Premier League Predictions
“I’m confused by 24,” the supercomputer suddenly admits to me on a Thursday lunchtime. “The long-running early 21st century espionage drama series?” I reply, “Yeah, some of the later seasons could have done with a stricter script editor, that’s for sure!”
“No,” our machine mechanically counters, “What I mean is that Chelsea have played 23 games before Matchday 23, while some clubs haven’t even reached 20 games yet. This is illogical to me, a supercomputer.”
I’ll not go into details but, believe me, the machine was made to understand that Chelsea are in the Club World Cup so have played their 24th game before Matchweek 23 while many other sides are still reeling from the series of Omicron-fuelled match postponements that have left the Premier League table looking, let’s be honest, achingly cool with myriad possibilities. HOWEVER, all being well, we should have a full and fat 10-game matchweek this weekend, so let’s get stuck in and see just what is going to happen.
Losing five games in a row against Watford is some going but it’s what Norwich have been up to. After the dismal failure of the “three-horse title race,” the battle at bottom is refreshingly tight and it feels like at least one of the sides in this game will be back in the Championship next season. That makes this clash on Friday night monumental, and the supercomputer is heavily backing Watford, despite Claudio Ranieri’s side not having won a league game since they de-Solskjaered Manchester United in November. Norwich have only won one Premier League game fewer than Chelsea since the start of November, but that’s very much three-horse race chat, to be fair.
Saturday in the Premier League commences with a heritage clasico, Everton against Aston Villa at Goodison Park. Sixteen league titles between the clubs but none since the Channel Tunnel project was begun; you never know what forces you’ll unleash when you start to dig. One player who loved a “dig” from range was “Stevie G” and, although he is now retired and not allowed to shoot in the Premier League, he’ll be on the touchline at Everton, facing up to his old rival Big Duncan Ferguson, who is, of course, temporarily in charge of the listing Merseysiders. The supercomputer, in a suspiciously diplomatic and nervous manner, makes this a very even game, with Everton, boosted by war cries and sheer relief, slight favourites at 39.0% but it also points out that passion will be running at 110%!
Manchester United at three o’clock in the Premier League used to mean a comfortable stroll for the English record champions. But this season, due to philosophical entropy, things are different. United’s last two outings in this slot have seen them lose 4-1 at Watford and 4-2 at Leicester. They’re at home this weekend, though, and have to face David Moyes, a manager who has never won away at one of the Classic Big Four™ even when he was coaching one of them. That’s 64 games now, which is quite a lot. The supercomputer gives Moyes a one-in-four chance and his former employers a one-in-two of a won-a-3pm.
Leeds and Newcastle, who both employed James Milner before the iPad was invented, face off at Elland Road. If Leeds have vague relegation fears then Newcastle have vast, incredible relegation fears. Even if Eddie Howe’s team score first it’s rarely positive news: they’ve done so 10 times this season yet won just once. In that context, a supercomputer-approved 24.4% chance of a Newcastle win is promising, but let’s just wait and see what happens, right?
Who would win in a fight between a swarm of bees and a single wolf? The supercomputer doesn’t speculate and as usual Attenborough didn’t answer, so we’ll never know but we can see who out of Brentford and Wolverhampton Wanderers wins on Saturday. Wolves haven’t let in a first-half goal in their last 13 Premier League games, while Brentford did everything but score a first-half goal in their game with Manchester United in midweek. There’s barely the width of a bee between the sides, according to the win predictor model, with Brentford on a cool 34.5% and Wolves on an even cooler 35.5%.
Manchester City’s serene, horse-neutralising progression to the league title travels to Southampton on Saturday evening, with Kevin De Bruyne (nine goals and two assists in his last 14 Premier League starts) looking in ominous form. The supercomputer is incredibly confident that City will add another three points to their portfolio and gives Pep Guardiola’s side a hefty 73.5% chance of victory.
