When I first sat down to write this article, I had initially decided to focus mostly on a comparison of the Bulls’ young players/rookies Ben Gordon, Chris Duhon, Kirk Hinrich and Luol Deng. In researching however, I found myself becoming increasingly interested in Gordon and less on his comparison with the others. While there is still some comparison, the majority of the article is on Gordon alone.
Notes: The majority of the statistics used for comparison are from the 2003-04 season, for two reasons. First, the 2003-04 season provides a larger sample size for comparisons, and second and more importantly, I’ve been too lazy to input the numbers for 2004-05 yet this season. Some statistics are from 2004-05, and it is noted where they are used. All other numbers, outside of rookie numbers that are obviously from 2004-05, may be assumed to be from 2003-04. 2004-05 numbers current as of 2/13/2005.
Offensive Efficiency and Offensive Possessions
If you’re familiar with more detailed basketball statistics such as Offensive Efficiency (also called offensive rating by some; denoted hereafter as OE), Offensive Possessions (denoted hereafter as OP), and OP per minute (denoted hereafter as OP/M), as well as their usefulness and importance, you may want to skip down a few pages. I realize that not all of you that will be reading this do know much about them, so I’ll take some time to detail the formulas and explain their importance. I’ve also written some other articles on various basketball topics and statistics that include some of the same ideas found here with some more in-depth explanation and analysis than I cover in this article. You can read those other articles and find my spreadsheets at my basketball statistics website: http://www.oocities.org/torch772.
Offensive Efficiency, also known as Offensive Rating (generally noted as points scored per 100 possessions as opposed to points per possession) by some, is the most important offensive statistic when comparing a team or player’s effectiveness offensively. It is a formula for deriving exactly what it sounds like it should: the number of points a team or player averages per possession. Other statistics, such as FG%, are much too blunt for determining effectiveness, as they do not take into consideration factors such as free throws and three point shooting. There are other metrics, most notably effective FG%, popularized by 82games.com, but it still does not take into account free throws, making it less useful because free throws are generally the most important and effective way for a player or team to score points.
It makes sense then, that if you’re attempting to find the number of points a player scores per possession, that you take their PPG and divide by their possessions, or (PPG/OP). This is in fact the formula for OE. While this is immediately easy to understand, the overwhelming impact this formula has on a team’s Win/Loss record is not so easily realized. Last year in the NBA, the average NBA team had an OE of 1.041, or in words, the average NBA team last year scored about 1.041 points for every possession (or equivalently 104.1 points per 100 possessions). The average team also averaged about 89.5 OP, and if you multiply 1.041 by 89.5, you find that NBA teams averaged 93.2 PPG, which was in fact the NBA average. As a hypothetical example, if you were to take the 2003-04 Chicago Bulls’ OE of 0.969, over the same 89.5 possessions you find the Bulls would have scored just 86.7 PPG (they actually averaged 88.5ppg, due to the fact they also averaged a slightly higher than average 91.4 OP), an astounding 6.5ppg less than the average team (Chicago had the lowest OE in the NBA last year). I won’t go into the specifics, but as a general rule of thumb, 1 PPG of differential is roughly equivalent to 2.7 wins per season. That is, if one team outscores its opponents by 1 PPG, it can expect to be 2.7 games over .500, i.e. 43.7 wins. Chicago’s offense was 6.5 PPG worse than the average offense over the same number of possessions, which means if they played average defense against average teams and both used 89.5 OP, they would be 2.7*6.5 = 17.6 games under .500, or about 23.4 wins (due to their defense being about average and their OP being roughly equal to their opponents, this number does approximate their “expected” win value of 21 wins – they won 23. For more on expected win value check the website linked above. I also have spreadsheets with the actual OE/OP and other numbers for all of the teams last year, as well as more in depth explanation.) A change from 1.041 to 0.969 in OE would have cost them 17.6 wins. In fact, for a quick approximation, each .01 change in OE is roughly equal to .9 PPG, because of the fact that the average NBA team uses 89.5 OP, and 89.5*.01 ~ .9 PPG. The important thing to note however is that this is only true at the team level.
