CelticsStrong
Around the League => Around the NBA => Topic started by: Moranis on May 09, 2018, 10:00:22 AM
-
https://projects.fivethirtyeight.com/2018-nba-predictions/
They currently have the Rockets at 80% and the Warriors at 20% to make the Finals. I get this is based on stats, but that just seems crazy and really highlights the flaws in stat based models.
-
You don't evaluate models like that.
Obviously the numbers are crazy and nowhere near the truth but you don't disregard a prediction/estimation model because it got one or two very wrong.
I don't follow this site so maybe this has been a reccuring issue but the other predictions look pretty good.
-
Look at the ELO history of the Warriors, it tanked in the later half of this season once the Rockets secured home court and Curry was injured. They were missing a MVP caliber player and mailing it in as they had the two seed wrapped up.
The Rockets ELO actually did the same thing, but a lesser decline once they clinched HCA through out the playoffs.
-
I'm fine with any model that tells me the Cavs have only an 11% chance of winning the title.
-
You don't evaluate models like that.
Obviously the numbers are crazy and nowhere near the truth but you don't disregard a prediction/estimation model because it got one or two very wrong.
I don't follow this site so maybe this has been a reccuring issue but the other predictions look pretty good.
Before the playoffs, they had the Rockets at 65% of making the Finals, the Raptors were 2nd at 44%, followed by the Sixers at 23%, Cavs at 20%, Jazz at 10%, Warriors at 7%, and I suppose there is not point in going forward. So in sum, the team responsible for the most dominant 3 consecutive regular seasons ever (including 2 titles over those seasons), 538 predicted would make it to where they are at just 31%, with a 7% chance to go to the next round, and only a 4% at winning the title. heck they only had them at 62% to win their 1st round series against the Spurs (which is odd because the Spurs they had at 42%). They only gave the Warriors a 60% chance to beat the Pelicans.
I'm not sure they got the other predictions right.
-
I never paid any attention to it in the first place. I couldn't care less what the statistical model created by some moron says.
-
How did they do head to head this season?
-
How did they do head to head this season?
Houston won 2 games lost one game. All three games were competitive.
-
The idea that an ELO rating is garbage because it rates a team with 7 more wins, better SRS, Net Rating than another team means you're just mad that GSW mailed in much of their season and most of all that Curry got hurt.
The GSW are -160 favorites to win the series FWIW.
-
https://projects.fivethirtyeight.com/2018-nba-predictions/
They currently have the Rockets at 80% and the Warriors at 20% to make the Finals. I get this is based on stats, but that just seems crazy and really highlights the flaws in stat based models.
The flaw being that stats are based on actual performance and GS mailed it in during the regular season? I mean untill we figure out how to factor in coasting stat based models are always going to have problems.
-
The idea that an ELO rating is garbage because it rates a team with 7 more wins, better SRS, Net Rating than another team means you're just mad that GSW mailed in much of their season and most of all that Curry got hurt.
The GSW are -160 favorites to win the series FWIW.
The idea is that the model is garbage because it is incapable of understanding why something happened, not that it happened.
-
You don't evaluate models like that.
Obviously the numbers are crazy and nowhere near the truth but you don't disregard a prediction/estimation model because it got one or two very wrong.
I don't follow this site so maybe this has been a reccuring issue but the other predictions look pretty good.
Before the playoffs, they had the Rockets at 65% of making the Finals, the Raptors were 2nd at 44%, followed by the Sixers at 23%, Cavs at 20%, Jazz at 10%, Warriors at 7%, and I suppose there is not point in going forward. So in sum, the team responsible for the most dominant 3 consecutive regular seasons ever (including 2 titles over those seasons), 538 predicted would make it to where they are at just 31%, with a 7% chance to go to the next round, and only a 4% at winning the title. heck they only had them at 62% to win their 1st round series against the Spurs (which is odd because the Spurs they had at 42%). They only gave the Warriors a 60% chance to beat the Pelicans.
I'm not sure they got the other predictions right.
I frequently quibble with Moranis but I’m inclined to agree here. 7% for GS to make the finals is silly.
The model might be ok in most cases but that’s a silly projection. Even if Houston sweeps, it will still seem low to me, and I don’t expect Houston to sweep or even win. Looking forward to the WCF (and assuming we make it, the ECF obviously).
-
The idea that an ELO rating is garbage because it rates a team with 7 more wins, better SRS, Net Rating than another team means you're just mad that GSW mailed in much of their season and most of all that Curry got hurt.
The GSW are -160 favorites to win the series FWIW.
