.7 | 15 | -0.55 | 14 | -16.5 | 14 | -8.0 | 13 | -14.2 | 14 | 2.9% | 14 |
Marvin? MAAR-VINN Marvin? MARVIN? Okay, so this is what we now need to deal with:
- Not all "This guy is the best guy on your team" metrics match up with one another well
- Ranking by ordinals (1st, 2nd, 3rd, etc) skews the actual data -- so one guy could be way ahead in something, but only come one spot higher
- Being good at a bunch of stuff is great . . .
- . . . but being okay at a bunch of stuff with no weakness is better here.
And thus, Marvin is our "best" player. Which means to me that this method is cracked, because we know that is not the case.
Okay, let's pretend that this information is something we have to really look at though. We still get the idea that Burks, Hayward, and Favors are among our top players. Trey comes later, after highly efficient but seldom used Moose and Squirrel -- which almost makes sense. They are not as efficient as Jeremy Evans -- but he's behind Trey but still a solid rotation guy here. I don't think any of those points are things fans who look at the data will argue against.
In fact, I am sure a lot of fans wouldn't be upset if our #8 rotation was these guys. With back-up PG being played in a two guard lineup off the bench with Clark and Burks, right? Not bad. Kind of dynamic, right?
The bad news is that this buries Kanter. He's not had a great year at all, and it shows. Compared to the rest of his teammates, he's not shining that much.
Which works out great for the Jazz front office who will probably get to low-ball Enes this off-season. If anything, he could be the next RFA for this team, even if he has the worst RPM. Thanks ESPN. Thanks for making us feel less secure over-all by pointing this out. Even if you list Evans as a SF and don't know that Alec Burks isn't a point guard.