There are three games at 2pm on Sunday, which includes Arsenal and Burnley both embarking on their second league game of the year. Bottom side Burnley have conceded three goals fewer than fourth-placed West Ham, although they have played five games fewer. It’s that sort of season but there’s one constant: The bottom three teams are going to be playing in the Championship next season and Sean Dyche’s side can feel the rusty fangs of the hated trapdoor on their backs. The supercomputer has no good news for them, giving Arsenal a 56.4% chance of winning for the 13th time in their last 19 games against Burnley.
Also taking place in London at the same time but below the big river is Crystal Palace against Liverpool. Last season, Jurgen Klopp’s team won this fixture but then embarked on the longest spell of existential bad form Our League has seen for some time. Liverpool are the clear favourites on the win predictor, but 2021-22 Palace have matchwinners, if not always players who win matches. Trent Alexander-Arnold assisted both goals in Liverpool’s EFL Cup win at Arsenal on Thursday night and has as many assists in all competitions this season as Kevin De Bruyne, Jack Grealish, Harry Kane and Riyad Mahrez combined. Is that good? It certainly sounds good.
Completing the 2pm jamboree is Brighton’s trip to Leicester in the forward-thinking-managers-clasico. A meeting of two cities that used to contain nationally renowned velodromes, Leicester are currently stuck in a cycle of bad defending from set pieces, and Adam Webster’s 1,500-watt header against Chelsea in midweek suggests the Foxes could suffer further against Albion. The supercomputer is less sure, probably correctly calculating that the events of Wednesday evening when Leicester suffered the latest turnaround in Premier League history cannot happen again. That reminds me, I must install the funny_old_game.exe patch asap.
The Premier League weekend concludes with one of its most aggro matches as Chelsea take on Tottenham at Stamford Bridge. Famously (to experts) this fixture in 2016 saw the visitors pick up nine yellow cards, which remains a Premier League record for a team in a single game. Chelsea cruised past Spurs in the League Cup semi-final but Tottenham are the team with momentum in the league. Yes, they’ve only won one of their last 31 league games at Stamford Bridge but the supercomputer gives the home team only a 43.3% chance of winning, which is indicative of the slow malaise that has enveloped them as winter has progressed. “Maybe a trip to the Club World Cup will invigorate Chelsea,” the supercomputer muses. “Maybe it will,” I reply. “Maybe it will.”
Ok yes, 2022 has been pretty cool with the AFCON and the FA Cup and the Carabao Cup and the rearranged Premier League games from “2021” but what I’ve been craving is a good, hearty helping of an Entire Premier League Matchday. And if I’ve been eager, then what about our old friend the supercomputer, straining under 7mm of dust? The machine has missed scanning a set of fixtures and making a considered and industry-renowned judgement. Well, my friend, wait no more, because MD22 is upon us.
The weekend begins in Brighton on Friday night with a visit from the scourge of Sussex, the Crystal Palace football team. The south London side will travel down to Falmer using an A road and a motorway combo that often gives its name to this derby but which the supercomputer refuses to recognise. Brighton are rated at 38.5% to win, with Palace at 31.6%, which means the draw is a) 23 or b) 29.9%. Yes, it’s b).
Saturday starts with a fixture so massive that UEFA made it the Champions League final last season. That ended 1-0 to Chelsea, while the league encounter at Stamford Bridge ended 1-0 to Manchester City. For many fans at this game thoughts will turn to City legend Frank Lampard who scored six goals in 32 Premier League games for the club and also had an earlier spell with Chelsea. That’s ancient history for the predictor model, though, and it makes the reigning (Premier League) champions 58.1% to defeat the reigning (continental) champions, Thomas Tuchel’s side on a gruel-style percentage of 17.3% to win.
Wolves have only conceded 14 goals this season and still, theoretically, have a chance of beating Chelsea’s Premier League record of 15 conceded, set in 2004-05. That would mean Bruno Lage’s team not letting in a goal in the 19 games they have left which doesn’t sound likely but also, if you check Wolves’ recent results, sounds…likely?? Southampton are given a round 27.0% chance of winning this game which sounds fair enough until you realise that three times nine (9) is 27. Oh, supercomputer you cheeky unit.
