Prior to flying to Tokyo on Sunday, I downloaded the Blazers' "Production by Position" statistics from last season from 82games, thinking that rummaging through these statistics would be more fun than watching the wretchedly insipid movies to which airlines subject their passengers. The goal was to tease out something useful about the Trailblazers from the available data. This is the result from the 9 hour flight. I hope you find it satisfactory.
The production by position statistics indicate how the team fares with the same player in different positions, and they also give an idea of how well the player does personally in each position. The statistics can be useful to help decide whether we should, e.g, trade Sergio and promote Jerryd, or do something else. They can also explain fan behavior :-) As an example, the production by position statistics for the 2007-2008 season show a wash between Martell and Travis at the small forward, with the 2007-2008 Blazers outscoring their opponents 47% when Martell was on the floor playing small forward, and 45% of the time when Travis was the SF. I conjectured the ensuing debate raged perhaps because we couldn't tell the difference. As a test of this hypothesis, you'll see they show we can't distinguish Sergio and Jerryd this year at PG--and, viola, another debate is raging.
82games provides several types of production by position statistics. I use two.
The first, which 82games call on/off court statistics, measure how much a player helps (or hurts) the team while playing a particular position. These tables have 9 columns for each player: position, percentage of the overall minutes the player played in that position, the 48 minutes scoring rate for the team while he was on the floor at that position, the 48 minutes scoring rate of the opposing team while he manned that position, the total +/- for the player for the season at that position, the number of games in we outscored the opponent when the player was at the position (a bit misleadingly labled "Wins" by 82games), the number of games the opponent outscored the Blazer when the player was at this position (labeled "losses"), and the "Win Percentage", which is the ratio of games we outscored the opponent divided by the total number of games during which the player played the position.
For this discussion, I'm going to modify this format a bit, and consider all of the point guard contributors together, all the scoring guards together, etc, so the first column will be the player's name instead of position. I will also arbitrarily limit considerations to contributions for players who manned a position at least 8 or 10 percent of the time, considering anything less as statistically insignificant.
The second set of statistics document personal performance at each position. They are fairly self explanatory, like field goal attempts (FGA) and turnovers (T/O), and are the sort of thing discussed frequently, so I won't belabor their meaning.
1. Point Guard
Here are the on/off court statistics for our point guards last season:
% PG Blazers Opponent Outscore Outscore
PG Minutes Scoring Scoring +/- Opponent Blazers Win %
Blake 55 101.3 92.5 392 42 23 65
Bayless 11 100.3 96.5 37 13 18 42
Rodriguez 30 94.4 93.5 21 29 45 39
(any one know how to insert tabs in this editor?)
Rudy and Brandon also played point guard some, but too infrequently to consider.
At first glance nothing unexpected jumps out: the team ran well with Blake at the point, and otherwise less well.
The first thing that surprised me, however, is that the Win % was essentially the same with Jerryd or Sergio. My impression is Sergio did a lot better (he certainly did it a lot more).
It should be a concern that without Steve, we turn into a 33 win team. Given how well the team played while Blake was injured, we don't really expect that steep a drop-off, but the fact remains that's how the numbers stack up over the entire season. Someone will retort that either Roy or Rudy can play the point, but by now we have ample if not sufficient evidence saying neither can do so effectively on a regular basis; there is a lot more to playing the point than running the offense.
The second surprise is that while Blake is viewed as deliberate and Sergio up-tempo, the evidence demonstrates the reverse; the offense scored 6.9 points less a game with Sergio at the helm than Steve, while only 1 point less when Bayless became the PG. Part of this might be explained by Steve playing on the Black unit while Sergio with the White, but the White unit was supposed to be our up-tempo guys--it didn't work out that way; even with Travis and Rudy, they have trouble figuring out how to score with Sergio running the point.
The third surprise is that, while Jerryd is perceived as the better defender, team defense suffered much more when Bayless replaced Steve (the opponent's scoring increased by 4 points per game) than Sergio (opponent scoring increased by only 1 point). Perhaps this can be explained by theorizing that opponents exploited Bayless' rookiness. This theory makes some sense sense, as the the team could wrest the essentially same win percentage from a point differential of +0.9 points with the more experienced player as with the +3.8 point differential with the less experienced player.
