FiveThirtyEight giving the Celtics an 83% chance to win finals

[Nate505]

How many times do we freaking have to go through this? The concept of probabilities cannot be this hard to understand. They gave Trump a 30% chance of winning, which was much higher than any other model. And given Hillary was polling 4 pts higher and won the popular vote by around 4%, that was pretty freaking astute of them.

Hey 538 damage control representative.

It's not damage control, I remember their model showing Trump's chances climb up to 30% the week of the election and getting a sinking feeling in my stomach. 30% was wayyyy higher than anyone else was reporting at the time.

I l ike how people look at 30% and think the odds are just incredible they'll be broken. It's like these people have never played poker in their lives or any other game. That's about the same odds you'll hit an open ended straight on the river in hold 'em. I've played thousands of hands in my life, and let's say it's not that rare to see it.

Obviously I'd rather be on the 70% side, but it's not a very comfortable side to be on.

[Edrees]

This thread is a great window into people who think about probabilities vs people who think in binary terms. For the later this series completely invalidated the model for the former you need to look at overall performance of the model in every series and beginning of season predictions.

The performance of the model in the NBA finals should be weighted more heavily than other series for the same reason that the performance of the model in the regular season game should be weighted more than a pre season game. The model seems to break down the higher level of basketbal that is being played. That's a big problem if it only can predict lower level basketball.

Imagine a doctor who can spot cancer and treat it before it becomes a problem. That's a lot more improtant than being right at about why your nose is sniffling too much or why your 3 lbs overweight. The better doctors will be able to spot the more important problems. That's how a good NBA model should work. I would expect it to be more accurate the more important the series is. You can't equally weight a first round prediction with a finals prediction.

[sp6r=underrated]

This thread is a great window into people who think about probabilities vs people who think in binary terms. For the later this series completely invalidated the model for the former you need to look at overall performance of the model in every series and beginning of season predictions.

The performance of the model in the NBA finals should be weighted more heavily than other series for the same reason that the performance of the model in the regular season game should be weighted more than a pre season game. The model seems to break down the higher level of basketbal that is being played. That's a big problem if it only can predict lower level basketball.

Imagine a doctor who can spot cancer and treat it before it becomes a problem. That's a lot more improtant than being right at about why your nose is sniffling too much or why your 3 lbs overweight. The better doctors will be able to spot the more important problems. That's how a good NBA model should work. I would expect it to be more accurate the more important the series is. You can't equally weight a first round prediction with a finals prediction.

You can't say the model is broken on the basis of one series. That is just too binary,

[Bolivar]

I can on some level understand that they had a model that had Boston ahead. But that 83% is just so arrogant and low intelligence that it's good that people acknowledge it. 83% was a valid chance for something like those GSW with KD teams, or peak Lakers with Shaq & Kobe. Not some team with good regular season or pre-season RAPTOR stats or whatever, that hasn't won ****.

[Edrees]

This thread is a great window into people who think about probabilities vs people who think in binary terms. For the later this series completely invalidated the model for the former you need to look at overall performance of the model in every series and beginning of season predictions.

The performance of the model in the NBA finals should be weighted more heavily than other series for the same reason that the performance of the model in the regular season game should be weighted more than a pre season game. The model seems to break down the higher level of basketbal that is being played. That's a big problem if it only can predict lower level basketball.

Imagine a doctor who can spot cancer and treat it before it becomes a problem. That's a lot more improtant than being right at about why your nose is sniffling too much or why your 3 lbs overweight. The better doctors will be able to spot the more important problems. That's how a good NBA model should work. I would expect it to be more accurate the more important the series is. You can't equally weight a first round prediction with a finals prediction.

You can't say the model is broken on the basis of one series. That is just too binary,

I didn't say it was broken. Just said one should acknowledge the value of the model should be much more weighted on finals series than regular playoff series. It seems to be that weighting a conference semi finals or first round series equal to a finals series IS what's binary rather than allowing the nuance of weighted averages.

[Bergmaniac]

Schools really need to do a better job at teaching the basics of probability theory, so many people are completely ignorant about this, as this thread amply demonstrates.

