In game 7 it didn't matter if missing 3s is random or not. What's the brilliant strategy that someone besides Joe would have tried with a limited Tatum, Rob throwing up, Brogdon unplayable and Brown dribbling off his feet? If anything they should have tried to run plays for Grant and Hauser and hope they could hit some threes.
Celtics vs Heat ECF Redux Discussion Thread
- Thread starter DeadlySplitter
- Start date
But how do you know that the next shots won't fall?
Do you play poker? Do you stop trying to get it all in with AA, if you lost with them the last 5 times you played them? That would be wrong, yes?
I think both in poker and basketball, you try making the play with the highest expectation. That's all you can control.
He's a role player and there was basically nothing expected of him but I just want to point out that Grant was really goddamn bad in the last 3 games of this series. Any idea that involves Grant shooting more or being involved more is a bad one
Both bank loans and state election outcomes were connected. Teams rightly believe that each shot is independent from the other, because there's no evidence that there isn't.
Remember, people are often willfuly blind, have selective memories, impose their narratives on reality, in other words, they have many cognitive biases that affect their understanding of a situation especially if it involves randomness.
yes, as has been noted repeatedly, Rudy Gobert. Also a bunch of guys voluntarily turned it down to move elsewhere.
Jaylen is in the kind of spot where he might not get the offer because of a few things:
1. He's fringy, he made it in a really weak/injured year and wasn't really that close in past seasons.
2. The new CBA was basically built to say "yeah if you sign 2 guys to supermax you better be okay with probably a 3 year window and basically a mandatory rough stretch down the road.
3. He's clearly not the franchise player (most supermax guys are the best guy on their roster when they sign it)
The assumption is at a certain point in a game the 3s not falling within the game becomes more predictive of future 3s in the same game than the full season sample. Which is false.
Turning it down is one thing. Offering him less than what he's eligible for is another. JB is not Rudy Gobert.
JB already took one discounted deal. I've seen nothing to suggest that he will take a second. Contracts isn't just about money for NBA players, there are status/respect considerations. If Brad officially offers him less than the Supermax, I expect JB to be gone unless JB suddenly decides for whatever reason he's willing to take a discount.
I know you know all of this.
The probability of Boston shooting 81/267 from 3 when their “true” mean was 0.377 and each shot independent is 0.0072. The probability of Miami shooting 89/205 when their “true” mean was .344 and each shot independent is 0.003. The probability of those two things happening in the same series, if they’re independent of each other, is 0.00002. It should happen about once every 50,000 playoff series. That’s slightly less than the probability of a team coming back down 0-3 by three orders of magnitude.
I assume when you say there’s “no evidence” that shots aren’t independent of each other you’re referring to the work on the “hot hand” theory. I’m not talking about “clutchness” or “streakiness.” I’m taking about treating shooting variance as nested within game and within series. The work on hot hand (e.g., Camerer 1989, Gilovich et al. 1985, Tversky and Gilovich 1989, Ibañez et al. 2017) was done to answer a completely different statistical question, and none of their studies tested whether what I’m saying is true. The same problem holds for the game theory stuff that came out emphasizing three point shooting (e.g., Skinner 2012, Skinner and Goldman 2016). As far as I can find nobody has actually tested whether a model that incorporates random effects by game and/or by series better explains shooting outcomes. If you don’t understand the difference between the two questions then please don’t lecture me about cognitive biases.
that's an interesting question. I wonder if anyone has looked at this (or how they even would). I mean individual games could have specific variables that affect shooting - travel/tired legs is the first that comes to mind.
And they would have had 5 days to rest up if they won it in 6. Timelord gets sick, then he gets better. Brogdon probably dun. Tatum doesn't sprain the ankle.
I oversimplified. Of course there are things that will cause full season percentages to be less predictive. Injury, fatigue are the obvious ones and well documented. Tatum missing threes after rolling his ankle? Brogdon airballing with a torn tending in his shooting hand? Yeah, probably stop shooting those.
Confidence and rhythm are the ones that we think should vary game to game or shot to shot and are hard to quantify. The hot hand studies as a proxy for confidence seem to suggest the effect is weak if it exists at all. Rhythm? Using the eye test the Celtics bricked a lot of good, in rhythm flow of the offense shots. What do you do at that point?