If it really was a wash between Bayless and Rodriguez, why did Coach McMillan play Sergio 3 times as much as Rex? Here are the personal performance statistics for the Blazers who played significant minutes at PG last season:
PG Minutes FGA eFG% FTA iFG% Reb Ast T/O Blk PF Pts PER
Blake 55 14.2 .536 1.6 12 3.8 7.6 2.3 0.1 2.8 16.6 16.2
Opponent 16.3 .453 3.5 32 5.4 7.7 3.7 0.1 2.3 17.8 15.0
Bayless 11 13.8 .455 8.0 48 4.3 5.3 3.8 0.2 6.0 19.3 13.1
Opponent 16.1 .490 4.9 32 5.5 8.9 2.8 0.4 4.6 19.8 19.0
Rodriguez 30 13.0 .448 2.9 29 5.1 11.3 4.6 0.1 5.2 13.9 14.1
Opponent 15.0 .443 4.1 33 5.0 7.5 3.8 0.3 3.0 17.0 15.1
Three numbers on offense pop out here: Bayless shot inside 48% of the time, created a miniscule 4.3 assists per 48-minutes, and fouled at a rate that could make only Greg proud. On defense, the opposing point guards' eFG% jumped up to .490 and their assists peaked at 8.9. Nate had to pick his spots to play the freshman edition of Jerryd Bayless to avoid the offense becoming too predictable and the defense too pourous; coaching is in part finding a way to make players look their best, and Nate McMillan is a master at this game.
Jerryd's numbers have rookie written all over them. His opponent shot with a higher eFG%, and he gave up more assists and commited more fouls than the other two, all indicative of (at least) inexperience on defense. Casually perusing the 82games statistics for point guards, it seems like his 48% inside field goal attempts is much higher than average--other point guards go inside about a third of the time. The likeliest explanations seem to be he can get away with this even against NBA point guards; part of it has to be his lack of confidence in other shots. Jerryd's eFG% surprised me--while it isn't great, .455 is not all that bad--because this is much better than the average frequently cited on other EB posts. However, people have been citing overall instead of by-position numbers, and these don't tell the whole story; we'll Jerryd's shot selection strategy penalized him when he played SG.
This was a season of genuine growth for Blake. He shot well (.536 3FG%; up from .502 the prior season) and went inside a bit more often (12% iFG% instead of 8% from the previous year). His low turnover rate remained steady (an increase of 0.1 T/O per game from 2007-2008). His PER increased by 2.6 points, while that of his opponent declined by 2.4. Steve is not an elite player, but this year he ecame above average at offense and at least average at defense. Since he is one of two players on the team (the other is Greg) who played only one position, what you see here tells everything we need to know about him.
Sergio had a banner year compared to his sophomore season, getting close to his rookie performance. However, even though Sergio delivered assists at a much higher rate, Steve's mix of scoring and handing out assists was more effective in terms of win %. Sergio still needs to improve at scoring. In the turnover numbers, you can also see Coach McMillan's reluctance to use any PG but Blake. Sergio turned the ball over twice as often as Steve, and Jerryd 65% more often. These are significant differences given the pace at which the Blazers play.
Since there is a cottage industry comparing Jerryd and Sergio, it is worth sneaking a peak at Sergio's rookie season:
PG Minutes FGA eFG% FTA iFG% Reb Ast T/O Blk PF Pts PER
Sergio 06-7 19 13.6 .470 1.4 22 5.2 12.6 4.4 0.1 4.1 13.9 15.7
Opponent 14.6 .462 6.6 27 4.9 7.2 3.2 0.3 3.1 18.5 17.0
Rookie Sergio shot better, rebounded better, assisted better, and played more efficiently than rookie Jerryd. Having said that, Sergio has never recovered the...um...shooting touch he had as a rookie. So while Jerryd appears to be the ideal future point guard in terms of athleticism, he is still a long way from demonstrating he is the One. Sergio has to show he has learned anything during his first 3 years in the league; that's too harsh; his defense has improved.
Fans sometimes wonder about Rudy playing the point. He did so 2% of the time and was about as effective (or ineffective) as Jerryd and Sergio. This isn't enough time to form a definitive conclusion, but, given his stellar play at the 2, he looks like an SG from the stats, not a PG. And the statics on Roy playing the point have been uniformly negative after Brandon's rookie season.
