Statistical Analysis. Humor. Knicks.

Monday, November 24, 2014

Next(ish)-Day Notes and Errata: Timberwolves @ Knicks

Lots to grumble about after a disastrous first quarter sunk the Knicks into a hole that they couldn’t quite climb out of Sunday night.  I’m going to do something weird.  I’m going to totally ignore Andrea Bargnani.  Plenty of logical arguments were made before the season about the holes in his game and the attendant role that he should be afforded but the simple truth is that any attempt to bolster (or usurp) those arguments based on three games is likely to be gobbled up by the dreaded Small Sample Size Kraken.  If you had an opinion on  Bargs before the year, you probably have the same opinion.  There’s nothing new to say.  Let’s talk in a week or two when we have some meaningful numbers to back up our preconceived notions and anecdotal evidence.

Here there be meaningless statistics!

Here there be meaningless statistics!

Today, instead, I want to focus on something that we can demonstrate pretty clearly to be effective without relying on the same-sized data sample that suggests the 76ers to be a title contender.

In the opener, Mike Woodson started Ray Felton and Pablo Prigioni alongside one another and gave them 31 minutes of run as the backcourt tandem.  The Knicks won by 7 and their margin of victory was more than supplied by a 61-44 advantage in the 31 minutes of 2-PG time.  Woodson responded by burying this approach for the rest of the opening week: the guard pairing didn’t see the light of day in the Knicks’ one-point loss to the Bulls and played just ten (admittedly unsuccessful) minutes in the team’s 109-100 loss to the Timberwolves Sunday night.

In and of themselves, the positive results from opening night mean almost nothing and the negative results from the two subsequent games mean only incrementally more.  But the two-PG thing isn’t some flight of passing (count it!) fancy, it was the mainstay of the Knicks’ approach last season and once Prigioni supplanted Kidd in the rotation in January, it guided the Knicks out of their mid-season doldrums and to the finish line of the very 54-win season upon which the Knicks hope to build.  Weirdly, Woodson disappeared the lineup in round two of the playoffs, marginalizing Prigs in a hopelessly misguided effort to go big and beat the Pacers at the style of play in which they specialize.  And while it’s difficult to isolate the specific effect of this decision (it’s hard to imagine the Knicks finding a way to take that series with a hobbled Chandler and a worse-than-invisible JR Smith, anyway) it is patently true that in the past 10 months, the Knicks’ success has roughly tracked with their utilization of a two-creator backcourt.  To wit (stats per nbawowy.com):

Prigs-Ray Chart

So, check it.  Not only do the raw scoring differential numbers emphatically suggest playing Felton and Prigioni together, but the underlying data you would expect to be most favorable to the contrary point of view simply does not exist.  The Knicks actually turn the ball over less with both point guards on the court, despite a more pass-heavy offense.  And any notion that the Knicks have to abandon this approach in favor of “going big” goes out the window with the revelation that the Knicks actually allow fewer shots at the rim and a lower percentage on those shots when Prigioni and Felton play together than when they don’t.  The only minor hit comes on total rebound rate.  If the plan is to throw away all this other good stuff to get 0.8% more rebounds then that basically speaks for itself.  Do I think all of these things mean Felton and Prigs make our interior defense better?  Maybe, maybe not (staying in front of penetrating guards before they can get a head of steam is a major component of interior defense).  But it certainly isn’t materially hindering it.

I’m not naive, I know that there are plenty of factors that go into the Knicks’ lineup decisions that have nothing to do with these kinds of numbers.  Getting rookies and new acquisitions involved is important, as is maintaining team morale, managing egos, and placating an increasingly eccentric (to be kind) front office.  Whatever.  Fine.

But what the hell, guys?  We found something that works and rode it to 54 wins and the 2-seed in the East last year.  We got knocked out in the same series we went away from it, even if it wasn’t necessarliy BECAUSE we went away from it.  We brought back almost every important player in that rotation and have since won a game emphasizing the two-PG approach and lost two marginalizing it.  This is meaningful information.  The Small Sample Size Kraken lays dead at our feet.  This IS our best backcourt.  Melo, JR, Amar’e, Bargs, Shumpert, Hardaway — they’re all part of a separate argument.  Give the two point guards all the minutes you can and figure out the rest as you go.

32 comments on “Next(ish)-Day Notes and Errata: Timberwolves @ Knicks

  1. DRed

    I generally agree-but I do think it’s going to be important to limit pablo’s minutes. We do need to find a good plan B.

