Proposed method to track performance of our starters

Discussion in 'Houston Rockets: Game Action & Roster Moves' started by Unstable, Oct 30, 2012.

Tags:
  1. Unstable

    Unstable Member

    Joined:
    Oct 29, 2012
    Messages:
    294
    Likes Received:
    22
    Hi all, here's what I will be trying to do to gauge if our starters are performing on par (for the moment I am treating par as equivalent to average as we have a young team just starting if they can hit 0.5 I'd be happy enough)

    I will be tracking each position - PG SG SF PF and C in comparison to the averages I have calculated based on the stats of the starters of the other teams, I will further sub divide the tracking to divisional, conference and nba levels.

    I am currently basing the stats on reported seasonal averages from nba.com, here are the abstracted figures for reference

    Div - FG% - 3P% - FT% - RPG - APG - SPG - BPG - TO. - PPG
    PG. - 0.454 - 0.340 - 0.793 - 2.6 - 5.1 - 1.0 - 0.1 - 2.1 - 11.8
    SG. - 0.455 - 0.351 - 0.801 - 3.2 - 2.8 - 1.2 - 0.4 - 2.0 - 14.1
    SF. - 0.459 - 0.340 - 0.780 - 6.0 - 1.5 - 1.3 - 0.7 - 1.5 - 12.0
    PF. - 0.471 - 0.304 - 0.796 - 8.5 - 2.1 - 0.8 - 1.0 - 1.9 - 17.7
    .C. - 0.493 - 0.131 - 0.684 - 7.8 - 2.2 - 0.7 - 1.3 - 2.2 - 12.4

    Con - FG% - 3P% - FT% - RPG - APG - SPG - BPG - TO. - PPG
    PG. - 0.453 - 0.362 - 0.835 - 3.2 - 6.0 - 1.2 - 0.1 - 2.5 - 13.8
    SG. - 0.445 - 0.367 - 0.813 - 3.5 - 2.9 - 1.1 - 0.4 - 1.8 - 13.8
    SF. - 0.453 - 0.347 - 0.784 - 5.3 - 2.1 - 1.2 - 0.7 - 1.8 - 14.0
    PF. - 0.505 - 0.214 - 0.742 - 8.6 - 1.9 - 0.7 - 1.0 - 1.8 - 15.5
    .C. - 0.543 - 0.071 - 0.659 - 7.6 - 1.4 - 0.6 - 1.3 - 1.9 - 11.2

    NBA - FG% - 3P% - FT% - RPG - APG - SPG - BPG - TO. - PPG
    PG. - 0.446 - 0.350 - 0.811 - 3.2 - 5.8 - 1.2 - 0.2 - 2.5 - 13.7
    SG. - 0.439 - 0.361 - 0.817 - 3.4 - 2.8 - 1.1 - 0.3 - 1.7 - 13.8
    SF. - 0.446 - 0.331 - 0.778 - 5.2 - 2.3 - 1.1 - 0.6 - 1.8 - 14.1
    PF. - 0.492 - 0.201 - 0.724 - 7.4 - 1.6 - 0.7 - 0.9 - 1.7 - 13.5
    .C. - 0.532 - 0.077 - 0.680 - 7.8 - 1.4 - 0.7 - 1.3 - 1.8 - 11.3

    These are baseline figures (and yes they will change as the season moves along but I am only like to recalibrate after say 20 games) what I plan to do is to measure our starters stats to these baselines - basically to divide them to get a figure take for example if asik shoots 0.6 his score for FG% would be 0.6/0.532 at NBA level which means he would score above 1 points. Overall there being 9 items tracked a passing score for a player is around 9points if they go over this they are doing fine, if they go under they should buck up.

    Comments and suggestions are welcomed. I will run this for a whole and see how it goes! We should be ready to roll tomorrow! Best wishes to our team!
     
    2 people like this.
  2. clubberclyde

    clubberclyde Member

    Joined:
    Jun 26, 2012
    Messages:
    1,975
    Likes Received:
    56
    [​IMG]
     
    1 person likes this.
  3. clubberclyde

    clubberclyde Member

    Joined:
    Jun 26, 2012
    Messages:
    1,975
    Likes Received:
    56
    In all seriousness, it sounds like a good idea.
     
  4. kevC

    kevC Contributing Member

    Joined:
    Jul 19, 2007
    Messages:
    5,155
    Likes Received:
    660
    Since you're just doing box score stats, why not just use PER and cut all this hassle?
     
  5. jayhow92

    jayhow92 Member

    Joined:
    Jan 15, 2011
    Messages:
    6,382
    Likes Received:
    290
    Lol. I'd rather track floor impacts than just look at stats. Things like help defense, man defense, defensive rotations, taking good shots, location of those shots, etc. Stats can be very misleading. It can show that guys like Jefferson and Monta Ellis are actually good.
     
  6. roxxy

    roxxy Member

    Joined:
    Jul 4, 2012
    Messages:
    5,120
    Likes Received:
    162
    Neat. Look forward to checking out your results.
     
  7. Unstable

    Unstable Member

    Joined:
    Oct 29, 2012
    Messages:
    294
    Likes Received:
    22
    HollInger's PER calculates player efficiency and is rather loaded towards offense. I am looking for a simpler method to gauge if our players are performing on par - slightly different angles of it. Please note that my method may not be great at telling how good a player is in comparison to another, but it would be useful to let us know and start to see if any of our players is trending up or down
     
  8. kevC

    kevC Contributing Member

    Joined:
    Jul 19, 2007
    Messages:
    5,155
    Likes Received:
    660
    PER is just a summarized box score. Literally all of your stats are incorporated in it and weighed accordingly to take out the guess work. It is also normalized so that the average is 15.0 so it serves what I think is your purpose well, how well do our players compare to the average.
     