Summary: It is not headline making news, but there is no one we can count on behind Steve to play point guard. We need to draft, trade for, or sign a point guard. We do not have this level of exposure at any other position except PF.
2. Shooting Guard
Here is the table summarizing the on/off court effectiveness of our shooting guards. Ready for a surprise?
% SG Blazers Opponent Outscore Outscore
SG Minutes Scoring Scoring +/- Opponent Blazers Win %
Fernandez 46 98.9 91.8 275 44 33 57
Roy 47 100.0 93.2 266 39 34 53
Bayless 4 85.5 100.5 -60 9 22 29
So by the metric of how often the team won with him as SG, Rudy was a bit more effective last season than Roy!. He had a better +/- as a shooting guard in fewer minutes. Of course he often played against the other team's scrubs, but he was also one of the closers during the final part of the season. This is amazing, because Brandon is pretty amzing himself. Rudy's SG game is borne out by the minutes at shooting guard, and it looks like Nate used them interchangably, at least in terms of minutes. What a luxury!
But wait. There's something wrong. Everyone knows Brandon Roy is legitimately All-Star, All-NBA, All-everything else, and Fernandez was not even close for rookie of the year. What's going on? The shooting guard statistics won't tell you; we'll see the answer later. There is not a lot of point in talking about Brandon as a SG. We know what he accomplished. Case closed. Let's move on.
While Bayless did not play enough SG for his statistics to be very meaningful, they are interesting. Both the team offense and team defense pretty much collapsed when he came in as the 3. People talk about him being a SG by nature, but this is not supported by the data. We'll try to understand why this happens when we examine his personal SG numbers:
SG FGA eFG% FTA iFG% Reb Ast T/O Blk PF Pts PER
Fernandez 15.5 .552 3.6 16 5.1 3.9 1.9 0.3 2.5 20.1 18.1
Opponent 14.7 .485 3.9 25 3.8 3.7 2.5 0.1 4.0 17.5 15.0
Roy 21.5 .512 7.7 34 5.8 7.1 2.6 0.3 1.9 28.5 26.4
Opponent 16.6 .527 3.5 26 5.5 4.2 2.2 0.4 3.8 20.4 17.1
Bayless 14.0 .214 6.5 48 4.0 6.7 4.7 0.0 5.5 10.5 0.7
Opponent 14.5 .457 3.0 24 5.5 4.7 1.7 0.2 2.5 16.0 14.5
No one can say enough about Roy; he was superior to his opponent nearly across the board: +4.2 FTA, +8% iFG, +0.3 rebounds per 48 minutes, 2.9 assists, 1.9 fewer fouls, 8.1 points per game, and +9.3 in PER. Since he is one of the primary ball handlers, I expect Coach McMillan uses Roy's T/O number as the upper bound of what can be tolerated; if Jerryd or Steve or Sergio can't get within shouting distance, why let them have the ball?
Rudy's personal numbers are not shabby, either; he exceeded Brandon in eFG% (on both offense and defense, turnovers, and personal fouls. This 48 minute point production at 20.1 points is sterling. By comparison, here Brandon's corresponding SG numbers from his rookie season:
SG FGA eFG% FTA iFG% Reb Ast T/O Blk PF Pts PER
Roy 06-07 17.2 .500 12.8 26 6.2 5.1 2.8 0.2 3.2 21.0 18.0
Opponent 16.2 .557 5.4 33 4.9 4.2 2.0 0.2 3.5 21.7 19.0
So rookie Rudy outdid rookie Brandon in eFG% (by +0.052 on offense, by +0.072 on defense), turnovers (by +0.9 on offense), blocks, personal fouls (by 0.7 on offense, 0.5 on defense), scoring differential (by +3.3 points per 48), and PER (by +0.1 on offense and +4.0 on defense). The 06-07 model of Roy played for a more disfunctional team, and while Rudy isn't exactly your usual rookie, I am not acusing Rudy of having a better rookie season, but only pointing out this still bodes well for the future.