I know it is likely JB won't take less than the supermax, I just don't think Brad will care if that's not the valuation he and Wyc put on Jaylen. I also think.... JB kind of is Rudy Gobert.
When he signed that deal Rudy had won 2 DPoY trophies, had made 4 straight 1st team All-Defense, made All-NBA three times he was far more personally accomplished than Jaylen... now you could argue he was at a position that is at time undervalued.
Also, we really need to stop with the revisionist history on this board that Jaylen "took a discount" last time... he decidedly did not, he took what he was offered because his agent didn't think he would get a better offer sheet as a RFA. He was coming off 13/4/2 the previous season. You can look back using google for articles around the time he signed, there was a pretty big split on whether it was an overpay, I don't see much of anyone who thought it was a discount, basically the 2 camps were:
1. Overpay
2. Worth investing in his potential to breakout.
Jaylen ended up underpaid by the end of his deal because he improved, but that's not taking a discount.
This is where scouting and video evaluation should come in. There are probably going to be tells in a players shot that might indicate how they will perform on a given day. We take this as a given with pitching (to use a different sports example).
While I don't agree with the "If threes aren't falling completely change your offense by the 2Q" crowd there IS something to be said about shooting rhythm and confidence within a game or a short series. Ask self honestly at pretty much any point in that series if you'd expect 44% shooter Al Horford or 35% Caleb Martin to knock down their three-pointers. Over the course of a season I agree but streaky does exist as made shots lead to all other kinda positive things on both ends of the floor.
Addressed these and I agree. The eye test matters for those components. But in this case they don't to me explain the shooting collapse. And it must also be said there are definitely negative effects to an offense from passing up open shots you'd normally take which is something coaches mention a lot.
This discussion is enlightening as to why I get so frustrated reading the discussions lately. Treating shots as coin flips ignores so much about the mechanics of shooting, shooting touch, scoring, etc. There are days when you try to shoot a basketball and the hoop could practically double in size and you couldn’t get it through. It’s really weird but it happens, even to the best of them. Basketball just isn’t a math problem and continuing to shoot 3s because of belief in mean reversion ignore so much about the realities of playing basketball. And I didn’t even start in on the mental impact of shooting while down a bunch having missed a bunch already. In my opinion there’s just no way to compare something as complex and difficult as shooting is to games of chance.
Are you talking about the odds of one team shooting precisely 81 of 267 and the other 89 of 205? It wouldn't be at all surprising that those exact numbers are very unlikely to come up in a specific series. But it doesn't seem wildly unlikely that a true talent 37.7% 3 point shooting team would hit at around a 30% rate over 7 games or a 34.4% one would shoot a little over 40%.
I'm thinking more about fatigue, which certainly affects shooting - but I have no ideas on how one might go about quantifying that. Particularly in my experience while athletes pay super close attention to their bodies, they are also in denial about what they might or might not "have" on any given evening barring an outright injury.
The concept of "rhythm" seems to be even more nebulous than fatigue.
I know that people are really looking into the science of performance, even the field is currently in its infancy.
Within a single game there's a lot of evidence that observed shooting percentage drops fairly quickly compared to expected (based on the player, shot type, closest defender, location) as you get to around the 15th shot of the game for that player. Which is a pretty good proxy for fatigue.
Or just, like, watch any overtime game.
Those are the cumulative probabilities of hitting 81 or fewer threes in 267 chances and 89 or more in 205 chances for Boston and Miami respectively, assuming all of the shots are independent of each other. It is wildly unlikely, that’s my point.
There's a lot more than variance that the three-point numbers are not capturing. Jimmy Butler is an obvious example. If you look at his basketball reference page, he looks like a horrendous three-point shooter after being respectable from distance early in his career. However, he has always been an excellent free throw shooter, which usually strongly correlated to three-point skill. Looking closer, Butler has doubled his three point shooting rate from the regular season to the playoffs in each of the past two years. Basically regular season Butler refused to shoot threes unless it was a desperate late shot clock situation and his numbers crashed accordingly. In the playoffs, he has a quicker trigger and is taking better looks. That's why something like Second Spectrum's expected effective field-goal percentage is going to be artificially low for a team like the Heat. The numbers say a Butler three is a bad play, because he's filled the stat sheet with junk numbers from a relatively meaningless regular season.