At the player level, the impact due to a change in OE is not so easily measured. First, however, it is probably best to explain how to calculate OP. OP is a measure of how many possessions a player used that resulted in some sort of offensive attempt, a 3pFGA, 2pFGA, or FTA. In general keeping of basketball statistics, FG% and FGA lumps 2pFG% and 3pFG%, 2pFGA and 3pFGA together respectively. It does not however keep a general number of offensive possessions, or any numbers that we can easily add together to arrive at the number. The problem lies with FTA, as there are several situations that complicate the issue. In general, 2 free throws are awarded in place of a shot, but sometimes 3 are awarded, and sometimes 1 is added to a made basket, which together should still be considered just 1 possession, and sometimes technical free throws are added which should not be counted as possessions at all, as well as a few other situations that need to be considered. The question then becomes what to divide FTA by to arrive at a good approximation for the number of OP used in taking free throws, and the best approximation seems to roughly be FTA/2.5. What this gives us then is (FGA + FTA/2.5) = OP.
Now, when considering the impact of OE at the player level, we need to separate out the number of OP the individual player uses, as well as the player’s OE. Then as a comparison, take the average NBA player’s OE and find the number of points the average player would score given the same number of OP. The resulting difference gives a decent approximation of the number of PPG that player’s team PPG would change by. As an example, Brent Barry of the Supersonics last year had an OE of 1.357, an absolutely outstanding number (Barry in fact had the highest OE of any player last year), and an OP of 8.0 per game. The NBA league wide average was 1.041 OE. Now, if we take Barry’s 1.357 OE and multiply it by 8.0 OP, we get 10.856 PPG (he in fact averaged 10.8 PPG, due to some small round off error). An average NBA player over the same number of OP would have averaged 8.328 PPG, a difference of 2.528 or 2.5 PPG.
Important Note: At the team level, defense can be determined using opponent OE and OP, however, at the individual level defensive statistics are not kept and thus nothing can be said about a player’s defense with respect to the NBA average until such statistics are recorded. Thus, all of the numbers I generate in this article that talk about player PPG differences are strictly offensive numbers only. While Barry may be more valuable offensively than an average NBA player, nothing can be said about his overall value because the other half of determining worth is unknown.
Recalling that 1 PPG is worth 2.7 wins per season, we find that Barry’s 2.5 PPG is worth 2.5*2.7 = 6.75 or about 6.7-6.8 wins per season over an average NBA player. That is an incredible number. The only drawback to Barry’s numbers however was that he merely used 8.0 OP per game, had he used more, he would have been much more valuable. The advantage of comparing Barry’s OE and OP to that of an average player over the same number of OP is that we can begin to talk about how valuable a player is. If you hypothetically were to have another player, let’s call him player A, that had the same OE as Barry of 1.357, but used 20.0 OP per game, and compared him to the average player, you would find that player A was in fact worth 6.3 PPG more than the average player, or an incredible 17.0 wins more per season. A team made up entirely of average NBA players and player A would be expected to win about 58 games. What you find then, is that player worth is not just reliant on OE, but also over how many OP the player has that OE. As another example, given two hypothetical players, player A that has a 1.100 OE over 10.0 OP and player B that has a 1.080 OE over 20.0 OP, you find player A is worth .6 PPG and player B is worth .8 PPG over and average NBA player. Despite having a lower OE, player B is in fact worth more because he is more efficient than the average NBA player over more OP than player A. (Peja Stojakovic was in fact the most valuable player by a large margin, worth 11.66 wins due to his high OE and high OP.)
Extending this idea, you find then that not only can a player be more valuable by having a higher OE than average over a larger number of OP than a player that has an even higher OE, but that a player that has a questionable shot selection (i.e. lower than average OE) over many OP in fact may hurt his team substantially more than a player with a low OE over a low number of OP (like many role players). The two players that hurt their team the most were Jamal Mashburn and Chris Webber, had they played full seasons. Luckily for the Hornets and Kings, neither player played a full season. For a player that did not play a full season, multiply the resulting wins by the fraction of the season they played. As an example, over a full season, Mashburn would have cost the Hornets 7.28 wins, but he played in just 19 games. To get the number of wins he cost the hornets over those 19 games, multiply -7.28*(19/82) = -1.7 wins. Despite the fact that Mashburn and Webber did not have the lowest OE of all players last year, their combination of low OE and low OP ended up making them the least valuable.
