PC Board OT thread

Post #1481 » by **Clyde Frazier** » Wed Aug 19, 2015 5:27 pm

https://www.reddit.com/r/nba/comments/3hl51s/shit_nba_fans_said_in_the_1980s/

Cool reddit post where someone scoured usenet archives looking for posts about the NBA from the 80s. You can also search it yourself here:

https://groups.google.com/forum/#!forum/net.sport.hoops

Post #1482 » by **Goudelock** » Thu Aug 20, 2015 1:49 pm

Clyde Frazier wrote:https://www.reddit.com/r/nba/comments/3hl51s/****/

Cool reddit post where someone scoured usenet archives looking for posts about the NBA from the 80s. You can also search it yourself here:

https://groups.google.com/forum/#!forum/net.sport.hoops

My favorite post courtesy of Mark D Tischler:

I'll tell you why the NBA is so boring!

1. Zone defenses are not allowed --> detracts from defense
2. 3-to-make-2 free throws --> thwarts comebacks
3. Bad shot selection
4. One-on-one play
5. 3-point shots --> further encourages bad shot selection
6. 24-second clock --> encourages bad shot selection and one-on-one play
7. No necessity for coaching, therefore there is none -- it's sandlot bball
8. Refs don't call charging --> discourages good defense
9. Refs don't call traveling (better this year I've heard)
--> allows for dunks from mid-court
10. 20-second timeouts --> to easy to get out of trap situations
11. too many playoff teams --> who cares about the regular season anymore

So people have been arguing about the same things for almost 30 freaking years. Oh, and some people back then thought that there was too much 1on1 play back in the 80's. Interesting....

PS: I'm going to post this link on the GB and see what other people think of this.

Post #1483 » by **Clyde Frazier** » Thu Aug 27, 2015 10:25 pm

[tweet]https://twitter.com/Deadspin/status/637002585136844800[/tweet]

Very sad news. Great ambassador for the league after he retired. Also didn't realize he was only 58.

RIP, Chocolate Thunder.

ceiling raiser · Post #1484 » by **ceiling raiser** » Mon Aug 31, 2015 1:41 am

What're everyone's thoughts on the Spurs-Suns series in 07?

Normally I don't give much thoughts to conspiracy/rigging theories, but G3 (with Donaghy on the crew) was bad. The suspension was iffy too, though I'm just speaking from memory.

SA had a great team that year, but was Phoneix robbed of a title? Series could have went either way and even if they likely beat Utah and Cleveland games aren't played on paper, but if I was a Suns fan, that series wouldn't sit right with me, even eight years after the fact.

Post #1485 » by **Quotatious** » Mon Aug 31, 2015 1:50 am

fpliii wrote:What're everyone's thoughts on the Spurs-Suns series in 07?

Normally I don't give much thoughts to conspiracy/rigging theories, but G3 (with Donaghy on the crew) was bad. The suspension was iffy too, though I'm just speaking from memory.

SA had a great team that year, but was Phoneix robbed of a title? Series could have went either way and even if they likely beat Utah and Cleveland games aren't played on paper, but if I was a Suns fan, that series wouldn't sit right with me, even eight years after the fact.

I find it hard to believe in any conspiracy theory that implies the refs trying to be in favor of the Spurs. From a marketing standpoint, Suns would likely generate more profit.

ceiling raiser · Post #1486 » by **ceiling raiser** » Mon Aug 31, 2015 1:53 am

Quotatious wrote:
fpliii wrote:What're everyone's thoughts on the Spurs-Suns series in 07?

Normally I don't give much thoughts to conspiracy/rigging theories, but G3 (with Donaghy on the crew) was bad. The suspension was iffy too, though I'm just speaking from memory.

SA had a great team that year, but was Phoneix robbed of a title? Series could have went either way and even if they likely beat Utah and Cleveland games aren't played on paper, but if I was a Suns fan, that series wouldn't sit right with me, even eight years after the fact.

I find it hard to believe in any conspiracy theory that implies the refs trying to be in favor of the Spurs. From a marketing standpoint, Suns would likely generate more profit.

Fair point. As a follow-up, do you give a suggestion of a 'lone wolf' Donaghy affecting the series on a meaningful level any credence?
[youtube]https://www.youtube.com/watch?v=fvkKdXLwt0U[/youtube]

Post #1487 » by **Clyde Frazier** » Mon Aug 31, 2015 1:55 am

fpliii wrote:What're everyone's thoughts on the Spurs-Suns series in 07?

Normally I don't give much thoughts to conspiracy/rigging theories, but G3 (with Donaghy on the crew) was bad. The suspension was iffy too, though I'm just speaking from memory.