The Heat aren’t the first #8 seed to play for the NBA title. The first this century, but last century isn’t that far in the past. (Allow me to introduce you to the ‘99 Knicks)
You'd want to correct for shot location, the person shooting, shot clock, and closest defender.
Thanks boss I’ll get right on that.
The point I’m making is the Celtics had a disastrously bad series shooting threes AND the Heat had a disastrously good series. Under the assumption that each shot is independent of every other shot there is simply no way to expect the outcome we observed. Whether it’s 1 in 50,000 or 1 in 10,000 or even 1 in 1,000 doesn’t matter. The three point shooting differences were not independent events drawn from a random sample represented by the teams’ season averages.
If you thought there were, say, a 1-in-3 chance that those shooting outcomes would happen, would you have game planned for the Heat differently than if you thought it was 1-in-50,000? Some teams appear to think that shot percentages are not nested within game or within series - the idea that shots are independent events is apparently treated as gospel here. By ignoring that non-independence, you overestimate your confidence in the outcomes. It’s a more complicated “math problem” than people are treating it, where just
bombing away from three on one end and sagging off “mediocre” shooters behind the arc at the other is probably not the best strategy once you account for the nested nature of shot outcomes. I am deliberately remaining agnostic on what could cause that non-independence - people have suggested fatigue, team or individual player strategy, shot selection, injury; it’s probably all of those and more. That’s the whole purpose of mixed effects models is to account for the fact that many things are currently unmeasurable but still contribute to our overall confidence of a particular outcome.
Last edited:
If I understand you correctly, you are arguing that the true 3 point % of both Boston and Miami is specific for the series and different than their overall shooting % throughout the season. If this is the case, what's the mechanism supposed to be?
Are you arguing that for 7 specific games Miami has higher true shooting percentage and Boston a lower one? That doesn't make sense to me. Why would people's true shooting percentage change in a particular series? *
I think a better argument is that Miami has a way of forcing Boston to take worse shots while Boston is giving Miami easier shots, but that's not what the stats say. Last time I checked, Boston had more wide open shots and Miami fewer. Why on earth should Miami's true 3% should be higher in wide open shots in these particular series? If there's a stat that we shouldn't expect to be least affected by particular factors that can be found in a specific series is wide open shots without pressure from the opponent.
Last but not least, I don't know how you got to finding a probability of 0.0072 for Boston's shooting to be 81/267. From my perspective, 267 shots is a sample representing Boston's true shooting percentage that we can find if let the team shoot an infinite number of 3% shots under game conditions. From my knowledge of basic statistics, a sample of 267 will have a high margin of error; like I know that if we poll 267 voters for the next election, the margin of error for that poll would somewhere between 5-10%. So Boston and Miami deviating 7% from their true shooting percentage is a pretty extreme outlier, but not outside the realm of possibility.
It should go without saying that the regular season numbers are also a sample. However it's a far larger sample, with a far smaller margin of error, which makes it more reliable one So if you ask me which sample is more likely to represent the team's true shooting ability, choosing the regular season numbers is a no brainer, especially if it pertains to wide open shots.
* Injuries strike me as the only factor that can legitimately be affecting 3% shooting performance. Then again injuries should also be affecting the alternatives. So for example, if you think that Brogdon's and Brown's 3 point % will be down because they are injured, you should expect that their success rate on midrange shots and drives to the basket to be similarly affected. I guess in that case, the best argument you can make is that injuries will affect their two point success rate less than their three point success rate. This might be true, but what also would be probably true is that the effect would be marginal.
After game 3, we expected Boston to rollover, after they lost the lead in game 6, we expected them to lose and when they won in the last milisecond, we expected them to win game 7. Which is to say, there's a lot of things in that series that pointed towards rhythm and confidence pushing the series in a particular direction, only to find out that the series went towards the completely opposite one.
I am really struggling with this.
