FiveThirtyEight giving the Celtics an 83% chance to win finals

[ITYSL]

the model for this series is dumb. big fat 0% maybe heat would have given them trouble.

Thanks, poster with an infinite quiver of incorrect takes.

[Capn'O]

538 is based on ELO and that's based on beating good teams with good net rating.

So many teams in the RS approached the Celtics post-January like they were facing the November Celtics and got blown out.
Five Thirty Eight reflects that and doesn't take into account things like championship level experience or choke jobs(i.e. Ayton/Ben Simmons).

538 has two models: the ELO model and the RAPTOR model.

Their RAPTOR model gives the Celtics an 83% chance of winning and their ELO model gives the Celtics a 68% chance of winning.

The Celtics have played amazingly well both offensively and defensively overall for a long period of time and 538 RAPTOR weighs that heavily.

Most betting sites are giving the Warriors a 55-59% shot to win the title.

When I first started diving into metrics, my friend who was teaching me had me start with baseball (2014). And he clued me in to something called the Mike Trout rule. And the Mike Trout rule was simple: if you are making an overall metric, and Mike Trout isnt at the top of it, you **** up

As the years have gone on, I've done something similar except with Vegas. If your predictive model is in wild conflict with Vegas... you **** up. But as I said on the other thread, predictive models for individual series in continuous clock sports are a trash concept to begin with - unlike oft misused advanced metrics like PER and RPM, I dont think they carry much value at all, other than to generate clicks by giving an (or any) answer)

I would kill to know how Vegas does their oddsmaking so accurately across such a wide scope of sports.. but I kinda get the feeling that you'd actually have to, to find out

Warriors aren't a bad analogue to Mike Trout. If they're in the mix and you're giving them a 17% chance something is awry.

[hoosierdaddy34]

Spot on. Lol The guy you were quoting literally said they gave Trump 30 percent to win on election day. Pretty sure he won. They're awful

Sent from my SM-S908W using RealGM mobile app

If my statistical model tells me there’s a 67% chance I don’t roll a 5 or 6, and then I roll the dice and a six lands, is the statistical model broken?

538s job is to assess probability and they do an amazing job. They aren’t fortune tellers and they don’t have BIffs sports almanac

In your example the model could be improved. Clearly you could better predict the outcome with more and better quality data.

Sent from my SM-S908W using RealGM mobile app

Clinton won the popular vote by like 3.5%. Trump had to thread the needle on the EC map to pull off the victory. 30% seems generous. How many other candidates won the popular vote by over 3% and lost the EC map? I’m sure it’s not many.

[falcolombardi]

538 is perenially in love with the celtics

they had them as overwhelming favorites in 2020, were high on them in 2021 too

there is somethingh about the celtics that their formula absolutely loves even when they suck

[falcolombardi]

538 has two models: the ELO model and the RAPTOR model.

Their RAPTOR model gives the Celtics an 83% chance of winning and their ELO model gives the Celtics a 68% chance of winning.

The Celtics have played amazingly well both offensively and defensively overall for a long period of time and 538 RAPTOR weighs that heavily.

Most betting sites are giving the Warriors a 55-59% shot to win the title.

When I first started diving into metrics, my friend who was teaching me had me start with baseball (2014). And he clued me in to something called the Mike Trout rule. And the Mike Trout rule was simple: if you are making an overall metric, and Mike Trout isnt at the top of it, you **** up

As the years have gone on, I've done something similar except with Vegas. If your predictive model is in wild conflict with Vegas... you **** up. But as I said on the other thread, predictive models for individual series in continuous clock sports are a trash concept to begin with - unlike oft misused advanced metrics like PER and RPM, I dont think they carry much value at all, other than to generate clicks by giving an (or any) answer)

I would kill to know how Vegas does their oddsmaking so accurately across such a wide scope of sports.. but I kinda get the feeling that you'd actually have to, to find out

Warriors aren't a bad analogue to Mike Trout. If they're in the mix and you're giving them a 17% chance something is awry.

i have heard people do this thingh with lebron and jordan in basketball too

we dont have plus-minus data on jordan, but lebron almost always seems to appear on top of plus-minus metrics and jordan/lebron always seem to appear around the top of boxscore ones

[cpower]

538 is perenially in love with the celtics

they had them as overwhelming favorites in 2020, were high on them in 2021 too

there is somethingh about the celtics that their formula absolutely loves even when they suck

must be smart's hair

[NBA4Lyfe]

The FiveThirtyEight RAPTOR player model has been really high on the Celtics throughout the entire playoffs considering the Celtics just shy of an all-time great team and gives them an 83% chance to win the championship.

