Glasgow Tims Supercomputer

Jan Žižka

Well-Known Member
Doesn't get much right. :))



Celtic are still projected to win the Scottish Premiership come the end of the season despite a 2-0 home loss to rivals Rangers.

The Hoops had been expected to run away with the title, with an initially predicted eight point gap when the campaign comes to an end. That was according to statistics site FiveThirtyEight who crunched the numbers and tipped the Hoops for a win.

Following Neil Lennon's outfit succumbing to Connor Goldson's double, the points value may have changed using the complex algorithm but the end result has not. Celts are still backed to seal their tenth title in a row - but just by ONE point.

Glasgow Times:


The innovative algorithm uses expected goals scored and conceded, transfer spend and previous season's results data to predict the outcome of various leagues and tournaments around the sporting world.

The FiveThirtyEight predicts Hoops will win the league title by one point, on 93 in front of Gers on 92. In its complexities, the algorithm even takes into account playing games this season behind closed doors and, as such, reflects a 60 percent reduction in home field advantage.

Celtic's goal difference has been predicted at +69 compared to the previous +74 while Gers are +64 compared to previous +59, a sway of just five goals.

St Mirren were initially predicted to be at the foot of the table come the end of the season, with Jim Goodwin's men relegated but Hamilton have slipped into the team expected for the drop.

 
Doesn't get much right. :))



Celtic are still projected to win the Scottish Premiership come the end of the season despite a 2-0 home loss to rivals Rangers.

The Hoops had been expected to run away with the title, with an initially predicted eight point gap when the campaign comes to an end. That was according to statistics site FiveThirtyEight who crunched the numbers and tipped the Hoops for a win.

Following Neil Lennon's outfit succumbing to Connor Goldson's double, the points value may have changed using the complex algorithm but the end result has not. Celts are still backed to seal their tenth title in a row - but just by ONE point.

Glasgow Times:


The innovative algorithm uses expected goals scored and conceded, transfer spend and previous season's results data to predict the outcome of various leagues and tournaments around the sporting world.

The FiveThirtyEight predicts Hoops will win the league title by one point, on 93 in front of Gers on 92. In its complexities, the algorithm even takes into account playing games this season behind closed doors and, as such, reflects a 60 percent reduction in home field advantage.

Celtic's goal difference has been predicted at +69 compared to the previous +74 while Gers are +64 compared to previous +59, a sway of just five goals.

St Mirren were initially predicted to be at the foot of the table come the end of the season, with Jim Goodwin's men relegated but Hamilton have slipped into the team expected for the drop.


Hope you’re Keeping well Jan.

Bloody brilliant this eh ?

Big major prediction for the season & they couldn’t get the first game right !!!
 
I think some of the information in that is wrong. Pretty sure they have Aberdeen on 10 points just now instead of 20.
 
How does Eduoard and Ntcham not giving a flying %^*& factor into ‘a complex algorithm’? Or Lennon’s boozing? or the year on year deterioration of Mongo? or Duffy being unable to play out with a back 4?

Some may say this entire ‘story’ represents a new low for the once proud profession of sports journalism.
 
Doesn't get much right. :))



Celtic are still projected to win the Scottish Premiership come the end of the season despite a 2-0 home loss to rivals Rangers.

The Hoops had been expected to run away with the title, with an initially predicted eight point gap when the campaign comes to an end. That was according to statistics site FiveThirtyEight who crunched the numbers and tipped the Hoops for a win.

Following Neil Lennon's outfit succumbing to Connor Goldson's double, the points value may have changed using the complex algorithm but the end result has not. Celts are still backed to seal their tenth title in a row - but just by ONE point.

Glasgow Times:


The innovative algorithm uses expected goals scored and conceded, transfer spend and previous season's results data to predict the outcome of various leagues and tournaments around the sporting world.

The FiveThirtyEight predicts Hoops will win the league title by one point, on 93 in front of Gers on 92. In its complexities, the algorithm even takes into account playing games this season behind closed doors and, as such, reflects a 60 percent reduction in home field advantage.

Celtic's goal difference has been predicted at +69 compared to the previous +74 while Gers are +64 compared to previous +59, a sway of just five goals.

St Mirren were initially predicted to be at the foot of the table come the end of the season, with Jim Goodwin's men relegated but Hamilton have slipped into the team expected for the drop.

@academical holds 538 in very high regard, they are losing a bit of credibility with this.
 
FWIW 538 is a decent stats modelling site. They’ll have no bias one way or the other. Their stats are predominately based off the previous results so it makes sense why they have the scum edging it. I fully expect as the season proceeds and their numbers update, we’ll start to see them predict that we will win the league. And then the papers will probably stop mentioning them ;)

I should also add that other sources, using purely this season’s data, are predicting we win the league by 10+ points.
 
FWIW 538 is a decent stats modelling site. They’ll have no bias one way or the other. Their stats are predominately based off the previous results so it makes sense why they have the scum edging it. I fully expect as the season proceeds and their numbers update, we’ll start to see them predict that we will win the league. And then the papers will probably stop mentioning them ;)

I should also add that other sources, using purely this season’s data, are predicting we win the league by 10+ points.
It's not much of a prediction tool if has to be updated after every game.
 
It's not much of a prediction tool if has to be updated after every game.
Surely it’s better that they keep their stats accurate? It would be pointless if they rigidly stuck to their pre-season predictions when they get more games and therefore more data to make note informed decisions.

Here’s one for everyone. At the start of the season, they projected a 78% probability that the scum would win the league. One full round of fixtures for us later and that’s plummeted to 54%... don’t you think it’s interesting that the media aren’t reporting on that side of things?
 
does the algorithm account for the dressing room split that is bound to come out shortly?
 
Guys, don't say a feckin word but I managed to get hold of the "code" they used to build this super computer program.

You can recreate this program if you have a ZX81 as well. Here is the code:

10 PRINT "Celtic 93"
20 PRINT "Rangers 92"

Try not to share it folks, we dont want them to know we know.
 
Why England Lose was very critical of using transfer fees in projecting league positions. Does it have a good track record of being useful/correct?
 
The Evening Times would report an expected alien invasion for hits / sales.

This is just about as low as I've seen them sink, and they've sunk A LOT.
 
Back
Top