FiveThirtyEight gives Rams 39% chance to make playoffs, less than Seattle

[sjm1582002]

I'd like to see what 538 had "predicted" for the odds of Trump winning 302 electoral votes.

I bet it was way less than 1%.

Reminds me of when all the pollsters blew it with their "it's too close to call" prognostications right before Reagan buried Carter in a massive landslide.

That's 2 of out 7 completely blown calls.

I bet my blindfolded dart thrower has a 50/50 chance to equal that lousy track record.

 

[Rmfnlt]

Maybe they should get a way???

Or STFU.

I agree with those that have said that, taking into account their caveat that their analysis may have flaws, totally invalidates the predictions.

Basically saying that we have little idea whether the car we built will make it to 100 miles before falling apart... but here it is! Why would anyone buy that? And why would anyone buy the admittedly flawed product this firm is promoting?

In other words, why would anyone buy their BS?

 

[PhillyRam]

They actually gave Trump his best odds of winning and it was still less than 10% I believe if I remember correctly.

 

[RamFan503]

It's kind of weird to take a swipe at them on elections in light of:
https://fivethirtyeight.com/feature...rump-a-better-chance-than-almost-anyone-else/

So they were WAY off but closer than those they chose to brag about being farther off than them. Color me unimpressed.

It really begs the question though. Apparently they have the capability to adjust their political predictions within hours. So why exactly can’t they even factor in changes that occurred months ago in the NFL?

Sorry Jrry. It’s a crap product and even if they get lucky and hit on a few, it’s still crap. Defend it all you want.

They actually gave Trump his best odds of winning and it was still less than 10% I believe if I remember correctly.

20% and then at the last moment upped it to 29%. It appears that they were one of the leastest wronger options.

 

[RamFan503]

Basically saying that we have little idea whether the car we built will make it to 100 miles before falling apart... but here it is! Why would anyone buy that? And why would anyone buy the admittedly flawed product this firm is promoting?

This analogy spot on.

 

[jrry32]

So they were WAY off but closer than those they chose to brag about being farther off than them. Color me unimpressed.

It really begs the question though. Apparently they have the capability to adjust their political predictions within hours. So why exactly can’t they even factor in changes that occurred months ago in the NFL?

Sorry Jrry. It’s a crap product and even if they get lucky and hit on a few, it’s still crap. Defend it all you want.

20% and then at the last moment upped it to 29%. It appears that they were one of the leastest wronger options.

That's not really how statistics work. For example, I may have a 10% chance of correctly predicting a certain number of Week 1 NFL games correctly. If I manage to do so, it doesn't mean my percentage chances were wrong.

 

[jrry32]

Or STFU.

I agree with those that have said that, taking into account their caveat that their analysis may have flaws, totally invalidates the predictions.

Basically saying that we have little idea whether the car we built will make it to 100 miles before falling apart... but here it is! Why would anyone buy that? And why would anyone buy the admittedly flawed product this firm is promoting?

In other words, why would anyone buy their BS?

You're not buying it. It's free. And it's not like a car. There's no danger or inconvenience to anybody if they're wrong. What they gain is the ability to see the results and try to improve their model based on the flaws they identify. Hell, that happens surprisingly often with video games, which people actually buy. When they're released, gamers play it and identify bugs. The company then patches the bugs.

 

[OldSchool]

So I’m confused, if they take their last rating the previous year and regress backwards how does Seattle end up higher than us? Sounds like they’re eating Seattle was better than us to end last season which again doesn’t make sense. There are multiple teams rated higher than us that this applies to just adding to the confusion.

 

[1maGoh]

So I’m confused, if they take their last rating the previous year and regress backwards how does Seattle end up higher than us? Sounds like they’re eating Seattle was better than us to end last season which again doesn’t make sense. There are multiple teams rated higher than us that this applies to just adding to the confusion.

Regress toward the average. Below average comes up closer to average. Above average goes down. Bad word choice on their part.

 

[TSFH Fan]

This should be fun:
Here's 538's election night Forecast: https://projects.fivethirtyeight.com/2016-election-forecast/

-----

As for the football predictions, it's sport, in other words, fun and games. We're not talking gospel here. As far as I can tell the 538 people aren't putting money where their ELO is. If people don't like the forecasts, oh well. No need to buy into it. If people think they can do a better job at predicting, do it: https://projects.fivethirtyeight.com/2018-nfl-forecasting-game/

-----

Regression toward the mean -- https://en.wikipedia.org/wiki/Regression_toward_the_mean

 

[Karate61]

You're not buying it. It's free. And it's not like a car. There's no danger or inconvenience to anybody if they're wrong. What they gain is the ability to see the results and try to improve their model based on the flaws they identify. Hell, that happens surprisingly often with video games, which people actually buy. When they're released, gamers play it and identify bugs. The company then patches the bugs.

Bad analaogy. You can't repair a forecast or work the bugs out of a prediction once its occurred. You got one shot at it.

I think the best analogy to fivethirtyeight is FAKE NEWS! Lying by Omission!

 

[jrry32]

Bad analaogy. You can't repair a forecast or work the bugs out of a prediction once its occurred. You got one shot at it.

I think the best analogy to fivethirtyeight is FAKE NEWS! Lying by Omission!

It's the statistical model that I'm discussing. The predictions are simply what it spits out.

 

[RamFan503]

So I’m confused, if they take their last rating the previous year and regress backwards how does Seattle end up higher than us? Sounds like they’re eating Seattle was better than us to end last season which again doesn’t make sense. There are multiple teams rated higher than us that this applies to just adding to the confusion.

Yeah..., I don’t think you are. I think they may be however.