Week 33 A 4-2 win at Elland road over Wolves, moves the White up to 10th in the league on 41 points. Nottingham Forest disastrous continues, as they lose 3-1 at home to Burnley, with Burnley reducing the gap between themselves and Forest to 5 points Everton poor run of home draws continues as the Merseyside derby ends all square at 2-2, allowing Fulham to take pole position at the top of the table , following their 2-1 victory away at "Keith Andrews" Brentford. Spurs defeat at home to Brighton means they slip out of the European qualifying positions, with Chelsea taking advantage and moving up to 5th. Manchester City's win over Arsenal puts them 5 points clear of the drop zone, whilst Arsenal remain 2 points of relegation, level on points with Scum.
Week 34 A 1-1 draw away to Bournemouth sees Leeds slip one place to 11th in the league. A nail biting game at the Stadium of Light ends with victory at Last for Nottingham Forest as they overcome Sunderland 3-2, and moves Forest out of the relegation zone, and replaced by Scum, after they were held to a goalless draw at home to "Keith Andrews" Brentford. Manchester City move above Leeds after they earned all 3 points in a 7 goal thriller at Turf Moor, with Burnley now 6 points adrift at the bottom of the league and realistically 9 points from safety. A victory for Burnley in the next game away to Leeds United is a must to keep the dream alive of maintaining Premier League football next season. Fulham were held 2-2 at home to Aston Villa, although Everton were unable to capitalise upon this, following their 3-3 draw away to West Ham.
Week 35 With just 3 Premier league games remaining, Burnley soiree came to an end when following a 1-0 defeat away to Leeds, their relegation was confirmed. Leeds move 6 points clear of the relegation zone. Scum's 3-0 victory against Liverpool moves then up to 14th and Nottingham Forest drop back into the bottom 3. Arsenal defeat at home to Fulham, puts them in 18th place, 1 point above the drop zone, whilst Fulham maintain their 2 point lead over Everton, who triumphed 4-0 at home to Manchester City. Spurs move above Liverpool on goal difference following their 4-2 win at Aston Villa, as the race for the final Champions League qualification place hots up.
When the season starts I'm sure a man of your calibre can provide a running comparison of this v the actual season.
Well I'll be running and attempting to improve my software program throughout the season. The biggest issue with it is trying to get the weightings right for the start of the season, as the first 8 games have a major effect over the predictions for the rest of the season, as the predictions are based on league position and points gained from a database of over 10,000 games, which is probably still a small sample size. Of course me being me, I'm more than happy to leave myself open for ridicule, so yes comparisons will be posted as the season progresses I'm relatively confident that Fulham and Everton won't be vying for the title when the actual season starts. I'm surprised that my program has Everton doing so well, given that they lost 3 of their first 4 games, thus there were no expectations that they would then go on a run of 31 games without defeat
It was an honest question, just out of sheer interest. No attempt to leave you open to ridicule, especially given all of the effort yhat you put into this site.
i probably didn't word that as well as I probably intended. Think for the past 7 season whenever I've done my prediction every result of which ever league we are in, I always do a review at the end of the season to show how good or more likely poorly I did. Thus I openly ridicule myself with stuff like this
Week 36 Despite 1 3-1 defeat at Spurs, Leeds were confirmed to be almost guaranteed to be playing Premier League football next season, with Bournemouth losing 2-0 away to league leaders Fulham, and Forest losing at home to Newcastle 3-2. Whilst that leaves both clubs 6 points behind Leeds with 2 games to play, Forest play Bournemouth on the last game of the season and due to a superior goal difference it is unlikely that either of these clubs will catch the Peacocks. Already relegated Burnley played with heart but were beaten 3-2 at home to Aston Villa and Everton's 3-1 win at Crystal Palace, keeps them 2 points behind Fulham. Palace find themselves sucked into the relegation survival race, having only gained 1 point out of the last 12, and are level with Scum, both clubs 2 points clear of the drop zone.