Meanwhile it rates the Warriors as more of a playoffs team with only a 17% chance to win.

https://projects.fivethirtyeight.com/2022-nba-predictions/?ex_cid=rrpromo

five thirty eight does NOT take into account the officiating. Their have been times in the west where teams could match the warriors( 2016 thunder and 2018 rockets) and they were screwed over by the refs. This year the warriors have gotten a favorable whistle in both the memphis and mavs series

But i agree the celtics have the better players and should win, i just dont see it happening when the league is hellbent on steph curry getting a finals mvp

[hugepatsfan]

538 is perenially in love with the celtics

they had them as overwhelming favorites in 2020, were high on them in 2021 too

there is somethingh about the celtics that their formula absolutely loves even when they suck

Because the Celtics' problems over the years have largely been too many in game meltdowns. But not total game meltdowns. Like they just forget how to play basketball for a few minutes at a time. They're like a baseball team that wins one game 15-0. Then the next game they lose 2-1 with a runner thrown out at home and an error that allowed a run for the other team and those two bone-headed plays turn a win to a loss. That gets you to 1-1 over those two games. But a statistical model is going to smooth that out and just say 16-2 score over a two games stretch, wow that's dominant. Just keep taking those two games and repeating it over and over and that's the Celtics' seasons of the last 3 years lol

Second half of these year it finally seemed to click for them. But the playoffs have shown some ugly habits of just going through some mind numbingly bad stretches. For most of the MIL and MIA series they looked like the waaaaaay better team but they went 7 games because they pissed away games with some of the worst quarters or few minute stretches in games you'll ever see a playoff team play. And in games where they were otherwise great in besides those snippets. The statistical models smooth all that out. To a stat model, that small stretch is a blip over the larger sample size. But in actual games they flip wins to losses.

[dhsilv2]

By those figures you’d expect the Celtics to be a juggernaut but it doesn’t pass the eye test. When I watch the Celtics its clear they have some exploitable flaws

Also 538 had Hillary as like a 90-99% favorite over Trump depending on what point of the race you look at

Why do people keep saying this? This isn't true...not to get into politics but 538 gave Trump a very real chance at winning and even wrote in depth about that before the election.

[soxfan2003]

538 is perenially in love with the celtics

they had them as overwhelming favorites in 2020, were high on them in 2021 too

there is somethingh about the celtics that their formula absolutely loves even when they suck

2021 they were fairly low on the Celtics based upon what I am seeing. Rated them middle of the pack among all nba teams.

In 2021, they were probably sleeping on the lakers a little not realizing how good Lakers role players were in terms of fit with Lebron/Davis. And the Lakers added strength with Lebron/Davis playing more minutes whenever needed.

But to be fair to 538 with a healthy Hayward for 7 games instead of Hayward was perhaps 80% for the final four games of the series, Boston may have beaten the Heat. Miami won that series in 6 but Boston actually outscored them by 1 or 2 points in the series with Hayward just playing the final 4 games after missing most of the playoffs. Hayward's passing/shooting or at least the threat of it was important for beating the defense Miami was throwing at Boston.

When the Heat got to the finals, injuries hit Bam, obviously a very key player for Miami, and Dragic, a starter who missed almost the entire series. The Heat still took the Lakers to 6 games based upon some great performances by Jimmy Butler.

Healthy Boston may have had a better shot than I thought at the time. I was never a believer in Kemba Walker but the 538 projections make more sense if they had him as a very good player or above.

[Pachinko_]

I would kill to know how Vegas does their oddsmaking so accurately across such a wide scope of sports.. but I kinda get the feeling that you'd actually have to, to find out

Are we sure it's not just people who read everything and have sources on top? Why does it have to be an algorithm? Could be some combination of the two.

Plus they also have to adjust the odds anyway depending on which side the betting is coming from, to avoid going bankrupt. Which by definition makes their predictions great for making money, but bad for assessing teams. I mean for how long did they have Lakers & Nets as favourites for the Finals? That must have been because of one-sided betting.

