1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

AllBuffs.com Intro To APBRmetrics

Discussion in 'Colorado Basketball Message Board' started by Goose, Nov 1, 2011.

  1. Goose

    Goose Hoops Moderator Club Member Junta Member

    Joined:
    Aug 25, 2005
    Messages:
    10,560
    Likes Received:
    1,545
    Welcome to the AllBuffs "Intro To APBRmetrics" area, or "How You Too Can Be A Stat Nerd". We're going to keep it simple, but this is going to discuss some stuff that we'll be discussing throughout the season when it comes to talking about the BasketBuffs. No more use of points per game for you. From now on, you're going to use words like "efficiency" and "tempo".

    HISTORY

    Everyone here has heard of "Moneyball", the book (now movie) of Billy Beane and how his use of Sabermetrics changed baseball. APBRmetrics is the basketball equivilent. Dean Oliver and John Hollinger both took Bill James' ideas for baseball and ushered them in to basketball using the internet. Basketball was more open to the idea of advanced stats than baseball was, with the Houston Rockets taking it to a new level in 2006 by hiring Daryl Morey as their GM.

    TERMS WE WILL USE

    The first four terms we will use are Dean Oliver's "Four Factors". These are considered the building blocks of APBRmetrics and efficiency. In other words, the better you are at these four things, the better your team is.

    eFG% (Effective Field Goal Percentage): eFG% = (.5*3FGM + FGM) / FGA

    eFG% was developed to take into account the fact that the standard FG% doesn't account for the extra value per shot for a 3 pointer. For example, shooting 33.3% from three point range over six shots would equal 2 makes/6 attempts and yield 6 points for the team. To get the same 6 points shooting from two point range, a team needs to shoot 50% (3 makes /6 attempts). To account for this, "effective" Field Goal percentage was born. An easy example from last year is
    Shannon Sharpe. Here are his numbers from last season:
    FGMFGAFG%3PM3PA3P%TS%
    2969.420115.067.428

    So, for Shannon, we have the following:

    (.5 * 1 + 29)/69 = 29.5/69 = .4275

    Another example is
    Levi Knutson:
    FGMFGAFG%3PM3PA3P%TS%
    158312.50681171.474.636

    (.5 * 81 +158)/312 = 198.5/312 = .636

    TO% (Turnover Percentage): TO% = Turnovers/Possesions

    This one is pretty self explanatory. Basically, if you have the ball 20 times and turn it over 3 times, your turnover percentage is 15%. A turnover is a wasted possesion. The less wasted possesions, the more likely you are to score. Just think of this as the basketball equivilent of dating.

    OR% (Offensive Rebounding Percentage): OR% = OR / (Offensive Rebounds for You + Defensive Rebounds for the Opponent)

    This is the percentage of shots that you miss that you're able to get the rebound on. It basically is saying "hey, it's ok if you miss the shot if you're going to get the rebound". Now KenPom makes a good point in that this stat does NOT include "team rebounds" (where two or more guys fight over the ball), so it doesn't always line up perfectly with box scores, but it's pretty consistent overall. Let's use last year's game at Kansas State for an example:
    CU Off. Reb10
    KSU Def. Reb14

    Using the formula, we get:

    10 / (10+14) = 10/24 = 41.7%

    And that right there tells you why CU won that game.

    FTRate (Free Throw Rate): FTRate = FTA/FGA

    This metric tells you how often you're shooting free throws, which is more valuable in the college game than it is in the pros. The more you can get to the line, the more "free" points that you get. Let's use last year's Texas game for an example:
    CU FTA32
    CU FGA60

    Using the formula, we get:

    32/60 = 53.3%

    And that's how we won the Texas game.

    Now along with the four factors, there are three other terms you're going to see us use regularly. They are much easier.

    Offensive/Defensive Efficiency = Points Scored * 100 / Possessions

    This is just telling you how many points your team would score/give up if there were 100 possessions in a game. You will not see possessions listed in the box score, so I usually just rely to the people who are much smarter than I like Ken Pomeroy to inform me of these stats. For our example, we'll use a hypothetical game in which there are 67 posessions (which is right around the national average from last year):
    Colorado75
    Arizona72

    For offensive efficiency, we would have:

    75*100 / 67 = 111.9

    For defensive efficiency, we would have:

    72*100 / 67 = 107.5

    The national average last year was 101.3, so this would be considered a great game for CU's offense, but a below average one for the D.

