| EADT reports 'supercomputer' prediction on 00:09 - Jul 29 with 1571 views | SteveH |
| EADT reports 'supercomputer' prediction on 08:58 - Jul 28 by DanTheMan |
Be interested to hear the methodology on how you generated the numbers, if possible. |
First the caveat - making predictions before a ball is even kicked is unlikely to produce accurate results.
In previous seasons, I started with general historical data, with % home wins, % away wins and how many wins/draws/losses are required to make the play offs. This always shows a significant home advantage. It also shows that even the top teams WILL lose during the season. The champions need to reach 95 points (average), which means they need to obtain 57points at home and 38points away. For play offs 78 points (H-47, A31). I would then rank the teams 1 to 24 and speculate whether Ipswich were over/under performing.
This year, I used the information learned, but the ranking was based on statistical evidence - even though I know this will be flawed!
The details I used
Goal scoring ability - with some adjustment for newly promoted teams and those relegated. (Ipswich ranked 8th :( )
Goals conceded - again with adjustments (Ipswich ranked 2nd)
Goal difference - this creates a completely different data set and helps to get a good average. (Ipswich rank 5th)
Pre-season betting - this gives an indication of the level of confidence in the clubs ability to win. Teams with larger fanbases rank better than those with smaller followings. Equally, I look at betting on Relegation as this is less likely to be biased.
The average, gave me a ranking order - which was then added into the data - which in turn changed the order slightly.
Based on these figures I was able to calculate a score for each fixture (with advantage to the home side, disadvantage to the away team).
The difference between the two figures determines the result. At this stage that is win/lose by 1 goal, or 2, or 3 or even 4 goals.
To run the simulation I also included a random element that effectively means that the team can do up to 50% better or worse than the ranking score dictates.
Of course as soon as the first goal is scored the prediction would change - especially when the predicted result does not match the actual result. Once the season starts other variables can be used, which will also change the prediction. |  |
|  |
| EADT reports 'supercomputer' prediction on 02:48 - Jul 29 with 1551 views | BondiBlue |
| EADT reports 'supercomputer' prediction on 16:15 - Jul 27 by The_Flashing_Smile |
I think this 'supercomputer' has crunched the old stats of these old pros and extrapolated from that, without any thought about their age/speed to gel. |
Age/speed to gel/motivation. They're all coming down from top end championship, even premier league teams on creaking hamstrings for a last payday. Accrington will make mincemeat of them. Mid-table, and that's only because mcgoldrick's a football genius. | |
| |
| EADT reports 'supercomputer' prediction on 08:54 - Jul 29 with 1507 views | GeoffSentence |
| EADT reports 'supercomputer' prediction on 00:09 - Jul 29 by SteveH |
First the caveat - making predictions before a ball is even kicked is unlikely to produce accurate results.
In previous seasons, I started with general historical data, with % home wins, % away wins and how many wins/draws/losses are required to make the play offs. This always shows a significant home advantage. It also shows that even the top teams WILL lose during the season. The champions need to reach 95 points (average), which means they need to obtain 57points at home and 38points away. For play offs 78 points (H-47, A31). I would then rank the teams 1 to 24 and speculate whether Ipswich were over/under performing.
This year, I used the information learned, but the ranking was based on statistical evidence - even though I know this will be flawed!
The details I used
Goal scoring ability - with some adjustment for newly promoted teams and those relegated. (Ipswich ranked 8th :( )
Goals conceded - again with adjustments (Ipswich ranked 2nd)
Goal difference - this creates a completely different data set and helps to get a good average. (Ipswich rank 5th)
Pre-season betting - this gives an indication of the level of confidence in the clubs ability to win. Teams with larger fanbases rank better than those with smaller followings. Equally, I look at betting on Relegation as this is less likely to be biased.
The average, gave me a ranking order - which was then added into the data - which in turn changed the order slightly.
Based on these figures I was able to calculate a score for each fixture (with advantage to the home side, disadvantage to the away team).
The difference between the two figures determines the result. At this stage that is win/lose by 1 goal, or 2, or 3 or even 4 goals.
To run the simulation I also included a random element that effectively means that the team can do up to 50% better or worse than the ranking score dictates.
Of course as soon as the first goal is scored the prediction would change - especially when the predicted result does not match the actual result. Once the season starts other variables can be used, which will also change the prediction. |
I reckon you would have been well up for this as a lad
https://boardgamegeek.com/boardgame/26794/logacta-chart-soccer
I loved it, somehow Ipswich won everything |  |
| |
| EADT reports 'supercomputer' prediction on 10:55 - Jul 29 with 1483 views | blueasfook |
| EADT reports 'supercomputer' prediction on 12:47 - Jul 28 by The_Flashing_Smile |
...and apparently I'm the one who starts arguments on here!
Anyhoo, nothing wrong with having second half of the season play off form. That's the best time to have it!
Reality is Derby are nothing like Paul Jewell's Ipswich. He wasn't starting with 5 players for one. |
Do you still sometimes cry about Paul Jewell leaving? |  |
| |
| EADT reports 'supercomputer' prediction on 11:02 - Jul 29 with 1463 views | clive_baker |
| EADT reports 'supercomputer' prediction on 16:47 - Jul 27 by Stu_boy |
Think Peterborough will be pushing hard this year, certainly better than the 10th place predicted with their attacking talent |
Agree. P'Boro 10th is a ridiculous suggestion. |  |
| |
| Sheff Weds, Bolton, Barnsley, Derby, Peterborough, Pompey and MK Dons!! on 11:03 - Jul 29 with 1462 views | unstableblue |
Include us, and that is 8 teams, who are very very capable of making top two and top six.
