Hollinger Celtics Preview

[saltlover]

I think Hollinger is somewhat of a slave to PER and his models.  And that's okay, because he made them and he's getting paid because of them.  But PER can't account for things that aren't publicly tracked.  For instance, if I had the time, I'd love to see the fg% of shots that come off of Rondo passes, vs. other's passes, vs. the dribble.  My gut instinct says that Rondo is amazing at creating better shots for people, but I don't know.  Maybe he just gets a lot of assists because he passes the ball a lot.  It's not tracked.

Similarly, there may be things that Jeff Green does which the Celtics place value on, but do not show up on a statline.  Furthermore, they may see things that Jeff Green does well under certain situations, and they have the ability to put him in those situations better than the Thunder did, and better than they did mid-season a couple years ago.  Perhaps those situations are unique to the Celtics, or perhaps Jeff Green being in the wrong situation was unique to his time in OKC, since Kevin Durant was a similar but much better player, and thus was not complimentary to Green.  If the Celtics are unique in their ability to utilize Green, then they did indeed overpay.  If the Celtics are not unique, then they may not have.  Unless he performs well, is traded, and continues to perform well, we'll never know if they didn't overpay.  If he doesn't perform well, we'll know they did overpay.

Also, Hollinger's model probably hates Jeff Green, because he was an average player, and he missed a year, which the model will assume was a serious injury with a lengthy recovery time.  In basketball, these are typically ACL tears, which rob players of effectiveness for multiple seasons.  Assuming that his heart ailment was truly a one-shot deal, he won't have problems with speed, agility, and strength, like those recovering from ACL's would, but just stamina, which will be mitigated due to his reserve role.

In what (admittedly little since he was paywalled) I've read from Hollinger, I've never seen him point out why his model might be wrong at times.  That bothers me.  It may or may not be wrong about Jeff Green, but it is certainly imperfect, and he needs to articulate those imperfections better.

[Q_FBE]

Hollinger, the ESPN mathematician is add it again apparently. His statistics were based over the full 66 games and I admit, the Celtics were horrible the 1st half of the season. Over the last half they made a significant surge with KG at Center. Ultimately the Celtics lost at Miami in seven games when Ray Allen missed the critical 3 pointer with 2:23 remaining in Game 2 to put the Celtics ahead 97-89. The Celtics could have easily closed this out in Five Games last year and we would all be singing a different tune.

Lets see - we lost Ray Allen (to our current arch rivals) and Greg Stiesma and got back Jeff Green (presumably upgrade), Jason Collins (upgrade), Darko Milicic (unknown), rookies Sullenger, Fab Melo.

Maybe Hollinger is using his spreadsheet and computer program to make a mis diagnoses of how NBA teams will do this year.

[sofutomygaha]

Hollinger gets a bad rap. What he's trying to do it very hard and he generally limits himself to trying to tell people what the numbers say.

To his credit, he laboriously acknowledges that PER is an offensive stat. In his scouting reports, he always dedicates a paragraph to what the +/- and synergy numbers say about a player's defensive value, and he discusses how well those very imperfect measures pass the eye test. He also frequently warns about the difference between good assists and bad assists by frequently pointing out the shooting percentages of the shot recipients and the home/road differentials that might suggest inflated assist totals (see Chris Paul's near MVP season).

Critics are right to point out that his evaluations are unforgivingly performance-driven. A player can flash certain abilities or demonstrate potential, but Hollinger won't have anything to say about it until they've been putting it to effect for a year or more. Hollinger acknowledged that Green was playing out of position in OKC and his '11-12 report reflected that. The difference of outlook comes from the fact that there isn't enough data on Green at SF to judge that he is worth $8M at that position.  Look at what he says- he doesn't say "Jeff Green is a bad player," he says that Green hasn't done anything yet to prove that he's more than a passable combo forward off the bench.

Hollinger's schema is notoriously hard on college players who haven't produced yet, even though he and everyone else knows how important it is to identify potential and not to always trust ultra productive college players with the wrong style of game.

I think that you should take Hollinger's stuff for what it is- an extremely useful way to look at what a player has accomplished. We obviously expect a lot more from Green than what he's accomplished so far, but you'd have to concede that his resumé so far is not that of an $8M small forward. He's completely right. We traded for him and we're paying him because we're betting that he will beat the projections- if he does you can be sure Hollinger will be hard at work showing people how he did it.

[More Banners]

$9M is the size of the hole in the roster if Jeff Green doesn't sign.

We might have the best rotation of any team we've seen in my lifetime, and Green makes that possible.  Since we're going for all the marbles, you try to carry the biggest guns you can muster, and Green was surely the best player we could possibly add, and we had $9M to give him.  Fine.

He's a good bet as a sub-star due to his versatility; he could play next to just about anyone.  No reason not to lock him up.

Perhaps $7M would have been better, but in context of what we're trying to do, if the only quibble on the roster is an extra $2M to Green, that's not a wild gamble, especially when you have an aging star 3 that has go go against Carmelo and Gerald Wallace in his division and Lebron every playoffs.

[saltlover]

Honestly, if he's not using statistical projections, and rather relying on a metric that only looks at what players have done in the past, to evaluate a contract, he's doing it wrong.  But his comments about Jeff Green look foolish in their hyperbole.  "No evidence that he's actually any good?"  Come on.  "Major health risk?"  He'd played in almost every game prior to last season, and suffered from a serious but not chronic condition that has been completely addressed.

Was this a risky deal?  Yes, it was.  But comments like "Can they amnesty him yet?" are complete overkill, and help to confuse the less informed readers (since Green's contract is ineligible for the amnesty provision.)  It's these statements, based upon an metric with a lot of flaws.  For instance, it rewards players who take more shots.  In fact, much of the drop in Green's PER from the 09-10 season to the 10-11 season can be explained by the fact that he took fewer shots, even though his TS% increased.  My criticism isn't that PER is flawed -- I think it has it's uses.  My criticism is that he makes unapologetic pronouncements from it as if it weren't flawed, and this hurts his credibility specifically, and the credibility of advanced sports statistics at large.

[bdm860]

Hollinger, the ESPN mathematician is add it again apparently.

Oh I see that pun you snuck in there. 

Completely off topic, but before my post, has there ever been a page of a thread on Celticsblog with so many words and no pictures? 5 posts and 4 different posters, several paragraphs of hundreds to thousands of words but no quotes, no avatars, no sigs, no embedded images, no youtube clips, nothing?  Just looks kind of weird.  Sorry for posting and ruining this page of monochromaticity with my quote, avatar, and sig.




Carry on.