Bayless' was the team's 3rd SG by minutes, and his SG numbers aren't encouraging. The tactic of attacking the rim appears to no longer work against larger SGs; 48% of his shots are still inside, but his eFG% plummets to a miserable .214; his FTAs are way down as well. The opponent SG can shoot over him, so their eFG% goes way up and iFG% goes way down; this is corroborated by Jerryd's improved PF numbers as an SG. What is interesting is Jerryd's assists are up, too. It appears being swarmed by stronger, bigger opponents caused Jerryd to find teammates to bail him out, to the degree where he finally begins to look like a PG! This is corroborated (and offset) by a spiraling turnover rate at SG.
Travis and Nick also played SG, but the minutes logged were too small to consider.
Summary: Unlike point guard, the statistics show the Blazers have SG covered, with two very strong players. It would be useful to have a solid backup who can hold his own in the event of injury or the need to use Roy or Rudy in other ways, but we are already better off than nearly any other team in the league.
3. Small Forward
For at least the past two seasons we have fretted about bleeding from the SF position. But this year 82games By Position Ranking says we had the 7th most effective corps of small forwards in the League. What? 7th best? In the whole league? Are you kidding? What happened?
Of course, we got Nicholas. Ah yes; that has to be it. We got Nick, right? Well, that's part of the answer. But the truth is:
% SF Blazers Opponent Outscore Outscore
SF Minutes Scoring Scoring +/- Opponent Blazers Win %
Roy 25 104.0 92.0 253 48 24 67
Batum 35 100.7 92.4 244 43 35 55
Outlaw 27 93.7 93.9 -10 36 42 46
For the second year in a row, Roy was a stand-out at SF. Say whatever you like. If Roy plays SF, we win the scoring battle 2/3rds of the time. Unfortunately, his 25% SF minutes shows this only works when the matchups are right, but the evidence is becoming overwhelming that Roy can play SF with a vengence.
Nick was the surpise of the season, and the table bears this out. Neiher scoring or defense falls off substantially when we transition the SF from Roy to Batum; the point differential falls only 3.7 points, from 12.0 to 8.3. It is interesting that this amount of point differential accounts for the 12% winning percentage differential when the rookie plays.
I am so used to the mantra that Travis is a mediocre SF that I was surprised by how effective he was this year. The offense falls off a cliff when Travis plays SF, but I was impressed that the defense didn't--are we talking about the same Travis fans love to abuse? This made me review his SF numbers from the 2007-2008 season: the offense improved from 85.7 ppg when Travis played the SF to 93.7 this past season. This is a massive jump, even though the team offense is still defective. Travis is on his way to becoming a respectable backup option at SF.
Here are the individual SF statistics:
SF FGA eFG% FTA iFG% Reb Ast T/O Blk PF Pts PER
Roy 22.2 .511 9.4 33 6.6 5.7 2.3 0.5 2.4 30.2 27.1
Opponent 13.9 .395 4.3 35 5.8 3.1 1.9 0.4 4.1 14.0 8.8
Batum 11.9 .541 1.8 34 7.2 2.5 1.7 1.3 4.5 14.3 14.8
Opponent 15.1 .514 3.9 32 5.8 3.1 2.4 0.4 3.0 19.3 16.4
Outlaw 17.0 .499 3.9 16 6.1 1.8 2.1 0.8 3.0 19.6 14.0
Opponent 15.7 .514 4.6 27 6.5 3.1 1.9 0.7 3.5 20.0 17.1
You can see Roy dominating at SF in a way he doesn't at SG: FGA (+8.3 per 48 minutes over the opponent SF), eFG% (+.116 over the oponent SF), FTAs (+5.1), Asts (+2.6), PFs (+1.7), Pts (+16.0 per 48 minutes), and PER (+18.3). Playing him as SF seems to throw the opponents' defense into chaos and make their offense stagnant. Say whatever you want. Repeat the mantra: Roy is not a small forward. Let the Blazers assert it. Let Brandon say it himself. But people, these numbers are getting into Wade territory, CP3 territory. Brandon Roy has become an All-Stud small forward, and this is the best kept secret in the NBA today.
So what about Nick? Maybe the eFG% of his opponent is a bit high, but that is probably because he guarded the opponent's best player instead of his own man. The FTA number, the opponent's assists, Nick's turnovers, the personal foul differential--all of these suggest a rookie. But rebounding? Already better than his opponent.
