League File | Standings | Schedule | League Leaders | Free Agents | Coaches | CSLO | D-League Standings | D-League Leaders | Player Potential Database

PB22 Default Ratings

Post Reply
User avatar
KW
CSL Champ 2017
Posts: 10659
Joined: Fri Jul 11, 2014 5:18 am
Location: CO
Contact:

PB22 Default Ratings

Post by KW »

I compiled a google sheet with the default rosters that come with PB22 to get an idea if there are differences compared to what we're used to. You'll see the percentiles of all rostered NBA players from that file. There are a few outliers in the default file (Gary Payton II and a few others have some wack ratings that stand out), and those outliers are very obvious (i.e. the 100 RARate for a PG, when the 90th percentile is just 11).

To create a few benchmarks:


The 90th percentile means you're roughly top 8 or 9 at that skill at your position. Probably the baseline for "Elite".

60th or 70th percentile is baseline for what would be considered ok for a starter.

50th is the median. There's a lot of crap at the ends of benches, so 50% or below should be considered an actual weakness for a rotation player.

I'd consider < 30th percentile as a real problem (for some ratings).

https://docs.google.com/spreadsheets/d/ ... sp=sharing

PG

Image

SG

Image

SF

Image

PF

Image

C

Image

--

I have a few notes regarding the league transition and what ratings clearly look different than what we have. I should point out that the default ratings are also far from perfect, and I am not suggesting we should blindly follow every difference we see to match the defaults.

I unfortunately cannot do this for our league file, because I don't have that data. But having created roughly 40% of the last 4 draft classes, I can comfortably spot some differences.

1) FG_RA

FG_RA is much much lower, across the board at essentially every position, as expected. This checks out, as interior ratings were purposefully nerfed during the 19 transition because everybody shot like 75%+ from the restricted area, and some over 90%. From my minimal testing in PB22, it is still on the high side for most players, but within reason, but I'll let the admins chime in on what they've seen. FG_ITP ratings seem pretty close between the CSL and the default roster.

2) Dunk Rate / RA Rate

This is where I think things are entirely different from what we're used to. The staggering thing here is that almost no PGs and a few SGs dunk or get to the restricted area at all. I thought this would be serious issue, but guards in my default roster tests are shooting just fine. More on this in part (4)

SFs and PFs tend to have a bit lower DunkRate but higher RARate than in the CSL (much higher for PFs).

We probably have DunkRate in the ballpark for C but our RARates are way way way down if 48 is just the 10th percentile and the median is 69, same goes for the high end of PFs. This makes a lot of sense for those of us with PFs and Cs who shot a ridiculous amount of 3-10 ft shots... the game is likely expecting a way higher RARate to send your big guy to the rim against the shot blockers, as a result it was probably way easier for defenses to seal off the drive and force your big guy into a pull up jumper, a post up, or a pass?

3) Shot blocking / Defense

The default rosters have really low block ratings, much lower than the CSL's nerfed numbers. Only 27 players in my last test recorded at more than 1bpg ... which is actually right around the NBA number in 2021-22 of 24. This is probably for the simple reason that players are taking way more threes, so nowhere as many shots for a rim protector to contest. Also our CSL players are all wet noodles in the paint, so that's why people were averaging 4+ blocks in our preseason. I don't believe the CSL should turn into the NBA and stop valuing defense, so I don't condone dropping blocks to this degree, but it's a pretty clear contributor to our current issue with interior offense.

I'm pretty confident the CSL has better DEF ratings at every position than the default roster. Another contributor to offensive issues. Also noting pretty low discipline, which would certainly be an issue in the CSL's drive-heavy game, but not in the default roster.

4) Post action/floor range

Almost no PGs post up (excluding outliers, every PG/SG is below 5 PostUp, vast majority below 3) and a floor range of more than 15 POST is pretty high. My theory is that the CSL guards are getting owned in the post by bigger defenders, and that drags them down. ATB is where most everything start for guards and wings, which makes basketball sense. Their scores in close appear to be mostly on drives from the perimeter.