  2. Nick C.

    I don’t disagree, but I ask what were the numbers with Kidd and either of them. Kidd is/was a PG right?

  3. Owen

    Hearty agreements….

    I guess I hadn’t really realized how good Prigioni was last year. He was solidly above average in most respects and if you care to look at Wins Produced (caveat caveat caveat) he was spectacular.

    Prigioni’s minutes should be an all you can eat situation, that’s my takeaway….

    Felton was solidly average with a lubricating skill set. And they do work well together and with Melo and Tyson.

    Good piece.

    Now about Bargs…..

  4. Robert Silverman

    I don’t disagree, but I ask what were the numbers with Kidd and either of them. Kidd is/was a PG right?

    You can see the Knicks top five-man units from last season here. Kidd plus another PG grades out quite well.

  5. Hubert

    So, with a day to think about it, maybe calling Mike Woodson an imbecile was a wee bit strong.

    Even when I despise Woodson (which has become too frequent recently) I still agree with Ephus when he says:

    I fear that replacing Woodson would be likely putting a sledgehammer to a sheet of ice, and that the shards of the team equilibrium (i.e. effort levels from ‘Melo, JR Smith, Felton and stability of Chandler, MWP and Amar’e) likely would not be successfully re-assembled.

    Firing him is only going to make us worse because we wouldn’t find a better coach to replace him with. He’s the last person hired by the old regime still in place. I would imagine if he were fired, our candidate list would include Maverick Carter, Leon Rose, and World Wide Wes.

    So yes, we’re stuck with him. And hopefully he is just using these early games to figure things out and won’t stick with some of our more miserable lineups when the sample sizes get bigger.

    I despise that he seems to be oblivious to his best lineups and is worst lineups. And I despise that, when in a pinch, he doesn’t go to what has proven to work but instead turns to things like size and experience. He’s a mediocre to above average coach. And we’re probably a mediocre to above average team, which is probably what really ails me.

    Just please, no more Martin-Chandler lineups. Please.

  6. Kevin McElroy Post author

    Guys,

    Ran numbers it’s 111-106 per 100 w/ Felton and Kidd and 115-104 per 100 w/ Prigs and Kidd. Prigs lineups are uniformly awesome. He has been our most important guard in calendar 2013 (Felton’s decimation of Boston notwithstanding).

  7. thenamestsam

    The Small Sample Size Kraken is better than everything that is not the Small Sample Size Kraken.

    As for the topic of the article, I’m in complete agreement. The Knicks had a winning formula last year. Now there’s nothing wrong with trying to find some additional formulas that work as well, because obviously having the flexibility to be successful with different types of lineups is a positive attribute. But going away from the winning formula entirely in the search for those other successful lineups is absurd. Give the lineup that works its minutes, and then spend the additional minutes tinkering with other things. This doesn’t seem that complicated.

  8. johnlocke

    Nice article. I’d like to believe that Woody is using these games to experiment. That two point guard lineup was undoubtedly good. What I think Woody has been trying to do is to see if a different lineup that includes Bargnani could be even better, so far the returns have said NO and my eyes say HELL NO, but I do think it’s important to tinker early and settle into the best lineup. Also JR returning will have a dramatic impact on lineups since he’s playing 28+ mins per game. I think we need to wait until JR comes back before deciding that Woody is a numbskull who doesn’t look at data. Especially since his GM is focused on data / metrics. And yes, I was similarly pissed off against Indiana but we weren’t going to win that with JR and Chandler playing like they played anyway…
    http://nypost.com/2013/10/01/knicks-president-mills-looks-to-get-analytical/

  9. flossy

    Look, the 2 PG lineup is clearly effective, but Pablo Prigioni is 36 years old. The Knicks need to figure out a way to play effective basketball with a more traditional lineup, even if their bread and butter is and always will be 2 PG units. It’s not asking all that much–practically all other NBA teams, even the very good ones, play the vast majority of their minutes with a single PG on the floor.

  10. Nick C.

    Owen, that’s pretty bizarre. Dolan = vintage Steinbrenner circa 1980s – suspension.

    Kevin, Bob thanks for the data.

