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

[tomas]

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


By: Cameron DaSilva | 1 hour ago

Last season, for the first time in more than a decade, the Los Angeles Rams made the playoffs. They did so by winning 11 games and taking home the NFC West title, supplanting the Seahawks atop the division.

It was a campaign that caught everyone off guard, one that almost no one predicted for the Rams. Yet, as good as they were in 2017, the Rams should be even better this season after adding Ndamukong Suh, Marcus Peters, Aqib Talib and Brandin Cooks.

Analytics website FiveThirtyEight isn’t quite as high on Los Angeles this season as fans and media members are. According to the site, the Rams have just a 39 percent chance to make the playoffs with a projected win total of only 8.2, which ranks 15th in the NFL.

To deflate fans’ hopes even more, FiveThirtyEight gives the Seahawks a better chance to make the playoffs (48 percent) and win the division (37 percent) than the Rams. Seattle’s projected win total is 8.9, nearly a full point higher than the Rams’.

Brace Hemmelgarn-USA TODAY Sports

Based on Seattle’s crumbling defense and lack of stability on offense, it would seem the Rams are the runaway favorites in the NFC West. If anything, the 49ers seem like the next-best team in the division, not the Seahawks. However, FiveThirtyEight puts San Francisco’s win total at 7.3 games, less than the Cardinals’ projected total (7.4). The 49ers have a 25 percent chance to make the postseason, while Arizona’s odds sit at 27 percent.


No matter how you cut it, FiveThirtyEight is going against the grain out West. The Rams appear to be Super Bowl contenders, but the site believes it’s more likely that they miss the playoffs than making it.

https://theramswire.usatoday.com/2018/09/05/nfl-rams-playoff-odds-seahawks-fivethirtyeight/

 

[Rmfnlt]

projected win total of only 8.2

As soon as they start counting fractions of games in the won/loss column, I'll pay attention to these bozos.

Until then? I'm gonna "deep six" FiveThirtyEight (whatever the hell that stands for).

 

[rking4441]

What were their odds during the 2016 elections? I think the Rams and their fans will be just fine!

 

[yrba1]

I like FiveThirtyEight for their political analysis but this is rather unusual for them. Jay Boice is the one who published this here:

https://projects.fivethirtyeight.com/2018-nfl-predictions/

Curious to hear what variables he factored in having a low chance of playoff berth for a talent-stacked team.

 

[Malibu]

Ridiculous statement. 538 is a hack political polling site who clearly predicted that Hillary would kill Trump.

If the reporter spent anytime analysising the Rams it would be obvious to him why the Rams should be minimum of 2 games better this year just the acquisition of 4 probowl defensive stars alone should be reason enough.

Just real lazy reporting.

 

[Prime Time]

Ha Ha Nate Silver. He's a terrible prognosticator. I'm surprised he hasn't already quit in disgrace, changed his name, and gone to live on a desert island somewhere. I'm not a gambling man but if I was I'd do the opposite of whatever he predicts and retire rich.

 

[hotanez]

Ha Ha Nate Silver. He's a terrible prognosticator. I'm surprised he hasn't already quit in disgrace, changed his name, and gone to live on a desert island somewhere. I'm not a gambling man but if I was I'd do the opposite of whatever he predicts and retire rich.

lol at his hair

 

[bluecoconuts]

I like FiveThirtyEight for their political analysis but this is rather unusual for them. Jay Boice is the one who published this here:

https://projects.fivethirtyeight.com/2018-nfl-predictions/

Curious to hear what variables he factored in having a low chance of playoff berth for a talent-stacked team.

These types of statistical analysis are becoming much more popular after the success of 538 in the 2012 election (which was the event that basically put him on the map for most of the population)... Part of the issue with them (as they openly admit) is that they're slow to react to changes as they look heavily on previous datapoints. That's why for political polling they may be slow to react to a shift in a particular race.

When it comes to sports, that means that they often will look at previous seasons to try and judge where a team might be. Hence why so many of these types of "season predictions" are showing higher than expected win totals for teams like Seattle (who are trending down quickly) and lower than expected win totals for teams like the Rams (who are trending upwards).

