FiveThirtyEight's round-by-round probabilities
- Thread starter Plinnius
- Start date
A nonymous
Active member
- Joined
- Jan 17, 2017
- Messages
- 477
- Reaction score
- 169
- Points
- 43
We all know how accurate FiveThirtyEight was in their prediction of a certain event this past November
(this is only meant to be a joke)
Sent from my iPhone using Tapatalk
(this is only meant to be a joke)
Sent from my iPhone using Tapatalk
cncmin
Well-known member
- Joined
- Nov 20, 2008
- Messages
- 22,623
- Reaction score
- 6,892
- Points
- 113
I don't know what your implication is here, but 538's prediction last November was actually reasonably good, as it iterated close to 11/8. A 52% slot for the Gophers winning Thursday and an 18% chance that they win Th/Sa does not bode well for the team, if this is accurate. Of course, 18% is still very far from zero.
I think this poses a far bigger question, given the analysts' consensus that MN/MTSU is a virtual toss-up, even in a #5/#12 game - why the hell was MTSU slotted as a #12 seed?!
If by great chance you mean 1% chance, then yes
A nonymous
Active member
- Joined
- Jan 17, 2017
- Messages
- 477
- Reaction score
- 169
- Points
- 43
If you read my post, it clearly said I meant it as a joke. Reading the comments on the FiveThirtyEight page on election night, people seemed to think that Nate Silver's model prediction of ~72% (iirc) of Clinton winning was a virtual guarantee she would win the election. I just meant to poke fun people who don't understand probability.
Sent from my iPhone using Tapatalk
I wondered this too. People keep saying what a joke it is that the Gophers got a 5 seed, but it almost seems like Minnesota/MTSU could have been a 6/11 or 7/10 game, but they only had a 5/12 slot available.
mcal24
Well-known member
- Joined
- Oct 6, 2012
- Messages
- 1,966
- Reaction score
- 26
- Points
- 48
The whole election thing with 538 bothers me, I know the poster above was joking but people don't understand probability. If I tell you there is a 50% chance of flipping a coin and getting heads and then you flip tails, am I wrong?
Don't know how I feel about these projections though... 23% chance for Gonzaga to reach the finals seems high. It's all random though, no matter how much time you put into predictions you'll still get a good chunk of games wrong. I'm sure he'd do a great job in projecting a 10 game series between two teams, but in one game its impossible to guess.
Don't know how I feel about these projections though... 23% chance for Gonzaga to reach the finals seems high. It's all random though, no matter how much time you put into predictions you'll still get a good chunk of games wrong. I'm sure he'd do a great job in projecting a 10 game series between two teams, but in one game its impossible to guess.
BarnBoy
Active member
- Joined
- Jan 12, 2009
- Messages
- 4,218
- Reaction score
- 21
- Points
- 38
They were actually one of the only media outlets that gave Trump about a 20% chance and were ridiculed by other media for it.
(Edit: Just seeing the other posts about 538)
Are you sure about that? When I looked it said EIGHT PERCENT!!! Better percentage than both Minnesota, the #5 and Cincy, the #6 seeds, who were at 1% and 5%.
Dayton, the #7 seed, was at 8% to make the Final Four.
Cincy, the #6 seed, was at 5% to make the Final Four.
Minnesota, the #5 seed, was at 1% to make the Final Four.
Butler, the #4 seed, was at 9% to make the Final Four.
Absolutely NO RESPECT. Gotta love it.
BarnBoy
Active member
- Joined
- Jan 12, 2009
- Messages
- 4,218
- Reaction score
- 21
- Points
- 38
Now let's hope we don't go out a lay a big egg.
How on earth could you give MTSU a 10 seed? Or even an 11 seed? Their resume is HORRIBLE. I don't think they would have gotten in if they hadn't won their conference tourney.
Last year, AFTER beating MSU, they were still a 6 pt underdog vs Syracuse, but lost by 25!!!
mcal24
Well-known member
- Joined
- Oct 6, 2012
- Messages
- 1,966
- Reaction score
- 26
- Points
- 48
They really don't seem that good. 18 of their 33 games are against teams ranked lower than Rutgers in the RPI. Wouldn't it be nice to play Rutgers for over half your schedule?
They're also 2-1 against top 50 teams, with their best RPI win being UNC Wilmington from the Colonial Conference (UNCW doesn't even have an RPI top 50 win).
Sent from my XT1635-01 using Tapatalk
Look again because Dayton has a 1% chance at the final four according to that site. And a 7% chance at the round of 32.
What website are you talking about?
GopherWeatherGuy
Well-known member
- Joined
- Oct 24, 2013
- Messages
- 15,644
- Reaction score
- 14,240
- Points
- 113
Move your cursor, it's actually Wichita St. that has the 8% chance.
7% chance at the Sweet Sixteen for Dayton, 31% chance at the round of 32. Wichita State is a pretty heavy favorite over Dayton for being the "underdog" seed.
A nonymous
Active member
- Joined
- Jan 17, 2017
- Messages
- 477
- Reaction score
- 169
- Points
- 43
If you read the methodology of how 538 determines the probabilities, you will see that they heavily weight the location of the game. In addition to their own Elo rating (which rates MTSU as better than the gophers), they combine other computer rating systems as well as the tournament committee's S curve ratings and PRESEASON polls. Since they so heavily weight distance traveled for the games, they probably think MTSU is actually a better team than UMN but we are more likely to win because the game is closer to home for us. Kind of a strange way to do things, but statisticians come up with these things.
Personally I don't think 538's modeling is very rigorous statistically. There are a lot of assumptions that go into those models which probably aren't valid. They also use some inputs that seem off base (preseason coaches polls????).
I wouldn't go out basing my bracket off their predictions. Too much statistical noise. The thing about basketball games is that every game is unique in ways that can't be accounted for in these models.
Sent from my iPhone using Tapatalk
Personally I don't think 538's modeling is very rigorous statistically. There are a lot of assumptions that go into those models which probably aren't valid. They also use some inputs that seem off base (preseason coaches polls????).
I wouldn't go out basing my bracket off their predictions. Too much statistical noise. The thing about basketball games is that every game is unique in ways that can't be accounted for in these models.
Sent from my iPhone using Tapatalk
Now he even has me reading it wrong
Same website, I was wrong. We are both wrong in part. I was reading Wichita St's percentages, which is WORSE, seeing as they are the #10 seed. But the 7% you saw for Dayton was their percentage of making the Sweet 16, not the round of 32.
So I apologize for not taking more time to get my information right, but my point is pretty much the same, almost every team is getting more respect than the Gophers.
Wichita St, the #10 seed, was a 8% chance of making the FF.
Seton Hall, the #09 seed, was a 1% chance of making the FF & a 5% chance of making the Elite 8.
Arkansas, the #08 seed, was a 2 % chance of making the FF & a 5% chance of making the Elite 8.
Dayton, thethe #7 seed, was a 1 % chance of making the FF.
Cincinnati, the #6 seed, was a 5 % chance of making the FF.
Minnesota, the #5 seed, was a 1 % chance of making the FF & a 4% chance of making the Elite 8.
So FOUR of the FIVE teams seeded immediately higher than the Gophers in their bracket are given a better chance of advancing deep in the tourney than the Gophers, all except Dayton.