As this article is actually about evaluating Gordon, it’s probably best to talk about how to determine how good he is. How valuable a player is to a team is not as important to us then as how good that player is. OE and OP determine how valuable a player is, but OE alone measures how good a player is – in a perfect world where each shot and play is independent of the next. The only problem is Basketball is not the same as Baseball, where a lone batter faces a lone pitcher and each player’s at-bat is almost entirely independent of the next player’s at-bat. Many things factor into OE: how good the other players on the team are, how good at getting open shots for teammates the point guard is, what position the player plays, etc. All are factors that affect OE. The problem is that many of these factors are quite difficult to measure objectively. One of the most important factors is OP per minute, or OP/M. There is a correlation between OE and OP/M such that as OP/M increases, OE tends to decrease both in value and variance. These both make logical sense, as a player takes more shots over a set time period, the player’s shot selection must become less rigid, the defense begins to focus on the player more, and the value of the player’s OE decreases. In addition, as you take more shots, the sample size increases, and variance decreases. This has two important ramifications, first that players with low OP/M can have wider ranges in their OE. Hypothetically, a player with a low OP/M and a true OE of 1.050 might one year have an OE of 1.150 and the next an OE of .950, where a player with a high OP/M and a true OE of 1.050 might have an OE of 1.100 one year and 1.000 the next, as an example. Thus when attempting to gauge a player’s true OE, a player with low OP/M is much more difficult to accurately gauge. An example of this would again be Brent Barry, who has a career 1.209 OE, but last year had a 1.357 OE. The average NBA player had an OP/M of .366, while Barry had an OP/M of .260. (The league leader was Allen Iverson with an OP/M of .640)
This brings us to the second important implication, because NBA teams have a limited OP/M – the NBA average was 1.865 (~ the player avg.* 5) – the higher a player’s OP/M is, the lower his teammates OP/M must be to compensate, thus indirectly raising the OE of the player’s teammates, or in reverse, the lower the teammate’s OP/M the higher the player’s OP/M must be to compensate, lowering his OE (See: Tracy McGrady, Orlando, for an example of the reverse effect). This means a player with a high OP/M will indirectly raise his teammate’s OE, while a player with a low OP/M will lower his teammate’s OE. Taken together, OE and OP/M give a relatively easy way to measure players against each other. To summarize, OE tells us how effective offensively a player is, and OP/M tells us something about how wide a range their OE may fall in and also indirectly raises or lowers the player’s value. To give a baseball analogy, one can think of OE like a pitcher’s ERA or WHIP and OP/M like IP. A pitcher with a low ERA is good, but a pitcher with a low ERA and high IP is rare. You’ll often see relievers or closers that have a very low ERA one season and a high one the next, where a starter’s ERA varies much less year to year typically. Also, just as my earlier example, a starter with a 3.50 ERA is often seen as much more valuable than a reliever with a 3.00 ERA because he pitches so many more innings at a better than average ERA.
Ben Gordon
