Predictions for Jeremy Lin

[xkonk]

Lin has come out of nowhere to put up pretty impressive numbers in his last four games. I'm curious what people think about two different Lin-related issues:

1) How does he play from here on out? Is this just a hot streak by a new player and the league will figure him out? Or is this the 'real' Lin? He wasn't especially impressive in any of his time prior to these four games, with either the Warriors or the Knicks. On the other hand, he's only played more than 20 minutes in a game two or three times prior to this streak and the last four games represent about a third of his career playing time.

2) What effect will Carmelo's return have on the Knicks? I can see two potential reasons to believe things get even better for Lin when Carmelo gets back in the line-up: Carmelo will eat up some of Lin's usage, and lower usage should lead to better efficiency. Also, Carmelo is supposed to make his teammates better, a la the big debate around trade time last year. Those two reasons are likely conflated. But, if Lin is expected to play better when Carmelo returns, it's hard to see how the Knicks wouldn't continue to play well. On the other hand, I for one don't put much belief in either of those claims. I see Carmelo returning and the Knicks' play declining as he eats up possessions that Lin would put to better use.

[Mike G]

To throw a wet blanket on this warm and fuzzy story, Lin has had 4.3 TO/36 min. this season.
He had 8 in his first 54 min. (5.3 TO/36) and 17 in the last 4 games (4.0)
So, he's gotten better.

[J.E.]

I think Carmelo returning should be OK as long he's not taking up any minutes that would have otherwise gotten to Fields, who is playing out of his mind according to all things +/-.
The much bigger problem will be Stoudemires' return because he is really having an awful season. Without him the Knicks defense looks quite good

[deepak]

I think Carmelo returning should be OK as long he's not taking up any minutes that would have otherwise gotten to Fields, who is playing out of his mind according to all things +/-.
The much bigger problem will be Stoudemires' return because he is really having an awful season. Without him the Knicks defense looks quite good

I actually think Stoudemire will benefit quite a bit (offensively) from having a real PG who knows how to run a pick and roll properly.

[Crow]

It would take either some skilled technical work and a comprehensive search- or alternatively, a bit of target hunting and simple methods, using accumulated knowledge of the league and knowing where to look, to undercover other cases of guys who had a "burst" of some level early in their careers (of a week or a month on some metric, such as winshares per 48) that was notable / dramatic and then track what happened from there and construct a sample or comprehensive distribution of how many of those players turned that burst into an outstanding, very good or good or even average career.

The more advanced the technical skills you have, the easier you could handle a bigger dataset and probably faster (though design time and design approval and design change are costs); but you might be able to develop enough of what you need in a few hours without fancy programming and you could always work a little longer if needed.

Is it better to have a lot of thoughtful basketball analysts (& hours) in an operation and a few techies or a few thoughtful basketball analysts (& hours) and a lot techies (& hours)? There is some relatively efficient zone with a mix of both but you better not be short on either ingredient. My impression is that GMs, assistant GMs and Coaches and assistant Coach are probably often short on time and might need help with analysis, especially if the techies' noses are kept at churning out the technical product. What percent of analysis is getting done using available public and proprietary information? Do any teams claim to near full initial digestion of all information or near full secondary / tertiary use of and priorization of that data and analysis? For any who are not making that claim or not making that claim with a strong case demonstrating it to be true, wouldn't you be looking to add analytic capability, in addition to / or apart from more tech guys to produce more data?

For those in a position to "know" or have an informed idea, what is the current average and marginal rate of return on analytic staff in general and for "techies" vs "basketball analysts"? Does your team believe that it has sufficiently explored the talent base and spent up right up to the optimal limit for these functions? Are you sure, completely sure? To be point that it isn't worth "risking" another $50,000 or more in pursuit of goals that owners and fans attached ten of millions of dollars of value to?

[xkonk]

One person, at least, seems ready to put Lin's (future) jersey up in the rafters: http://fivethirtyeight.blogs.nytimes.co ... -fluke/?hp

[Crow]

So Nate Silver did something with a powerful but publically available data source and presumably in a few hours or a lot less. I am not a fan of all of the analysis or phrasing in the writeup but it is a example of quickly obtained burst analysis. No new tech design, approval and implementation. No advanced programming language skills needed by this analyst in this case using this method. What is the information he developed worth? What is it worth to be scouring the net for such information? Would a comprehensive in house tech solution by worth a lot more? How much more? How much are you willing to pay? How much time are you willing and able to give to reading and rooting through the output? Do you want to do "enough" or do more than anyone else to maybe get edge? How valuable is any little bit of edge?