  11. thenamestsam

    I’d like to believe that Woody is using these games to experiment. That two point guard lineup was undoubtedly good. What I think Woody has been trying to do is to see if a different lineup that includes Bargnani could be even better, so far the returns have said NO and my eyes say HELL NO, but I do think it’s important to tinker early and settle into the best lineup.

    I’ve got no problem with tinkering, but it’s not like there isn’t time for tinkering and still playing some 2 PG lineups. Nobody is saying that Felton and Prigs should play 48 minute each, or that Beno should come in every time one of them goes out. Even a heavy 2 PG game like the Bucks one was only 31 minutes, and that was probably too many minutes for Prigs to be sustainable. So lets say an upper cap is something like 24 minutes a night for those two sharing the floor together. Realistically if we want to keep Prigs minutes sub-30 it should be a little less than that. That still leaves plenty of time to look at other lineups, experiment with going big, whatever you want to look at.

    But playing a lineup that you know works very well for 0 minutes in what was eventually a 1 point loss just does not make any sense to me.

  12. JK47

    Yeah, I mean, didn’t we just see Jason Kidd disintegrate before our eyes after being asked to play 2,000 minutes? I love Pablo, but I don’t think it’s smart to play him for 25+ minutes per game like they did with Kidd last year.

    Beno hasn’t looked like a world-beater so far, but the guy has played almost 15,000 minutes of relatively competent NBA basketball, and probably wouldn’t kill us if he made his way back into the rotation. The more Beno plays, the more the Knicks can do the two-PG thing.

  13. flossy

    But playing a lineup that you know works very well for 0 minutes in what was eventually a 1 point loss just does not make any sense to me.

    I think (read: hope) Woodson is looking at JR Smith’s suspension period as an extended preseason, where he can dick around with different lineups to see if anything involving Andrea Bargnani actually works, without having to commit. The fact that Bargamel rode the pine and we had our best 5 from last year on the floor down the stretch against Minnesota leads me to believe that Woodson knows what works, and is throwing some other stuff at the wall until JR Smith comes back. If Woodson is still doing this in December, January, etc. given relative health, then THAT will make no sense.

  14. Owen

    Man, just realized John Wall is making the max. Unreal.

    I would not want to be an NBA gm and be forced into that deal….

  15. Hubert

    If we’re going to tinker (and we should), one of the things I’d love to tinker with is removing Felton from the equation. I am very curious if it’s the two PG’s that make the lineup tick, or simply Prigioni. Because when it’s two PG’s, he is the PG. Felton does not run the offense in the two PG lineup.

    The reality is that by the time he came into the own, Pablo mostly played with Felton. Would it work just as well with Shumpert? And I’m curious if we actually get more passing from Felton when he is playing the 2nd PG role. Remember that ridiculous 20-2 run in game 2 vs Indiana? That was Prigioni and Kidd’s corpse. No Felton was needed.

    So yeah, I wonder if it’s the two PG’s that make it tick, or if we should really just be calling it the Prigioni at PG lineup. These are the kind of things I would love to see us tinker with as opposed to the big lineups.

  16. DRed

    If we’re going to counter teams that go big by embiggening ourselves, why not try playing Cole Aldrich? He’s big.

  17. Hubert

    johnlocke there is a really funny (and not in a good way) quote from Woodson in that article you posted:

    “Tracking different lineups who play well together, that’s all still up in the air,’’ Woodson said. “Sometimes you take a lineup that plays 10 minutes together, it looks good on paper but [bleep] the lineup that’s played 20 minutes together is winning games.’’

    Yikes.

    If you want to give him the benefit of the doubt, maybe by “all still up in the air” he means small sample sizes. But I got the distinct impression it meant he doesn’t think looking at the point differentials of certain lineups is meaningful. And that if he plays 5 guys for 35 minutes and they get outscored by 8, and plays 5 guys for 13 minutes and they outscore their opponents by 25, I think he would attribute the win the unsuccessful lineup based on that statement.

  18. DRed

    Bargnani is crushing the boards. He’s hauled down 85.7% of the 2.3 rebound opportunities he’s generating per game. Only 7 out of the 316 players in the NBA who have 2 or more opportunities per night can say they are doing better than that. And that’s factorial.

  19. Nick C.

    “Tracking different lineups who play well together, that’s all still up in the air,’’ Woodson said. “Sometimes you take a lineup that plays 10 minutes together, it looks good on paper but [bleep] the lineup that’s played 20 minutes together is winning games.’’