The idea is that the model is garbage because it is incapable of understanding why something happened, not that it happened.
Yeah, that is what I'm getting at. I'm not sure the fix, but any time a model is based on something which is clearly flawed, its predictive factors disappear. If they were just using the model to say who had a better season, then it would be fine, but they use the model to predict future events, and in that case what happens in a season should factor in, but can't be the only thing. The Warriors coasted, took time off, had players out with injury, and frankly just didn't care. This is something that happens a lot with the better teams, especially as the seasons of success start getting larger. You can't predict the playoffs based solely on the regular season for that reason. Now maybe if they took prior year's performances, especially in the playoffs, and then somehow combined that with the current season, the predictive nature of the model would make more sense (and would thus disregard the Warriors poorer statistical season).
-
When Chris Paul has been healthy and playing full strength, Houston has been the best team in the league all season. It is not surprising that a model based on past performance would have them as the best team.
It is also not surprising that the Vegas odds favor the Warriors. Remember that Vegas odds reflect how they think people will bet, not who they think will win.
Should be a great series. I think Houston can beat them although I would not put it at 80% likely. Perhaps more like 55% likely.
-
Hopefully, the model is constantly being tweaked to try to gain accuracy. Also, it's not an all or nothing; it is a helpful tool to ask "why" and look for reasons to agree or disagree with the model. Making a wrong prediction doesn't necessarily invalidate the entire model.
If a statistical model told you that if you flipped a coin 10 times you should get 5 heads and 5 tails, but then you actually flipped the coin and got 7 heads and 3 tails...would you throw out the model?
-
I'm fine with any model that tells me the Cavs have only an 11% chance of winning the title.
TP, but you don't need a model to tell you that. ;)
-
When Chris Paul has been healthy and playing full strength, Houston has been the best team in the league all season. It is not surprising that a model based on past performance would have them as the best team.
It is also not surprising that the Vegas odds favor the Warriors. Remember that Vegas odds reflect how they think people will bet, not who they think will win.
Should be a great series. I think Houston can beat them although I would not put it at 80% likely. Perhaps more like 55% likely.
Just bumping this thread as there is now a lot different context for this discussion with the series 2-2 and Houston having home court. This still feels like pretty much an even call to me with Houston having a slight edge due to home court.
I have not had a chance to actually watch any of the games and I rely a lot on what I see vs box scores but what I saw during the regular season told me Houston is a really good team when they have a full speed Chris Paul. Warriors are really good too of course but Paul seems to be neutralizing Curry enough and that leaves Harden to do his thing and Durant to do his thing.
I have always thought Green and Thompson were overrated; they are good players, just not great. Houston does not have that much after Harden and Paul though. I hope Houston beats them.
-
They are dubious at even politics, at times and misinterpret data due to their skews a lot. I would not trust them on sports. That being said I like to read the site sometimes.
-
This prediction business is fickle. Nate Silver went from a star to a scrub in one night (presidential election). I think he's somewhere in between- a good data analyst that will often get things wrong, like other people.
-
This prediction business is fickle. Nate Silver went from a star to a scrub in one night (presidential election). I think he's somewhere in between- a good data analyst that will often get things wrong, like other people.
For people who literally do not understand probability because they only understand the binary (wrong/right) this is true.
-
Hopefully, the model is constantly being tweaked to try to gain accuracy. Also, it's not an all or nothing; it is a helpful tool to ask "why" and look for reasons to agree or disagree with the model. Making a wrong prediction doesn't necessarily invalidate the entire model.
If a statistical model told you that if you flipped a coin 10 times you should get 5 heads and 5 tails, but then you actually flipped the coin and got 7 heads and 3 tails...would you throw out the model?
100%.
The idea that a model needs to "understand" something is like asking a hammer to understand something. It is just a sophisticated and useful tool. You as a person looking at the model need to use it appropriately.
Given that its a relatively straightforward ELO system with adjustment for opponent, rest, etc just means its only a little different than point differental, SRS, etc.
The Rockets are really good and while I do think the Warriors are the better team overall they are now looking like they might pull off the upset.
-
Hopefully, the model is constantly being tweaked to try to gain accuracy. Also, it's not an all or nothing; it is a helpful tool to ask "why" and look for reasons to agree or disagree with the model. Making a wrong prediction doesn't necessarily invalidate the entire model.
If a statistical model told you that if you flipped a coin 10 times you should get 5 heads and 5 tails, but then you actually flipped the coin and got 7 heads and 3 tails...would you throw out the model?
100%.