Norwich vs. Everton is a fixture steeped in the betrayal of Mike Walker, manager of negative-goal-difference-but-top-three-in-1992-93 Norwich City, who then walked out on the Canaries the following season to take up a glamour role at Everton. It had been only six and a half years since Everton had been champions of England, so it made sense, but some people in the cathedral city have never forgiven Walker, or indeed anyone called Mike. In 2022 neither side is in great shape; Norwich haven’t scored for six games, Everton are 15th and in a rut. But in the Norfolk flatlands even a rut can seem like a mountain, and the predictor model this gives Everton a 52.0% chance of victory. Do it for Mike Walker.
Relegation six pointers don’t come meatier than Newcastle against Watford. The one-win Magpies who, like a medieval baron, have taken Wood from their rivals to bolster themselves and weaken the enemy, against Watford FC. Eddie Howe outsmarted the Hornets in 2014-15 to win the Championship by one point and who knows, he may do the same next season, but in the meantime he’ll have to contend with the fact that Newcastle have only won one of their last 10 Premier League games against Watford. It’s a tough sport and the supercomputer offers Claudio Ranieri’s side – who have scored only two fewer Premier League away goals than Manchester United this season – a solid 30.0% chance of victory. Tune in and wallow in the drama.
Talking of teams in trouble, Burnley are pushing their “yeah it looks bad but we’ll come good in the spring and you’ll look stupid” script to the absolute limit while Leicester have injuries and form issues of their own, and etc and so on. Look, the harsh truth is that Brendan Rodgers’ team are favourites (44.1%) but Burnley (27.7%) are up against it and that is how they like it. Expect fireworks! (not actual fireworks).
“You can’t win anything with kids,” I explained to the supercomputer, making, at best, a weak excuse for a dismal managerial performance in the over-60s north Oxfordshire darts league. By coincidence that was also the verdict of 1990s pundit Alan Hansen when Aston Villa defeated Manchester United at Villa Park in 1995. United though, bolstered by that very youth, went on to win the title that season, and Villa haven’t beaten them at home in the league since. But the supercomputer cares little for 90s soundbites and with Steven Gerrard’s supersquad on the rise it gives United only a 44.3% chance of extending that historic run in Birmingham.
West Ham defeated Leeds last weekend in the FA Cup and they are predicted to do so again this weekend in the Premier League. That’s just the way it goes.
Anyone wondering how Liverpool, shorn of AFCON talent, would do against a Premier League team got an answer on Thursday night as the Reds failed to break down 10-man Arsenal at Anfield in the League Cup semi. They take on another side form the capital on Sunday, Brentford, and the predictor model gives Jurgen Klopp’s side a hefty 69.7% chance of winning, based on logic. Using narrative as an alternative you have to give the plucky Bees a small-to-reasonable chance of remaining unbeaten against Liverpool this season.
The weekend concludes with the north London derby at Tottenham Hotspur Stadium. Two clubs who have to accept Chelsea being European champions while scrapping to get back into the competition themselves. Arsenal haven’t won away to Tottenham since March 2014, when Bukayo Saka was 13. Does the supercomputer think it’s time for that run to end? Sort of: it makes this game a 41-29-31 chance, which, if you think about it, is the perfect distillation of a local derby. Well done, I say to the supercomputer happily. You’re a real football fan now.
The end of a year is a time for reflection for some. But not in the supercomputer community! Not only do these proud machines have no emotional attachment to the system of months and years we cling to in the hope that recurrent rituals will add meaning to our existence, but they also understand that a season runs from August to May and we need to get the games in. Thick and fast. And lo, onto matchweek 22 we go.
Premier League football in 2022 starts with the sort of fixture that can take New Year resolutions, crush them into a tiny metallic cube and then bury that cube 16 miles beneath the ocean floor. It’s Arsenal against Manchester City, a fixture that has ended in an away win in the last four seasons, with City opening the scoring inside the opening two minutes in the last two. It’s hope against inevitability, it’s youth against truth, it’s City on 53.9% to win.
Tottenham’s Harry Kane is the Premier League’s top scoring player on New Year’s Day but Watford have the in-form forward in the division in the form of Emmanuel Dennis, who has been involved in 10 goals in his last eight Premier League games. He should have been about to go to AFCON to play for Nigeria but some sort of registration loophole means he is not. Doesn’t seem fair, and it isn’t fair. The supercomputer gives Spurs about a 50% chance of winning by the way. Fairer.
