--- In APBR_analysis@y..., "McKibbin, Stuart" <smckibbi@c...> wrote:
> I'm charting 5 categories of stats: 1) Offensive rebounds, putbacks
> rebounder and second chance points. I do not count recovering a
> as an offensive rebound. 2) In-the-act FTA's, technical FTA's, and-
> bonus FTA's, and threepoint FTA's. 3) Steals, fastbreak attempts
> fastbreak points off steals 4) Blocked shots, who recovers them,
> attempts off blocks, points off blocks 5) Offensive fouls
> First, apparently I undercount steals and blocks. In the games I've
> the boxscores say the Lakes had 123 steals and 115 blocks, and the
> 104 steals and 66 blocks. My numbers are Lakers 109 steals and 86
> opponents 86 steals and 45 blocks. I only catch about 72% of the
> 86% of the steals. It's a fast paced game, and little tip blocks in
> are hard to pickup on, but I didn't think my numbers were THAT bad.
> course, it could simply be that the official scorekeepers are
> liberal in handing out gold stars. One question: Because my numbers
> far off does that limit the usefulness of my stats?
Good luck replicating NBA numbers. Their scorers are quite arbitrary
in awarding steals and blocks. As an example, the Toronto block
shot "record" was rescinded a few days later when people questioned
it. A lot of steals and blocks are awarded without any logic I see.
I tend to undercount them as well.
Does this affect the usefulness? Yes, somewhat. For instance, if
your analysis says, for example, that 80 steals lead to 100 points
and we then say that a steal increases offensive efficiency to
approximately 1.25 pts/poss, but the league actually recorded 100
steals, the conclusion is clouded. If we go to use the overcounted
NBA numbers, we could easily be overstating the value of an NBA-
> Some preliminary numbers from the 14 games. Lakers 109 steals, 67
> attempts, 83 points. Opponents 86 steals, 59 fastbreak attempts, 82
> So if we assume the teams averaged 1 point for each non-fastbreak
> steal that means Lakes 109 steals, 125 points; opponents 86 steals,
> points. Totals: 195 steals, 234 points or 1.2 points per steal. I
> can't extrapolate this to the league at large (Laker games may not
> representative) but it is interesting.
Would be nice if you didn't assume 1 pt for non-fastbreak attempts.
Also note that if you use the official count on steals, you have 227
steals and your estimate of 234 points off of it -- pretty close to 1
pt/assist. That's where the potential undercounting is a problem.
Maybe "true" steals are worth 1.2 pts, but NBA official steals may be
worth only about 1 pt.
> Blocks: I counted 131 blocks for the games, the shooting team
> of them. So a block deprived the offense of the ball just 57% of
> Moreover, those 75 defensive recoveries resulted in very few
> attempts yielding just 33 points. It seems the blocked shot isn't
much of a
> fastbreak starter.
Good info. A blocked shot is primarily a defensive contribution.
> Offensive rebounds: My numbers matchup with the boxscores pretty
> you add in the offensive recovery of blocked shots). What is
amazing is how
> much better the Lakers are than their opponents at the rebounder
> the ball back in the hole. Lakers 150 off rebs (plus 6 off team
> shots by the rebounder, 48 putback baskets. Overall 176 points from
> second chance opportunities. The opponents had 156 off rebs (plus
> team rebs) 75 shots by rebounder, but only 29 putback baskets.
> points in 167 second chance opportunities.
Hmmm. Mixed results. I have generally done incomplete looks at the
issue and almost always with _good_ teams. Those good teams suggest
that scoring off an offensive rebound is a little more efficient than
scoring in general. Maybe it ain't true with mediocre or poor teams.
Useful to know that approximately 1/2 of all offensive rebounds are
shot by the rebounder. I can definitely use that. I've always
assumed 1/5th would be, knowing that it doesn't make a huge
difference to my analysis. I need to review this stuff a little
later to see how it all adds up.
> Last thought, the following stats show how unique and outstanding a
> Shaquille O'Neal is: 54 off rebs, 33 putback shots by him making
> He personally scored 57 points directly after his grabbing the
> teammates chipping in for another 12 points. In the 14 games, I
> 27(!) of his FTA's as "and-1's" (17.4% of the 155 FTA's he shot in
> games), the next highest on the Lakers is Kobe with 7.
I estimate total possessions from free throws as 0.4 * FTA. So
0.4*155 = 62 total possessions from his free throws. The "and 1's"
do not count as a possession, which is why I use 0.4, not 0.48 or
something closer to 0.5 (which would account only for technical
fouls). From your numbers, 155-27 = 128. 128 foul attempts on two-
shot fouls means 64 possessions. So I'm off by 2 possessions, which
could be some of the technical fouls shot by the Lakers.
Let me take a guess then at how many FTA's Kobe had:
(x-7)/2 ~ 0.4x
x ~ 35
That is very low, knowing that he has shot 144 in 16 games.
He "should have" shot many more "and-1's", probably on the order of
20. The person who should have the next most and-1's is Devean
George or Robert Horry, with maybe 4-5.
How many total "and-1's" have their been for the Laker team? How
many total technical foul shots for the team?
I need to check to see how variation from the 0.4 would affect my
offensive efficiency ratings. I think your info slightly raises
Kobe's pts produced, his scoring possessions, and his total
possessions. So, it would slightly help my calculated numbers for
For some of you that
> calculate offensive efficiency, that helps him, right? He had 37
> fouls in the 14 charted games, 9 are offensive (24.3%). For those
> that calculate defensive efficiency, that helps, too, right?
Keep counting. If 25% of all fouls are offensive, that is a useful
piece of knowledge. Or if it's 20-30%, that helps.