We fizzled out last term under McKenna, whilst other teams such as Bolton kept momentum and question marks remain about goals. Blame it on left wing back, blame it on striker availability/form... but we fizzled out. Last day against Charlton was an opposition who didn't turn up.
Add in Wycombe Wanderers, Charlton... and basically 3rd and play-offs would still be a bloody good achievement! but then alas we're in a lottery.
VERY VERY excited about the season and tomorrow... but HMS P!ss The League we are not going to see |  |
| |
| EADT reports 'supercomputer' prediction on 11:41 - Jul 29 with 1429 views | The_Flashing_Smile |
| EADT reports 'supercomputer' prediction on 10:55 - Jul 29 by blueasfook |
Do you still sometimes cry about Paul Jewell leaving? |
Do you think you'll ever stop being a tvvat? |  |
| Trust the process. Trust Phil. |
| |
| EADT reports 'supercomputer' prediction on 11:45 - Aug 22 with 1329 views | Parsley |
Just thought I'd have a look at this again to see how the predictions have changed.
Ipswich now favourites to win the league (29% probability)
48% chance of top 2
36% chance of finishing 3rd-6th
16% chance of finishing outside the playoffs
A lot can still change but very encouraging how things are shaping up
https://projects.fivethirtyeight.com/soccer-predictions/league-one/ |  | | |
Login to get fewer ads
| EADT reports 'supercomputer' prediction on 12:03 - Aug 22 with 1249 views | GeoffSentence |
It says we will get 90 points
Way off.
100 pts 100 goals. | |
| |
| EADT reports 'supercomputer' prediction on 12:09 - Aug 22 with 1222 views | GeoffSentence |
Based on their offensive and defensive ratings, they think we are on a par with Hull. | |
| |
| EADT reports 'supercomputer' prediction on 12:20 - Aug 22 with 1180 views | vilanovablue |
| EADT reports 'supercomputer' prediction on 12:09 - Aug 22 by GeoffSentence |
Based on their offensive and defensive ratings, they think we are on a par with Hull. |
I'm not claiming its merits btw :) |  | | |
| EADT reports 'supercomputer' prediction on 12:45 - Aug 22 with 1110 views | GeoffSentence |
| EADT reports 'supercomputer' prediction on 12:20 - Aug 22 by vilanovablue |
I'm not claiming its merits btw :) |
Just an observation, Hull being a lower-mid table Championship team at the moment. That's where they think we stand based on their rating system. Doesn't sound unreasonable to me. | |
| |
| EADT reports 'supercomputer' prediction on 12:48 - Aug 22 with 1094 views | LegendofthePhoenix |
only another 67 points for us then, that is an average of 1.63 pts per game for the remaining 41 games. What exactly is super about that supercomputer? |  |
| |
| EADT reports 'supercomputer' prediction on 14:00 - Aug 22 with 1028 views | GeoffSentence |
| EADT reports 'supercomputer' prediction on 12:48 - Aug 22 by LegendofthePhoenix |
only another 67 points for us then, that is an average of 1.63 pts per game for the remaining 41 games. What exactly is super about that supercomputer? |
The GS1 SuperBrain has been set to calculating this and its final report is that the promotion party will be 10th April at Cheltenham. | |
| |
| EADT reports 'supercomputer' prediction on 15:05 - Aug 22 with 973 views | Kropotkin123 |
| EADT reports 'supercomputer' prediction on 11:45 - Aug 22 by Parsley |
Just thought I'd have a look at this again to see how the predictions have changed.
Ipswich now favourites to win the league (29% probability)
48% chance of top 2
36% chance of finishing 3rd-6th
16% chance of finishing outside the playoffs
A lot can still change but very encouraging how things are shaping up
https://projects.fivethirtyeight.com/soccer-predictions/league-one/ |
Yeah, I don't think this computer is so super. It is currently 75th out of 76 entries in the 1-24 prediction league |  |
| |
| EADT reports 'supercomputer' prediction on 15:17 - Aug 22 with 930 views | BlueBoots |
| EADT reports 'supercomputer' prediction on 16:18 - Jul 27 by The_Flashing_Smile |
Absolutely. 10 games will tell us much more - and with Derby, especially, no-one can make a solid prediction. They could be anywhere from flying to struggling. |
I'll make a solid prediction for you...I'll break into Carrow Road and take a selfie of myself having a dump on the centre circle if Derby get 89 points this season... |  |
| |
| EADT reports 'supercomputer' prediction on 16:51 - Aug 22 with 840 views | Pinewoodblue |
Looks like you can download the data and do your own thing. If you have the right skill set. |  |
| |
| EADT reports 'supercomputer' prediction on 17:29 - Aug 22 with 776 views | mylittletown |
a) GIGO
b) Their first prediction - ITFC to beat Bolton was wrong, so presumably we now finish on 78 points? |  | | |
| EADT reports 'supercomputer' prediction on 20:29 - Aug 22 with 707 views | Nthsuffolkblue |
Where did the same people place us in our first season down as a matter of interest? How accurate were they? |  |
| |
| |
By continuing to use the site, you agree to our use of cookies and to abide by our Terms and Conditions.