Leicester’s game with Norwich is off, meaning that Jamie Vardy – injured anyway – will likely reach February with fewer than 10 Premier League goals, as some experts predicted. Talking of predictions, the cutting edge model fancies an even contest between Crystal Palace and West Ham at Selhurst Park. Palace have only won two of their last 13 Premier League games against the Hammers and the supercomputer does (38.8%) favour them slightly but like a display eagle at a country fair, it could go either way.
Sunday January 2, an incredibly bleak phrase, has four games at 2pm. There’s Brentford against Aston Villa, a fixture where only fans of those two sides probably know how good the Bees have been against the Villans recently, unbeaten in their last seven. Actually one being that also knows this is 2022’s Supercomputer of the Year, the supercomputer, which makes Brentford narrow favourites at 36.9%. This one could be a minor classic!
Brighton have never won away at Everton but they’d never scored a goal at Chelsea until this week and that one’s been ticked off so just watch them go at Goodison. The home side will be rested and have history – too much history if anything – on their side but Albion are looking like they’re headed into one of their fabled “good spells” so the supercomputer might have to learn a lesson from this one.
Newcastle make the long trip to Southampton to play in “The Waning Playing Career of Kevin Keegan clasico” and once again, the model sees this as being a very close game. Hats off to the fixture computer (a distant cousin) for scheduling such an egalitarian start to the year. Five of the last nine games between these sides have been level and if it happens again it won’t be the most outlandish thing to happen in 2022.
Leeds against Burnley is, like it or not, a relegation six-pointer. Sean Dyche has defeated snow this season but has won only one Premier League game. Traditionally Leeds dominate Burnley, the Clarets being one of only two clubs in their league history that Leeds have a +50% win rate against. Another three points would go down extremely well in this part of Yorkshire, and send Burnley scurrying back across the moors with gloom in their hearts. The supercomputer agrees, giving Leeds a pleasant 45.6% chance of victory.
Super Sunday concludes with Chelsea against Liverpool, aka the two members of the three-horse title race to have pulled up lame halfway around the racecourse. Whoever wins will still believe they can gallop up behind Manchester City, and in recent years Liverpool have enjoyed their trips to Stamford Bridge, just as Chelsea have tucked in at Anfield. The win predictor model concurs that this is an away win buffet, but one where the stakes are much higher than the sausage rolls.
The first matchweek of 2022 comes to an end on Monday with Manchester United against Wolves. The visitors haven’t won a league game at Old Trafford since 1980, back when the best supercomputers could offer us was Pac-Man. Still, if you like to see famous figures from the 1980s gobble up chances then United’s new frontline of Ronaldo and Cavani could be the gaming experience you need. The (our) supercomputer makes United a 49.3% against a team who have the joint-second best defence in the Premier League right now, which feels about right. Put some coins in the machine, 2022 is up and running.
Even a respected supercomputer can barely cope with a zero day gap between Premier League matchdays, especially when there are sudden postponements to consider and the sheer narrative of a Premier League Christmas. It’s relentless, and no-one ever really stops to consider the welfare of an advanced win prediction model that is working night and day to forecast matches in Our League. You may have only just got over the pulsating medium-quality action at Newcastle on Monday night, but here are the next games already formed.
No three words cause more confusion in England than “Bank Holiday Tuesday” but if Christmas Day is on a Saturday then, I’m afraid, here we are. This theoretical day off has allowed sport-hungry administrators to schedule three Premier League games at 3pm, namely Watford vs West Ham, Southampton vs Tottenham and Crystal Palace against Norwich.
Starting at Vicarage Road, our beloved computing friend makes West Ham the favourites, which makes sense, given Watford have lost four of their last five Premier League home games. But hang on, West Ham have lost three of their last four away games. Can both teams lose a game? No. Can both teams be on the London underground network yet only one of them be based in London? Very much so.