This post is written trying to convince the reader that Brown kind of is Gobert, except actually worse and less accomplished. Do you really believe that?
Gobert is literally played off of the floor in the playoffs by any team that has a small ball 5. He has absolutely 0 offensive game so he can't punish anyone in the post.
Gobert was also 2 years older than Jaylen is now when he signed his Super Max and there's a long history of guys his size not aging well...especially when they have very little actual basketball skills.
Isn't the better comparison Bradley Beal? I think Jaylen is better than Beal and Beal got the full super max (in addition to being 3 years older than Jaylen).
Most players that have the ability to get the Supermax get it....because they are very good players and teams don't want to lose them. The one guy that didn't had multiple mitigating circumstances that don't apply to JB. They aren't similar situations or players
With all due respect to your perspective, the answer to the question “what is the probability of getting 81 or fewer successes in 267 independent trials if the true probability is 0.377” has been know for over 300 years and is a component of any intro stats class. The answer is there’s a 0.7% probability. You can Google a binomial probability distribution calculator and see for yourself.
The reason polls have larger margins of error is, among other things, that they account for the fact that voting outcomes are not independent. Which is exactly what I’ve been trying to say.
Gobert's playoff struggles mostly came after he got the supermax (also much of them had as much to do with the team around him as his play, Mitchell and Gobert couldn't share the floor).
I agree most players eligible either get offered it or get traded. Gobert signed for less, Kemba allegedly would have signed for less but CHA offerred him way less.
But as to Jaylen... he's one of the least personally accomplished guys every to be eligible, and he's one of the few who have a clearly better teammate, and he's the first to have it happen with the new CBA agreed. There are a number of factors that conspire to make him a guy who is more likely than most to not get the full offer, in the sense that usually the threat of trading a guy for $0.85 is devastating, but less so when you're talking about your #2.
I get the Beal comp, in that Beal was also very very fringy, probably worse than Jaylen..... that deal was also widely acknowledged at the time to be one of the worst deals in the league at the time of signing, and came from ownership being scared of losing a "Star" and not having any real interest in competing. I'm not sure "Beal got a supermax instead of getting traded and the team is a joke of a 12 seed and he has no value in trade when he did before the extension" is a good argument for offerring Jaylen a supermax.
If they think he's worth it or tradable for more in a year on that deal they'll offer him the full $290M, but when a lot of national guys (Lowe, Marks, etc.) are discussing that they think he'll get the Gobert offer, it's something to consider as a real possibility, and if Jaylen says he'd rather leave money on the table by going elsewhere than sign for a penny less than the most they can offer... he'll probably be on a new team.
In both of your answer you gave me no due respect.
As for the second sentence, outcomes in infividual states has nothing to do with what I am saying, because I am talking about voting preferences of individual voters.
Beal is worse than Jaylen. I think that's something that is a very common held belief.
You know what you're talking about and you follow this stuff but you leave out things when it fits your narrative. I think we both know that the real reason the Beal contract received so much scorn was because he also got a no trade clause. The thinking was he might deserve a supermax. Again, he was also 28 and not 26.
If Jaylen gets the Supermax, I don't think anyone bats an eye. There's also been scuttlebutt, by the same people you mentioned as getting the Gobert supermax, that Jaylen would have a "robust market" if they wanted to trade him.
Shooting in basketball isn’t like flipping a fair coin or rolling a fair die. Sometimes you miss a lot, for example, because your mechanics are off, and until THAT gets fixed, yes, it has some predictive quality about it. Unlike a coin or die.
I think that was like a cherry on top of the shit sundae. Everyone thought the money was terrible because he's not worth it, and to add a no trade clause was clownish. The contract still sucks without it.
But more to CD's point, dumb teams doing dumb things is not a reason to think it's the "normal" state of things, and particularly because of the penalties that are in the new CBA. For example, GSW probably don't extend Poole if they know that they're going to lose their taxpayer MLE, the ability to package contracts, future ability to trade picks, and all of these other things that handicap their flexibility. Before, it was just money to billionaires, but going forward Max contracts to non elite players are going to be incredibly debilitating from a team-building standpoint.