  9. durvasa

    durvasa Contributing Member

    Joined:
    Feb 11, 2006
    Messages:
    28,021
    Likes Received:
    1,174
    So are you going to weight all of those stats equally, then?
     
  10. Unstable

    Unstable Member

    Joined:
    Oct 29, 2012
    Messages:
    294
    Likes Received:
    22
    For a start no weighting, but may be considering different weights later e.g. PGs will have higher weightage on assists (or maybe both guards) and forwards and center get weighted for rebounds

    Lets give it a spin first anyways I am only comparing PG to PGs I.e. each role to its counterparts so weightage may not matter
     
  11. Fierguard

    Fierguard Contributing Member

    Joined:
    Jun 12, 2002
    Messages:
    216
    Likes Received:
    5
    It is not a bad idea... but you need to weight it somehow per position. For example, if Lin is .2 in blocks, I do not care. The instance of the individual event is still so small it is meaningless over the course of a season.

    Also you might want to weight by percentage over... Example, if Patterson is shooting .493 versus say .545, that should mean more.

    Let me stop there before this too complex.

    I would love to see the big data setup that the Rockets are using. Sigh... inner nerd.

    Fierguard

    Bleed Red Rockets Red.
     
  12. lookabove

    lookabove Contributing Member

    Joined:
    Jul 25, 2012
    Messages:
    619
    Likes Received:
    16
    You can use WS or WP.
     
  13. Unstable

    Unstable Member

    Joined:
    Oct 29, 2012
    Messages:
    294
    Likes Received:
    22
    Maybe - I always thought WS or WP is more a career-wise tracking device.

    I am looking for a generic assessment tool of whether our starting players are performing up to expectations (current expectations being the average as a norm) in their roles.

    Take for example the performances of JLin's preseason stats of
    0.283 - 0.231 - 0.6 - 2.3 - 6 - 2.3 - 0.3 - 2.5 - 6.3 would translate to a score of 0.6+0.7+0.7+0.7+1.0+1.9+1.5+1.0+0.5 or 8.6 at the NBA level (for PGs) - which is slightly below the expected average of 9.0 but not of great concern since I would take a score of between 8 to 10 to be a decent performance level. Anything above 10 would be good performance, anything below 8 would be a bad performance. Anything above 12 would be WOW, anything below 6 (I think the player should be sent down to D-league).
     
  14. iconoclastic

    iconoclastic Member

    Joined:
    Oct 10, 2007
    Messages:
    5,978
    Likes Received:
    278
    Watching the games?
     
  15. Wapzoe

    Wapzoe Member

    Joined:
    May 22, 2012
    Messages:
    717
    Likes Received:
    40
    I like this, good effort!

    Gonna see just how well Lin and Asik live up to their contracts (they should project to be above the average based on their contract amount)
     
  16. Azadre

    Azadre Contributing Member

    Joined:
    Jul 7, 2001
    Messages:
    8,751
    Likes Received:
    270
    Track the performance by looking at the Wins.
     
  17. Unstable

    Unstable Member

    Joined:
    Oct 29, 2012
    Messages:
    294
    Likes Received:
    22
    Will try to, I am in Singapore, it would be work time for me whilst they are playing :( but will try and post my analytics after the game when I see the numbers out on NBA site. Please note I will only be tracking starters... but you are welcome to try and calculate for other players (just note that you need to reverse the division for TurnOvers (it being opposite value to the other factors).

    Generic method is to take player stats and divide by averaged baseline (for TO you take the averaged baseline divide by player stats instead) - and you round it to 1 decimal place and then add up all 9 factors - if the player you are interested is hitting 8 to 10 - they are doing pretty decently.

    Please note decent play does not necessary translate to a win though....
     
  18. Unstable

    Unstable Member

    Joined:
    Oct 29, 2012
    Messages:
    294
    Likes Received:
    22
    Okay fellow clutchfans, I think I will call this method of analysis the Peer-Averaged Comparison Score or PAC-Score for short.

    As a trial of the methodology let's analyze the recently conluded game of Cleveland Cavaliers against the Wizards, we know Cleveland won but did all its starters played well in their roles? If we simply took the standard NBA +/- method we would see

    Irving +23, Waiters +13, Gee +14, Thompson +20 and Varajao +7 - seems like all the starters were positive and doing fine right?

    But if we were to use PAC-Score a different picture emerges
    Irving 8.7, Waiters 8.1, Gee 4.7, Thompson 9.7 and Varajao 15.3

    In fact, the value of Varajao was under-rated by the +/- scoring (V actually had a wow game) and the fact that Gee wasn't really pulling his weight got hidden by it - the other starters were decent.

    I can't wait to try this analytics after today's game for our starters and see what it can tell us!
     
    1 person likes this.
  19. jbasket

    jbasket Contributing Member

    Joined:
    Jan 26, 2012
    Messages:
    3,511
    Likes Received:
    316
    I like it, but I ultimately feel like the good old "eye-test" can have similar results. One can check how good a starter is doing by watching him play, very simply. Good luck though.
     
  20. gtmkcp

    gtmkcp Member

    Joined:
    Sep 4, 2012
    Messages:
    500
    Likes Received:
    7
    Nice. Yeah Varajao had a great game, so wierd that his +/- is much lower than the other starters. But yeah, sometimes stats don't mean much, watching the game can tell another story from just stats. Still looking forward to your stats for our starters.
     

Share This Page