There are many other quick cuts at "bursts" I or others could put into B-R searches, if I or they had a reason to. How many teams can say they have already fully explored the topics of "bursts" and estimated and actual performance after bursts? Is it simply a techie "solution" or a search by some means then a lot of "basketball analysis"?

[Crow]

It appears Lin is playing great 1 on 1 defense and / or getting lucky. Overall team defense is neutral with him on / off using the raw data. Modestly positive on the defensive RAPM, prior informed and not.

[Jeff Fogle]

I don't know...how can an article like that not mention that Lin was facing the 30th ranked defense, the 26th ranked defense, the 21st ranked defense, and then an exhausted Lakers team that was playing on a b2b after overtime with a very poor defensive point guard. How can that NOT be in the discussion?

Remember asking similar questions a year ago when NS wrote about "The Carmelo Effect" without mentioning the possibility that Denver's home edges at altitude (getting to face more tired defenses) or Karl's ability to run a well-designed offense may have influenced the data. (Great that we can now go look at Carmelo's quint-essentials from the last few years!).

It's like the thing from the ESPN stat guys that was posted True Hoop a few days ago about the 76ers. They mentioned the +10.0, but didn't mention that Philly has had an extremely home heavy schedule, and a very friendly schedules in terms of weak opposition (ranked 29th at the time in SOS by Sagarin). Philly wouldn't be at +10.0 against a balanced schedule...and they subsequently fell below the threshold with home losses to SA and LAC (which made their schedule even more home heavy!). Will Philadelphia look like championship material once their schedule evens out? That question should at least be asked in any assessment of their championship potential.

Very frustrating that stuff running in the NY Times or at ESPN leaves out important context when making evaluations.

At least Lin's streak stopped at four...which helped lop off 12 of the top 15 guys on the list who were at 5 or better. Would have liked to have seen a qualifier on heights too...how often do guys in his height range rise up to be all-time greats or franchise players...how much more common is it percentage-wise for them to settle into something that doesn't carry franchises.

Adjustments for early schedule strength and the metaphorical "broad shouldered-ness" for carrying a franchise would probably lead to more reasonable assessments. And, it's unlikely Lin would have topped the 50% in each game threshold vs. a more representative set of defenses anyway. At least less likely.

It is a fun story to follow though. Fun kid to root for. His wikipedia page says he wants to be a priest some day. May be seeing a Michael Chang type replay...

[J.E.]

I don't know...how can an article like that not mention that Lin was facing the 30th ranked defense, the 26th ranked defense, the 21st ranked defense, and then an exhausted Lakers team that was playing on a b2b after overtime with a very poor defensive point guard. How can that NOT be in the discussion?

Right. It reminds me a little bit of Jennings going crazy (55pts) in '09. Many people portrayed him as the next Jordan. Might have wanted to take into account that it was against the Warriors, which were #1 in pace, #29 in DRtg; anyone who had seen the game might have also noticed that GSW went under *every* screen in that game. Basically they didn't care about how many points Jennings would score, at all.
Lin's situation is a little different, but he certainly didn't face the toughest opponents as of yet

[Statman]

For those in a position to "know" or have an informed idea, what is the current average and marginal rate of return on analytic staff in general and for "techies" vs "basketball analysts"? Does your team believe that it has sufficiently explored the talent base and spent up right up to the optimal limit for these functions? Are you sure, completely sure? To be point that it isn't worth "risking" another $50,000 or more in pursuit of goals that owners and fans attached ten of millions of dollars of value to?

I don't know - but I DO know that if the Timberwolves would have had me (or probably any number of "analytical" guys) on their payroll - they wouldn't have drafted a 26 year old that some scout had seen play a couple times on the other side of the world. Heck - I would have had that guy rated in his leagues and tourneys with probably thousands of other guys across the world in various skillsets - taking into account pace/stats/team quality/etc of everyone. No scout can "see" that many players.

I've exchanged emails with Mark Cuban in the past - last season he told me they had no need for another "stats" guy because they had Roland Beech. I didn't ask - but I thought "seriously, you have ONE 'stats' guy? Really?" He also told me he had zero interest in the college basketball player analytics I do - since he doesn't follow college basketball.