[Rastas]

lm with them
We going to see a couple more big blow outs and an at best 4 - 1 or a sweep for the Cs.
Hate both teams but the head tells me this is a really bad match up for the Warriors.


Sent from my RMX3031 using RealGM mobile app

[xdrta+]

lm with them
We going to see a couple more big blow outs and an at best 4 - 1 or a sweep for the Cs.
Hate both teams but the head tells me this is a really bad match up for the Warriors.


Sent from my RMX3031 using RealGM mobile app

You can make big money if you bet the Celtics to sweep. $100 would bring back $2000, last I saw.

[Capn'O]

When I first started diving into metrics, my friend who was teaching me had me start with baseball (2014). And he clued me in to something called the Mike Trout rule. And the Mike Trout rule was simple: if you are making an overall metric, and Mike Trout isnt at the top of it, you **** up

As the years have gone on, I've done something similar except with Vegas. If your predictive model is in wild conflict with Vegas... you **** up. But as I said on the other thread, predictive models for individual series in continuous clock sports are a trash concept to begin with - unlike oft misused advanced metrics like PER and RPM, I dont think they carry much value at all, other than to generate clicks by giving an (or any) answer)

I would kill to know how Vegas does their oddsmaking so accurately across such a wide scope of sports.. but I kinda get the feeling that you'd actually have to, to find out

Warriors aren't a bad analogue to Mike Trout. If they're in the mix and you're giving them a 17% chance something is awry.

i have heard people do this thingh with lebron and jordan in basketball too

we dont have plus-minus data on jordan, but lebron almost always seems to appear on top of plus-minus metrics and jordan/lebron always seem to appear around the top of boxscore ones

Yup, those two would get you there too.

This might be analogous to when Mike went to play baseball. He came back, was still great but not quite the same and the the Bulls had lost some key pieces. The models might have looked at that and said "y'know we're still looking at a very good team but both Mike and the Bulls have begun their decline. Also... Dennis Rodman??? Whoa." Then Mike had a proper offseason, the Bulls retooled, and all of a sudden you're looking down the barrel of the 96 Bulls.

I think we have a similar situation here. The Warriors have been hung up for a few years but all of a sudden everyone's back, firing on all cylinders, and they have a new splash brother. If the Celtics can beat them healthy it would be a huge accomplishment.

[LAL1947]

FiveThirtyEight giving the Celtics an 83% chance to win finals

So we not finna talk about FiveThirtyEight giving the Celtics an 83% chance to win finals?

Fixed!

[FNQ]

I would kill to know how Vegas does their oddsmaking so accurately across such a wide scope of sports.. but I kinda get the feeling that you'd actually have to, to find out

Are we sure it's not just people who read everything and have sources on top? Why does it have to be an algorithm? Could be some combination of the two.

Plus they also have to adjust the odds anyway depending on which side the betting is coming from, to avoid going bankrupt. Which by definition makes their predictions great for making money, but bad for assessing teams. I mean for how long did they have Lakers & Nets as favourites for the Finals? That must have been because of one-sided betting.

Never said it was an algorithm.. in fact I've been arguing against the idea that numerical data is good for making accurate predictions, in that very post you clipped

Yes, the odds are influenced by betting, but they also have to start somewhere, and that start is extremely accurate.. when the actual event is near. Things like preseason rankings, across any sport, are almost entirely based on gambler's beliefs because there are an infinite amount of variables in any sports season. But when it comes to a series, especially a high profile series like a playoff series, you can throw that out. Because its not only Vegas influencing it, its the pro gamblers. Won't catch them making huge bets in the preseason, but you will in the postseason

[Big J]

538 is perenially in love with the celtics

they had them as overwhelming favorites in 2020, were high on them in 2021 too

there is somethingh about the celtics that their formula absolutely loves even when they suck

Probably because all of the MIT grads that make these models are huge Celtics homers.

[mateos]

Seeing how boston acts when closing games and also seeing how concentrated gsw on both defense and offence, i find this statistics ridicilous.

GSW will win it.

[robbie84]

Seems like alot of dumb money will be betting GS. Not that they should be huge dogs, but dumb money like casual Connor McGregor fans betting on him like he's invincible.

[B-easy]

One of the interesting things about probabilities is even after an event happens we can never know the "correct" probability. If GSW wins with a 20% probability that doesn't mean the 80/20 split was "wrong".