IF Nets want to max next season wins...

Home for all your discussion of basketball statistical analysis.
Post Reply
Crow
Posts: 5567
Joined: Thu Apr 14, 2011 11:10 pm

IF Nets want to max next season wins...

Post by Crow » Fri Aug 03, 2018 11:47 pm

Trade (for guards) or don't play Russell, LaVert and RHJ much (maybe more for RHJ). Start Dinwiddie, Harris, Crabbe, Faried, Davis. Bench of Napier, Graham, Carroll, Allen. Tell Coach Atkinson to pick 5 main lineups and to knock off the way way excessive dink lineups.

If they want to tank, either do the same as last season and / or trade the fairly good players first. Let the bad ones linger, play at least moderate time in bad lineups and then go off contract.

Crow
Posts: 5567
Joined: Thu Apr 14, 2011 11:10 pm

Re: IF Nets want to max next season wins...

Post by Crow » Sat Aug 04, 2018 12:03 am

How ridiculous was the lineup management? One lineup used much over 1.5 minute per game (and only for 4 minutes). Only 4 lineups used over 1 minute per game. Top 10 most used lineups only accounted for about 25% of total minutes. 621 dink lineups used less, mostly way less, than 1 minute per game for season.

sndesai1
Posts: 132
Joined: Fri Mar 08, 2013 10:00 pm

Re: IF Nets want to max next season wins...

Post by sndesai1 » Sat Aug 04, 2018 12:21 am

why is it necessarily a bad thing to use "dink" lineups?

i've seen you harp on that a lot, but i don't instantly see why it's a bad thing.

it may be that i just don't know what that term actually means

Crow
Posts: 5567
Joined: Thu Apr 14, 2011 11:10 pm

Re: IF Nets want to max next season wins...

Post by Crow » Sat Aug 04, 2018 1:12 am

I define dink lineups as lineup used less than 1 minute per game for the season or less than 2% of total time. Dink equals small.

I "harp" on it because dink lineups usually do worse than non-dink lineups... and because teams usually want to do better and probably could if they played more tested, more proven, on average better lineups.

In this case the Nets 5 most used lineups got 18.7% of the total minutes and were on average plus 5.1 points per 100 possessions. The other 620 dink lineups got 81.3% of the time and were on average -5.8 pts / 100 poss.

Unless you are trying to tank (and they probably were totally or to some degree), this pattern is a bad thing. Play very lightly understood small minute lineups that on average do badly and you are going to do badly and probably worse than you could have if you played better tested, better performing lineups for more minutes.

If Atkinson says he was trying his best and is being honest, then these splits are very troubling to me.

Now better performing lineups are not guaranteed to do better in more minutes but it is at least somewhat better bet than for poorly performing lineups, on average.

Marks says he wants to win more next season. So they can pay attention to the data (for them and almost everyone)... or not.

Crow
Posts: 5567
Joined: Thu Apr 14, 2011 11:10 pm

Re: IF Nets want to max next season wins...

Post by Crow » Fri Aug 10, 2018 5:25 pm

So, still agnostic about dink lineups or any more willing to accept the data about their negative impact? I could provide even more team cases where they far underperform the most used lineups but I haven't see much sign of that data having affect.

sndesai1
Posts: 132
Joined: Fri Mar 08, 2013 10:00 pm

Re: IF Nets want to max next season wins...

Post by sndesai1 » Fri Aug 10, 2018 7:24 pm

i don't really see dink lineups performing worse than non-dink lineups as proof that a coach shouldn't experiment.

in fact, it seems like they recognize a lineup isn't providing much and so they play it sparingly and try something else.


without increasing the # of minutes your best players play, how do you get your best lineups more minutes?

Crow
Posts: 5567
Joined: Thu Apr 14, 2011 11:10 pm

Re: IF Nets want to max next season wins...

Post by Crow » Fri Aug 10, 2018 9:30 pm

You stop playing as many dink, lightly tested lineups and play bigger minute, more "proven" lineups. Straightforward.

How much "experimenting" should coaches do? There are coaches who spend 60-75% of total regular season time on dink lineups. That is beyond experimenting to me. That is wasting the opportunity to possibly learn about hardly any lineups to even a modest degree. I'd rather know more about 5-10 key lineups to the level of 200 plus minutes each than 200-500 lineups to the level less or way less than 10 minutes for season (which is what most coaches are doing and consequently not doing the former). I've previously indicated acceptance of using dink lineups for roughly 20% of total time. Try 100 different dink lineups if they want. Do they really need 500-700? I don't think it is necessary.

Sure, experiment with lineups, early in season, when you are way ahead, way behind or anytime you are comfortable with the consequences of not playing the best that you could play. But if you are fighting to be decent or great or as high as you can be on playoff seed I wouldn't put experimenting above executing.

Post Reply