The idea that a model needs to "understand" something is like asking a hammer to understand something. It is just a sophisticated and useful tool. You as a person looking at the model need to use it appropriately.
Given that its a relatively straightforward ELO system with adjustment for opponent, rest, etc just means its only a little different than point differental, SRS, etc.
The Rockets are really good and while I do think the Warriors are the better team overall they are now looking like they might pull off the upset.
Right. As an interpreter of the stats, you get to take any prediction model in context.
So, the "general consensus" is GSW is the best team, but 538 had Houston favored, 80-20. So, if that runs counter to the eye test, what is this predictive model seeing that we aren't, and/or what are we seeing that the predictive model isn't.
Likely, the predictive model is seeing how good Houston's defense secretly is (despite the offense getting the hype), how surprising good Capella is (his impact I think outpaces his budding reputation), and assigns some historically determined value to home-court advantage.
Now, it is also missing the fact that the Warriors may have "coasted," or, at the very least, had some very very good currently healthy players miss a lot of time due to rest and injuries.
So you can take this as learning something we might have missed about Houston, but also us knowing something vital that the model might not know, and make our own predictions.
But as I've mentioned before, just because a model says with 1 dice role, you have a 5/6 chance of not rolling a 1, then you go and roll a 1, does not mean the prediction model was "wrong" or "Broken."
-
Hopefully, the model is constantly being tweaked to try to gain accuracy. Also, it's not an all or nothing; it is a helpful tool to ask "why" and look for reasons to agree or disagree with the model. Making a wrong prediction doesn't necessarily invalidate the entire model.
If a statistical model told you that if you flipped a coin 10 times you should get 5 heads and 5 tails, but then you actually flipped the coin and got 7 heads and 3 tails...would you throw out the model?
100%.
The idea that a model needs to "understand" something is like asking a hammer to understand something. It is just a sophisticated and useful tool. You as a person looking at the model need to use it appropriately.
Given that its a relatively straightforward ELO system with adjustment for opponent, rest, etc just means its only a little different than point differental, SRS, etc.
The Rockets are really good and while I do think the Warriors are the better team overall they are now looking like they might pull off the upset.
Right. As an interpreter of the stats, you get to take any prediction model in context.
So, the "general consensus" is GSW is the best team, but 538 had Houston favored, 80-20. So, if that runs counter to the eye test, what is this predictive model seeing that we aren't, and/or what are we seeing that the predictive model isn't.
Likely, the predictive model is seeing how good Houston's defense secretly is (despite the offense getting the hype), how surprising good Capella is (his impact I think outpaces his budding reputation), and assigns some historically determined value to home-court advantage.
Now, it is also missing the fact that the Warriors may have "coasted," or, at the very least, had some very very good currently healthy players miss a lot of time due to rest and injuries.
So you can take this as learning something we might have missed about Houston, but also us knowing something vital that the model might not know, and make our own predictions.
But as I've mentioned before, just because a model says with 1 dice role, you have a 5/6 chance of not rolling a 1, then you go and roll a 1, does not mean the prediction model was "wrong" or "Broken."
Game 5 in Houston could make 538 look very smart. I think Golden State takes it, but I certainly wouldn't bet on it (I did bet Houston +9 last night, shoulda bet the ML, oops).
-
At the beginning of the playoffs it had us with 5% or 1% chance of making the finals
-
I don't think we should disregard 538. I think they should be considered like any other "valued added" (ie not Twitter) source of information, regardless of their conclusion. Nate Silver and his staff are respected, though people (myself included) gave them more credit than they were due. I think this is b/c they nailed the Obama election.
I think the Warriors will beat the Rockets. I'm not sure there's a model in the world useful in predicting whether a player hits or misses a clutch shot at the end of game 7.
-
Well if the Rockets win Game 5 at home... then what? The Warriors will be in a precarious situation where they have to win 2 in a row, including a road win.
-
I stopped reading 538 when they had Hilary winning convincingly only to have Trump win.
-
What are the best models that predicts sports outcomes? Is 538 one of them?
-
What's 538?
-
If a statistical model told you that if you flipped a coin 10 times you should get 5 heads and 5 tails, but then you actually flipped the coin and got 7 heads and 3 tails...would you throw out the model?
I'd want to know why I got an unexpected result. Random probability or a specific reason, or set of reasons?
Maybe the coin you had was weighted. Maybe it was being manipulated by the thrower. Who knows?
-
538 is junk. ESPN analytics as a whole are mostly laughable.
-
if the warriors lose a single playoff series during the next 5 years, they should all retire. That's just shameful. That team is historically stacked.