5) Spacing and well rounded actions/ranges

Look at any guard, and you'll see a pretty good mix of DriveShot/DriveKick/C&S/PullUp actions, and floor range heavily skewed to ATB, with most of the rest being mid range and maybe some corner. Bigger wings/forwards/centers you can add in the post action and range. The bigs that literally would never attempt a 3 (Gobert, Capela, etc) are basically the only ones that don't have much in ATB/COR or PullUp/C&S action. Those guys also rarely ever drive, and almost exclusively post up, some of them will catch and shoot or have some mid-range.

All of them appear to shoot ok, except those who have teammates that don't space the floor- biggest example is Embiid, who shot a disastrous 38% in my last test, as Simmons (who has big man tendencies playing PG, appears to be an awful combo despite great ratings), Maxey (only 2 3PA/game, lots of driving), and Harris (a lot of short and mid range) didn't give him adequate spacing, and Embiid only shot 11% of his shots at the rim. Trying to operate in a crowded paint clearly does not appear to work here.

User avatar
KW
CSL Champ 2017
Posts: 10659
Joined: Fri Jul 11, 2014 5:18 am
Location: CO
Contact:

Re: PB22 Default Ratings

Post by KW »

My conclusion: There's work to be done.

I believe some normalization of ratings is going to be needed in our league to keep up with the new engine, particularly a boost to FG_RA, and potentially re-rating Dunk Rate and RA_Rate to match. Floor ranges, particularly ATB for guards and wings have to be adjusted across the board, and most of the league will have plenty of ball actions to work on to get players to be well rounded, rather than just one trick ponies spamming post ups, drives, or threes.

User avatar
Dennis
CSL Champ 2028
Posts: 10356
Joined: Fri Jul 11, 2014 10:19 am
Location: Düsseldorf
Contact:

Re: PB22 Default Ratings

Post by Dennis »

Interesting findings! Yeah only admins can exactly know if we are off. But very detailed and smart analysis. I can’t say for sure what is going wrong as I have no idea about shooting inside ratings despite some scouts with deviation, but what I expected to be in place was at least off to default file.

Btw, given shooting ratings had to be changed from dds3 to dds19 and now again needs a big overhaul it really would be the best to share what the true shooting ratings look like because they won’t be in line with what we scouted at some point anymore. Especially without deviation and not as letter grades because letter grades won’t be helpful at all to evaluate draftees or players in general.
Boston Celtics Hall of Fame
PG G. Dragic [7] - Jersey Retired <> 2013 - 2015
PG R. Westbrook [0] - Jersey Retired <> 2016 - 2019

Fox - Johnson - Bridges - Hlinason - Whiteside
CSL Champions Image 2028

User avatar
Silogical
CSL Champ 2026
Posts: 3758
Joined: Thu Aug 06, 2015 10:24 am
Contact:

Re: PB22 Default Ratings

Post by Silogical »

Its good info but the default league looks broken. The 17th best offensive team shot .503 efg% Last yr in DDS19 the best offensive team shot .500 efg%. I dont think we should try to mimic this league's ratings.

User avatar
bt
CSL Champ 2020, 2023 & 2027
Posts: 5793
Joined: Fri Jul 11, 2014 6:12 am
Contact:

Re: PB22 Default Ratings

Post by bt »

That could also mean CSL DDS19 is broken and more likely so given how preference work and tailoring players who probably shouldn't be altered (guards with 100 post) can stuff things up.

It's a good insight but doesn't really matter anyway as it all depends on what admin want to accomplish here. If it was a pure conversion, you would make everything as similar as you can and be done with it but this seems more like a conversion with fixes so depends on what they plan to fix. Inside play for guards, etc, but does that mean fixing bigs then? It also means a ton, maybe all players are going to be different. How different, who knows but fixing different areas could be a bit of work to get it right.

