1. Welcome! Please take a few seconds to create your free account to post threads, make some friends, remove a few ads while surfing and much more. ClutchFans has been bringing fans together to talk Houston Sports since 1996. Join us!

Efficiency Ratings and the Rocks -- How many people are actually playing?

Discussion in 'Houston Rockets: Game Action & Roster Moves' started by dwmyers, Nov 8, 2002.

  1. dwmyers

    dwmyers Member

    Joined:
    Sep 30, 1999
    Messages:
    288
    Likes Received:
    3
    Efficiency rating is a stat I first saw used with a guy named Martin Manley (who called it a performance rating) and basically it is all the offensive stats counted as a point, all missed shots and turnovers counted as a point, and the result divided either by (a) number of games played, or (b) 48 minutes.

    Numerically it is:

    (Pts Scored + Rebounds + Assists + Steals + Blocks - Missed Shots - Missed Free Throws - Turnovers) / (Games)


    To get a feel for it, scoring about 30 in ER over a season is an All-Star level performance, and rare. A 15 in ER is about average. On the nba.com site, you can get the top 50 Efficiency ratings for players, as well as the top 50 Efficiency ratings for players per 48 minutes. Last I checked, the Rocks have 3 players in the top 50, which puts them among the better players in the league:

    1. Steve Francis, 2, 30.75 ER
    2. Cuttino Mobley, 44, 16.75 ER
    3. Kelvin Cato, 46, 16.25 ER

    I was curious, so I calculated the ERs for the rest of the Rockets, through 4 games:

    Griffin, 12.75
    S Hawkins, 7.5
    M Norris, 5.5
    J Collier, 3.5
    Y Ming, 3.0
    K Thomas, 0.5
    B Nachbar, 2.0
    G Rice, -1.0

    Conclusions? The Rockets have Steve playing heroically well. A 30 for a point guard is getting to Magic Johnson range. But he almost has to, given the lack of performance at the 3 spot. Who is doing it there? Almost no one.

    It looks as if the Rockets have 4 offensive performers and otherwise have pretenders. The Rockets vaunted depth is nowhere to be seen.

    Now, the caveats. ERs overvalue missed shots, cause for a missed shot to become a change of possession, it takes both a missed shot and a defensive rebound by the other team. If a possession, as John Hollinger says, is worth 1.02 points, then ER overvalues secondary offense as well. These kinds of statistical measures are also primarily an offensive measure. They don't give a feel for the defensive side of the ball.

    Dave.

    Note. edited the formula cause I dropped assists from the formula when I first wrote this message. That might be what affected some comments below. - dwm.
     
    #1 dwmyers, Nov 8, 2002
    Last edited: Nov 8, 2002
  2. DearRock

    DearRock Member

    Joined:
    Jan 4, 2001
    Messages:
    2,139
    Likes Received:
    0
    Biggest caveat of all is the fact that so many player were injured, suspended, or new. The same rating taken after the next five game, barring any new siginificaant injuries, will show much improved number for KT and Rice. I do not know what you need to score to show you are deep, but I maintain, when healthy, this team is deep.
     
  3. robbie380

    robbie380 ლ(▀̿Ĺ̯▀̿ ̿ლ)
    Supporting Member

    Joined:
    Aug 16, 2002
    Messages:
    23,996
    Likes Received:
    11,175
    kind of a silly stat...i guess thats because efficiency doesn't always equate to actual worth. shaq last year was a 25.02.

    and right now aaron mckie (#9) is more efficient than tim duncan (#10). then kurt thomas is #12. and gilbert arenas is #15 while jason kidd is #18. not really a good stat.
     
  4. dwmyers

    dwmyers Member

    Joined:
    Sep 30, 1999
    Messages:
    288
    Likes Received:
    3
    Just checked out Shaq's #s. I got:

    (1822 + 716 + 200 + 137 + 41 - (1229 - 712) - (717 - 398) - 171)/67 =

    28.47

    for Shaq's ER in 2001-2002. In Shaq's case, it's the value of a missed free throw that factors into his lower ER. Assuming he completed 70% of his free throws, then

    398 / 0.7 = 569 chances and

    (1822 + 716 + 200 + 137 + 41 - (1229 - 712) - (569 - 398) - 171)/67 = 30.69

    As far as Kurt Thomas being currently higher than Jason Kidd, even a dog has his day, and perhaps these guys with high stats have had a good day or two in those four.

    Offhand, ER, as a composite stat, does show things that points scored does not (like the value of a player such as Jason Kidd). I particularly like looking at secondary offense, which is ER with all the shooting terms removed:

    (rebounds + assists + blocks + steals - turnovers ) / games

    For Shaq that number would be:

    (716 + 200 + 137 + 41 - 171) / 67 = 13.78

    which is damned close to half his total ER.

    I've looked at corrected composite formulas as well. The problem with John Hollinger's offensive formula is that he seems to be multiplying pace factors into his offensive stats. Otherwise it seems to account for things like the "true" value of a missed free throw (he counts it as .42 point).