    Pace

    Pace is the last term we will use and it just basically tells you how many posessions a team has in a game. This is used for the efficiency stats mentioned above.

    OTHER LINKS

    There is a wealth of info out there, however it all starts with KenPom.com. Basically, if you can use KenPom for the basic info, you're set. Everyone should bookmark his site now.


    Basketball Reference also has a college basketball section. I prefer the site for NBA info, but the college basketball stuff isn't bad either and is worth a bookmark.
     
  2. boulder77

    boulder77 Certified Organic Club Member

    Joined:
    Jun 14, 2010
    Messages:
    2,166
    Likes Received:
    355
  3. Goose

    Goose Hoops Moderator Club Member Junta Member

    Joined:
    Aug 25, 2005
    Messages:
    10,560
    Likes Received:
    1,545
    And that's why we're posting this. I promise it's not as tricky as it looks, so if there are any questions, please feel free to ask them. Buffnik and I are working on game previews for this season that will use some of these stats, so we wanted to get this up now so everyone could look over it in case they had questions.
     
  4. wsp4820

    wsp4820 Sally Club Member Junta Member

    Joined:
    Jul 8, 2005
    Messages:
    24,225
    Likes Received:
    1,555
    Someone call the cops, cuz Goose is killin it.
     
    Goose likes this.
  5. CUFan

    CUFan Welcome back Club Member

    Joined:
    Jun 22, 2010
    Messages:
    6,956
    Likes Received:
    464
    Rep. Thank you. These are interesting stats. I look forward to seeing them as the season progresses. Initially I am excited to see the free throw and rebounding ones. Free throws and rebounding are so important to a successful college team, I am glad to have what looks like a good way to measure them.

    This type of stuff reminds of what the guy at Purple Row does for the Rockies. I like that kind of analysis and definitely appreciate the effort.
     
  6. Buffnik

    Buffnik Real name isn't Nik Club Member Junta Member

    Joined:
    Mar 20, 2009
    Messages:
    80,597
    Likes Received:
    16,220
    Great stuff, Goose.

    If you look at the 2010-11 season, the Buffs were exceptionally good with offensive efficiency. Where we struggled was with rebounding efficiency and defensive efficiency. If we can make major strides in those areas, we can absorb the offensive losses from last season.

    Basically, you have to look at it like a coach would.

    Do we have someone as efficient as Alec Burks (#6 in the nation in offensive efficiency among players who played top minutes)? No.

    Do we have a 6th man as efficient as Levi Knutson (#6 in the nation among all players in offensive efficiency regardless of minutes)? No.

    So how do we make up for that?

    1. We rebound better to get more possessions than our opponent.
    2. We limit the other team's offensive efficiency so that our own offense doesn't have to be as good.
    3. We limit turnovers so that we don't waste possessions.
    4. We cause more turnovers in order to force wasted possessions.

    These things are all complementary. If you control the defensive glass, you don't give up easy points (better d efficiency) and you're more likely to get easy points yourself off fast break opportunities (better o efficiency). Forcing turnovers work the same way.

    We have an opportunity to be just as good as last season, or even better, if the team completely buys into Boyle's system and finishes Top 3 in the conference in rebounding and defense.
     
  7. dio

    dio Admin Club Member Junta Member

    Joined:
    Dec 21, 2007
    Messages:
    30,109
    Likes Received:
    2,546
    Holy ****e this is awesome

    Sent from my ADR6350 using Tapatalk
     
  8. Sexton Hardcastle

    Sexton Hardcastle Club Member Club Member

    Joined:
    Jul 11, 2005
    Messages:
    17,472
    Likes Received:
    367
    :bowdown::bowdown::bowdown::bowdown:
     
  9. absinthe

    absinthe Ambitious but rubbish. Club Member Junta Member

    Joined:
    Dec 1, 2006
    Messages:
    25,881
    Likes Received:
    1,499
    Stat nerds strike, well done Goose.
     
  10. Burrito Palazzo

    Burrito Palazzo huff my smug Club Member

    Joined:
    Oct 8, 2010
    Messages:
    31,661
    Likes Received:
    4,351
    Jesus, Goose. What's apbr stand for?
     