So, with all that other stuff out of the way, let’s focus on Gordon. When I first started looking over stats for the article, Gordon’s OE and OP/M jumped out at me, and I was a bit surprised at how high they were. Through 47 games, Gordon is averaging 22.9 MPG, 11.5 FGA, 2.7 FTA, 12.6 OP, and 13.8 PPG. The numbers seem pretty good, but not particularly unusual, although the MPG seems a bit on the low side. It turns out that the OP and PPG for 22.9 MPG is very high. Gordon is averaging on the season a 1.099 OE and a .549 OP/M. In fact, among NBA guards last year, only one other player had both a higher OE and higher OP/M: Kobe Bryant. There were a good handful that had a higher OE, but of those, only Kobe had a higher OP/M as well. In fact, in the entire NBA only 2 players had both a higher OE and OP/M. One was Kobe Bryant. Maybe you’ve heard of the other: reigning MVP Kevin Garnett. That’s right, only Bryant and Garnett in the entire NBA had a higher OP/M and OE than Gordon does now. Not Shaq, not Duncan, not McGrady. That’s not to say Gordon is a better or more valuable offensive player – Shaq and Duncan had significantly higher OEs with only somewhat lower OP/Ms, McGrady a significantly higher OP/M and somewhat lower OE. But looking at both, only Bryant and Garnett were better. That’s saying something, especially considering Ben is a rookie. In fact, when I compared his numbers to the other rookies of 2004-05 that are doing anything, no other rookie was close. And when I say close, I mean no other rookie is even in the same ballpark as Gordon. Here’s a list of the rookies doing anything this year, sorted by OP/M (the discrepancy between the table’s value of Gordon’s OE and OP/M and the ones I listed above are due to Yahoo! rounding stats to one decimal place - the 1.099/.549 are derived numbers from game logs, which are the exact and correct numbers. The numbers for all other players are based only on Yahoo!’s rounded values):
|
Last Name |
First Name |
Team |
MPG |
FGA |
FTA |
PPG |
OE |
OP |
OP/M |
|
Gordon |
Ben |
Chicago |
22.9 |
11.5 |
2.7 |
13.8 |
1.097 |
12.6 |
0.549 |
|
Telfair |
Sebastian |
Portland |
10.7 |
4.2 |
1.4 |
4.3 |
0.903 |
4.8 |
0.445 |
|
Deng |
Luol |
Chicago |
28.0 |
11.5 |
2.5 |
12.5 |
1.000 |
12.5 |
0.446 |
|
Okafor |
Emeka |
Charlotte |
35.8 |
13.9 |
4.2 |
14.8 |
0.950 |
15.6 |
0.435 |
|
Smith |
JR |
New Orleans |
20.1 |
7.3 |
1.1 |
7.0 |
0.904 |
7.7 |
0.385 |
|
Harris |
Devin |
Dallas |
15.3 |
5.0 |
1.2 |
5.6 |
1.022 |
5.5 |
0.358 |
|
Udrih |
Beno |
San Antonio |
14.0 |
4.5 |
1.2 |
5.5 |
1.104 |
5.0 |
0.356 |
|
Smith |
Josh |
Atlanta |
23.3 |
6.1 |
2.6 |
7.7 |
1.078 |
7.1 |
0.306 |
|
Howard |
Dwight |
Orlando |
32.1 |
7.2 |
5.1 |
10.5 |
1.136 |
9.2 |
0.288 |
|
Childress |
Josh |
Atlanta |
23.9 |
6.0 |
2.1 |
7.2 |
1.053 |
6.8 |
0.286 |
|
Igoudala |
Andre |
Philadelphia |
32.2 |
6.6 |
2.6 |
8.7 |
1.139 |
7.6 |
0.237 |
|
Duhon |
Chris |
Chicago |
24.2 |
5.2 |
0.7 |
4.9 |
0.894 |
5.5 |
0.226 |
Only Udrih and Igoudala have a higher OE, Udrih by just .005. Both however have substantially lower OP/M, and Udrih is playing just 14.0 minutes with 5.0 OP per game, and his numbers are likely to vary widely. Igoudala’s OP/M meanwhile is .237, one of the lowest for any guard playing 30+ minutes, and 57% lower than Ben’s. In addition, Gordon went through a very poor stretch to start the season, however, in his last 30 games he has upped his averages to 1.158 OE and .576 OP/M. Care to guess the number of players with a higher OP/M and higher OE in the entire NBA last year than Gordon’s over his last 30 games? That’s right, zero. In fact, Gordon’s numbers were so eye popping that I decided to compare them to other rookies from previous years. I looked at the rookies since 1992 (I cut it off at 1992 arbitrarily because it was a good class and I didn’t want to do the work for more years), and I picked about 15 or so I thought had the best chance of besting Gordon from each year, about 180 or so all told since 1992. Of the 180 or so rookies I looked at, exactly zero had a higher OE and OP/M than Gordon does this season. Below is a list of the 25 or so closest rookies (with Hinrich thrown in just for comparison):