Additionally, an issue with this specific model, is that they don't have a way (and therefore don't at all) of accounting for the offseason that teams have. They could track all the changes but it's too much work (they would need to determine how much each player helps or hurts their teams chances at winning on any given day) to be worth it for them most likely. So what they do instead is simply just revert the teams 1/3rd of the way back to the mean, meaning that bad teams will always trend up and good teams will always trend down at the end of the model, and a team can trend up regardless of their offseason or a team could trend down regardless of their offseason.

So as far as this model knows, the Rams are, recently, a bad team that had a good year last year and got worse over the offseason.

Every football fan knows that the first two points are correct and the last is wrong. But the model aint a football fan.

So basically these model predictions are stupid, at least for the NFL.

 

[RamFan503]

What is your favorite colour? Right... you are out of the playoffs. The correct answer was puke green. That’s worth 3.87692 losses.

 

[Karate61]

Part of the issue with them (as they openly admit) is that they're slow to react to changes as they look heavily on previous datapoints.

They must think Fisher is still coach!

 

[RamFan503]

They must think Fisher is still coach!

Or maybe Linny?

 

[thirteen28]

View: https://www.youtube.com/watch?v=soXMCkoWfQo

 

[sjm1582002]

538 in action.

 

[jrry32]

Y'all need to chill. 538 does a good job. They even admit the flaw in their preseason predictions:
At the start of each season, every existing team carries its Elo rating over from the end of the previous season, except that it is reverted one-third of the way toward a mean of 1505. That is our way of hedging for the offseason’s carousel of draft picks, free agency, trades and coaching changes. We don’t currently have any way to adjust for a team’s actual offseason moves, but a heavy dose of regression to the mean is the next-best thing, since the NFL has built-in mechanisms (like the salary cap) that promote parity, dragging bad teams upward and knocking good ones down a peg or two.

That's why Seattle isn't getting hammered for their shitty off-season and we aren't getting credit for our great one. We play a very difficult schedule this year if you base it on last year's results.

 

[Karate61]

We don’t currently have any way to adjust for a team’s actual offseason moves,

Maybe they should get a way???

 

[rams1fan]

I read their methodology. Basically at this point there algorithm uses last years numbers. It does not take into account any off-season moves.

 

[RamFan503]

Y'all need to chill. 538 does a good job. They even admit the flaw in their preseason predictions:
At the start of each season, every existing team carries its Elo rating over from the end of the previous season, except that it is reverted one-third of the way toward a mean of 1505. That is our way of hedging for the offseason’s carousel of draft picks, free agency, trades and coaching changes. We don’t currently have any way to adjust for a team’s actual offseason moves, but a heavy dose of regression to the mean is the next-best thing, since the NFL has built-in mechanisms (like the salary cap) that promote parity, dragging bad teams upward and knocking good ones down a peg or two.

That's why Seattle isn't getting hammered for their crappy off-season and we aren't getting credit for our great one. We play a very difficult schedule this year if you base it on last year's results.

So in other words - damn near useless. If they admit that their methodology is flawed, what is the use in putting it out there? If your system pulls down good teams and lifts up bad teams based on limited information and a known absence of data, it is at best fatally flawed. @flv nailed it IMO.

Seriously. If you are going to put something out there as a predictor, it should at least attempt to accurately predict.

 

[RamFan503]

Y'all need to chill. 538 does a good job.

And no WE don't and no THEY don't - as far as the NFL is concerned.

 

[jrry32]

So in other words - damn near useless. If they admit that their methodology is flawed, what is the use in putting it out there? If your system pulls down good teams and lifts up bad teams based on limited information and a known absence of data, it is at best fatally flawed. @flv nailed it IMO.

Seriously. If you are going to put something out there as a predictor, it should at least attempt to accurately predict.

You put it out there to continue developing it and trying to work out the kinks.

And no WE don't and no THEY don't - as far as the NFL is concerned.

Yep, they do.