Tottenham’s revival under Antonio Conte has been stylishly impressive, with the vocally dominant coach even coaxing a couple of goals out of Harry Kane. Spurs’ trip to Southampton is Conte’s seventh game in charge, yet only his second away from home. Tottenham haven’t played on the road in the Premier League since November 7, so you have to wonder how they’ll cope with a dressing room that doesn’t contain NFL posters. Then again, they generally smash Southampton, including last season’s fixture, when Harry Kane assisted Son Heung-Min four times. The win predictor loves stuff like this and gives Spurs a 50.1% chance of winning, a result that would give their chances of finishing in the top four another big boost.
Crystal Palace lost their way against Spurs on Boxing Day, and the absence of the suspended Wilfried Zaha is a corollary of that wretched performance. But Norwich at home is just the sort of thing doctors order. Even real doctors. Not only are the Canaries rock bottom of the Premier League, but they haven’t won any of their last 16 games at Selhurst Park, while they’ve scored just 10 goals in their last 27 Premier League away games. In a fight you’d heavily back an eagle to destroy a canary and it feels like a very similar situation here. The supercomputer agrees, giving Palace a 56.7% chance of swooping to victory.
Bank Holiday Tuesday is completed by a fixture that nearly always delivers; Leicester against Liverpool. Two seasons ago it was this game, at this time of year, that convinced even the biggest doubters that Liverpool were about to end their long, long wait for a league championship. Jurgen Klopp has two games worth of Mohamed Salah to use before he leaves for AFCON and you’d fancy a relatively rested Liverpool to try and make the most of Leicester’s defensive issues. But then Leicester could do similar to Liverpool, and that’s why we love the game.
There are two games on Wednesday, starting with Chelsea taking on Brighton at Stamford Bridge. Chelsea have never lost to the English channel-adjacent club and almost certainly don’t plan to this week. That said, Thinking Graham Potter’s side have lost just once away from home in the Premier League this season and can be frustrating opponents, what with their ability to pass the ball accurately, and the supercomputer recognises this, making Chelsea favourites of course, but only at 54.5%, which, for the European champions at home, is something of a robotic snub.
Starting 45 minutes later than that game in west London is Brentford against Manchester City in further west London. The last time these sides played in the top-flight was back in 1937-38 when City became the first and only reigning champions to get relegated. You’d like to think such a thing was technically possible this season, except it’s not, because City already have 47 points and are SAFE. Yes, the Brentford supporters will clap and sing but the win predictor gives City a 68.8% chance of winning and you have to accept that will almost certainly happen.
The final two Premier League games of 2021 occur on Thursday, starting with Everton taking on Newcastle at Goodison Park. Eddie Howe may have marked this game on a special calendar as “the final match before the transfer window opens” while Rafa Benitez may be thinking “well, I’ll get a decent reception from the away fans anyway.” In truth, neither of these sleeping giants™ should be struggling but this is football and they are. Everton are supercomputer’s favourites but not massively so. A troubled draw perhaps? I’ve seen them happen and so have you.
We end the year with Manchester United taking on Burnley. It’s unlikely Sean Dyche will ever use the 4-2-2-2 formation but he’ll have rigidly taken notice of United’s struggles on Monday night at Newcastle. In 1994-95 United conceded only four goals at home all season; so far this campaign they’ve let in 22. Time moves on, fashions change, but Burnley will always be a “difficult prospect”. How difficult? Why not tune in and find out (by enjoying some classic late December Premier League action). “Happy New Year…” I say to the supercomputer “… oh by the way, will Excel ever be able to handle dates before January 1, 1900?” As usual, there’s no satisfactory response so the power is switched off. Millennium bugs and all that, you can’t be too careful.
Some say Boxing Day football is the most sacred tradition this country has left, but the pandemic has shown no respect for it, already carving off three of the 10 fixtures into a postponement-shaped bin. Until we see the remaining seven games kick off with our own eyes it’s tough to second guess exactly what’s going to happen, but that’s where our result-predicting multi-simulation supercomputer comes in. No virus is going to bring this machine down. Not that sort of virus anyway. Here’s how it thinks Sunday and Monday will go.
