Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: For 10 World Cups, my model's 2 favorites had the champion every time

papers.ssrn.com

Post reply on HN
Print view View original

11–20 of 53 posts

Re: Show HN: For 10 World Cups, my model's 2 favorites had the champion every time

#11

Post by swiftcoder »

derdi wrote: post #10

> Applied prospectively to the in-progress 2026 World Cup from the Round of 32, the model identifies Argentina (28.0%) and Spain (21.1%) as the leading championship candidates. Seems weird to wait to run the "prospective" simulation until the World Cup is already in progress. Although it seems that the model also needs to use "the actual bracket and group-stage performance". So it's not prospective?

It predicts likely winners based on the round of 32 performance (plus prior data). That's still "prospective" with respect to the finals

Re: Show HN: For 10 World Cups, my model's 2 favorites had the champion every time

#12

Post by mcphage »

dmurray wrote: post #6
mcphage wrote: post #3

How does this paper not even mention the word "overfitting"?

The abstract does say > limitations, principally the small number of tournaments available for validation and the risk of in-sample weight selection But I agree this model is no more valuable than Paul the Octopus.

That almost makes it worse—like they're vaguely aware that training too heavily on too small a data set makes badly trained models, but are unaware that it has a name and is an actual identified problem.

Re: Show HN: For 10 World Cups, my model's 2 favorites had the champion every time

#13

Post by derdi »

swiftcoder wrote: post #11
derdi wrote: post #10

> Applied prospectively to the in-progress 2026 World Cup from the Round of 32, the model identifies Argentina (28.0%) and Spain (21.1%) as the leading championship candidates. Seems weird to wait to run the "prospective" simulation until the World Cup is already in progress. Although it seems that the model also needs to use "the actual bracket and group-stage performance". So it's not prospective?

It predicts likely winners based on the round of 32 performance (plus prior data). That's still "prospective" with respect to the finals

Yes. I don't like phrasing this as being prospective for the World Cup as a whole. It's for the knockout stage. (Which the abstract says! But the title doesn't.)

Re: Show HN: For 10 World Cups, my model's 2 favorites had the champion every time

#15

Post by yardie »

Does the model account for the blatant favouritism in the refs? We used to laugh about it before but as the cameras have gotten better it has become a lot more visible. And in this case, is turning the tournament into a bit of a joke.

-- Egypt was robbed.

Re: Show HN: For 10 World Cups, my model's 2 favorites had the champion every time

#20

Post by glimshe »

swiftcoder wrote: post #11
derdi wrote: post #10

> Applied prospectively to the in-progress 2026 World Cup from the Round of 32, the model identifies Argentina (28.0%) and Spain (21.1%) as the leading championship candidates. Seems weird to wait to run the "prospective" simulation until the World Cup is already in progress. Although it seems that the model also needs to use "the actual bracket and group-stage performance". So it's not prospective?

It predicts likely winners based on the round of 32 performance (plus prior data). That's still "prospective" with respect to the finals

Which is very reasonable. You estimate odds after seeing teams playing with the actual squad selection at that period in time. Otherwise I'd dismiss the predictions as lucky guesses in a row.
Post reply on HN