And lastly Mr. Outlaw. Ok; his numbers aren't sterling, but they say we didn't lose very much when he stepped onto the court as our small forward--the oppoment SF shot a little better, shot a few more free throws, got slightly more rebounds, turned the ball over slightly less, and blocked a few more show--but only a little better than Travis. You can see Travis' aversion to contact in the iFG%, where his opponent abuses him with 27% of his shots inside to Travis' 17. But this is a team low for opponents SFs (by a significant margin), and suggests that maybe Travis is learning how to defend drives to the rim. His stats give the signature of a great role player. Travis has become a useful citizen in Blazerland.
One more observation: I believe the jury has rendered a unanimous verdict: when both Brandon and Rudy are on the floor together, make Roy the SF and Rudy the SG. Again, this only works when the matchups allow for it, but it maximizes the teams' chance of success more often than not.
Summary: The SF position shows a surprising route to the Blazers' success. Have a great player at the starting position, a second guy off the bench who can tread water, followed by a world beater to finish them off for dead.
4. Power Forward
Truth in Advertising Alert: LaMarcus became my favorite Blazer this season, so you can filter any bias in the analysis as needed.
% PF Blazers Opponent Outscore Outscore
PF Minutes Scoring Scoring +/- Opponent Blazers Win %
Outlaw 19 107.4 96.1 181 41 22 65
Aldridge 68 98.3 91.4 387 45 33 58
Frye 8 88.0 91.1 -21 16 27 37
So Travis made a bigger positive impact as a PF than my man? We outscored the opposition 65% of the time while Travis is on the court as a PF? Sweet. Maybe we (the fans) have dismissed the Travis option too soon. He played the PF only 19% of the time, again suggesting Travis at PF is an option we can effectively exploit only when the matchups are right, but a +11.3 scoring differential and outscoring the opponent 65% of the time can't be sneezed at.
LaMarcus played a team high 68% of the PF miutes, the largest percentage for any player at any position. This is because only other alternative available besides Travis seemed to be Channing. Frye didn't distinguish himself particularly as a PF, with our team outscoring its opponents only 37% of the time while he was our PF--that translates to 30 wins a season. As in Steve's case, these numbers say LaMarcus really needs help.
Speaking of LMA, what did he do for the team when he was on the floor as our PF? We scored nearly 7 points a game more when Aldridge manned the PF. Blake was the only Blazer with a bigger +/- throughout the season. LaMarcus accumulated his +/- in the West, which is insanely deep at the 4. 82games By Position ranking says we had the 2nd most effective PF corps in the league last season. The PF minutes numbers say this success can be attributed almost exclusively to LaMarcus. "All animals are equal, but some are more equal than others."
Here are the individual PF numbers:
PF FGA eFG% FTA iFG% Reb Ast T/O Blk PF Pts PER
Outlaw 20.2 .518 7.5 22 8.8 1.7 1.8 1.9 3.7 26.9 23.0
Opponent 17.0 .546 5.2 38 9.2 2.7 1.9 1.2 6.7 22.7 19.9
Aldridge 20.3 .485 5.1 33 9.5 2.7 1.9 1.3 3.1 23.7 21.6
Opponent 15.8 .503 3.9 36 8.6 2.2 2.2 0.8 5.2 18.9 15.5
Frye 19.8 .438 2.8 16 7.7 2.2 2.0 0.6 5.4 18.9 11.2
Opponent 16.0 .477 4.7 36 8.9 3.2 1.9 0.1 4.9 19.2 15.8
The opposition PFs could score on Travis and LaMarcus relatively easily (eFG% of .546 and .503, respectively), but they had their own trouble on defense, fouling our quicker guys--speed kills as they say. Travis' PF PER of 23.0 is astonishing. LaMarcus' PF PER of 21.6 is not bad, either, considering he did so much of the heavy lifting night in and night out.
Given all the abuse he receives from a few of the fans for not being Karl Malone, I find it interesting that LMA's iFG% is only slightly below the average for an average PF, and that he outrebounds his opponent by +0.9 rebounds per 48 minutes, a signifcant margin. The cant about him being a spineless rebounding wuss is not borne out by the evidence. If you look at the numbers for our other bigs, none rebound as well in the PF position as when they play center (even Joel averaged only 15.9 rpg as a PF). It appears lack of rebounding at the PF is something structural in the way the Blazer offense works. Somehow plugging in Paul Millsap or Brandon Bass as LaMarcus' backup is unlikely to change this by itself.
