New Site - CraftedNBA

[dtkavana]

All plus minus and advanced stats have been updated.

[Crow]

Firefox.

[dtkavana]

I was able to replicate and fix that issue. It was a firefox specific issue.

Thanks for pointing that out.

[Mike G]

Using this awesome site, one can quickly get a "most improved" list based on previous career RPM and PIPM vs those numbers for the current season.
(RPM-MYRPM)*Min/48 = points added by a player above and beyond his career rates; assuming 100 poss/48 minutes.
Same for PIPM, the two are averaged to get column 1.

Code: Select all

pts+  improved       Tm    RPM    MYR   PIPM   MYP
84   De'Aaron Fox   SAC   2.05  -3.54    1.8  -1.2
58   Marc Gasol     MEM   4.93   1.30    3.7   1.8
54   Jerami Grant   OKC   1.37  -2.73    2.2   0.2
49   Danny Green    TOR   3.57   0.96    4.2   1.9
49   Noah Vonleh    NYK   2.73  -1.60    2.2   0.8

47   Nikola Vucevic ORL   4.96   1.32    4.6   3.2
47   Pascal Siakam  TOR   4.02   0.76    4.0   2.4
45  Emmanuel Mudiay NYK  -0.27  -5.31   -0.3  -2.5
43   Paul George    OKC   6.75   3.95    4.5   3.1
43 D'Angelo Russell BRK   0.98  -2.59    0.3  -0.7

Oddly enough, of players with >350 minutes this year, the total RPM improvement is minus-451 points; total PIPM points added is -203. A small shortfall but still odd.
I'd think players who have improved would get more minutes, while those regressing should get fewer, in general. Even if 55% of players are (per minute) un-improving, the net (xMin) should be positive.

Biggest dropoffs x minutes played:

Code: Select all

pts+   unimproved      Tm     RPM   MYR    PIPM   MYP
-85 LaMarcus Aldridge SAS   -3.00   3.12   -0.8   1.4
-58   Chris Paul      HOU    3.48   8.17    2.6   5.0
-58   Klay Thompson   GSW   -1.55   2.07   -0.2   1.5
-58   Clint Capela    HOU   -1.26   3.30    1.2   2.5
-55   Karl-A Towns    MIN    1.59   4.82    0.7   2.9

-54   Trevor Ariza    WAS   -2.82  -0.65   -3.9  -0.2
-53   Kyle Anderson   MEM   -0.95   3.57   -1.1   0.6
-50   Eric Gordon     HOU   -2.23   1.82   -0.9   0.9
-45   Otto Porter Jr. WAS    0.87   5.17    1.1   2.6
-43   Cedi Osman      CLE   -4.78  -1.12   -3.7  -2.8

[Crow]

With MY being multi-year?

[Mike G]

Yes, and I believe the multi-year value includes the current season for PIPM but not for RPM.
Rookies therefore don't have a MY-RPM; and their PIPM are a bit different from their MY-PIPM (?)
Some veterans also don't have a MY-RPM -- Ilyasova, D Rose, Augustin, Canaan -- so maybe they are/were 0.00 ?

I've only played with these numbers a bit, and clarifications from the author would be great.

[dtkavana]

That's really good stuff Mike. I may use something like that for the Most Improved Player on the Awards page that I am working on right now.

I have All-NBA, All-Defense, and All-Under 23 in addition to a few individual awards like MVP.

Hope to have this out before Christmas Day.

[dtkavana]

You are correct with the MYPIPM including the current season and the MYRPM not including it.

As for the few veterans without a MYRPM, I will have to double check that as I may have missed those guys.

With regards to the rookies PIPM vs MYPIPM I am unsure how that works and that would be a question for Jacob Goldstein.

https://twitter.com/JacobEGoldstein

[Mike G]

Ironically, the longer De'Aaron Fox continues to out-perform his previous season, the less "improved" he will register regarding PIPM-MYPIPM. His year2 numbers will raise his MY and diminish the difference.
This might be OK, if you think a 2nd-year player "should" improve, and therefore disqualify from MIP consideration.

[dtkavana]

I am excited to announce the release of a few new features this afternoon.
Please play around and let me know what you think. Thanks!

1. Team pages now feature a new lineup - "Versatile" - which aims to produce a teams most playoff ready/versatile lineup and spit out their projected +-
http://craftednba.com/teams/indiana

2. Another addition to the Team Page is the "Young Core", which uses @FiveThirtyEight 's CARMELO system to spit out a team's best young (under 26) players.

3. The final addition to the Team Page are some Quick Stats as well as the Four Factors.

4. ...and lastly, on the Individual Player pages, we have added some Combine Measurement data. We currently have data for about 250 players and hope to expand to nearly all active players in the next few days.

[Crow]

Enhancements are good.

Would you consider adding actual raw plus minus for the team lineups (versatile and closers and young core is enough have data)? A step beyond that would be to calculate projected lineup data for the team discrete stats and compare that to actual data too (showing actual - projected differences). The player stats can't be expected to 100% thru put to actual lineup returns but the comparison results can help identify the skill overlaps / excesses that lead to diminishing returns as well synergy returns not anticipated by simple summation of player data. Player data would have to standardized to same number of minutes or possessions. Perhaps 6 minutes or 13 possessions (roughly equivalent) or maybe 10, 15, 20 or 25 possessions.

[Crow]

The category pages have a few customizable filters and positions but it would be even more excellent if there were more filter options (discrete stats, demographics, salary) and perhaps the ability to pull-up the discrete stats on a page for all those who qualify. It would also be nice to pre-calculate the averages for all who qualify (on the metrics and the discrete stats). You can find median by sorting a column but mean is important information too.

[Crow]

On the best offensive players category page if you use 700 minute and plus 1.5 on Combo Offensive Plus Minus (the availability of which I want to praise), there are 36 such players. On CDPM, 16 above 1.5. On overall CPM, 39 are over plus 2.

I dunno if it is me but these numbers seemed to change moment to moment.

[Crow]

Would you consider adding other plus minus metrics, like http://canzhiye.com/rapm/ , http://nbashotcharts.com/rapm ,
and others by Kevin Kerrigan, bbstats, ElGee, Ben Taylor, etc.?

May need quality control but prompting closer examination would be a good thing.

[Crow]

It might be a little late, but it would be valuable to preserve metric values at least half season mark and allow comparison of first half to second half season. I am going to do for RPM and maybe other metrics but it would handy to do it for all, cleanly, regularly.