The Mavs are supposedly one of the most advanced teams in terms of analytics. Take it for what it's worth.

As for Lin - he rated highly in my ratings in college - despite weaker comp. He had a very high FT rate, with very good rebound and assist rate for his position and pace- with a great steal rate. Those are the 4 things I look at first in the college stats when it comes to an NBA future for a guard - since the rebounding/steals often correspond to quality athleticism/quickness - and FT rate to getting to the hole creating contact and creating you own shot. So - while fairly surpised - he didn't appear outta nowhere on my radar, I suspected a couple years ago he would be better than many low level NBA PGs.

[bchaikin]

As for Lin - he rated highly in my ratings in college - despite weaker comp. He had a very high FT rate, with very good rebound and assist rate for his position and pace- with a great steal rate... I suspected a couple years ago he would be better than many low level NBA PGs.

the problem is that players from div I tier 4 schools (the lowest conferences competition-wise like the atlantic sun, big south, ivy, metro atlantic, mideastern, northeast, patriot, southern, southland, summit, SWAC) rarely - and i mean rarely - make it to the nba. only 4 in the last decade have played major minutes in the league (stephen curry, george hill, kevin martin, and jason thompson - this year andrew goudelock is playing for the lakers and reggie williams for charlotte), and i believe only ten or so from div I tier 4 schools were even drafted in the last decade...

lin played at harvard for 4 years (06-07 to 09-10), and only the stats for his last 2 seasons were of any notice (ages 20-21). those stats those 2 years were very good for a div I tier 4 player, but not outstanding. when you compare those stats to the stats of other 20-21 year old div I tier 4 players, they don't compare to those of say a stephen curry (curry averaged 34.0 pts/40min, lin 20.4 pts/40min at that age) or reggie williams (much better scorer/rebounder), and many had similar passing and scoring rates as lin, with lower/better turnover per touch rates...

even after playing in 29 games for golden state last season, i doubt many had lin on their radar. he's just making the most of an opportunity now. odds are he'll soon be able to afford a bigger couch...

out of curiousity how did you have keith benson rated (now in the D-league)? i ask as i look at his college stats (oakland) and see little difference to those of say jason thompson (rider), both went to div 1 tier 4 schools, and both are 6-11. looking at each's last 2 years in school they had similar scoring, shooting, and rebounding rates, yet benson was by far the better shot blocker (by over 30%) with a lower turnover rate. yet thompson was a # 12 overall pick (2008) but benson only a mid 2nd rounder (2011). do you suspect benson will be a good pro, better than thompson?...

[mystic]

I've exchanged emails with Mark Cuban in the past - last season he told me they had no need for another "stats" guy because they had Roland Beech. I didn't ask - but I thought "seriously, you have ONE 'stats' guy? Really?" He also told me he had zero interest in the college basketball player analytics I do - since he doesn't follow college basketball.

That sounds rather like Cuban trying to be polite. I don't think that the Mavericks have just "one stats guy", but rather one leading stats guy in Roland Beech. The Mavericks are still getting the stuff from Wayne Winston too, so, what exactly could you offer what they don't have already? Another opinion on how to interpret those numbers? I find it hard to believe that this will lead to an improvement.

And the Mavericks are usually picking late and usually finding nice players for the respective positions they are looking for. So, I would suspect that they have a scouting system for college players in place and are happy with the results. Your analysis might just not add anything new to their perspective.

As for Lin - he rated highly in my ratings in college - despite weaker comp. He had a very high FT rate, with very good rebound and assist rate for his position and pace- with a great steal rate. Those are the 4 things I look at first in the college stats when it comes to an NBA future for a guard - since the rebounding/steals often correspond to quality athleticism/quickness - and FT rate to getting to the hole creating contact and creating you own shot. So - while fairly surpised - he didn't appear outta nowhere on my radar, I suspected a couple years ago he would be better than many low level NBA PGs.

And the funny thing is that Donnie Nelson was the first GM to approach Jeremy Lin and offering him the spot on the Mavericks summer league roster. The Mavericks were the team bringing Lin into the spotlight, which led to his contract with the Warriors.

http://sports.espn.go.com/dallas/nba/co ... id=5387547

If financial offers and plans for Lin's development are relatively equal, he seems to have an affinity for the Mavs and Nelson, who was the first NBA executive to latch onto him in April at the Portsmouth Invitational, and the only one to offer Lin a summer league roster spot.