That's probably what they're working on and we'll just live with it. In all the conversions I've done or taken part of, after a season, everyone pretty much forgets and the new system is the norm going forward.
sacramento kings

tremont waters | skylar mays | nigel johnson
matisse thybulle | justin james
robert franks | darius bazely | john butler
donta hall | xavier tillman | patrick williams
jon collins | jonathan isaac | norvel pelle | mfiondu kabengele


CSL CHAMPION Image 2020, 2023, 2027

User avatar
KW
CSL Champ 2017
Posts: 10659
Joined: Fri Jul 11, 2014 5:18 am
Location: CO
Contact:

Re: PB22 Default Ratings

Post by KW »

Silogical wrote: Mon Aug 01, 2022 5:47 am Its good info but the default league looks broken. The 17th best offensive team shot .503 efg% Last yr in DDS19 the best offensive team shot .500 efg%. I dont think we should try to mimic this league's ratings.
Offense in this league is far below the NBA. I'm not saying we should do everything to mimic the NBA either, but 22 does not appear broken. The real life NBA's eFG% by team ranges from 49.4% to 55.6% with league average being 53.2/%. Only tanking teams shot below 51. We've been stuck in the 90s and early 2000s with PB3 and PB19, it's a golden opportunity to move towards a modernized league.

User avatar
Silogical
CSL Champ 2026
Posts: 3758
Joined: Thu Aug 06, 2015 10:24 am
Contact:

Re: PB22 Default Ratings

Post by Silogical »

KW wrote: Mon Aug 01, 2022 12:25 pm
Silogical wrote: Mon Aug 01, 2022 5:47 am Its good info but the default league looks broken. The 17th best offensive team shot .503 efg% Last yr in DDS19 the best offensive team shot .500 efg%. I dont think we should try to mimic this league's ratings.
Offense in this league is far below the NBA. I'm not saying we should do everything to mimic the NBA either, but 22 does not appear broken. The real life NBA's eFG% by team ranges from 49.4% to 55.6% with league average being 53.2/%. Only tanking teams shot below 51. We've been stuck in the 90s and early 2000s with PB3 and PB19, it's a golden opportunity to move towards a modernized league.
I would like the CSL to mimic the NBA. Its just with DDS offense can overpower defense. If the admins can figure out how to mimic the NBA while keeping everything balanced great. I dont want another 4 seasons where 1 thing stands out as OP over the rest.

Also where is the extra % coming from in the NBA? Im sure its all from 3 point shooting.

An idea for testing would be a preseason league with a good prize at the end so people try to win. The 15th pick in the next draft or something people would try for.

User avatar
Silogical
CSL Champ 2026
Posts: 3758
Joined: Thu Aug 06, 2015 10:24 am
Contact:

Re: PB22 Default Ratings

Post by Silogical »

If anyone wants to break this down with our preseason numbers and/or DDS19 numbers. https://www.basketball-reference.com/le ... oting.html

I think buffing both blks and FG_RA just makes those two more powerful while canceling each other out.

User avatar
KW
CSL Champ 2017
Posts: 10659
Joined: Fri Jul 11, 2014 5:18 am
Location: CO
Contact:

Re: PB22 Default Ratings

Post by KW »

Silogical wrote: Mon Aug 01, 2022 12:54 pm I would like the CSL to mimic the NBA. Its just with DDS offense can overpower defense.
so just like the NBA?

User avatar
Silogical
CSL Champ 2026
Posts: 3758
Joined: Thu Aug 06, 2015 10:24 am
Contact:

Re: PB22 Default Ratings

Post by Silogical »

KW wrote: Mon Aug 01, 2022 1:09 pm
Silogical wrote: Mon Aug 01, 2022 12:54 pm I would like the CSL to mimic the NBA. Its just with DDS offense can overpower defense.
so just like the NBA?
balance first, mimic 2nd. Both great.

User avatar
emplep7
Commissioner & CSL Champ 2018
Posts: 6688
Joined: Fri Jul 11, 2014 5:06 pm
Contact:

Re: PB22 Default Ratings

Post by emplep7 »

In theory we try to mimic the NBA as close as possible, but with how our ratings have been historically catered around the engine that always proves to be really difficult during the transition. Add on that we typically see different results from our internal testing once 30 humans start setting strats and tinkering and it leads to what we are seeing now.

Blocks are going to be restored to previous levels with some slight reductions at the top (the BLK rating is more powerful in this engine, but previous levels will be mostly inline with what we saw in the CSL in DDS19). There was no noticeable difference in shooting %s when testing both increased and decreased BLK ratings.

RARate is being tested at different levels as well as FG_RA, but there's a very fine balance so as to not make these too OP so we are being very careful with any adjustments here.