    I don’t understand what that second half even means. It is just weird. Why wouldn’t you track to see what works? All I can say is I wouldn’t want him cooking my dinner.

  20. Hubert

    I think the second part means that if he played AB 35 minutes and Knicks won, but they got outscored by 10 when AB played, and the only reason they won is because the Knicks were plus 20 in the 13 minutes AB sat, that he would attribute AB being part of the lineup that won the game. Because AB played, and they won. And tracking which lineup made them win is still up in the air. So don’t look at who actually made us win, look at who played when we won.

  21. Hubert

    By the way, some pretty cheap tickets on stubhub if anyone is interested in going tonight. I just got me some seats in the 100’s for $45.

  22. Donnie Walsh

    “Tracking different lineups who play well together, that’s all still up in the air,’’ Woodson said. “Sometimes you take a lineup that plays 10 minutes together, it looks good on paper but [bleep] the lineup that’s played 20 minutes together is winning games.’’

    I don’t understand what that second half even means.

    I can’t figure out what the bleeped out expletive could be. Is he cursing the lineup that is winning games? (And why is “games” pluralized??)

  23. KnickfaninNJ

    I think the second part means that if he played AB 35 minutes and Knicks won, but they got outscored by 10 when AB played, and the only reason they won is because the Knicks were plus 20 in the 13 minutes AB sat, that he would attribute AB being part of the lineup that won the game. Because AB played, and they won. And tracking which lineup made them win is still up in the air. So don’t look at who actually made us win, look at who played when we won.

    Hubert, I know this sounds weird but statistically this is not totally irrational. There could be interactions between when Barg’s is playing and when he is not. For example, if the he played on the starters and we did less worse than against the starters than we would have without him, then our second unit’s advantage could become enough to win the game for us. Or the opposing coach could have been forced to alter his lineups in way unfavorable to the other team because of Woodson starting Bargnani. If we had many games of data some with Bargnani playing and some without we could test this hypothesis, by looking at the games where he was totally out and the ones he was in and comparing. We don’t have this data so we have to have fun speculating.

  24. DRed

    I’m pretty sure Woodson is saying that sometimes you have lineup A that certain data suggests will play well together, but on a given night it is lineup B that actually wins you the game. When he says “[But] we got things going on with Mr. Dolan that can help us” I have no idea what he’s talking about, unless he’s talking about a coup.

  25. KnickfaninNJ

    When he says “[But] we got things going on with Mr. Dolan that can help us”

    I didn’t know he said this. That sounds like a trade might be in the works. We have a lot of shooting guards. It would make sense to trade one of them for a backup center. But I don’t think any of our guards are worth a backup center on the trade market except maybe JR and I don’t want to trade him or find it likely the Knicks do want to trade him. So who knows?

  26. johnlocke

    LoL at all the different ways one can read a quote. My take was that Woody meant – and I could be giving him benefit of the doubt — that: lineups of small sample sizes in terms of minutes PER game may not be as instructive, a group that plays 10 minutes per game together over the course of 80 games has amassed 800 minutes together, so you could say that it’s “proven” that that lineup is effective, but if that same lineup had played 30 minutes per game instead of 10, you can’t argue that the results would have been the same or better for the team.

    I would agree with that and add that lineup data is useful but it obscures a lot of really critical information (especially for smaller samples) such as the score of the game when the lineup is used — garbage time, or close game in the 4th quarter? the lineup that that lineup is facing (starters or the bench players) just as two examples..

    “Tracking different lineups who play well together, that’s all still up in the air,’’ Woodson said. “Sometimes you take a lineup that plays 10 minutes together, it looks good on paper but [bleep] the lineup that’s played 20 minutes together is winning games.’’

  27. Brian Cronin

    By the way, some pretty cheap tickets on stubhub if anyone is interested in going tonight. I just got me some seats in the 100?s for $45.

    Shit, I should have looked. Every game was so high before the season that I didn’t even think to look now. Fairweather fans = good prices on tickets at StubHub. I’ll be sure to keep an eye for future games. Thanks for the head’s up!

  28. Brian Cronin

    Oh true, but even against the Bobcats the ticket prices were way too high.

    Honestly, I sort of dig attending games I know that the Knicks will win. It makes me happy to watch them win. ;) My wife even keeps that in mind whenever she buys me tickets for Christmas or whatever. She’ll ask people at her office, “Which team sucks?”

Comments are closed.