Really? You mean now, or when his contract was signed? I don't like Beal that much because I think defense is important, but the guy was just recently second in the the league in scoring. I think there are many who would have considered him better than Jaylen.
I was referring to as of now. I think at the time it would have been more of a coin flip but I also think the age and positional value favors Jaylen which would cause less people to denigrate the contract
I know people dislike these charts but they do a lot of work.
It feels like Brown is perceived as the better all around player (or at least the higher upside player) and has been for some time.
Can you give a quick tutorial about how to read these charts? I see them posted all the time but have no idea what the x and y axis values mean or how to interpret.
Also... the real key is everyone thought WAS was crazy not to trade him when they had good offers. Now that situation was somewhat different in that you can talk yourself into Jaylen because in part this is a true title contender, so your return needs to not be a big step back (where WAS was garbage and had no reason not to make a trade).
There isn't going to be a perfect comp for Jaylen, the pool of Designated vet eligible players is small, and each situation is unique, there are also new factors this year.
I think DARKO is really good, I would say you should also probably include one of the others though, because each one has some blind spots, and with Jaylen in particular DARKO is a massive outlier.
LEBRON had Beal as the better player at the year each signed his contract:
They're in a similar tier of player certainly. I'd probably prefer Jaylen now given age differences, but in terms of contract... similar.
From the website:
To be clear, the actual DPM isn't that important as much as the career trends. But in short these are simply player impact metrics.
And @Cellar-Door is correct that other stats might support Beal as the superior player but I feel pretty comfortable in saying that up to now, Brown was perceived as the higher ceiling player, if only because of his defensive advantage.
Bradley Beal is a good NBA player that is paid like a great NBA player, and gets additional flak because he is miscast as a no. 1 on a bad team. Jaylen is a similar player, a little younger and a little better overall, but still in the same tier but he’s had surplus value because he is on a very good contract. The second he signs that Super Max he’s going to be subject to the same kind of scrutiny as Beal. Paying a top 25 player as if he were a top 10 talent is a dangerous use of resources.
Last edited:
https://apanalytics.shinyapps.io/DARKO/
Here is the basics of DARKO, it's projection and box score based.
I'd also say, I think games is the worst of the charts, I would use either age or season:
I think Tatum in game 6 of the Sixers series proved you can be terrible shooting and it can turn on dime.
1) I don't play poker, and I find poker talk to be among the most boring things on Earth. No offense intended.
2) Basketball players aren't robots. Sometimes players have bad stretches in games and rally (see: Tatum's game 6 vs. Philly). But a lot of times they simply have bad games, period. I'm never going to be against Tatum shooting good-look 3s no matter what his in-game FG% may be. But most other guys... yeah, if you're 0-5 then you gotta stop.
Ive brought this up before but I highly recommend folks to read about aleatoric versus epistemic uncertainty. Explanatory article here: https://www.stat.berkeley.edu/~aldous/157/Papers/Fox_Ulkumen.pdf
The distinction is between inherently random systems (aleatoric uncertainty) versus systems which are deterministic but which we have imperfect insight into. These come in varying degrees depending on the level of abstraction you approach the modelling problem. The important part is that the extent of aleatoric uncertainty bounds how accurate any model of the process can possibly be. I cannot make a model that does better than 50% on the outcome of a coin flip (assuming a fair coin and assuming we dont have access to the minute physics of the flip, the coin's trajectory through the air, etc).
However, in the case of shooting a basketball I can start with a model that assumes all players shoot 36% from three and then improve the model by adding relevant information the simple model does not have access to, like which player took the shot, the location of the shot, the nearest defender etc etc.
This argument has basically been about the extent to which shooting is aleatorically random versus epistemically random. I personally feel there is still a high degree of aleatoric randomness, assuming we aren't running an insane physics simulation tracking a player's physiology, shot mechanics etc. and even if we had access to confidence, however you represent that, and rhythm etc there will still be a substantial degree of variance.
The distinction is between inherently random systems (aleatoric uncertainty) versus systems which are deterministic but which we have imperfect insight into. These come in varying degrees depending on the level of abstraction you approach the modelling problem. The important part is that the extent of aleatoric uncertainty bounds how accurate any model of the process can possibly be. I cannot make a model that does better than 50% on the outcome of a coin flip (assuming a fair coin and assuming we dont have access to the minute physics of the flip, the coin's trajectory through the air, etc).
