EPTPascal Siakam is outplaying Serge Ibaka. Jacob Poetl is outplaying Jonas Valanciunas. Problem is Ibaka and JV combine to make over $40M/year while Poetl and Siakam make 10% of that.
If this keeps up Casey has got some tough decisions to make.
NBA Season 2017-2018 - Page 37
| Forum Index > Sports |
|
JimmyJRaynor
Canada17815 Posts
Pascal Siakam is outplaying Serge Ibaka. Jacob Poetl is outplaying Jonas Valanciunas. Problem is Ibaka and JV combine to make over $40M/year while Poetl and Siakam make 10% of that. If this keeps up Casey has got some tough decisions to make. | ||
|
Twinkle Toes
United States3605 Posts
| ||
|
JimmyJRaynor
Canada17815 Posts
RAPM is based on Adjusted-Plus-Minus or APM. APM is based on Plus-Minus. Plus-Minus began being used in the 1960s by hockey coach Scott Bowman. Basketball later adopted it. Plus-Minus has been continually refined and improved over the past 40+ years and RAPM is the latest cutting edge development in the refinement of plus-minus. here is a look at Adjusted Plus Minus and standard boxscore Plus Minus. https://fansided.com/2014/09/25/glossary-plus-minus-adjusted-plus-minus/ Once you have Plus-Minus and Adjusted-Plus-Minus or APM under your belt you can move on to RAPM. here is a detailed look at RAPM. https://squared2020.com/2017/09/18/deep-dive-on-regularized-adjusted-plus-minus-i-introductory-example/ | ||
|
Twinkle Toes
United States3605 Posts
| ||
|
JimmyJRaynor
Canada17815 Posts
![[image loading]](https://i.imgur.com/FajxUU3.jpg) during Jones 39 minutes on the court the Timberwolves scored 22 more points than they surrendered. Thus, he was "+22" during his 39 minutes. During Gibson time on the court 15 more points were scored than were surrendered. So Gibson was "+15". If a player was on the court during a time when the team surrendered 12 more points than they scored that player would be "-12". Once you understand basic Plus-Minus you can move on to Adjusted Plus/Minus or APM. Understanding the philosophy behind APM is difficult for 10 year olds who know nothing about systems of equations and unknowns. Nylon Calculus' simplified explanation of APM requires Grade 9 math. To completely understand exactly how APM works requires 2 years of university level Linear Algebra. Those cheesey "math for sociology majors" courses won't do it. I'm talking some pretty hard core Algebra. | ||
|
ZenithM
France15952 Posts
The flaw of these metrics is that if 2 guys share the floor most of their playing time, you're not going to know who's doing the real work. | ||
|
Twinkle Toes
United States3605 Posts
And how would you respond to ZenithM? | ||
|
JimmiC
Canada22817 Posts
| ||
|
Twinkle Toes
United States3605 Posts
On November 27 2017 22:17 JimmiC wrote: Not to mention how does it account for communication on the d end (switches whos the problem) and systems. Pretty sure they keep the formulas top secret so people cant see the obvious flaws and then some people who cant critically think about the numbers (which show HUGE changes when plauers change teams) treat them like an accurate ranking and player comparison tool. Instead of just another stat that provides some info but has its own flaws. Morey said in an interview that it is the most important property right now in team management, and the backdoor draining of sports statisticians is actually more intense that superstar acquisitions. | ||
|
Twinkle Toes
United States3605 Posts
| ||
|
JimmiC
Canada22817 Posts
| ||
|
JimmyJRaynor
Canada17815 Posts
5th year in a row ESPN's Real Plus Minus has done a good job capturing the performance of every Toronto Raptor. When it was first introduced I didn't pay much attention to it. However, after 3 years of good measurements i dug into the work of Jeremias Engelmann and Dan Rosenbaum. These guys have done some good work. On November 27 2017 19:33 ZenithM wrote: Goddammit, this again. The flaw of these metrics is that if 2 guys share the floor most of their playing time, you're not going to know who's doing the real work. meh, he asked. i answered. that phenomenon is known as co-linearity. it is an issue with APM but not so much with RAPM. fixing the colinearity issue is a very difficult problem. However, the "Regularization" aspect of RAPM has managed to solve this problem. a great example of how RAPM fixes colinearity can be seen with Demar Derozan and Kyle Lowry of the Toronto Raptors for the past 4 years. These 2 players are together A TONNE. And yet, Lowry's defense is much much better than Derozan's. It is reflected in RAPM. On November 27 2017 22:40 Twinkle Toes wrote: And JJ, how does this RAPM etc. explain why Harden and CP3 can be productive even with high usage rate, all on a similar position, but Westbrook and PG13 and Me70 cant even if they are playing well? performance of 2 players used over a very short time period can not be projected by the metric. of course, no other metric can. On November 27 2017 23:00 JimmiC wrote: Right, the actual analytics they use not the made for espn one espn plopped on their site. If it was even 50% effective as certain people think imagine what a team like the rockets would pay to have it only for themselves. 50%? maybe its 65.4453% effective. ESPN paid Engelmann for it. Engelmann has worked for other NBA teams for years. The developer of APM worked for NBA teams for 10+ years. That'd be economics prof Dan Rosenbaum. His work is published.. peer reviewed. And is such a good stat it has formed the foundation of new stats like RAPM. Its a good stat. It is limited. APM, RAPM, and xRAPM offers a different slice of reality than standard boxscore stats. | ||
|
JimmiC
Canada22817 Posts
| ||
|
JimmyJRaynor
Canada17815 Posts
On November 27 2017 23:18 JimmiC wrote: If you actually believed this and posted like it I would have way less problems. But that is not how you treat you. If someone changes drastically (Kyrie) you make up a reason (effort) if it fits your narrative you treat it liek the gospel. Then when people call you on it "they wouldn't understand the math like you do". I've said it is a fine stat and worth looking at, just that it's not gospel. Then you argue, then write basically that LOL. i've told you its limited and named its limitations several times. you need grade 9 math for the Nylon Calculus article on APM. One other guy went through the article as well. All any stat can do is measure on field performance. Humans change due to off-field//off-court events. When Tim Raines decides to turn into a cocaine addict his performance dramatically declines and no stat is able to predict this. Does this mean every baseball stat is useless? nah. Baseball stats are fine at measuring current performance. | ||
|
JimmiC
Canada22817 Posts
| ||
|
JimmiC
Canada22817 Posts
| ||
|
JimmyJRaynor
Canada17815 Posts
On November 27 2017 23:32 JimmiC wrote: Poor example for multiple reasons, Baseball is by far the easiest game to statistically analyze because of the interactions. Every other sport is searching for a way to measure it as accurately as baseball but the complexity is so much higher. its a perfect example because it illuminates how even the best, most accurate baseball stats can predict for off-field events such as drug use. Same thing happens when a player starts using steroids. They experience a dramatic, totally unpredictable, improvement in their power hitting. No stats can predict for off the field events so don't start expecting them to do so. xRAPM, RAPM, APM, and their variants won't solve this ever. Nor will any other stat. | ||
|
JimmiC
Canada22817 Posts
| ||
|
JimmyJRaynor
Canada17815 Posts
On November 27 2017 23:37 JimmiC wrote: So you believe the only limit to those stats is outside factors such as drugs and peds? no, i've already told you several of the limits of this particular statistic. check my previous posts. in particular, examine my comments on Kyle Anderson. | ||
|
JimmiC
Canada22817 Posts
| ||
| ||