[sp6r=underrated]

The performance of the model in the NBA finals should be weighted more heavily than other series for the same reason that the performance of the model in the regular season game should be weighted more than a pre season game. The model seems to break down the higher level of basketbal that is being played. That's a big problem if it only can predict lower level basketball.

Imagine a doctor who can spot cancer and treat it before it becomes a problem. That's a lot more improtant than being right at about why your nose is sniffling too much or why your 3 lbs overweight. The better doctors will be able to spot the more important problems. That's how a good NBA model should work. I would expect it to be more accurate the more important the series is. You can't equally weight a first round prediction with a finals prediction.

You can't say the model is broken on the basis of one series. That is just too binary,

I didn't say it was broken. Just said one should acknowledge the value of the model should be much more weighted on finals series than regular playoff series. It seems to be that weighting a conference semi finals or first round series equal to a finals series IS what's binary rather than allowing the nuance of weighted averages.

The model correctly pegged Boston as a legit contender when this board was writing them off completely. That is much more impressive to me than the failure of Boston in the finals. If you want to convince me their model is broken given me an alternative model that did a better job

[Profound23]

I assumed this was to keep bettors from betting on the Vegas favorite. Anyone who bought into this lost money.

[Pointgod]

Yeah this deserved a bump. 538 was always better at election predictions. Some of their sports models are... questionable

538 shows the limitations of advanced stats in general when evaluating basketball. It doesn’t take into account experience, adaptability, coaching, match ups, hustle, heart and the weight of the moment. They literally just crunch a bunch of stats without context.

[Flash Falcon X]

This is why statistics like this are so goddamn stupid. How could you even use numbers to predict an outcome of a Finals series especially when there's so much intangible factors at hand.

[MrGoat]

538 has a seriously terrible basketball model

They also predicted Clinton would win and we ended up getting four year of Trump. I would say that its flawed in more than basketball.

They gave Trump a much higher chance than almost all of the other models. That's part of how they initially got their reputation. They were giving him like a 25% chance when most others were giving around a 1% chance which made them look good. Later they released some of their ridiculous sports models which I've never taken too seriously, 83% for Boston on May 30th. After just watching this series you can't seriously tell me that if they were to run that series 100 times Golden State wouldn't win the series way more than 17 times.

[SOA]

Don't think it calculated Tatum's disappearing act in the Finals. Whether that was Wiggins defense, Tatum's hurt shoulder, Tatum's lack of aggression, etc.

Models and real life are not the same thing as flipping a coin. Emotions, feelings, injuries, adjustments come into play in a series that can never be put into a magic formula.

But it doesn't mean the model is wrong. As Harry Callahan once said: a man has to know his limitations. And all mathematical models have limitations.

[Edrees]

You can't say the model is broken on the basis of one series. That is just too binary,

I didn't say it was broken. Just said one should acknowledge the value of the model should be much more weighted on finals series than regular playoff series. It seems to be that weighting a conference semi finals or first round series equal to a finals series IS what's binary rather than allowing the nuance of weighted averages.

The model correctly pegged Boston as a legit contender when this board was writing them off completely. That is much more impressive to me than the failure of Boston in the finals. If you want to convince me their model is broken given me an alternative model that did a better job

Wait was the model predicting Boston as a contender when they were below .500? Because everyone here was saying Boston was a real contender once they got their season going and especially when the playoffs started.

[Dupp]

This is still a great prediction since it was made when the celtics were .500 and ended up making game 6 of the finals.

[ChartFiction]

simple algo didn't account for the fact Curry turns it up whenever he wants

[GunnerWRX]

To be honest, I think the Celtics got lucky even be in the Finals. Winner takes all, but they took 7 games to beat the the Bucks without Middleton, and 7 games against a hobbled Butler. The Warriors beat them 3 straight to eliminate them.

The Celtics are a young team, they will and need to improve. But if they feel entitled and think "oh we are the East champ and we will see you again in the Finals next season". Sorry, many teams don't reach the Finals again. The Magic with Shaq. Even Mark Jackson talked about it - he reached the Finals 1 time over 17 seasons.

I would say the road to reach the Finals again next season will be harder, and they need to work harder / improve more, if their ambition is a championship.