The numbers demonstrate Frye just doesn't cut it for us as a backup PF in the Blazers' system. This is hard to say, because I really like Channing, and thought he had turned the corner last summer. Wrong again.
Summary: While Travis can excel in the right situation, there is no one else behind LMA at PF; LaMarcus does as amazing a job pulling the sleigh as anyone on the team. We need to show our apprciation by getting him some help.
Last are the centers. Here are their on/off court numbers:
% PF Blazers Opponent Outscore Outscore
C Minutes Scoring Scoring +/- Opponent Blazers Win %
Aldridge 7 120.1 103.8 96 31 14 69
Oden 33 100.4 94.1 170 36 24 60
Pryzbilla 49 96.2 89.0 294 37 40 48
Frye 10 91.6 103.1 -100 15 32 32
Wait a minute. There's that LaMarcus guy again, with the kind of numbers fans love. We outscore the opposition 69% of the time he plays center? We average 120.1 points per 48 when this happens? Sign me up. Maybe smallball isn't too bad.
That Oden guy may be pretty good, too. You can see how much he changes the game just by being on the court; even though the team scoring differential is "only" +6.3 ppg with Greg as our center, we outscore the opponent 12% more often than with Joel, who owns a higher differential of +7.2 ppg. Given Joel's guady +/-, it surprised me was that we only broke even (at 48%) when Joel played center. Frye did not hold his own as a backup center, and the team was man-handled whenever we were forced into this option.
The personal statistics are as follows:
C FGA eFG% FTA iFG% Reb Ast T/O Blk PF Pts PER
Aldridge 16.1 .500 7.6 36 11.6 1.4 1.9 1.0 5.8 21.8 20.2
Opponent 12.6 .595 4.9 63 12.9 2.2 1.4 1.9 6.3 18.7 22.8
Oden 12.8 .564 8.3 90 15.5 1.1 3.2 2.6 8.7 19.7 20.3
Opponent 12.3 .552 5.1 59 9.2 1.8 2.2 2.8 6.8 17.5 17.5
Pryzbilla 6.6 .623 4.1 92 17.6 0.6 1.9 2.4 5.3 11.0 17.5
Opponent 14.4 .515 4.5 51 10.9 1.8 2.4 1.4 4.4 18.3 17.5
Frye 15.1 .450 2.0 20 10.0 1.0 1.7 1.3 8.0 15.4 11.6
Opponent 12.6 .541 7.9 60 11.9 2.7 2.3 2.5 4.6 19.4 22.9
If nothing else, all of our centers are really expert at fouling. And this is not just Greg.
The data says the Blazers' centers shoot less than any other position: PGs about 13 times a game; SG about 18 times a game; SF and PF roughly 20 times a game each; C 12. This is bizarre, but it is probably a relic of the days when Joel was all we had. Perhaps this is related to having good options scoring at SG, SF and PF, but some of it has to be our weak set of PGs; they can't get the ball into the post in a way that is always useful. The entire team seems to have problems executing the pick-and-roll at both ends of the court, and this shows up as dismal FGA numbers for the centers. The exception is LaMarcus, who is a shot magnet.
The numbers also show that LaMarcus struggles a bit to hold his own as a center, as he is out rebounded and shoots a much lower percentage of inside shots than the opposing center. His opponent grabs more rebounds--but his own rebounds are way up from when he plays PF. However, the FTA--man, those are Bayless like FTA numbers--and PF differential is squarely in his favor, suggesting he compensates for his relative lack of bulk with speed.
News Flash: looking at the iFG% and Reb numbers, Joel and Greg positively man-handled their opponents inside. Greg's PER of 20.3 is unreal for a rookie. No wonder he was such a disappointment.
Summary: Overall, the corps of centers didn't do nearly as well as the SG/SF/PF position, as we were ranked a "disappointing" 16th in the league. The only things holding us back are LaMarcus' defense against bigger centers, Joel's offense, and Greg's inexperience. We believe at least the last one of these will change for the better next season. But you can see the SF recipe of three really good players at one position leading to a higher ranking in the future. The difference from the SF case is here is all three horses will eventually be running wild.