However, in the case of shooting a basketball I can start with a model that assumes all players shoot 36% from three and then improve the model by adding relevant information the simple model does not have access to, like which player took the shot, the location of the shot, the nearest defender etc etc.
This argument has basically been about the extent to which shooting is aleatorically random versus epistemically random. I personally feel there is still a high degree of aleatoric randomness, assuming we aren't running an insane physics simulation tracking a player's physiology, shot mechanics etc. and even if we had access to confidence, however you represent that, and rhythm etc there will still be a substantial degree of variance.
Well, you can't actually know that for sure.
This is what we know.
Ok, so why? Because the shots are not independent events? (Your assumption.) Because the teams' season averages are not the right way to figure out true odds in a 7 game series featuring the same teams, with each event bringing each team closer to a season ending event (making strategy or decisions different from the regular season)? Something else. I don't think we have any idea. Just guesses. Maybe some of which are testable.
I'm not sure I understand what exactly your concept of "nested" means, but it seems to be a hypothesis that you view as better to explain the results we saw.
I mean, I'm fully onboard with the idea that it is an error to imagine we saw an extraordinarily unlikely thing happen that should allow us to say boy were we insanely unlucky. But that doesn't mean we should simply reject out of hand the idea that shooting percentages do matter, that although clearly something else was going on in this series to make our sense of "true" odds off there still is some loose or medium correlation and validity to treating shots as independent events, and that if we could really understand all the variables and dynamics we might well conclude that actually there was a high degree of unlucky in the series that we should not expect to see with a higher n.
What this thread feels like to me are people trying to deal with a sad sports result by consoling themselves that the team got unlucky. I don't have a problem with rejecting as flawed the questionable statistics or methods they use to get there. We should. Using math poorly to make emotional arguments rarely leads to a good place. But I also think it kind of misses the big picture point. Because I'm not sure that any of this actually says they are wrong in the big picture sense. They might be. Or there may have been some serious unlucky that happened, just not as a significant as the flawed method indicates.
Ok say you’re trying to game plan for a seven game series. You model a bunch of different approaches to your offensive and defensive schemes. In each of the models, you assume that three point shooting are entirely independent events drawn from a binomial distribution with true mean of 0.37 (or whatever the season average was).
Now say instead of doing that, you say each game is going to have it’s “true” random shooting, which itself is drawn from the binomial w 0.37. So Game 1, the team’s true shooting is actually 0.34. And Game 2, the team’s true shooting is actually 0.42. Then the three point shots they take in that game are independent events, but the probability of making them are more similar to each other than the overall season average.
Now let’s say you’re designing a team, figuring out how many bigs you need, how many 3-Ds, etc. And you simulate a bunch of consecutive seven game series. Instead of treating every single shot as an independent event, each series has its own “true” shooting percentage (again taken as a random draw from the season average) and each game within that series has a new true shooting percentage (taken as a draw from the series average).
That’s what nested means. It’s entirely uncontroversial. An upper division stats class hammers this problem home. In any simple experimental design, you almost always end up with nested groups - samples that are more like each other than they are to the overall population, usually for no measurable reason. (That doesn’t mean the reason doesn’t exist, just that it’s too hard to measure or control for and it’s not necessarily of interest for the time being).
When you ignore nested problems and treat all events as independent, you get overly confident in the outcome. You assume that, say, a team that hits 50% of 3 pointers one night can’t possible do it again. Or that a team that hit a season low from three one night won’t do it again the very next game. Or that a candidate who appears to be down significantly in the polls can’t possibly win.
