Posted by shoe3 on 1/20/2022 7:45:00 PM (view original):
Posted by cubcub113 on 1/20/2022 4:41:00 PM (view original):
Posted by shoe3 on 1/20/2022 3:49:00 PM (view original):
Posted by Basketts on 1/20/2022 2:32:00 PM (view original):
Posted by shoe3 on 1/19/2022 11:07:00 PM (view original):
Posted by Basketts on 1/19/2022 6:36:00 PM (view original):
Pretty good for a team clearly not in the top 5.
I will say that they are 19-2 with Brooks starting at center though ;)
Yeah, you know, the guy with fewer rebounds per 40, more turnovers per 40, and a whopping .535 to .525 Fg% advantage. He was the difference. :)
I was half joking since I agreed with you earlier in the conversation that the whole Brooks/Paquet debate was minor in comparison to bad rng, and that Paquet was more than capable of starting for a contender. However, the breakdown was pretty much what I expected if you were trying to eek out some offensive efficiency:
10.8 - Brooks (Reb per 40)
11.9 - Paquet (Reb per 40)
15.4 - Brooks (Pts per 40)
10.4 - Paquet (Pts per 40)
2.2 - Brooks (TOs per 40)
1.7 - Paquet (TOs per 40)
1.2 - Brooks (Blocks per 40)
1.7 - Paquet (Blocks per 40)
I personally would take that extra 5pts per 40 on better shooting for the cost of 1 reb per 40. The TOs seem negligible considering Brooks had a larger offensive burden. Predictably, he did take a hit in blocks, and that might be enough to push Paquet higher than Brooks on a +/- comparison, but it was the sacrifice he chose to make and I think it was the correct call.
The difference in raw points is mostly a function of distribution (possibly some IQ at play here, in getting off a few more shots over the course of a season I suppose), and we have to consider that the extra points came from attempts that would have otherwise come primarily from better scorers. That’s even more true at the end of the year when promises were off the table.
Anyway, the difference in FG% was pretty marginal, and when you factor in those possessions from TOs and REBs (not to mention the apparent and hidden BLK factors) the switch was a wash, at best. Brooks is an excellent player. So is Pacquet. There’s no reason to think switching the starter would have any impact, and the actual numbers bear that out pretty well.
What did have an impact, I think, was getting Herrod regular minutes. I can only speculate, but considering where he is on the big board now (#20), it looks like perhaps the idea was to bury him to see if he could be kept/dropped to the projected 2nd round, and “on the fence” status, which was scrapped after the rough start. Understandable attempt, though I probably would have made a different call personally.
If you want me to break out statistical analysis:
Brooks 25 minutes, Paquet 19 minutes
Brooks 15.4 Pts per 40
Paquet 10.4 Pts per 40
Brooks 57.6 efg%
Paquet 54.3 efg%
The team is 57.7 efg% by the way. So not only is Brooks scoring 50% more but he's also far more efficient (basically a point per 33 possessions). I have some modeling that suggests that if you want Pacquet to get up to team efficiency, you need to drop him down to about 6-7 points per 40 (which is what we call a non-scorer).
So at team efficiency, Brooks will be scoring about 2.4x as much as Pacquet, which is just a massive difference. And Brooks is actually playing a few more minutes than Paquet in this sample (meaning he'd be a bit more tired if their STA were the same), so round it up to 2.5x.
So to summarize:
Brooks can score 2.5x more points at team efficiency on a team that *badly* needs scoring
Brooks draws 29% more fouls per FGA (from STA and IQ advantage)
Brooks can rebound roughly as well (+2/3 letter grade of IQ, +1 REB, -6 ATH basically cancels)
Paquet has a tiny (if even existent) defensive advantage at the 4 and a more discernable one at the 5 (+35 SB, +6 ATH, -4 DE, -2/3 letter grade IQ)
Brooks has a massive STA advantage (90 v 75). This is one of the reasons why he is so much of a better scorer. Switch the STA and they are pretty comparable IMO.
This is not a glaring difference, but starting Brooks over Pacquet and playing him 15 minutes is probably worth about 1 point extra per game IMO. They just aren't that close. My algorithm has 12 extra minutes of Brooks over Pacuqet worth to be 0.1 points excluding STA, so basically flipping the start (even not considering STA) is worth 15 points per season of starting Brooks over him. But 15 STA is just massive, especially since Rowle loves to run uptempo.
I have some questions.
**First, what formula are using? My understanding is that for players who don’t shoot 3s, efg% is just fg%. The reason to use the figure is to figure out at what rate a player who only takes 2point shots needs to hit to outperform a 3-point shooter, so you only break out the “e” when comparing post to perimeter. But since you’re coming up with figures that are not just these two players’ fg%, curious what you’re actually doing.
**Second, why are you using efg% as though it means efficiency? It’s a part of the picture, but by no means tells the whole story.