    Dave
     
  5. codell

    codell Member

    Joined:
    Aug 26, 2002
    Messages:
    19,312
    Likes Received:
    715
    IMO, in the formula, TOs should be multiplied * 2. I feel a turnover is worth more than a missed shot.
     
  6. GATER

    GATER Member

    Joined:
    Jun 25, 2000
    Messages:
    8,325
    Likes Received:
    78
    codell -
    To give a TO a value of 2 assumes that every TO results in a score...obviously not true. IMO, 1 is too low but 2 is too high.

    On the other hand, if an assist can only be given for a made shot, why is the formula value 1 instead of somewhere between 2 and 3 (some assists are 3 pt'ers).
     
  7. codell

    codell Member

    Joined:
    Aug 26, 2002
    Messages:
    19,312
    Likes Received:
    715
    Actually, I should have clarified my post. I didnt mean that TOs should be doubled because of the other teams potential to score after the TO occurs.

    I was more thinking along the lines of: TOs = missed shot opportunies for us (i.e. 2points).

    Im sure my theory is full of flaws and subjectivity. Ive just always though of it like, if a player scored 30, but has 7 turnovers, then in my mind, hes really +14 in the scoring department instead of +30, which is how some would see it. Goes back to the arguement of, well Steve and Cuttino are ball hogs, but they are averaging 65 pts and shooting 50%. But look at their TOs, which is around 8. When people prop up Steve and Cuttino, they fail to look at some of the negative stats sometimes.
     
  8. dwmyers

    dwmyers Member

    Joined:
    Sep 30, 1999
    Messages:
    288
    Likes Received:
    3
    This is the deal. A turnover is a change of possession. The two guys I know who have looked into this (Martin Manley, John Hollinger) valued a possession as 1.0 and 1.02 points respectively. Understand, that a turnover is counted as a steal on the other side, so the total point change is 2 points. 1 for the team with the steal, -1 for the team committing the turnover.

    A missed shot + an offensive rebound is a continuation of a possession. It makes sense that an offensive rebound is worth as much as a missed shot, but opposite in sign.

    A missed shot + a defensive rebound = a change of possession.
    The sum should be worth 1 or 1.02 points.

    An assist + a shot = 2 (or 3) points. How much is the assist worth? John Hollinger argues, hueristically, that a shot consists of:

    1. getting open
    2. getting the ball
    3. making the shot.

    If each of these three steps are weighted equally, and assuming a 2 point shot, then an assist (step 2) is worth about 0.7 points. But I don't see any good way to get at the value of an assist.

    Back to missed shots. If a missed shot is worth -0.7 points, then

    1. an offensive rebound is worth 0.7 points to the rebounder.
    2. a defensive rebound is worth 0.3 points to the rebounder.

    If a missed shot is worth -0.5 points, then offensive and defensive rebounds are worth the same and both are worth half a point.
     
  9. GATER

    GATER Member

    Joined:
    Jun 25, 2000
    Messages:
    8,325
    Likes Received:
    78
    I pass you the ball and you make the shot. Your ER is 2 (or 3). My efficiency rating is .7 - just on an intuitive level, there seems to be something wrong with the ratio.
     
  10. codell

    codell Member

    Joined:
    Aug 26, 2002
    Messages:
    19,312
    Likes Received:
    715
    Heres my argument though. The value of a turnover isnt always translated in stats or in a box score.

    Turnovers can be momentum killers/momentum changers/momentum preventers. So to me personally, the value of a TO, relatively, is exponentially greather than just 1 or a change of possession.

    See, I can live with Francis/Cuttino lack of passing/ball movement/getting their teamates involved if it werent for the turnovers. Their turnovers arent just from forced shots, but also from poor quality passing. So even at times when they do try and get their teamates involved, they dont make a good pass. Furthermore, even if they do get the pass to a teamate without turning it over, the player receiving the pass is not in the optimal position to score. Now whether thats due to the passer making a bad decision on when to pass it or due to the player receiving the pass not putting himself into the right position, is of course, highly debatable sometimes.

    A great stats in my eyes is, turnover to assist ratio. For his career, Francis TO/A ratio is roughly 1.7 to 1. Gary Paytons is roughly 3 to 1. Now, before some of you jump on me for this comparison, yes, I know they arent the same type of player and Payton has better shooters and blah blah blah. Im not comparing them directly, just using Payton purely as an example.

    So to sum up, when you see Francis and Mobley not pass as much as they should YET STILL have a high number of turnovers due to bad passing and/or poor decision making when they actually do pass, it is totally couterproductive with regard to establishing offensive rhythm. At times, their ISO tendencies AS WELL AS their passing/ball handling (when resulting in a TO), kill team momentum and offensive "flow".

    Thats why I feel TOs are worth more and consequently, should be discussed much much more on here when hyping up Francis and Cuttino or when using a forumla like the one that started this thread.
     