  11. Buffnik

    Buffnik Real name isn't Nik Club Member Junta Member

    Joined:
    Mar 20, 2009
    Messages:
    80,597
    Likes Received:
    16,220
  12. Darth Snow

    Darth Snow Hawaiian Buffalo Club Member Junta Member

    Joined:
    Feb 1, 2008
    Messages:
    56,904
    Likes Received:
    4,916
    ?
    [​IMG]
    [​IMG]
     
    Burrito Palazzo likes this.
  13. jgisland

    jgisland Club Member Club Member

    Joined:
    Aug 31, 2007
    Messages:
    2,468
    Likes Received:
    471
    Ok, be forewarned, I like basketball and stats a little too much. I have been playing around with the kenpom.com stats for the last month or so (also beware, the site is going from free to $20 a year in the next week). But here is something I put together for last years Pac 10 stats + CU and Utah.

    TeamConfAdjTempoAdjOEeFG%TO%OR%FTRateAdjDEeFG% DEFTO% DefOR% DEFFTRate DEF
    Arizona 5P1066.2117.354.118.833.743.696.448.618.629.236.6
    Washington 7P1071.1117.853.716.737.631.594.746.420.631.944
    ColoradoB1267.9117.352.516.934.639.3102.450.619.332.435.7
    Washington St.P1068.5106.551.618.228.2369446.12132.832.6
    CaliforniaP1068.8109.350.819.530.147.198.950.418.228.633.8
    UCLA 7P1066.6107.250.32235.744.393.246.31831.333
    Arizona St.P1063.9102.850.11927.82999.749.1203339.3
    Southern California 11P1063.1106.149.918.230.435.492.146.218.528.436.1
    StanfordP1065.6103.24920.43337.797.649.121.330.136.3
    OregonP1067.5105.648.217.830.430.898.151.722.432.735.8
    UtahMWC68.3100.947.720.128.836.799.347.415.732.238.7
    Oregon St.P1070.2100.747.321.833.643.6100.953.823.432.437.3
     
    Goose and Buffnik like this.
  14. Buffnik

    Buffnik Real name isn't Nik Club Member Junta Member

    Joined:
    Mar 20, 2009
    Messages:
    80,597
    Likes Received:
    16,220
    Great work, jgisland.

    Three things that jumped out at me:

    1. Tempo is fast in the Pac-12.
    2. Washington is extremely well-coached.
    3. Our defense was awful last year.
     
  15. tante

    tante Club Member Club Member

    Joined:
    Jul 21, 2009
    Messages:
    25,787
    Likes Received:
    1,359
    if I am reading it correctly, Washington was really aggressive on defense (high defensive FT rate) so they must have a deep bench to keep the offensive numbers up even with the foul trouble some guys would be in.
     
  16. jgisland

    jgisland Club Member Club Member

    Joined:
    Aug 31, 2007
    Messages:
    2,468
    Likes Received:
    471
    1. I ran the #'s on this. The B12 adjusted conf tempo last year was 67.358 and the P10 was 67.15 last year. So very similar pace of play. FYI - of all major conferences the B10 by far plays the slowest pace, 64.573.

    2. Yes they are, and they should be really good again, even with losing Isiah Thomas. They picked up a Frosh Tony Wroten who seems to be the real deal.

    3. Awful, Atrocious, miserable....pick your negative adjective, that is what our defense was.
     
  17. jgisland

    jgisland Club Member Club Member

    Joined:
    Aug 31, 2007
    Messages:
    2,468
    Likes Received:
    471
    Pretty much, they played 8 guys pretty heavy minutes which is a pretty deep rotation in college basketball. They averaged 20.2 fouls a game which was the highest in the P10 and way up there nationally.
     
  18. Goose

    Goose Hoops Moderator Club Member Junta Member

    Joined:
    Aug 25, 2005
    Messages:
    10,560
    Likes Received:
    1,545
    Allsome work here.

    EVERYONE REP THIS POST!
     
  19. Goose

    Goose Hoops Moderator Club Member Junta Member

    Joined:
    Aug 25, 2005
    Messages:
    10,560
    Likes Received:
    1,545
    Wisconsin might have single handedly brought the Big 10 down from 67 to 64 by themselves.
     
  20. torerobuff

    torerobuff Dancing is forbidden Club Member

    Joined:
    Oct 18, 2010
    Messages:
    2,554
    Likes Received:
    565
    Kenpom is going paid? that sucks. I don't use it enough to justify paying for it, but I do enjoy going there once a month or so during the season.
     