Again, the actual reason for the non-independence doesn’t matter (although as folks have speculated above, it would probably be really useful if you COULD measure it - then you might be able to say “tonight we move away from the 3 point offense.”) The math is kind of shitty to deal with but it’s completely doable. When you run the models you just account for the fact that three point shooting within games, and within series, are more like each other than expected by random chance. Those are called random effects. You can include other stuff too - opponents defense, expected shot quality, player usage, etc - but that’s not what I’m talking about. Those are fixed effects. They’re measurable. They’re repeatable. The lack of random effects is partially what sunk the banks and what sunk Clinton, and I have a strong suspicion is causing problems for game planning for a bunch of NBA teams. Again, I think if the Celtics had a reasonable expectation that the three point shooting would go as it did, they might have approached the series differently. Non accounting for the non-independence within games/series made them too confident in shooting percentages regressing. That is purely speculation on my part but given it’s been a problem in other sectors for so long (and given that apparently not a lot of people here have even heard of random effects?) makes me think it could legitimately be a problem in the NBA.
I see -- you're viewing it as a game planning problem. And I think that's legitimate.
I was viewing the discussion here more as emotional fans trying to protect themselves psychologically. I think the ultimate question here is -- were the Celtics unlikely? For some that makes them feel better. It also could matter for purposes of offseason moves, but that's not really how I perceive the discussion here.
Was the three point shooting that actually occurred something you would expect if you were omniscient and could know all the variables? If the same series was simultaneous played in a massive number of alternative universes where everything else is the same would the results all be the same, or would we expect variance? I mean, I guess if literally everything with the same, I guess you'd get the same result. Maybe it's all deterministic, but trying not to get too philosphical. You know, you start at time zero with everything being the same and then you just let it happen. Maybe it would always work out exactly the same. But if there were variance, how does what happened on real Earth match up with the averages. Was it an outlier?
We don't know. But maybe season shooting is some evidence that it was an outlier. And if that makes people feel better, I'm ok.
I think good coaches treat each game in a series as evidence, such as it is, but also bring their sense of history, intuition and feel into it, because it's a proxy for what we don't understand. I think that no matter how good you think your model is, there just aren't enough events in sports to be super confident. And all it takes is one blown variable to make you make the wrong decision. I mean, maybe truly highly scientific sports like formula 1 it's very different. But it would have to be a really sophisticated model in basketball. There is just so much going on.
I was viewing the discussion here more as emotional fans trying to protect themselves psychologically. I think the ultimate question here is -- were the Celtics unlikely? For some that makes them feel better. It also could matter for purposes of offseason moves, but that's not really how I perceive the discussion here.
Was the three point shooting that actually occurred something you would expect if you were omniscient and could know all the variables? If the same series was simultaneous played in a massive number of alternative universes where everything else is the same would the results all be the same, or would we expect variance? I mean, I guess if literally everything with the same, I guess you'd get the same result. Maybe it's all deterministic, but trying not to get too philosphical. You know, you start at time zero with everything being the same and then you just let it happen. Maybe it would always work out exactly the same. But if there were variance, how does what happened on real Earth match up with the averages. Was it an outlier?
We don't know. But maybe season shooting is some evidence that it was an outlier. And if that makes people feel better, I'm ok.
Yeah, it's pretty complicated though and you could overcook it, in real time, right? You could make the exact wrong decision in the other direction. I mean unless you were highly confident in your model.
I think good coaches treat each game in a series as evidence, such as it is, but also bring their sense of history, intuition and feel into it, because it's a proxy for what we don't understand. I think that no matter how good you think your model is, there just aren't enough events in sports to be super confident. And all it takes is one blown variable to make you make the wrong decision. I mean, maybe truly highly scientific sports like formula 1 it's very different. But it would have to be a really sophisticated model in basketball. There is just so much going on.
That money on the contract was not thought of as terrible when he signed it. There was a question of does he deserve the super max because of the team that he was signing it to and the fact that they were going nowhere. Most people thought that the smart thing to do was to lock him into a super max and then, if they couldn't make moves to be competitive, trade him to a team and get value back to blow it up. Again, that's why the no-trade clause made the contract a joke. They lost all flexibility with him.
If CD's point was what you just said it was then I 100% missed that. The original point he made involved Rudy Gobert in a comparison that only made sense because he wants Jaylen to takes less than the supermax and Gobert did that. The situations and the players aren't remotely comparable.
I like Jaylen, you guys don't. That much is clear.
The CBA is prohibitive to signing two super max contracts, that is also very clear.