|
Last Name |
First Name |
Team |
MPG |
FGA |
FTA |
PPG |
OE |
OP |
OP/M |
|
Iverson |
Allen |
Philadelphia |
40.1 |
19.8 |
7.2 |
23.5 |
1.036 |
22.7 |
0.566 |
|
Anthony |
Carmello |
Denver |
36.5 |
17.9 |
6.4 |
21.0 |
1.026 |
20.5 |
0.561 |
|
Gordon |
Ben |
Chicago |
22.9 |
11.5 |
2.7 |
13.8 |
1.097 |
12.6 |
0.549 |
|
Robinson |
Glenn |
Milwaukee |
37.0 |
17.6 |
6.2 |
21.9 |
1.091 |
20.1 |
0.543 |
|
James |
Lebron |
Cleveland |
39.5 |
18.9 |
5.8 |
20.9 |
0.985 |
21.2 |
0.537 |
|
Mourning |
Alonzo |
Charlotte |
33.9 |
14.4 |
8.1 |
21.0 |
1.190 |
17.6 |
0.520 |
|
Oneal |
Shaq |
Orlando |
37.9 |
16.1 |
8.9 |
23.4 |
1.190 |
19.7 |
0.519 |
|
Abdur-Rahim |
Shareef |
Vancouver |
35.0 |
15.2 |
6.5 |
18.7 |
1.051 |
17.8 |
0.509 |
|
Brand |
Elton |
Chicago |
37.0 |
16.1 |
6.6 |
20.1 |
1.073 |
18.7 |
0.506 |
|
Van Horn |
Keith |
New Jersey |
37.5 |
16.9 |
4.9 |
19.7 |
1.045 |
18.9 |
0.503 |
|
Carter |
Vince |
Toronto |
35.2 |
15.3 |
5.4 |
18.3 |
1.048 |
17.5 |
0.496 |
|
Webber |
Chris |
Golden State |
32.1 |
13.6 |
4.7 |
17.5 |
1.130 |
15.5 |
0.482 |
|
Stackhouse |
Jerry |
Philadelphia |
37.5 |
15.2 |
7.2 |
19.2 |
1.062 |
18.1 |
0.482 |
|
Hill |
Grant |
Detroit |
38.3 |
15.2 |
7.3 |
19.9 |
1.098 |
18.1 |
0.473 |
|
Duncan |
Tim |
San Antonio |
39.1 |
15.7 |
5.9 |
21.1 |
1.168 |
18.1 |
0.462 |
|
Francis |
Steve |
Houston |
36.1 |
14.5 |
4.7 |
18.0 |
1.099 |
16.4 |
0.454 |
|
Pierce |
Paul |
Boston |
34.0 |
13.5 |
4.1 |
16.5 |
1.090 |
15.1 |
0.445 |
|
Stoudamire |
Damon |
Toronto |
40.9 |
16.1 |
4.2 |
19.0 |
1.069 |
17.8 |
0.435 |
|
Wade |
Dwyane |
Miami |
34.9 |
13.1 |
5.1 |
16.3 |
1.077 |
15.1 |
0.434 |
|
Marbury |
Stephon |
Minnesota |
34.7 |
13.0 |
5.0 |
15.8 |
1.053 |
15.0 |
0.432 |
|
McDyess |
Antonio |
Denver |
30.0 |
11.6 |
3.2 |
13.4 |
1.040 |
12.9 |
0.429 |
|
Ilgauskas |
Zydrunas |
Cleveland |
29.0 |
10.7 |
3.7 |
13.9 |
1.141 |
12.2 |
0.420 |
|
Gasol |
Pau |
Vancouver |
36.7 |
13.0 |
5.8 |
17.6 |
1.149 |
15.3 |
0.417 |
|
Billups |
Chauncey |
Boston |
25.4 |
8.9 |
3.5 |
11.1 |
1.078 |
10.3 |
0.406 |
|
Hardaway |
Anfernee |
Orlando |
36.8 |
13.3 |
4.0 |
16.0 |
1.074 |
14.9 |
0.405 |
|
Baker |
Vin |
Milwaukee |
31.2 |
10.6 |
5.0 |
13.5 |
1.071 |
12.6 |
0.404 |
|
Hinrich |
Kirk |
Chicago |
35.6 |
10.8 |
2.2 |
12.0 |
1.027 |
11.7 |
0.328 |
Interestingly, the player that comes closest to both of Gordon’s numbers is Glenn Robinson. Even though Robinson has taken a beating the past few years due to off court issues, Robinson’s rookie year was considered by most a success I think, especially offensively. In fact, of the guards on the list, no other guard has a higher OE than Gordon’s 1.099, and only one, Allen Iverson, has a higher OP/M. However, using Gordon’s averages over the last 30 games, on this list Gordon would rank 4th overall in OE - 1st among guards - and 1st overall in OP/M. Only Duncan, Shaq, and Mourning had a higher OE. That’s pretty heady company. And while it’s not generally a good idea to compare stats based on partial seasons, the 30 game stretch does represent almost 2/3rds of Gordon’s season and coincides with some of the best basketball Chicago has played, which leads me to my next topic.