The scheduled early kick-offs have departed, so Boxing Day (by the way it’s called ‘Boxing Day’ because the expected goals industry have worked out that it’s the day of the year players feel most confident about having a shot from long range) gets underway with four games at three o’clock. Manchester City face Leicester City at the Etihad in a game between two of the four teams to have been champions of England since Manchester United last were. Last season, Brendan Rodgers’ team won this game 5-2 thanks to super striker Jamie Vardy. This time, Pep Guardiola is musing about strikes and player welfare, but the supercomputer expects he’ll still ask his employees to wear football kits and take on Leicester, and gives the league leaders a monstrous 71.6% chance of victory.
Talking of top-four clubs, Arsenal travel to Carrow Road hoping to continue their youthful charge at Champions League qualification. No-one knows why, but Norwich have never won a Premier League match on Boxing Day in seven attempts, while Mikel Arteta can now call on an array of youngsters so promising that you can actually hear opinions being revised about the club in real-time if you hold your ear to a glass and press that glass onto any football website. The win predictor puts Arsenal at a detox-healthy 61.6% chance of winning, the only concern being that the youngsters might be tired out and emotional after Christmas Day otherwise it might be even higher.
Tottenham love the concept of running around. When it arrived as an instruction from Harry Redknapp it was earthy and amusing; now it is coming from Antonio Conte is it is urbane and continental. Either way, when a man in a club suit is shouting at Spurs to run run run, they improve considerably. They take on Progressive Pat Vieira’s Crystal Palace on Sunday, I mean Boxing Day, a club who don’t tend to do very well on Boxing Day, although they are now managed by someone (Progressive Pat Vieira) who won all eight such games he took part in as a player. Spurs haven’t lost on Boxing Day since the iPhone came out and the supercomputer, though not an Apple machine, gives Conte’s side an almost 55% chance. Like Spurs, it loves running [simulations of Premier League matches that is!].
The three o’clock games wrap up with West Ham against Southampton at London Stadium. The Hammers have won more games against Saints than any other side in their Premier League history and with David Moyes’ side still brooding after their slightly unfortunate exit from the Carabao Cup in midweek, expect a festive, dramatic response from the Westfield Shopping Centre-adjacent club. The supercomputer agrees.
Steven Gerrard’s Aston Villa take on Chelsea at 17:30 on Boxing Day, having already given stern tests to both Manchester City and Liverpool in the short period under their new manager. Had Frank Lampard still been Chelsea boss the level of narrative surrounding this game could possibly have shut the league down for months, but instead it’s Thomas Tuchel’s task to try and coax some points out of his vast but not virus-free squad. Chelsea lost at Arsenal on Boxing Day last year and against Southampton the year before. Could former transfer target Stevie G make a hat-trick of festive despair and virtually shunt the Blues out of the title race? Our calming supercomputer still makes Chelsea favourites at 44.9% but think what sort of percentage they’d have had if this game had taken place earlier in the season.
Boxing Day concludes with the most even match of the round, according to the all-knowing machine. Brighton take on Brentford in a “I Can’t Believe It’s Not League One” spreadable clasico. There’s historic beef between the owners of the clubs, and there’s been some ducks in front of goal from Brighton in particular this season. The supercomputer makes the home side ever-so-slight favourites but the draw could also be as inevitable as bread sauce.
The matchweek concludes not on Boxing Day but on ‘Boxing Day,’ which is apparently what December 27 is called if December 26 falls on a Sunday. But only someone who hasn’t been out of the house since the mid 1990s would care about such arcane UK bank holiday minutiae, and let’s face it, if they hadn’t even watched any football since the 1990s then they’d probably still assume Newcastle United versus Manchester United is the biggest game of the season. It’s still important, of course, and with Ralf Rangnick’s team having had what seems like four months on the training pitch since they last played, they should be brimming with new ideas and jaunty shapes. Newcastle just need a win, any sort of win, and the supercomputer, like a sort of robotic Three Wise Men combo, offers them a gift. AKA the gift of hope. It rates Eddie Howe’s team at 23.2% to win the game, which seems sort of generous to me but then hey, it is Christmas after all.