Where we diverge is that I think they should still offer it to him because, even if he's overpaid by 1-3% of the cap, he's still a valuable trade asset. I am not sure if you, CD or both of you are saying that he's not tradeable if he gets the supermax? Or if he will just be overpaid? Will he somehow lose value? All of the reporting we have is that Jaylen will have a very big trade market if and when they decide to move him. One of those people being Zach Lowe, who was brought up earlier.
My messaging on this has been incredibly consistent. You negotiate with him and see if you can get him for less than the supermax. If you can't, you give him the supermax and kick the can down the road for a year or two and see where you are. Because this CBA is so onerous you probably have to try and trade him. If he doesn't want to stay here for whatever reason then you have to try and trade him or if someone knocks your socks off in a trade that makes you better this year, cuts salary and doesn't damage the future then you take it (the latter is extremely unlikely)
Almost everyone agrees with this. Ryan Bernardoni, who knows more about the cap than anyone here, basically laid this entire plan out and said that if they were to trade him this offseason they would very likely have to take a meaningful step back and no one wants that.
I don't know what a reasonable counter or argument is to this.
I would also like to see trade offers that are workable. @benhogan has laid out a few that work. Others involve players like Dejounte Murray, which are just outright laughable. I've looked at it myself and think that there are very few teams that match up (Memphis, Brooklyn, maybe New Orleans depending on how they feel about Ingram)...that's another factor to consider
Last edited:
I have no background in statistics so apologies if I'm misunderstanding, but it seems like a lot of this conversation is a redux of the talking head debates about analytics vs eye test or guys like CJM saying "I like math" vs guys like Barkley saying "you can't win as a jump-shooting team". I completely agree with you (as much as I understand) that its kind of ridiculous to assume game by game or series by series shooting will regress to season averages. Even accounting for the fixed effects you are talking about, and using the huge amount of tracking data and everything that is available these days, I agree there are really important things that cannot be incorporated into those models. People have talked about fatigue, which is probably important, and others like to talk about momentum, but I think the biggest factor has to be confidence. Anyone who has played a sport knows how important your mentality is in determining performance, and it is really hard to maintain the same mentality you have in a regular season game in a game 7. Basically, there's no way anyone can convince me that Jayson Tatum's shooting in elimination games, or "Game 6 Klay" are just predictable regression to a mean. There is definitely an intangible variable to shooting based on how you are feeling and probably some truth to all the old truisms like "you just have to see one go down".
TLDR: Joe Mazzulla "liking math" is all well and good but you can't just assume 3PA=good offense. There are nights when you have it and nights when you don't and a good team should be able to adjust
I think I’m doing both while trying to couch it in a discussion of strategy. Here’s the thing. If you treat a series as having its own true 3 pt shooting percentage (drawn as a random variable from the season average), and each game within it having its own true shooting percentage nested within that, the probability of the Celtics’ shooting performance goes from 0.7% to 11%. For Miami, it goes from 0.3% to 8%. Again, there’s nothing I’m changing about the team’s shooting abilities. I’m just treating the outcome of each 3 pt shot as being dependent on that particular night’s “truth” AND the series’ level “truth” rather than assuming every single shot is independent of every other one. That feels like a much, much better model to describe what actually happened. Saying the Celtics had a 1 in 10 shooting performance and the Heat had about a 1 in 13 shooting performance better captures it, doesn’t it?
Emotionally, that makes me feel better. Celtics had some bad luck that was made worse by Brogdon’s injury, but it wasn’t some insane inexplicable completely impossible shooting performance. Shit happens. I absolutely see the same thing a lot of people say about the impact of confidence, but I’m going to leave that as currently unmeasurable.
Strategically, this approach has very different implications for both a single playoff series and for team building. (Actually I wonder if the Bucks/Bud were working from the same flawed assumption - figuring if they drop, the opponents’ shooting has to regress.) The Celtics doubled down on shooting threes this season. They went out and got Brogdon and Gallinari. They pushed Hauser into the rotation. It was the major offensive identity. They went from 10th to 9th to 2nd on 3PA in the regular season. But if you end up having a much higher probability of a dreadful series from 3, and the opposing team is game planning for exactly that, maybe it makes sense to diversify the offense.