**Third, can you explain what “2.5x more points at team efficiency” means? What kind of claim are you making? One point per 33 possessions gets closer to what you mean in actual terms, I suspect (more on that later), so what’s the effective difference between your terminology here, and why are you using it? It *sounds* like you’re using a statistical concept to frame this a certain way, and I’d like you to explain that, especially given how different your use of efg% apparently is.
**Fourth, why are you focusing so hard on the points, and looking past the rebounding, turnover, and block discrepancies, which all go in Pacquet’s favor? In other words, in terms of whole picture efficiency, possessions matter too, right?
**Fifth, can you show your work on the one point per 33 possessions? It’s a dubious figure (because we’re talking whole picture efficiency). There is .016 difference in the fg% between these two players. You are essentially making the argument that Brooks on the floor for an extra 7-8 minutes per game is worth ~2 points (roughly 60 possessions per game), right?
**Finally sixth, and probably most important, did you miss the part about distribution? It is feeling like you’re getting caught up in this minute statistical analysis, and compartmentalizing something that really should be seen relative to whole team. Brooks did score a few more points, whether by higher distribution alone, or perhaps added *a little* by an IQ advantage. But at no point in the season was he the best scoring option on the team. When the discussion started, it was about the third option, and at the end of the season, unless there was a favorable matchup at the 5, there were normally 4 better scoring options (for a flex team) starting with him. So that higher distribution came at a cost. It meant Leonard and Han - who were both better scorers by the end of the year, by attribute - got fewer looks than probably was optimal. Distribution is zero sum, as you know, so it’s not just about Brooks and Pacquet here in a vacuum, team efficiency is about the whole team. To that end, and to OR’s point, wouldn’t it be great to get that +/- figure added?
**Anyway, the only other thing I’ll say is that stamina would definitely be important if this was a press team. On my FB/P teams, I’d be starting Brooks, for sure. On my teams that are flex/combo, Pacquet is an easy choice (the block makes him a clear winner, rather than the wash they are at flex/man). Because I suspect there is a GI/GO issue related to #5, I think you’re vastly overestimating how much value Brooks is getting with that stamina, when his backup is such high quality. If it’s a scrub, that stamina has a ton of value in any set, of course. But Pacquet is an absolutely elite defensive post player and rebounder, who can *absolutely* score enough to be an effective 3rd option in flex when the top players are doubled. Even more so at the end of the year when he wouldn’t have even had to do that much in a lineup with Leonard and Herrod.
1. You need to take into account free throws. Brooks is just so much fresher than everyone else, that when combined with A+ IQs he's actually a foul drawing machine.
2. no idea what else you want to include except turnovers. Brooks and Paquet have about identical BH/PA (12 BH in a big is roughly meaningless) so this isn't an issue
3. According to my decent calculations, if both Brooks and Paquet score at an efficiency equal to the rest of the team (which is roughly optimal) Brooks will end up scoring 2.5x as much as Paquet. That number isn't exact, but it's going to be somewhere between 2 and 3, very likely extremely close to 2.5. How do you get that? I described in more detail above but Brooks is scoring MORE per minute than Paquet at a significantly more efficient rate. Since usage and efficiency are inversely correlated in HD, if we were to artificially raise Paquet's efficiency we would have to drop his usage even further. And according to my house model raising someone's efficiency by 3 efg% mean you need to decrease their usage by about 36%, which changes Pacquet's usage from 10.4 points per 40 to 6.7, hence 6-7 points.
TLDR for people who don't want to weed through math:
If you increase a player's usage, you should expect his efficiency to decrease. When a player (Brooks) is not only scoring at a 50% higher rate, but also with 3 efg% more efficiency, he is a significantly better offensive option and is going to end up scoring at 2-3x the rate after balancing for efficiency. According to this measure of efficiency, Brooks was properly used and Paquet was moderately overused on the offensive end this year.
4.
Efg% and points are the only stats that should be used for analysis in this game, just because scoring is so complicated. We have really good rating formulas for rebounding (about 60% reb 30% ath 10% iq), defense (depends on set and position, but the0nlyis and I went through and made some a few years ago--defense is also really complicated, but sadly we can't track stats to see how well players perform defensively), and turnover creation and prevention. I don't even like using Efg% for analysis like this (I do this very rarely) because truly understanding ratings is always better than using noisy stats. To a trained Flex/Man eye, Brooks vs. Paquet offensively wasn't a question at high D1 from the start of the year, it's pretty clear that Brooks is going to be a viable scorer and Paquet really isn't. But it's nice when the stats completely backup your guesses based on the ratings.
5. 1 point per 33 possessions stats comes from the 3 efg% difference between Paquet and Brooks. But this isn't even considering that Paquet was scoring at 50% more usage than Brooks in 20% more minutes. If you were to use Paquet and Brooks equally, it would be much closer to 1 point per 16 possessions. This is only possessions where they finish it with a shot or are fouled though, which is probably only 10 times per game or so. So going from 10 Paquet possessions to 10 Brooks possessions probably nets you about .5 points in offensive efficiency per game in a vacuum. Brooks will also draw a ****-ton more fouls though, which probably more than cancels out his slight to moderate defensive disadvantage.\
6. While trying to be respectful, you're thinking about this completely wrong. Remember that I established this baseline for 57.7% team efg%. Brooks is right on that, so no, you don't want to be taking any shots from Brooks and giving them to the team. If Brooks' efg% was significantly below that 57.7% baseline, then yes he would be stealing from the team.
