I pledge to say no more about Seattle-area shooting guards. (At least until training camp)

Why would you use SG exclusively for #s? Is Crawford being classed at other positions somehow unrelated? Last season was only an anomlay is you use SG exclusively, as it skews the stats against Crawford.

I’m not a Berri disciple who says efficiency is everything. For example, David Lee plays a very different role from Eddy Curry; it’s not a straight comparison and just because Lee’s eFG% is higher doesn’t mean he’s a better offensive player. But Nate and Crawford play the same position, the play similar roles in the offense, they take a similar number of shots and a similar mix of outside shots and jumpers. Nate just makes more of them (except for slightly fewer free throws). It’s an easy, apples to apples comparison.

Shifting to shooting guard has nothing to do with his eFG%. You are correct that TO rates tend to be slightly higher for point guards than shooting guards, but Nate still isn’t turnover-prone at all. His rate is better than 2/3 of all NBA shooting guards (it would be top 10 for PGs). Crawford’s higher rate might be due to his extra ballhandling duties, but it’s not like he was our full-time ballhandler – he plays shooting guard, too.

re: defense, you’re the one who brought up the opposition SG #s on 82games, and Mike correctly points out that last season WAS an aberration for Crawford.

Robinson’s turnover rate per possesion declined significantly for the same reason his assist rate declined significantly. That is to say, Robinson was utilized as an off-guard more and relied on less for passing and playmaking. Hinrich, Hughes, Johnson and Crawford are all playmaking guards whose involvement in the game is much more significant than that of Robinson.

Robinson isn’t a playmaker. His assist numbers are mediocre. There is no possible argument for Robinson being a superior passer and playmaker for Crawford. It’s a major part of guard play and Robinson doesn’t have it.

Robinson is a marginally better rebounder than Crawford. He’s hardly a good rebounding guard like Richardson. Realistically, on a Knick team that was one of the best rebounding teams in the league and one of the worst for assists, is this of importance?

Robinson’s eFG% improved because of the aforementioned change from PG to SG. His assist numbers went down and every other statistic was virtually unchanged. Robinson still plays he same style and he’s shown no signs of improvement. Aside from your belief that all players who are 23 years old improve, I don’t find any evidence to support your theory that Robinson will improve.

Why are people using the positional stats to rate his defense? Why is Crawford’s time at SG the relevant statistic? If you’re using overall numbers, this season is hardly a statistical aberration.

Good point, but there is one caveat. Opposition PER (oPER) isn’t that reliable either. It doesn’t always pick the correct defender by position (sometimes last year Marbury defensed the SG with JC on the court), and it doesn’t account for help defense. It also doesn’t account for rotations on defense due to a breakdown. For instance a player’s defender may get past him, but he may neither get the assist nor score as the defense rotates and the ball swings around to an open player. You could argue the merits of +/- versus oPER all day. Personally, I like to look at both.

Nonetheless Crawford did post a respectable 14.8 oPER against other SG last year. However The year before his oPER was 18.0, and the year prior to that it was 18.2. So either Crawford became a good defender last year, or we’re looking at a statistical outlier. I’m inclined to believe the latter, but I’ll keep an eye on his oPER and his performance on the court

Well if his mean don’t produce, then neither does Crawford because his own production is marginally better than that of his opponent. So if that makes him a good defensive player, you have to consider him a bad offensive player.

“Robinson is a pure energy guy who makes copious mistakes when given heavy minutes…”

Despite his rep as a wildman, Nate turned the ball over on only 9.6 percent of his posessions, significantly better than average for a shooting guard (and much better if you want to consider him a PG). Among the guys who did worse were Kirk Hinrich, Larry Hughes and Joe Johnson.
Nate’s TO rate of 9.6 was also down by a third from his rookie year, a nice sign of improvement.

Crawford’s TO rate was 11.4.

“How is Nate Robinson better than Crawford? He can?t run an offense, he can?t pass, and he rarely plays any defense.”

He may not be a point guard (Crawford can’t run an ofense full-time, either), but his assist totals aren’t wretched and his A/TO ratio is only slightly worse than Crawford’s. (1.6 vs 1.3 / 1)

He’s also good rebounder, for a guard, which adds some value. And of course he is a much better shooter than Crawford.

There are plenty of other posts on the defense, and Crawford may be better than I thought, but there’s no evidence he’s actually good.

“Why are you so insistent that Robinson will improve? Do you have any factual basis to support this? Is the man going to somehow develop PG skills in the offseason? He certainly hasn?t ever shown any signs of change in his game.”

Nate improved in every category from his rookie year. Beyond that, virtually every player in the league gets better from age 23 to their late 20s. It doesn’t mean he will become a point guard, or a shot-blocking center, but he will almost certainly become a better shooting guard.

Doesn’t it say that opposing 2-guards did fine against Crawford? An eFG% 3 points higher than his, though shooting less than he did?

I never heard anyone accuse Frye of being a good defender…. until now. But not just fatkid… Jon Nichols has come up with a 2.0 version of his defensive composite ratings over at 82games.
http://82games.com/nichols2.htm

Frye actually rated above average.
- Balkman (87th percentile)
- Lee (74th)

Everyone else below average, though, Crawford, surprisingly to me, not by much. (45th percentile)
Collins (42nd percentile)
Richardson (39th percentile)
Curry (36th percentile)
Francis (35th percentile)
Jeffries (21st percentile)
Robinson (19th percentile)
Marbury (18th percentile)
Rose (3rd percentile)

The best defenders in league, by this rating, were Duncan, Joel Przybilla & Bruce Bowen. The worst “name” player – if you call it a name – was Kyle Korver.

This system is extremely complex, and has some of the same weaknesses as adjusted plus-minus, but it seems well thought out and a better defensive measurement than any I’ve seen. Thoughts?
http://82games.com/nichols2.htm

This comes back to the basic fundamental question of defense. Most players are either “good defenders” or “bad defenders” based on reputation and little else. Curry has been labeled a horrendous defender for years, but do other centers put up numbers against him? No, they don’t. Frye was deemed a good defender, but every PF in the league was having a great night against him.

And statistical +/- is a highly ambiguous metric. Is David Lee really an amazing defensive player, or does he tend to share floor time with other good defenders whilst playing against the second unit? Likewise, is Jeffries’ defense awful, or is he usually on the floor with weak defenders while matched against starters?