• ... [JohnH:] ... to ... team would score if ... This seems sound, to me. However, I was under the impression that average points-following-a-foul were not
Message 1 of 39 , May 27, 2003
View Source
• 0 Attachment
--- In APBR_analysis@yahoogroups.com, bchaikin@a... wrote:
>
[JohnH:]
> PFs ended up at 0.35 because that was the ratio of made foul shots
to
> personal fouls, minus the average points per possessions that a
team would score if
> the foul hadn't been committed.

This seems sound, to me. However, I was under the impression that
average points-following-a-foul were not significantly different
from average points-per-possession.

>
[BC:]
> lets use the 01-02 season as an example. there were 42619 FTM and
50497 PF
> for a ratio of FTM/PF of 0.85. courtesy of harvey pollack's guide,
there were
> 4553 offensive fouls committed that season, so now that ratio now
creeps up to
> over 0.90 FTM/PF. although a small minority of fouls can be
considered good
> fouls (as in fouling a poor FT shooter near the end of a game if
you are losing),
> and many fouls result in no FTA (non-shooting fouls committed
before the
> bonus and thus not resulting in any FTA), the bottom line is that
the only way for
> a player to score from the FT line (except the rare technicals,
double fouls,
> etc) is to be fouled, and thus - on average - every PF committed
results in
> the opponent scoring close to a point, somewhere around 0.90 or
more points per
> personal foul committed....
>
> are you somehow accounting for the discrepancy between this 0.90
and your
> value of 0.35 in your other coefficients? is this what you mean
by "...minus the
> avg pts/poss that a team would score if the foul hadn't been
committed..."?...

Logically, if a foul doesn't increase the likelihood of the other
team's scoring, it shouldn't be a negative factor at all. If .9 pts-
allowed results from the average foul, how can that even be
considered a net detriment to the fouling team?

I'm not convinced the majority of fouls are 'bad' fouls. If you
are playing tight defense, you are going to get called for some
fouls. Sometimes you will get away with a no-call.

Many fouls are decidedly good or neutral. Stopping Shaq from a dunk
(`1.25 vs. 1.98 probable pts) is better than letting him dunk.

>[JH:]
> Field goals ended up at 1.65 after I subtracted credit for an
assist. It
> varied by team; field goals by Utah players were worth the least
because they were
> most like to have credit split with the assister.

Is it assumed, then, that John Stockton's FG made are as likely to
have been assisted as Kirilenko FG are? I recall DeanO claiming to
have had a formula that distinguished between the players on a team -
- not just between teams.

>
[BC:]
> am i to understand that in your individual player rating system a
FGM by a
> player on one team may not be of the same value as a FGM by a
player on another
> team? if so i would have a hard time comprehending this as a
trustworty rating
> system as in any single game all FGM by players on both teams are
of equal
> value (aka a 2pter is a 2tper, and a 3pter is a 3pter, a FTM a
FTM)...

I think what John is saying is that individual credits can be forced
to add up to the final score, by splitting credit between the scorer
and the assister. Again, this makes sense to me.

>
[JH:]
> And now for our main event....
>
> As far as rebounds are concerned, I think the "a rebound is a
possession"
> logic is the main reason the Danny Fortsons of the world are
vastly overrated by
> most statistical methods.

Fortson seems to rank high in per-minute lists and low in per-game
lists. In the rare seasons when he's stayed healthy (maybe once),
he's ranked pretty strong in my (non-linear) rating system. He's
been a high-%, medium-strength scorer as well as a rebounder.

>
[BC:]
> what praytell is a rebound if it isn't getting your team
possession of the
> ball? no matter how you take the ball away from your opponent, if
you do take
> the ball away without your opponent scoring, how can those
separate events be
> valued differently? how can a steal (or a turnover that wasn't a
steal) and a
> def reb be valued differently?...

I'm solidly with JohnH here. A steal removes the ball from the
opponent AND gains the possession. A rebound is gaining possession
of what is essentially a loose ball.

I still don't follow how an offensive rebound is worth more because
it is rarer. There are all kinds of events rarer than a rebound,
which are worth whatever they are worth, and not based on their
rarity.

For sure, some guys get a lot of defensive rebounds from missed FT,
and these are almost 'givens'. But in the long haul, I have to
guess there would be even more total rebounds available from non-FT
situations.

>
[BC:]
> as for danny fortson how is he - and "his kind" - overrated? who
is
> overrating him? danny fortson is an absolutely superb rebounder
matched by only a
> handful of players in the past 3 years, who, unfortunately, has
played in less than
> 60% of his team's games in his six years in the league, and
commits way too
> many fouls to be a regular starter.

That's how I've generally assumed fouls should come into
consideration in player-ratings: They limit your minutes. Of
course, if your ratings are strictly 'per-minute', you may get wacky
results.

>
[JH:]
> Basically, there are two ways to get a defensive 'stop':
>
> 1) force a turnover,
> 2) force a missed shot AND get the rebound.
>
>> You're asking me to equate a portion of 2) with the entirety of
1). To me
> that makes no sense.
>

[BC:]
> from the standpoint of bookkeeping stats from a team perspective i
understand
> your concern, but from the standpoint of rating individual players
the second
> option above is cleary two separate events and the first is
clearly one, and
> the bottom line is that once a shot is missed, its a free ball and
either the
> offensive team can regain possession of the ball or the defensive
team can.
> and as you have cleary shown
> by your labelling of a steal and a turnover as equal and opposite
events and
> of absolute equal value equivalent to that of an avg team
possession of 1.02
> - because one team gains a team possession and one loses one, the
situation is
> identical for an offensive and defensive rebound - either you are
regaining
> team possession of the ball or you are losing it...

It seems that Bob is agreeing with John; am I right?
• I ve bought your book also. First I ve read the Prospectus of Hollinger and after that Basketball on paper. I v read reviews that said: the prospectus is good
Message 39 of 39 , Feb 1, 2004
View Source
• 0 Attachment