    #10 codell, Nov 8, 2002
    Last edited: Nov 8, 2002
  11. dwmyers

    dwmyers Member

    Joined:
    Sep 30, 1999
    Messages:
    288
    Likes Received:
    3
    I tend to think comparing Gary Payton and Franchise is a mistake. Their games are so different. For example, in their second years (I'm ignoring year 3 cause of Steve's known health problems then):


    Stats - Payton - Francis
    PPG 9.4 19.9
    RPG 3.6 6.9
    APG 6.2 6.5
    TPG 2.1 3.3
    ER 13.7 23.3

    Francis puts the ball in the hoop far better than Payton did at comparable points in their careers. He's a better rebounder and a comparable assist man. He has about 1 to 1.5 more turnovers a game.

    The thing is, Francis is so masterful at scoring we keep comparing him to 10 year veterans.
     
  12. codell

    codell Member

    Joined:
    Aug 26, 2002
    Messages:
    19,312
    Likes Received:
    715
    Again, my intention was not to compare them directly but use their stats as examples and emphasize that TOs are a huge detriment to Steve's game and to this teams overall play.
     
  13. Dr of Dunk

    Dr of Dunk Clutch Crew

    Joined:
    Aug 27, 1999
    Messages:
    46,648
    Likes Received:
    33,664
    dwmyers,

    The equation you are using has been in existence for around 30 years or so, I think. It is called the Tendex rating system. There is also a variation on it called the modified Tendex rating that weights the individual stats a bit more.

    Dean Oliver's Journal of Basketball Studies has some interesting reading if you're into stats and weighting of player performance based on equations.

    http://www.rawbw.com/~deano/ (click on the Methods link at the top of the page).
     
  14. lpbman

    lpbman Member

    Joined:
    Dec 12, 2001
    Messages:
    4,240
    Likes Received:
    816
    The stats tell me that the rest of the team is watching Steve play Micheal Jordan, instead of playing basketball

    Kenny and Yao aren't going to be pushing the ball up the court themselves... Steve is actually going to have to RUN an offense
     
  15. Lobo

    Lobo Member

    Joined:
    Sep 17, 1999
    Messages:
    499
    Likes Received:
    0
    I think the term efficiency rating is actually a bit of a misnomer, since it doesn't factor in minutes played. A better term might be impact rating .

    What the "ER" tells me is that we're pretty much a one-man team right now.
     
  16. Easy

    Easy Boban Only Fan
    Supporting Member

    Joined:
    Jul 23, 2002
    Messages:
    38,323
    Likes Received:
    29,863
    Don't forget this is less than 2 weeks into the season. It's not unusual for a good player to struggle in a few games and a mediocre player to be hot in some stretches.

    I think this rating make at least some sense in terms of showing how valuable a player is. Shaq's problem is his free throw. As dominant as he is, if it wasn't the fact that he suddenly got hot at the line during the WCF, the Lakers wouldn't be the champions right now. It'd be interesting to see Shaq's rating in the playoffs when his ft% went up.

    A lot of people have observed that Shaq is the only superstar you don't want to go to at crunch time because of his ft problem.
     
  17. Easy

    Easy Boban Only Fan
    Supporting Member

    Joined:
    Jul 23, 2002
    Messages:
    38,323
    Likes Received:
    29,863
    I think the fairest way to do it is for the assist player and the shooter share the points. If an assist is worth 1/3 of the scoring process, then the assist should worth .7 and the scorer should get 1.3 for a 2 point shot. But that would be very hard to calculate because you'd have to record how many shots are assisted and how many aren't. There'd be at least 3 more columns added to the statline.

    BTW, the reason why a possession is worth about 1 point is that the probability of scoring 2 points in a possession is about 50%. But in light of the NBA fg% in the past decade, that number should be considerably lower.
     
  18. dwmyers

    dwmyers Member

    Joined:
    Sep 30, 1999
    Messages:
    288
    Likes Received:
    3
    Team ERs after Game 5


    Steve Francis: 30.4 per game, 37.6 per 48 minutes
    Cuttino Mobley: 16.8 per game, 24.8 per 48 minutes
    K Cato: 16.6 per game, 28.5 per 48 minutes
    E Griffin: 13.4 per game, 19.5 per 48 minutes
    K Thomas: 9.3 per game, 18.2 per 48 minutes (!)
    --------------------------------------------------------------------
    J Hawkins: 6.4 per game, 12.0 per 48 minutes
    M Norris: 5.4 per game, 11.3 per 48 minutes
    Y Ming: 3.0 per game, 10.3 per 48 minutes
    J Collier: 3.5 per game, 10.8 per 48 minutes
    B Nachbar: 2.0 per game, 8.0 per 48 minutes
    G Rice: -0.5 per game, -1.3 per 48 minutes

    Top 5 looking much better now :>
     
  19. Easy

    Easy Boban Only Fan
    Supporting Member

    Joined:
    Jul 23, 2002
    Messages:
    38,323
    Likes Received:
    29,863
    Cato is more efficient than Mobley.
     
  20. Yetti

    Yetti Member

    Joined:
    May 20, 2002
    Messages:
    9,589
    Likes Received:
    529
    WE hope that the poor showing of our "DEPTH" will improve as the Season moves along and more players find their rythem and shot. They will of course have to demand that the ball be passed to them, when they are open.
     

Share This Page