The ball and floor actions are going to be critical and we know that any testing we do wont find what 30 human GMs will find through a full 82 game season so there's concern with making drastic changes.
Image

User avatar
emplep7
Commissioner & CSL Champ 2018
Posts: 6688
Joined: Fri Jul 11, 2014 5:06 pm
Contact:

Re: PB22 Default Ratings

Post by emplep7 »

Another consideration is that the default roster files are made under the assumption that you cannot change floor locations.

By increasing the shooting ratings AND allowing floor locations to be changed however we want it results in an unplayable game and some crazy shooting percentages.

As Ryan said, I do agree that the most important thing at this point is balance and trying to duplicate how players should be behaving based on the previous engine (for the most part). This conversion wont have shooting ratings revamped, but its possible in the next conversion that we can do that with the stipulation that floor locations are going to be set by the admin team for each player and will not be allowed to be adjusted.
Image

User avatar
KW
CSL Champ 2017
Posts: 10659
Joined: Fri Jul 11, 2014 5:18 am
Location: CO
Contact:

Re: PB22 Default Ratings

Post by KW »

For those who were interest in this... Jon and Andre were able to provide me with the percentiles of CSL ratings for use of draft creation. Thanks a ton!

I also wanted to compare them with the default PB22 rosters to see where our league differs. I've added a tab on the Google Sheet linked in the first post of this thread. I left off the extreme outliers (0th percentile and 100th percentile, mostly crazy numbers from the default rosters) because they aren't useful.


Image

Image

Where CSL is lacking:

1) LocATB - across the board, this league has very very low floor range for ATB. Not surprising considering the engine we're coming from. Spacing does seem to matter for the first time in the DDS universe and it's good to see. That does seem to mean pushing players you wouldn't necessary want shooting a significant volume of threes outside the paint. Those 70 POST guards should become a thing of the past and don't appear well-suited for this engine, they're going to have to slash from outside the paint.

2) FG_RA, across the board, but particularly worse on the top half of guards and wings (aka all the serviceable ones). Well, this is one reason why your 70 POST guards aren't working. When they get to the paint, they can't finish at the rim. Do we want them quite as powerful as the default rosters? Probably not, but a middle ground probably needs to be found.

3) RA_Rate, this one is mixed, as our guards are better than average, but the high ends of wings, and all PF/Cs are way below the defaults. Low RA_Rate means the players' drive, post move, etc is going to wind up more often as a FG_ITP attempt rather than a much more efficient FG_RA attempt. If your player is getting his drive stopped, post move not getting all the way to the rim, he is going to fire up a TON of FG_ITP shots, which will absolutely tank efficiency. Less skilled offensive players get burned particularly badly here. This one is not new, it's been a concern since PB19.

Where the CSL's ratings are higher than default:

1) LocPaint - any surprise here? The higher side of our guards and wings are still very much stuck in the paint game. A lot of teams still clearly have non-shooters at PF, evidenced by ridiculously high post numbers at PF basically across the board.

2) Guard/Wing scoring+passing - this league may be big man focused versus guard focused, but the data wouldn't lead you to believe that it has anything to do with the ratings. The median point guard in the CSL is +8 in SCR and +17 in passing than his PB22 roster counterpart, but also has +10 PASS ball action. The average SG in the CSL is +10 SCR +14 PAS with an identical PASS ball action to his PB22 roster counterpart. A wing in this league is more skilled offensively than the default, by a healthy margin. You can't say that about a CSL big man to the same degree.

3) Defense at all positions except PG - not a surprise at all. SG are a bit higher DEF, way higher STL, SFs way higher on steals, PFs and Cs in this league are much better shot blockers, and discipline is high across the board. High STL on guards and wings has always stopped drives at a great clip, high BLK will close down the paint. On a similar note, a lot of Cs playing PF and PFs playing SF in this league has skyrocketed already high rebounding numbers in the > 50th percentile.