The most important relationship here is the inverse causation between efficiency and usage, and the fact that a team is most efficient when each player's efficiency is the same.
(paragraph below is only for people who want my less simplified marginal analysis explanation)
Yes, I realize that I said above is completely false. You actually want each player's efficiency on his marginal shot to be the same (that extra shot he's "taking" from other players), so players who get more easy buckets and have a more elastic efficiency to usage relationship (their efficiency varies more given that same variance in usage) should have a higher efg% to maintain equal efficiency. In the real world, this is why you see big men like Capela/Adams have super high efg% and not be used more, they get tons of easy buckets to pad their stats. I actually don't think this type of marginal analysis is relevant in Hoops Dynasty. LP players have to work for their buckets almost as much as shooters. And even if there's a small difference, I take solace in the fact that shooters draw fewer fouls and are more useful at comparable marginal efficiency. I just estimate these very subtle facts effectively weigh each other out, unless a player draws a *shitton* of fouls. Then I'll push their distro a bit higher than the teams.
(done with mathy stuff)
6. Just like 97% of the community (and me until very recently) you're underestimating the value of stamina in sets like mo/man. It's not just that guys are way less useful when they are "tired" or "getting tired" but it is also that a guy with 90 stamina will be way more fresh, on average, during his four-minute stint to start the game when compared to the 75 stamina guy. Fresh v fairly fresh is a huge difference too! It's not just that Brooks can play "more minutes" than Pacquet, it's that he can play 2-3 more minutes but do it at a significantly higher energy level! There is sometimes a lot of time between when a player needs to be subbed out and when the next dead ball is!
6b. I just gave you the numbers to show Pacquet is not a third option on any top 5 D1 team. He was literally the 6th option on a fringe Top 5 team and was moderately overused to get there (-3% efg is kind of a lot). But let's discuss it from a ratings point of view instead because I think the majority of the community likes that more than analyzing the connection between efg% and usage%.
Warning: Extremely Controversial Stuff Incoming
I think the question we should be asking at this point is "Why is Pacquet such a bad scorer, cubcub, he has 99 ath and 69 lp?"
Let's find a low LP 4 actually was a successful 3rd scoring option on a fringe Top 5 Team...
Here:
Sean Talbot
Talbot Stats:
17.3 Points per 40
56.5 efg%
Pacquet Stats:
10.4 Points per 40
54.3 efg%
Note: Although Talbot played an easier schedule (SOS=13 vs Pacquet's SOS=3), Pacquet has a better PA supporting cast. We'll say it evens out, although I think Pacquet was in a far better situation.
Ratings Comp Between Them:
Pacquet -1 ATH -33 SPD +6 LP -23 PER +11 BH -4 STA -2/3 IQ
Note how the ATH/LP is very similar (Pacquet's is even better) and Talbot doesn't exactly have Brooks level stamina.
Why does Talbot blow him out of the water (70% higher volume and 2.2 efg%)? It's because the PER cancels out the LP and BH, and all that's left is a massive SPD and IQ difference. The only real bad blood I've ever had projected at me by the HD community was a user who called me a bit of a "cocky know-it-all" for preaching my "SPD is actually a really useful rating for scoring bigs" in the Discord and I've also had established users like Gil/0nly/Sportsbulls etc. tell me I'm misguided for thinking that speed is so important for bigs. But Talbot draws fouls at a rate wayyyy higher than Pacquet and ends up being an actual viable scorer, so there you have it.
Summary:
Q:
Why is Pacquet is not a viable high D1 scorer even though he has an ATH/LP combo of 100/70 in Flex/Man?
A:
He has no PER to increase his FG%, and he doesn't draw fouls because of his really bad spd/STA combo and his SO IQs.
Q:
How do you prove this if I don't trust cubcub's rating analysis?
A:
You can use the fact that there is an inverse relationship between volume and efficiency in HD to prove that if someone scores at a significantly LOWER volume with moderately LOWER efficiency, they are a much worse scorer. Pacquet had low volume this year (10.4 points per 40) and still was less efficient than the rest of his team by 3 efg%.
I think this is one of my better posts ever, by the way, for people who really want to squeeze that last point of strength from their team through player distribution.
_________________________
A common counterargument to everything I said above goes like this: "The players who draw the fouls don't actually take the free throws." (similar to "The players that get the "rebound" stat don't directly contribute to getting the rebound"). I don't totally buy that
for shooting fouls. From support ticket answers though, it sounds like they simulate a player taking a shot and the results are: make, miss, or fouled, meaning it is directly the players superior ratings that cause them to get FTs.