2021-22 NBA Rookie Thread (part 5)

[MotownMadness]

Read on Twitter

Hartenstein has some good stats for a 7 footer getting ready to hit UFA.

[Kurtz]

I'll look into it, but I don't know if it's really necessary. RAPTOR is a very good advanced stat but we're already using EPM as the main metric, which is very similar to RAPTOR (based on RAPM and incorporating player tracking data for example), and testing indicates EPM is a bit more accurate. RAPTOR also has some issues with smaller minutes sample size and doesn't handle roster discontinuity as well. But, if I can get the data easily, I might replace WS/48 with RAPTOR.

Just my opinion, but if you have a chart with several advanced stats, then it would make sense to sort by the advanced stat that most correlates with the eye test/logic. So if EPM ranks Herbert Jones at 2 and Jock Londale at 3, I would avoid sorting by that stat, and would try using one of the other ones that produce a more sensible ranking. Maybe VORP?

That would drop Herbert Jones all the way down to 3

Yes and perhaps that's where he belongs. But it would also drop Jock Landale with his 10 mpg out of the top 3.

[jordanwilliams6]

Read on Twitter

?s=21

[JonFromVA]

you need to include RAPTOR in your list

I'll look into it, but I don't know if it's really necessary. RAPTOR is a very good advanced stat but we're already using EPM as the main metric, which is very similar to RAPTOR (based on RAPM and incorporating player tracking data for example), and testing indicates EPM is a bit more accurate. RAPTOR also has some issues with smaller minutes sample size and doesn't handle roster discontinuity as well. But, if I can get the data easily, I might replace WS/48 with RAPTOR.

Just my opinion, but if you have a chart with several advanced stats, then it would make sense to sort by the advanced stat that most correlates with the eye test/logic. So if EPM ranks Herbert Jones at 2 and Jock Londale at 3, I would avoid sorting by that stat, and would try using one of the other ones that produce a more sensible ranking. Maybe VORP?

Back in the day, John Hollinger designed PER so it would pass the "eye test" or did we call it the "smell test", but at the time people didn't trust these advanced stats and dismissed one out of hand that didn't somewhat agree with their preconceived notions about who is good.

But advanced stats don't really exist to reinforce the eye test, in the case of EPM, etc, its trying to predict plus-minus accounting for teammates. If you want to check on these stats, you should go to the source and look at a players +/-, on/off data, lineup data, etc. Then perhaps you can make a case that Herbert Jones/whoever isn't contributing to winning so much as being taken along for the ride.

Or we could just admit a 23 and a 26 year old rookie have some advantages over their younger peers.

[JonFromVA]

Read on Twitter

?s=21

I can't help but point out that Isaac Okoro is a year younger than the Bull's rookie Ayo, and nearly 4 years younger than Thybulle.

[JonFromVA]

Read on Twitter

Hartenstein has some good stats for a 7 footer getting ready to hit UFA.

Hopefully he can find a team that will give him minutes. That is supposedly why the Cavs gave up his Bird rights, not just so he could join the Clippers and play even less mpg as a backup.

[Kurtz]

I'll look into it, but I don't know if it's really necessary. RAPTOR is a very good advanced stat but we're already using EPM as the main metric, which is very similar to RAPTOR (based on RAPM and incorporating player tracking data for example), and testing indicates EPM is a bit more accurate. RAPTOR also has some issues with smaller minutes sample size and doesn't handle roster discontinuity as well. But, if I can get the data easily, I might replace WS/48 with RAPTOR.

Just my opinion, but if you have a chart with several advanced stats, then it would make sense to sort by the advanced stat that most correlates with the eye test/logic. So if EPM ranks Herbert Jones at 2 and Jock Londale at 3, I would avoid sorting by that stat, and would try using one of the other ones that produce a more sensible ranking. Maybe VORP?

Back in the day, John Hollinger designed PER so it would pass the "eye test" or did we call it the "smell test", but at the time people didn't trust these advanced stats and dismissed one out of hand that didn't somewhat agree with their preconceived notions about who is good.

But advanced stats don't really exist to reinforce the eye test, in the case of EPM, etc, its trying to predict plus-minus accounting for teammates. If you want to check on these stats, you should go to the source and look at a players +/-, on/off data, lineup data, etc. Then perhaps you can make a case that Herbert Jones/whoever isn't contributing to winning so much as being taken along for the ride.

Or we could just admit a 23 and a 26 year old rookie have some advantages over their younger peers.

Well let's first start by tackling the low-hanging fruit and admitting that we shouldn't be pro-rating the results of a rookie playing 10 minutes/game against rookies playing 30 minutes. Let's call it the Boban Marjanovic rule. (this isn't a failing of the formula btw, we just need to set minutes thresholds if we're going to rank properly).

Do you have the EPM formula handy? I can't find it to scrutinize it.

[HiRez]

Just my opinion, but if you have a chart with several advanced stats, then it would make sense to sort by the advanced stat that most correlates with the eye test/logic. So if EPM ranks Herbert Jones at 2 and Jock Londale at 3, I would avoid sorting by that stat, and would try using one of the other ones that produce a more sensible ranking. Maybe VORP?

Back in the day, John Hollinger designed PER so it would pass the "eye test" or did we call it the "smell test", but at the time people didn't trust these advanced stats and dismissed one out of hand that didn't somewhat agree with their preconceived notions about who is good.