User avatar
hardenwithnod
Chairman of the Board
Posts: 7607
Joined: Thu Aug 11, 2016 3:26 pm
Contact:

Re: PB22 Default Ratings

Post by hardenwithnod »

Good stuff, that explains a lot on why so many guys just shoot at the 3-10 ft range and suck at it.
Alvarado/Mitchell/Tatum/Bender/Birutis

User avatar
Marcos_Beck
CSL Champ 2019
Posts: 7731
Joined: Fri Jul 11, 2014 5:15 am
Location: Brazil
Contact:

Re: PB22 Default Ratings

Post by Marcos_Beck »

So far, what I trully understand from this engine is:

it’s impossible to drive from the perimeter. Almost every drive in our league gets stopped (maybe because players are so thin their STP% is too high), could have something to do with higher DEF as well, but the thing is, if your players isn’t mainly on the post then he shouldn’t be driving at all because he’ll hardly reach the rim.
Is a drive the best way to score? Idk.
But surely getting all your drives stopped and shooting tons of 3-10 FGs isn’t as well.

So, players like Lonzo and Louzada who aren’t exactly snipers and have high handling, they should be driving and therefore I have them at 60% of their time inside the arc, so driving becomes easier. If there was a way to make players reach the rim while being on the perimeter, I would have them mostly at ATB to initiate offense, however the engine is flawed or our ratings aren’t really suited to the game and the number of 3-10 shots is just too high.
18-19 Chicago Bulls: CSL Champions
#1 D.Rose #11 J.Holiday #21 J.Butler #42 A.Horford #13 J.Noah

29-30 Philadelphia 76ers: CSL Champions
#1 L.Ball #15 M.Beasley #23 B.Bowen #21 M.Wagner #51 K.Towns

User avatar
Marcos_Beck
CSL Champ 2019
Posts: 7731
Joined: Fri Jul 11, 2014 5:15 am
Location: Brazil
Contact:

Re: PB22 Default Ratings

Post by Marcos_Beck »

Maybe we need better shooting ratings from ITP and MID so players don’t need to get to the rim at all times to score efficiently. Maybe other changes will make perimeter players specially more ATB heavy but right now with those ratings it’s just impossible
18-19 Chicago Bulls: CSL Champions
#1 D.Rose #11 J.Holiday #21 J.Butler #42 A.Horford #13 J.Noah

29-30 Philadelphia 76ers: CSL Champions
#1 L.Ball #15 M.Beasley #23 B.Bowen #21 M.Wagner #51 K.Towns

User avatar
Dennis
CSL Champ 2028
Posts: 10356
Joined: Fri Jul 11, 2014 10:19 am
Location: Düsseldorf
Contact:

Re: PB22 Default Ratings

Post by Dennis »

Im not sure I really agree on that. There are a a few things that work imo to make your team more efficient. At least that’s how we turned the corner.

1) post still insane valuable and even worse than before as it determines the amount of shots taken even more crazily for PGs as their pass pref got nerfed to a way too high level in the 3 to 19 transition for this engine. So less post is just not possible.

2) c&s + Post best combination. Yeah our ratings might off/lower compared to default file, however, that is what works best so far. As Mercos said, driving from the perimeter just doesn’t work at all.

3) My offline testing showed for my team that „post focus“ improved overall FG% by 3.5 (!)% in average.

My conclusion is therefore way different despite that our guys are super inefficient in 3-10 area. Not sure if correct but that’s at least what I found in my tests.

Sucks a bit as dds22 is therefore even more unrealistic than 19 and you really have to limit yourself as you only can do what works with the engine and can’t build a team like you would want it, but we have to work with that I guess.

The value of the louzadas, smiths and foxes really got a hit but you can at least somehow fix them if you accept that this engine is trash and work with what works best.
Boston Celtics Hall of Fame
PG G. Dragic [7] - Jersey Retired <> 2013 - 2015
PG R. Westbrook [0] - Jersey Retired <> 2016 - 2019

Fox - Johnson - Bridges - Hlinason - Whiteside
CSL Champions Image 2028

User avatar
hardenwithnod
Chairman of the Board
Posts: 7607
Joined: Thu Aug 11, 2016 3:26 pm
Contact:

Re: PB22 Default Ratings

Post by hardenwithnod »

Yeah, that's one thing I notice is that guys don't really drive from ATB right now. Or at least not efficiently.
Alvarado/Mitchell/Tatum/Bender/Birutis

Post Reply

Who is online

Users browsing this forum: No registered users and 0 guests