SA had a great team that year, but was Phoneix robbed of a title? Series could have went either way and even if they likely beat Utah and Cleveland games aren't played on paper, but if I was a Suns fan, that series wouldn't sit right with me, even eight years after the fact.

To put it simply, there's a good chance Horry decking Nash cost the Suns a title. Of course they could've still lost the series, but they competed very well with the Spurs, and they'd be favorites against Utah and Cleveland. Nash's legacy to the average fan would be forever changed.

I don't think Donaghy played a role in the series outcome.

GSP · Post #1488 » by **GSP** » Mon Aug 31, 2015 2:10 am

Clyde Frazier wrote:[tweet]https://twitter.com/Deadspin/status/637002585136844800[/tweet]

Very sad news. Great ambassador for the league after he retired. Also didn't realize he was only 58.

RIP, Chocolate Thunder.

A terrible shame. Hes the best dunker ever for a center IMO

ceiling raiser · Post #1489 » by **ceiling raiser** » Mon Aug 31, 2015 2:17 am

Hawk · Post #1490 » by **Hawk** » Tue Sep 1, 2015 6:40 pm

Guys, DRtg or DBPM, which stat do you prefer if you had to choose one?

Post #1491 » by **bondom34** » Tue Sep 1, 2015 6:43 pm

Dr Olajuwon wrote:Guys, DRtg or DBPM, which stat do you prefer if you had to choose one?

Neither. I wouldn't want to use either one unless absolutely forced, and then.......I guess D rating? Really not a fan of DBPM, I think OBPM is good, but defense isn't captured well by box scores.

ceiling raiser · Post #1492 » by **ceiling raiser** » Tue Sep 1, 2015 7:07 pm

Dr Olajuwon wrote:Guys, DRtg or DBPM, which stat do you prefer if you had to choose one?

I don't mind DBPM too much, but if we're talking about the play-by-play era DRPM works pretty well.

Kinda OT but recently I've been a bit soured on prior-informed RAPM. From a Bayesian POV adding a prior to a model helps improve your starting guess from 0->prior, but I think when we're comparing some of the top stars (which we do most commonly on this board), the fact that PI can lag behind one or more years (like was the case with Durant, and more recently seems to be the case with Davis; probably with others too) does not make me comfortable with using it as a base for my player comparisons. At this point I prefer a clean NPI model (either using a built-in package or direct computation using the linear algebra), and taking it for what it is. There are still colinearity issues (and noise, though I think if you don't take it as a straight ranking it's safe), but I think if you manually compare guys to their teammates (across several years if the core is stable), you can get around that and paint a pretty complete picture.

Post #1493 » by **bondom34** » Tue Sep 1, 2015 7:15 pm

Yeah DRPM I'm okay with, that's usually my go to defensive metric. Not a huge fan of any plus/minus without context, but DRPM is a decent quick rank.

fuzzy_dunlop · Post #1494 » by **fuzzy_dunlop** » Tue Sep 1, 2015 8:33 pm

so not sure if folks are aware, but talkingpractice have released their top 25 2014-2015 IPV values (RS+PO):
http://talkingpracticeblog.com/

Post #1495 » by **bondom34** » Tue Sep 1, 2015 9:41 pm

Thanks for that, haven't kept up w them in a bit.

Post #1496 » by **Quotatious** » Wed Sep 2, 2015 2:12 pm

"New opinions often appear first as jokes and fancies, then as blasphemies and treason, then as questions open to discussion, and finally as established truths."

- George Bernard Shaw

I feel like it could be the PC board's motto. It also seems like something that fpliii would put in his signature.

Post #1497 » by **Texas Chuck** » Wed Sep 2, 2015 2:56 pm

Really off-topic and wrong sport, but this is always one of the coolest plays in sports:

[tweet]https://twitter.com/Rangers/status/638942652486959104[/tweet]

The-Power · Post #1498 » by **The-Power** » Wed Sep 2, 2015 3:13 pm

Dr Olajuwon wrote:Guys, DRtg or DBPM, which stat do you prefer if you had to choose one?