But advanced stats don't really exist to reinforce the eye test, in the case of EPM, etc, its trying to predict plus-minus accounting for teammates. If you want to check on these stats, you should go to the source and look at a players +/-, on/off data, lineup data, etc. Then perhaps you can make a case that Herbert Jones/whoever isn't contributing to winning so much as being taken along for the ride.

Or we could just admit a 23 and a 26 year old rookie have some advantages over their younger peers.

Well let's first start by tackling the low-hanging fruit and admitting that we shouldn't be pro-rating the results of a rookie playing 10 minutes/game against rookies playing 30 minutes. Let's call it the Boban Marjanovic rule. (this isn't a failing of the formula btw, we just need to set minutes thresholds if we're going to rank properly).

Do you have the EPM formula handy? I can't find it to scrutinize it.

I’m already filtering out < 10 minutes and about half the max for games played as well. As the season goes on I may increase the filter to 12-15 minutes, but I like to see some of the under the radar guys. The minutes are listed there so you can just skip over those guys if you don’t think they’re worthy. You might think someone needs to be playing 30 minutes a game to be counted but someone else might think 10 is a reasonable cutoff…there’s no setting that works for everyone so I lean towards show more instead of show less.

[zshawn10]

Only positive I can think for having Green is that if we had Mobley or Cade, we probably don't have a chance at the top 5 again next year

[basketballRob]

Only positive I can think for having Green is that if we had Mobley or Cade, we probably don't have a chance at the top 5 again next year

Green is a year younger than those guys.

Sent from my SM-G950U using RealGM Forums mobile app

[bkseven]

Only positive I can think for having Green is that if we had Mobley or Cade, we probably don't have a chance at the top 5 again next year

Green is a year younger than those guys.

Sent from my SM-G950U using RealGM Forums mobile app

You realize he turns 20 in like 10 days right...? lol

[basketballRob]

Only positive I can think for having Green is that if we had Mobley or Cade, we probably don't have a chance at the top 5 again next year

Green is a year younger than those guys.

Sent from my SM-G950U using RealGM Forums mobile app

You realize he turns 20 in like 10 days right...? lol

It's considered his 19-year-old season because he was born after February 4th. This is the 20-year-old season for Cade, Mobley, Barnes, and Suggs.

Green could've waited for one year and came out when he was 20 like the others in the top 5.

Sent from my SM-G950U using RealGM Forums mobile app

[reanimator]

Green is a year younger than those guys.

Sent from my SM-G950U using RealGM Forums mobile app

You realize he turns 20 in like 10 days right...? lol

It's considered his 19-year-old season because he was born after February 4th. This is the 20-year-old season for Cade, Mobley, Barnes, and Suggs.

Green could've waited for one year and came out when he was 20 like the others in the top 5.

Sent from my SM-G950U using RealGM Forums mobile app

Cade and Green are only 4 months apart. Hardly a year and questionable if it at all factors into their development curve.

[srhcan]

Rookie tracker through Jan. 25.

you need to include RAPTOR in your list

I'll look into it, but I don't know if it's really necessary. RAPTOR is a very good advanced stat but we're already using EPM as the main metric, which is very similar to RAPTOR (based on RAPM and incorporating player tracking data for example), and testing indicates EPM is a bit more accurate. RAPTOR also has some issues with smaller minutes sample size and doesn't handle roster discontinuity as well. But, if I can get the data easily, I might replace WS/48 with RAPTOR.

what about PER?

[srhcan]

If you look by PER with at least 20 MPG then these are the top 5 rookies:

1. Evan Mobley: 34.4 MPG, 16.08 PER

2. Scottie Barnes: 35.5 MPG, 15.65 PER

3. Franz Wagner: 32.4 MPG, 14.77 PER

4. Chris Duarte: 28.9 MPG, 13.21 PER

5. Josh Giddey: 30.8 MPG, 13.09 PER

* Omar Yurtseven has highest PER but he only play 14.8 MPG so he did not make the cut
* Alperen Sengun has 2nd highest PER but he only play 18.2 MPG so he did not make the cut

[basketballRob]

Per is mostly a big guy stat.

Sent from my SM-G950U using RealGM Forums mobile app

[bstein14]

Cade missed training camp and the Preseason due to an ankle injury. If you remove his first few games in the NBA his stats shoot up quite a bit overall. If you recall, he started out the season shooting like 20% from the floor and 10% from 3 point land those first 4 or 5 games on a lot of attempts.

In his 22 games in Dec/Jan he's averaging 17.9 PPG on about 44% FG%, 39% 3pt%, with close to 5 RPG and 6 APG. I'd expect his numbers in Feb/March/April to build on those with close to 20 PPG 45% from the floor 40% from deep and 5+ RPG and 6+ APG as he finishes out his rookie season strong.

Cade does have the advantage of being a #1 usage guy on the team but that also does create more focus on him so he's certainly still learning on the fly and improving against NBA double teams.

It will certainly be interesting to see who they give NBA rookie of the month to in the East for January.

[JonFromVA]

Per is mostly a big guy stat.

BPM has supplanted PER as the preferred pure box score stat. HiRez had added it at my request, but it got pushed out by some of the newer stats.

Mobley still leads all rookies by a lot with his +1.3 BPM (Sengun is second at 0.2).

Cade had a nice January, but ihis BPM is just -1.5 at the moment.

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

[srhcan]

Per is mostly a big guy stat.

Sent from my SM-G950U using RealGM Forums mobile app

I dont think Hollinger invented it with the purpose to put big guys on top. PER is usually the standard nowadays

[JackTalkThai]

Interesting fact: there are only 13 players in the NBA that have been double teamed more than Cade Cunningham.