It doesn’t seem coincidental to me in the 3 prolonged win streaks Chicago has had this season, Gordon has an OE of 1.131, and an OP/M of .550, while in Chicago’s prolonged losing streak to open the season he averaged 0.872 OE and .546 OP/M. Gordon’s play, aside from a rather slow start, has been at or near all-star caliber level. Which brings me to the question: why isn’t Gordon getting as many minutes as he is capable of handling? Or at the very least, 30-32mpg which I’m confident Gordon is in physical shape to handle. Or even more to the point, why on God’s earth is Duhon getting more playing time than Gordon (or even allowed to step on the court for that matter)? I know some Bulls fans out there will argue that Skiles’ philosophy of team chemistry and togetherness, rah rah rah and all that is what is helping the Bulls win, and guys like Duhon and Hinrich who do the little things are what’s making this team go. While I can agree chemistry helps a team, the differences are much more tangible than some Bulls fans seem to believe. Subtracting Crawford and adding Gordon has accounted for the largest part of the turn around. Crawford was a black hole last year, with a 0.977 OE to go with a .504 OP/M. Crawford also averaged 17.7 OP last year, and replacing 13.7 of them with Gordon has improved the Bulls by 1.7 PPG alone, or about 4.6 wins. Meanwhile, Curry has improved by about 1-2 wins, and a healthy and also improved Chandler has also helped. Replacing guys like Ronald Dupree with guys like sharpshooter Eric Piatkowski has helped out as well. In fact, Chicago has improved its OE from a league worst 0.969, to a very respectable 1.042, a difference of +.073 OE, a gigantic leap. While everyone is talking about the improved defense, when you calculate the point change due to the OE change and correct it for the change in OP, you find that Chicago improved itself by about 5.1ppg, accounting for nearly 14 more wins (The Bulls are on pace to win 18-19 more games this year), and of the overall total from OE, Gordon in just 22.9 minutes has accounted for about a third of that. Chicago’s success has been due to its improved shooting, not defense, and Gordon is leading the charge. This brings me back to my original question: why again isn’t Gordon playing more? If one were to swap out the 5.5 OP Duhon is averaging per game at an absurdly bad 0.894 (no that’s not a misprint) for Gordon at 1.099, the Bulls would improve by about another 1.1 PPG, or roughly 3 wins. I just don’t see Duhon possibly being a valuable enough “glue” guy to make up this difference.
Normally, I tend to dislike predictions and projections, and try to stay as far away as possible from them. In the case of Gordon however, I feel it’s almost necessary because of the severe lack of coverage this guy is getting. Right now, he’s head and shoulders above the other rookies numbers wise, but probably won’t end up with the award. Gordon is certainly capable of scoring 20 points a game, right now. Optimistically, I expect sometime in his prime, Gordon will push 25 PPG scoring and will be considered the best player from this draft. I know many will think that is a rather ridiculous projection to make at this junction given his current numbers, but based on Gordon’s OP/M and OE I think it’s not only reasonable, but likely. As it currently stands, he projects to somewhere between 20 and 25 PPG based on his seasonal averages and right around 25 PPG based on his numbers over the last 30 games. While there will likely be some drop-off in scaling up to 35-40 minutes a night, I’m quite sure Gordon’s numbers will scale far better than most due to his high OP/M, and he appears to have a healthy work ethic and should improve as his career progresses. That is not to say Gordon is without flaws, he certainly has his fair share. Offensively, he needs to work on getting to the line, as his FTA/FGA ratio is rather pedestrian. He also needs to work on his turnovers, but I feel that his turnovers are less of a problem than some think, because they tend to look at his TO/MPG numbers, as opposed to his TO/OP numbers. This is a more balanced metric to compare Gordon to other guards with, as very few are dominating the ball as much as he does in his time on the court. He averages about .175 TO/OP compared to the .135 TO/OP of other shooting guards last year, about 29% more than normal. Gordon does however partially make up for this with better than normal offensive rebounding, which helps to offset the negative possession differential created by the excessive turnovers. I don’t want to get too much into defense, as that is increasingly in the realm of subjective opinion with the stats readily available, but I will say that his steals and blocks tend to be somewhat below average, but not alarmingly so. How important or how much work he does or does not need, I’ll leave to the reader to decide.
I’ve also included two graphs, the first is Gordon’s OE per game this season, with the Bulls’ seasonal win% shown for comparison. The black line is a moving 15 game average of Ben’s OE. The second is a straightforward graph of Ben’s OE this season.