As it was already mentioned, the box score is not great to capture defense regardless of which metric one uses. I tend to like individual DRTG more but only after putting some context into the numbers. What I prefer about the DRTG is the approach of using per 100 possessions and it also takes team-context into account. It shares credit evenly through all players whenever an opponent's possession ends without a box score event on the defense's side - that's a problem and an advantage simultaneously. It's favoring bigs for their box score production in terms of defense but since most bigs are more impactful than perimeter players on defense this can be an advantage (because it makes it right) as well. On the other hand, however, it probably doesn't "punish" big men with poor defense - which is worse for a team than a poor defensive guard due to the relative importance of their position - enough. They assume every player has the same defensive possessions per minute which undersells the impact (positive and negative similarly) of bigs; the difference in box score production between good and bad defensive big men might not be sufficient to see the differences; and of course the problem of stat-padders and players who are better on defense than the box score suggests maintains - but at least team-efficiency of defense helps to adjust for that a little bit, although this (sometimes seen as an inconvenient by-product) adjustment brings up its own problems. Some of these problems can be reduced, albeit certainly not abolished, by avoiding cross-positional comparisons.

DBPM doesn't give me much. It's something I look at for fun but I barely use it in any argumentation. It's even more prone to "manipulation" (= to be unjustifiably influenced) than DRTG due to the lack of team-effiency consideration (which can be a good thing at times, sure, but overall I look at it as a negative especially when it comes to the evaluation of defense through box score numbers) and people also often forget to consider minutes played (what BPM didn't do, as some might not know; per 100 possession stats obviously don't take minutes played into account as well but that should be common sense).

Fortunately we have new data available to evaluate defense by objective measures. Any defensive evaluation of historical players has to start with watching games and creating contexts for the numbers we have. The same is necessary in terms of evaluation of offense but the historical data for offensive production is broader and more useful which makes it less imperative for shallow analysis.

bondom34 wrote:Yeah DRPM I'm okay with, that's usually my go to defensive metric. Not a huge fan of any plus/minus without context, but DRPM is a decent quick rank.

May I ask why? You said you don't like plus-minus approaches without context and also mentioned that defense can't be adequately captured by the box score - yet you like a combination of both? Do you hope that throwing two - in your opinion - flawed approaches into one cup makes one trustful one? Do you consider the box score data to be the context DRAPM lacks? Or do you like DRPM simply because it agrees with your eyes the most?

Post #1499 » by **bondom34** » Wed Sep 2, 2015 3:35 pm

The-Power wrote:
bondom34 wrote:Yeah DRPM I'm okay with, that's usually my go to defensive metric. Not a huge fan of any plus/minus without context, but DRPM is a decent quick rank.

May I ask why? You said you don't like plus-minus approaches without context and also mentioned that defense can't be adequately captured by the box score - yet you like a combination of both? Do you hope that throwing two - in your opinion - flawed approaches into one cup makes one trustful one? Do you consider the box score data to be the context DRAPM lacks? Or do you like DRPM simply because it agrees with your eyes the most?

Because its the best of a bunch of flawed stats. Its not perfect, but I can tell by looking at it that its better than the other options. I don't consider box score to be context plus/minus lacks, I consider actual context to be context that's missing. For example, what role a guy plays (a guy who's a 3rd option or who isn't a high usage player or going against bench players vs. a primary/secondary option starter, think a guy like Korver vs. a guy like Jimmy Butler).

I don't entirely dismiss plus/minus (though I pretty much do NPI) if its used correctly, I just find many don't use it in a way it should be. Its a small piece of the puzzle along with the box score. My thing with defense is that its not really well captured in any way by box scores. I'd prefer a mix of RPM and player tracking (the Nylon Calc stats are great for this too w/ rim protection), as they can at least give a better idea to me of what effect someone has on a team's defense.

The-Power · Post #1500 » by **The-Power** » Wed Sep 2, 2015 4:00 pm

bondom34 wrote:Because its the best of a bunch of flawed stats. Its not perfect, but I can tell by looking at it that its better than the other options.

This is human and I get what you're saying. But one has to be cautious to not fall into confirmation bias.

bondom34 wrote:I don't consider box score to be context plus/minus lacks, I consider actual context to be context that's missing. For example, what role a guy plays (a guy who's a 3rd option or who isn't a high usage player or going against bench players vs. a primary/secondary option starter, think a guy like Korver vs. a guy like Jimmy Butler).

Well, what hinders you to consider actual context when looking at +/- data? By the way regression data takes quality of opponents (and teammates) into account by default, that's the point of adjustment in the first place. And a 5th option shouldn't look as impactful on offense as well by the metric. The problem isn't the approach (because it's the only way to adequately measure impact approach-wise), it's the potential noise.

bondom34 wrote: I'd prefer a mix of RPM and player tracking (the Nylon Calc stats are great for this too w/ rim protection), as they can at least give a better idea to me of what effect someone has on a team's defense.

I agree that player tracking has to be a large part of basketball-analysis in the near future (or let's say from last year on) and this is especially true for the evaluation of defense. Nevertheless, when it comes to earlier years we have to work with what we've got as unsatisfactory as it might be at times.