Week 37 With both Bournemouth drawing 3-3 at home to Manchester City, and Nottingham Forest drawing away at Scum 3-3, Leeds United despite losing at home 2-0 against Brighton, secured Premier League football next season, with 1 game remaining. With Nottingham Forest hosting Bournemouth in their last game of the season , Scum, Wolves and Crystal Palace, will be keeping an eye on that result, as those 3 teams still remain catchable. Everton's 5-0 thrashing of Sunderland, maintains their 2nd place, 2 points behind Fulham, who demolished Wolves at Molyneux 5-0. Everton's final game of the season is away to 4th place Spurs, whilst Fulham entertain 9th place Newcastle. Arsenal having flirted near the bottom of the league for nearly half the season, find themselves in the top half of the table after their 3-0 win over relegated Burnley.
Week 38 With Leeds already secure in their Premier league future, fans were disappointed with the display at West Ham, where the Whites lost 2-1, which saw Leeds end the season in 14th position. Relegated Burnley finished the season with a home win over Wolves , whilst Scum lost 6-1 away to Brighton. That goal scored by Scum was crucial as it guaranteed that they would remain in the Premier League courtesy of having scored 1 more goal than Wolves. With both Nottingham Forest and Bournemouth playing each other, only 1 of those sides had a chance of catching Wolves and avoiding relegation. However it was not to be as both sides fought for their Premier league futures, and the match ended in a thrilling 3-3 draw. Sunderland poor start to the season was forgotten after their final home game of the season saw them earn a respectful 1-0 victory over Chelsea, and saw the Black Cats end the season in 11th, although their local rivals Newcastle will have the bragging rights having finished one place higher up in 10th. Everton's 4-3 victory away at Spurs was not enough to catch Fulham who were crowned Premier League Champions following their 1-0 win over Newcastle. That defeat for Spurs however meant that Liverpool's victory over Brentford earned them the final Champions League place, with Spurs having to settle for Europa league qualification. Whilst this is only a bit of fun and I am well aware that any attempt trying to predict a seasons worth of football results and expect them to have any sort of accuracy, never mind a program written by a middle aged bloke, is prone to failure, there are a couple of things that need to be addressed. Obviously stats cant predict the future, they can just possibly show a trend for the past, providing of course you have enough data. All the above was calculated using over 12,000 games played in the Premier League since it began. Whilst that in itself may seem a lot, from a statistics point of view for the purposes of this experiment, I would say I would need about 8 times that many games, if I was to submit this as an University project. So the data set is small. Whilst a quick glance at the above league table looks fairly representative, obviously no one expects the likes of Everton or Fulham to be challenging for the title, and whilst the so called big teams can struggle ( as seen last season), I wouldnt expect Man City and Arsenals seasons to be portrayed anywhere like the program suggested. There is a big emphasis on form, as well as checking historic results based on previous meetings, points and league positions and that is difficult to simulate at the start of the season. It was only after I had started posting these, that I came up with a solution to reduce the random elements of those first 2 months of games, and those changes are not reflected in this. Finally we all know that if football was played on paper, predicting results would be easy, and we would all be winning Sky's Gillet Super Six every week, or the more sought after prize of bragging rights for winning Stonkins competition. Thus any attempt at trying to do a prediction program, no matter how accurate you are trying to be with your results, needs some element of random factor. The problem with that though is whilst you can take into account wealth, form , previous meetings, and use historic data to give some sort of basis for a prediction, how do you quantify the likes of significant changes to a club. I would say most people expect Brentford to struggle this season, due to both Thomas Frank leaving and replaced by an unproven manager. How can that be taken into account, without letting personal bias have any affect. I appreciate that not many people on here are that interested in the method behind stats (hey most arent interested in stats themselves ) or even this silly thread , but if any of those who are maths orientated have any ideas of a mathematical way to represent things like Frank going to Spurs, Wolves selling their best players etc, then I'm more than happy to listen to ideas and see if I can incorporate them. Anyways that is this thread done for this season, other than of course in respect of Brizzle's suggestion of comparing to how the real season unfolds, although given I've already identified issues, then if I manage to get one teams positions or points total I will be happy
So approx. 1/3rd of the season gone, so lets have a quick review. Anything highlighted in green, means that current position and points are within 3 points of my prediction, which I class as a success. Anything highlighted orange means that current position and points are out by at least 9, and thus deemed a total failure. Unfortunately after creating this thread, I spotted a flaw in the prediction method (there are of course still plenty of flaws left ), however with the updated method, I only recorded the final league table, so cannot make any comparisons at present. so 17 sets of data that I consider to be successful so far, 11 that are outright failures, and the remaining 12 being out enough to be considered wrong