"He said he loved my game and the way I played and my basketball IQ, but then he said, 'I think you're about one year away from the NBA, so I want you to come play for my D-League team,' " Lin said. "That's how it started. He didn't draft me because he still thought I needed a year to develop, but he wanted me to come play for them, get comfortable with the city, get comfortable with the organization."

Nelson was even right about that one year. So, what exactly would you have told the Mavericks regarding Jeremy Lin? They offered him a 1yr fully guaranteed contract and Lin decided to take the same offer from the Warriors. Maybe he and his manager thought that it is more likely to get playing time in Oakland than in Dallas.

[Statman]

out of curiousity how did you have keith benson rated (now in the D-league)? i ask as i look at his college stats (oakland) and see little difference to those of say jason thompson (rider), both went to div 1 tier 4 schools, and both are 6-11. looking at each's last 2 years in school they had similar scoring, shooting, and rebounding rates, yet benson was by far the better shot blocker (by over 30%) with a lower turnover rate. yet thompson was a # 12 overall pick (2008) but benson only a mid 2nd rounder (2011). do you suspect benson will be a good pro, better than thompson?...

My ratings had Jason Thompson ranked 44th nationally in '08 - with a 154 overall rating.

Benson was ranked 87th nationally in '11 with a 145 rating.

Now, while some players do make it in the NBA ranked that low (87th is really good - just not normally quality NBA future good) - it's fairly rare, especially for an upperclassman.

I thought Thompson was a bit of a reach as the 12 pick back then.

I believe Benson rated worse because Oakland was a bit faster paced, he made less of a statistical footprint in relation to his teammates and opponents than Thompson. 9 rating points isn't that far apart though.

[Statman]

That sounds rather like Cuban trying to be polite. I don't think that the Mavericks have just "one stats guy", but rather one leading stats guy in Roland Beech. The Mavericks are still getting the stuff from Wayne Winston too, so, what exactly could you offer what they don't have already? Another opinion on how to interpret those numbers? I find it hard to believe that this will lead to an improvement.

Nelson was even right about that one year. So, what exactly would you have told the Mavericks regarding Jeremy Lin? They offered him a 1yr fully guaranteed contract and Lin decided to take the same offer from the Warriors. Maybe he and his manager thought that it is more likely to get playing time in Oakland than in Dallas.

You asked what I would have "told" the Mavs about Lin, but that was not the point of my post (I mentioned the Timberwolves because of such an obvious example of a team being horribly underprepared at the time they should be the MOST prepared in terms of scouting and analytics of players from other leagues) - all I was saying is that Cuban had told me they had A "stats" guy. I was inquiring about some sort of entry level intern type position - hardly trying be some head of analytics (like my man DeanO). They very well may have some great team of "stats" guys - although I think some here would know some others other than just Roland if they did. I'm not even sure Winston works with them anymore, but he may. Anyway - I was just answering the question Crow had with my VERY limited info from one email about how many analytics guys do these teams really work with. I'm guessing, in many cases - VERY few - compared to scouts in the NBA and scouts/analytics guys in MLB.

And the Mavericks are usually picking late and usually finding nice players for the respective positions they are looking for. So, I would suspect that they have a scouting system for college players in place and are happy with the results. Your analysis might just not add anything new to their perspective.

What nice players for their respective positions have the Mavs drafted outta college? They got Josh Howard almost 9 years ago. They did flip Tractor Traylor for the Big German 5 years before that.

I know they often are drafting late - but as a Mavs fan I sure as heck hope they aren't happy with their results in terms of scouting/drafting college players.

Anyway - I, IN NO WAY, am saying the Mavs should have me in their employ and that I would do a better job than anyone else in place there. I do the stuff I do, I (and some others - thanks mom) think it's cool, I hope to start getting a site up to share what I do with more than the handful that have ever seen my stuff, and I have ideas on how to do some cool (imo) future projections. The thing I don't have is time, I have very little of it. If I ever am in the employ of a team or some other entity in which I actually could invest 40-80 hours a week (as opposed to about 1-3 hours a week now) in analytics, I personally think I could do some very interesting things that would make my employers happy. BUT, I would not expect others to think that about me - considering how little I've put out there compared to many of the great minds that visit this site and have created some very innovative analytic portfolios.