Isiah did things on the biggest stage that Cade (and certainly the team around him) isn’t doing just yet, but there are signs that he can be that kind of player.
Sure, but Isiah was not the ball-dominant kind of player around whom an entire offense was structured. As much as he was a legit MVP-level player, that team still had a different model and it’s just odd in a pinch-me-I’m-dreaming way to have a guy like Cade. Pistons fans know the degree if intracacy required to build a contender without a guy like that, and how it can be so much easier with a guy like that – add chumps like Boobie Gibson to Lebron’s team or Antoine Walker to Wade’s team, and off you go. So it’s so incredibly exciting to have a guy like that.
1 Like
BPM has a very specific way that it handles this concept and I would imagine the box score prior aspects of these models all have similar adaptations. The gist is that there’s a “creator” estimate that penalizes non-creator shots significantly higher than creator shots on the exact idea you’re talking about. It’s much easier to be efficient when you take easier shots. Can read about it here:
So, you can definitely say they don’t do this correctly but you’re likely wrong for a given all-in-one that no such adjustment exists. It’s a well appreciated concept among nerds.
I will say too, the only college-level implementation that I’m aware of (Evan Miya’s) definitely looks to me like it penalizes usage far too much. Especially at the college level on these meh teams, there’s literally nobody else to generate a shot so the marginal shot created by a dude w 35% usage on these teams is way better than the alternative even tho that shot is extremely inefficient. These guys are eating the cost of having terrible teammates in their estimates and it doesn’t make sense to me.
It’s one reason why I really like looking for very high usage guys w good FT shooting in up transfers even if their 2FG% and 3FG% isn’t so hot. They probably had to force up a ton of trash shots that they simply won’t have to deal with once their teammates can create a little for themselves and actually finish when passed to.
Read through the first half so far. Interesting read. Yeah I guess my complaint is probably that they do a bad job quantifying that factor. For example, if I’m reading it right, they have a set of estimated coeffecients based on your role (which caps at some amount?) instead of scaling on a gradient based on the usage. Also some other things I disagree with there. Their position coeffecients makes some bonkers assumption. For example, valuing ORBs differently based on “position” seems nuts to me. There are some stats I could buy it for, but building your model based on all these positional assumptions… feels very flawed.
Doesn’t really pass the sniff test when it thinks Kelly Olynyk and Marvin Bagley are significantly better players than Cade.
The coefficients aren’t arbitrary, they’re the result of models fits to very large sample RAPM. So IMO that’s the way to think about how these results end up that way. E.g. I’m not surprised at all that creator OREBs are worth more bc they’re much more likely to be actually incremental OREBs that wouldn’t have occurred otherwise. Rebounds are often taken from your teammates, tho this is less of a problem w OREBs than DREBs.
That’s not to say that I think they’re infallible and this version of BPM is much simplified so you really can pick it apart. 1.0 had a bunch of weird interaction terms that made it fit better out of sample but are pretty obviously artifacts of common player types as opposed to those things themselves somehow being worth X points above/below average. And that gets at how, fundamentally, what we would really want is a tally of each “thing” that a player does and accord it a point value like baseball can do with the average value of a single, double, out, etc. Basketball just doesn’t lend itself to that kind of atomization so you get proxies and wide error bars on your estimates.
The error bar thing in particular is worth keeping in mind. I tend to use these metrics to think about guys as “replacement level”, “rotation”, “starter”, etc. as opposed to taking anything right of the decimal very seriously. Even if something strikes you as unlikely I think it’s worth seriously considering the model is right but a lot of the time what it’s really saying is “yeah we can’t really figure out the difference between these two”.
Re: Cade specifically tho it’s really just that his TS% is bad and even adjusting it for his usage it’s still bad. I don’t think anybody is super worried that will be the case long term but it’s true now and a reasonable reason to take Barnes especially but also anybody else among the group of dudes who aren’t quite NBA average in their first year.
Killian Hayes coming out party…
Yeah Killian doing his best Cade impression and going to cost the Pistons their top 3 draft slot
1 Like
Livers played 43 minutes had a double double awesome stuff
7 Likes
These lineups were hilarious.
Pistons should trade the pick they have Jamarko Pickett… Fans are mad the Pistons won and I get it but the big guys for the Pistons hardly played.
I’m definitely more upset at Harden and the 76ers for being unable to beat the Pistons in the last game. Especially when it should have meant more to them than Detroit
Yeah this game was a true measure of who could tank harder. Feel like the Pistons got wrong-footed because they played a real lineup for 8 minutes, then adjusted to trash, while the Thunder went trash from the jump (is Maledon the only long term prospect they played?)
Killian Hayes took twenty five shots!!! What else do we need to know!?
Really want Killian to develop. He is fun to watch when he’s aggressive. And his vision is fantastic. Still haven’t seen enough from him to deter me from being open to drafting a backcourt player. I think Hayes might be perfect to lead a second unit if he continues to play like he has the last couple weeks.
2 Likes
Hayes is better off as a backup PG leading the 2nd unit to spell Cade IMO. He’s not a starter material
2 Likes
Whether he’s starter material in the future or not. The bench role is best for his development.
1 Like
And Killian is a fantastic defender.
His offense is coming along. I thought his 10 point, 3 steal game vs. Philly was more impressive than putting up what he did against OKC. That was not an NBA team they beat last night. OKC probably kept Simpson on the G-League squad as to avoid being competitive.
I’m also bummed that Casey played Cade and Bey. If you’re only going to play them 5-8 minutes, why bother? I’m sure in Cade’s case, it was because so many OSU fans were in attendance. Still, now Cade’s averaging 2 points a game for April and Saddiq’s streak of making a 3 in every game is over. 
1 Like
Paywalled, but looks interesting…
1 Like
Is it after next year they need to decide on extending Hayes?