RAPM 2021

[SpreeS]

http://nbashotcharts.com/rapm?id=-2146555570

Who could explane why these players are so low on RAPM list . All these players are holding own teams in PO spot in stronger west conference. Also they have great stats and play w/o other stars.

80 Curry 1.45
81 Jokic 1.41
84 Doncic 1.38
98 Lillard 1.25

How could Green be so low at DRAMP -0.23 310th? He is a engine of Warriors defense.

[leolozon]

It must be too dependant on how other players are performing. I've still haven't seen the argument for this stat.

I'm also not a fan when a stat takes +/- too much into consideration. A player's ability isn't directly dependant on the 9 other guys playing. If find it weird when people don't see the context. Who are you playing with? Who is replacing you when you're not on the floor? It matters.

[Muha_i_samolet]

+/- stats are 100% worthless when evaluating an individual player.

[ElectricMayhem]

1) Kawhi Leonard
2) Rudy Gobert
Ok, I see you RAPM. And I see you seeing me.
3) Garrett Tem....Oh, for Chrissakes, you just HAD to go and ruin a good thing!

[Wilfried]

Damned Hinkie!
You traded MCW away and now we're stuck with Ben Simmons

[art_tatum]

Cause single advanced stats based on +/- aren't worth much comparing players from different teams. For example curry is 80th in rapm but 3rd in rpm. Lol

[CptCrunch]

http://nbashotcharts.com/rapm?id=-2146555570

Who could explane why these players are so low on RAPM list . All these players are holding own teams in PO spot in stronger west conference. Also they have great stats and play w/o other stars.

80 Curry 1.45
81 Jokic 1.41
84 Doncic 1.38
98 Lillard 1.25

How could Green be so low at DRAMP -0.23 310th? He is a engine of Warriors defense.

RAPM isn't RAPM isn't RAPM. All RAPM are different.

RAPM datasets are wildly different.

RAPM's amount of regularization is different.

There isn't a standard way to calculate RAPM.

(I am not discounting this version, just stating that RAPM across 'analysts' are different).

[AbeVigodaLive]

It's why it's important to take every statistic with a grain of salt.

It'll be interesting to see which ones retain relevancy over time — and which new metrics are used to judge performance in the coming years.

In the end, the only thing I know for certain is that we'll have more to choose from to cherry pick our arguments for or against certain players.

[scrabbarista]

Zach LaVine is 12th on his team. So, the rest of the Bulls are carrying their All-Star... I guess?

[Quentin]

Any stat that uses +/- should be used to assess team lineups, not the player in the lineup.

[Lunartic]

1) Kawhi Leonard
2) Rudy Gobert
Ok, I see you RAPM. And I see you seeing me.
3) Garrett Tem....Oh, for Chrissakes, you just HAD to go and ruin a good thing!

Temple has been great for the Bulls, he and Thad Young are a reason why they are in the playoffs right now

[scrabbarista]

I gotta give credit to the stat creator for not fudging his results, at least. There's no way he'd get those results if he was just trying to make the stat look good.

It's also important to remember that this is dealing with an extremely small sample size. His three-year version ('18-'21) is:

1. George
2. Antetokounmpo
3. Durant
4. James
5. Leonard
6. Harden
7. Paul
8. Curry
9. Tatum
10. Embiid

I'm a huge supporter of +/- generally, but always with context and always as just a piece of the overall statistical puzzle. There's no one stat that will ever come close to being a definitive all-in-one. Personally, my favorite method of statistically evaluating players is a combination of +/- stats and box score stats (WS, VORP, etc..), with a dash of team success thrown in. I find this really brings a nice balance, and it's what I do for my MVP formula.

[scrabbarista]

The 5-Year version is:

1. Curry
2. Paul
3. James
4. Antetokounmpo
5. Butler
6. Leonard
7. Embiid
8. George
9. Gobert
10. Lillard
11. Harden
12. Lowry
13. Durant
14. Holiday
15. Tatum
16. Beverley
17. Jokic
18. Walker
19. Adams
20. Davis

Other than Beverley and Adams, those are all star players. I doubt you'll find a single commonly available stat of any kind that could give a top twenty list of this quality for the last five seasons. But it goes to show that even the best stats are still just a single chunk of information - just a single piece of a larger puzzle.

[Hook_Em]

OT but where do you find SRS ratings for this year?

[KqWIN]

There's a lot of people who seem uninformed on what RAPM is. Nothing wrong with that, but before you dismiss something so quickly you should know what it is. RAPM is an estimated +/- that is adjusted for your teammates and opponents. If any wants to know how this estimation is made, I'm happy to explain.

It may seem surprising that someone like Garrett Temple is so high and maybe someone like Draymond's defense is so low. But when you consider what goes into this, it shouldn't The Bulls are +6.8 with Temple on the floor and -8.5 without him. Obviously he's going to look good here. As far as Draymond's defense, the Warriors actually have a better DRTg with him on the bench. Again, shouldn't be a surprise that he doesn't rate out well.

Does this make it a useless stat? Is 3FG% a useless stat because Kent Bazemore's is higher than Curry's? Of course not. You really can't evaluate 3 point shooting without considering their percentage. But that doesn't mean 3FG% is everything.

[giberish]

IMO the big flaw (or limitation) with RAPM is sample size. Even a full 82 game season is a small sample-size yielding noisy results.

[TheSuzerain]

Zach LaVine is 12th on his team. So, the rest of the Bulls are carrying their All-Star... I guess?

Bulls have some hilarious lineup disparities. It's an intriguing storyline for the Bulls season.

[scrabbarista]

There's a lot of people who seem uninformed on what RAPM is. Nothing wrong with that, but before you dismiss something so quickly you should know what it is. RAPM is an estimated +/- that is adjusted for your teammates and opponents. If any wants to know how this estimation is made, I'm happy to explain.

It may seem surprising that someone like Garrett Temple is so high and maybe someone like Draymond's defense is so low. But when you consider what goes into this, it shouldn't The Bulls are +6.8 with Temple on the floor and -8.5 without him. Obviously he's going to look good here. As far as Draymond's defense, the Warriors actually have a better DRTg with him on the bench. Again, shouldn't be a surprise that he doesn't rate out well.

Does this make it a useless stat? Is 3FG% a useless stat because Kent Bazemore's is higher than Curry's? Of course not. You really can't evaluate 3 point shooting without considering their percentage. But that doesn't mean 3FG% is everything.

Solid analogy in the last paragraph. That's how I view +/- stats: essential, but (of course) not "everything."

[scrabbarista]

OT but where do you find SRS ratings for this year?

https://www.basketball-reference.com/leagues/NBA_2021.html

It's on the far right hand in each conference.

[dhsilv2]

The fact it's 2021 and people both dismiss +/- data, don't understand RAPM, and don't understand sample sizes is a true failure of the world's education system.

The sample size is way too small to at this point be looking at RAPM (any "vanilla" version). That's the reason ESPN uses their RPM as it uses other metrics to help get things right. It's why really a 2 year sample is what you'd want. But whatever.... it's 2021 this metric isn't new and that we have to continue to explain that you shouldn't take any stat and blindly rank players let alone a metric requiring more data...