  21. Goose

    Goose Hoops Moderator Club Member Junta Member

    Joined:
    Aug 25, 2005
    Messages:
    10,560
    Likes Received:
    1,545
    Not everything is. The basic stuff is going to be free still, but more in-depth and (I believe) the blog are.
     
  22. Buffnik

    Buffnik Real name isn't Nik Club Member Junta Member

    Joined:
    Mar 20, 2009
    Messages:
    80,597
    Likes Received:
    16,220
    Don't forget Penn State. When they played Wisconsin late in the year, I don't think either team got to 40 points.
     
  23. chipwich

    chipwich Administrator Club Member Junta Member

    Joined:
    Jul 8, 2005
    Messages:
    20,682
    Likes Received:
    1,646
    good stuff..
     
  24. jgisland

    jgisland Club Member Club Member

    Joined:
    Aug 31, 2007
    Messages:
    2,468
    Likes Received:
    471
    Anybody who is really interested in advanced bball stats I encourage you to read "A Starting Point for Basketball Statistics." This is a paper written by 4 leading researchers on advanced basketball stats (one of which is Dean Oliver, aka the godfather of bball statistical research and former Director of Quantitative Analysis for the Nuggets who now works for ESPN). Yes, it is a little math heavy at points, but if this isn't your thing you can skip over that and still get over the jist of it. They do a nice job of why they think this approach this works and touch on a few other advanced stats methods (adjusted +/- to name one).

    Paper is free to download here
     
  25. Ruckus

    Ruckus Peep my words. Club Member

    Joined:
    Jan 4, 2008
    Messages:
    16,544
    Likes Received:
    1,784
    I'm excited to meet this Goose character. Basketball can't come soon enough. Five days.

    -- I'm a Guy On A Buffalo --
     
  26. Nor Cal Buff

    Nor Cal Buff Well-Known Member

    Joined:
    Jan 12, 2009
    Messages:
    516
    Likes Received:
    65
    Good post Goose. It seems like the basketball forum here is going to have much higher level discussion than the football forum does. Looking forward to it.
     
  27. jgisland

    jgisland Club Member Club Member

    Joined:
    Aug 31, 2007
    Messages:
    2,468
    Likes Received:
    471
    This is early in the season so 1 game can drastically sway these numbers, but CU's offensive #'s are atrocious.

    EFG% 48% - 174th in the country
    TO% 24.4% - 247th in the country - CU turns the ball over once every 4 times down the floor.
    OR% - 16.9% - 334th in the country - Nearly the worst in the country, CU needs to hit the offensive glass
    FTRate - 38.8% - 153 in the country

    You aren't going to win games shooting EFG% of 48% while only grabbing 16.9% of offensive boards, it just isn't going to happen. I have to believe Boyle and crew are preaching this and I hope to see better rebounding today when CU plays Maryland.

    The good think is the defense is playing pretty well
    Holding the opposing team to 51.7 EFG, making them turnover the ball as much as they do at 24.4%. The rebounding numbers are a little more troublesome, the opposing teams are grabbing 32.8% of their offensive rebounds. Teams are not getting to the line against CU however, nearly the best in the country at a defensive FTRate of 13.3%.
     
  28. Darth Snow

    Darth Snow Hawaiian Buffalo Club Member Junta Member

    Joined:
    Feb 1, 2008
    Messages:
    56,904
    Likes Received:
    4,916
    Playing WSU really hurt... but we knew that was coming after SHT's performance in game 1. Losing Higgins and Burks's rebounding ability is killing us almost as bad as losing their scoring. The guards are giving us **** for rebounding (excepting Chen). I expected that from Nate, but Dinwiddie needs to start figuring out how big he is.
     
  29. jgisland

    jgisland Club Member Club Member

    Joined:
    Aug 31, 2007
    Messages:
    2,468
    Likes Received:
    471
    Burks rebounding ability was vastly underrated last year. He wasn't banging down low, but he had a great eye to see where the ball was going to come off the rim and would be right there. He was the 2nd best re-bounder by far after dre. Pretty impressive he out rebounded Relphorde and and Dufault so handily.
     
  30. jgisland

    jgisland Club Member Club Member

    Joined:
    Aug 31, 2007
    Messages:
    2,468
    Likes Received:
    471

Share This Page