My new app takes a 5-player unit (real or hypothetical) and searches the 2347 lineups that played 20+ poss this year to find its top 50 comps based on player types. Imagine new lineups and see how they might fit together!
See how it works
@NylonCalculus
:
18
92
425
Replies
A big thank you to
@ElGee35
@JacobEGoldstein
and
@bballport
for making some of the stats and data infrastructure that feed into this app.
0
0
17
@rosszzor
@NylonCalculus
Big time! It's fun to explore the 3-d player space with the plotly viewer.
0
0
1
@CrumpledJumper
@NylonCalculus
It's not possible to build a lineup of players from different teams, right?
1
0
2
@dangercart
@NylonCalculus
It is! On the team drop-down menu pick the first option called "All" and then you can input hypothetical lineups with players from any combination of teams.
It's much more fun that way!
1
0
5
@CrumpledJumper
@NylonCalculus
Thank you.
With the focus on WHAT players have done, this tool is best for conventional thinking. A separate or hybrid tool with a focus on performance results would be very valuable for unconventional thinking.
2
0
2
@bballstrategy
That would definitely be useful!
The main drawback I see to using performance results is that they would be more variable from year-to-year. eg. 3-point attempt rate is more stable than 3P% (I think), especially for players who don't get to play much.
0
0
0
@CrumpledJumper
@NylonCalculus
I once made 4D triangles using depth through size and saturation. Check out from
2
0
3
1
0
2
@CrumpledJumper
@NylonCalculus
no Klay Thompson in there?
Is the drop down list in the tool based on this season's lineups? Or is Klay somehow excluded?
Cool tool though, can't wait to play around with it!
1
0
1
@Punk_Basketball
@NylonCalculus
No Klay, because he doesn't have any stats from 2019-20 season:(
But you can plug in Danilo Gallinari as a substitute -- his stats are pretty similar to Klay's 2018-19 numbers.
2
0
5
@CrumpledJumper
@NylonCalculus
Just to offer my humble appreciation for your app and the work behind it!
1
0
2
