Hi everyone!
Welcome to the second edition of my EPL Predictions newsletter, where each week, I'll post detailed forecasts for each game.
As a general rule of thumb, each issue will contain an assessment of the previous week's forecasts, plus naive and biased predictions for upcoming games.
(you can learn more about the general assumptions of the model here)
Let's get to it.
End-of-season predictions
Liverpool's draw at Fulham has really damaged Klopp's title aspirations according to my model. The Reds began the season with a 35% chance of lifting the title, and those odds have dropped to 28% following the draw at Craven Cottage.
Other teams that have seen a big shift in their end-of-season projections are North London Rivals, Spurs and Arsenal, with both sides increasing their chances of a top-4 finish by more than 5 percentage points — mostly at the expense of United, which dropped their chances from 19% to just 10% following the weekend's result.
Naive Predictions
The naive predictions are computed using historical data only, and it's the forecast system that I use to simulate the end-of-season probabilities for the Premier League.
The following table showcases each game's forecasts under this model.
Biased Predictions
In this section, I present my biased forecasts, which, are based on both my hunch and historical data. For each match, I'll present a detailed viz with probabilities for different events within the match and a small write-up.
The following table shows the summary of my biased forecasts.
Arsenal vs. Leicester City
Yes. Arsenal started the game against Palace in style, and their back-line seems to have considerably improved with the arrival of Saliba and Zinchenko.
Jesus showed that he can be a real threat and I don't think it'll take too long for him to start getting his name on the scoresheet.
That being said, The Gunners showed some iffy moments during the second half against Palace and I think they don't have the maturity within the squad to fully control games, at least not yet.
If you forced me to predict a result, I'd say a draw at the Emirates.
Aston Villa vs. Everton
Not going to lie, there's very little room for me to talk about Villa, so I'll keep this one pretty short.
I didn't think Everton's performance at Chelsea was too bad, and Tarkowski seems to have added tons of security to Lampard's back-line – however, their attacking options do seem bleak.
I hate to disagree with my model, but I do think that a low-scoring fixture (under 2.5 goals) is the most likely outcome.
Brentford vs. Manchester United
Wow. Manchester United just can't seem to catch a break.
Watched the game against Brighton and it really frustrated me that the pundits couldn't stop talking about how poor Manchester were during the game instead of focusing on the massive display by Brighton.
Not a United hater, but I do have Toney on my FPL team so I hope they concede some goals on this one.
Brighton & Hove Albion vs. Newcastle United
Really, really excited for this fixture.
There are high expectations for both sides this season, and, so far, there's not enough data (at least the one I feed to my model) to reflect recent developments at both clubs; so I'm curious about how my model's forecasts learn from both sides during the season.
Just because FiveThirtyEight's model is predicting the hosts to win with a 49% chance (a bit too high in my opinion), I'll be rooting for anything other than a Brighton win.
Manchester City vs. AFC Bournemouth
Haaland hat-trick. That's it, that's the write-up.
Southampton vs. Leeds United
I was quite surprised that my model ended up favoring Leeds as the most-likely winner in this one.
FiveThirtyEight's model has a very different prediction than mine with a 32% chance for Leeds and a 44% probability of a Saints win, so I'll be keeping a close eye on the match.
Southampton have become quite "meh" in my opinion and I think they're the most likely to be relegated from last season's Premier League teams. If I had to choose a result though, I think it would be a 2-2 draw.
Wolverhampton vs. Fulham
Not much to say on this one, as I haven't watched a Wolves game in quite a while, but I'm rooting for Fulham. Loved the way they just took the game to Liverpool in the opener and I'd love for them to keep on the momentum and surprise us by staying in the Premier League this season.
Chelsea vs. Tottenham Hotspur
This one will be really interesting.
Spurs really struggle at Stanford Bridge and I think they'll play cautiously and should be happy to come out with a draw.
Chelsea, on the other hand, found themselves in a bit of trouble during some periods of the game against Everton and will be wary of Spurs' pace on the counter. Not sure what to expect on this one, but I wouldn't be surprised to see a low-scoring draw.
Nottingham Forest vs. West Ham United
Didn't watch the Forest game, but looking at the post-match stats they seem to have been caged against the wall for most of the game against Newcastle.
West Ham had some bright moments against City, but in the end, they appeared to have accepted defeat as soon as the second goal went in. Not excited to watch this one, and I expect a 1-0 or 2-0 win from David Moye's men on the road.
Liverpool vs. Crystal Palace
Liverpool must still be slapping themselves in disbelief after that first-half display against Fulham, so I'm expecting them to make up for the result against the Eagles at Anfield.
Palace played decently against the Gunners, and if not for a massive save from Ramsdale, I think they could've gone and snatched the draw on that one. However, I'm still expecting them to lose this one.
How did we do last week?
Sadly, we were beaten by 538's model in both the naive and biased predictions for the gameweek.
However, not by much. The biased and naive models had 3% and 5% higher error rates than the predictions made by the folks at FiveThirtyEight, respectively.
To be fair, most of the error from my model came from the Manchester United match, where I've already acknowledged that my forecasts are way too pessimistic on Brighton's chances (in general) for the season.
All things considered, I'm happy with the results and look forward to seeing what Matchday 2 has in store for us.
Hope you enjoyed today's issue! Let's see how we fare in this season's second gameweek, I'll post the model's error in next week's issue.
If you like my work, please help me by subscribing to my website and sharing my posts.
Catch you later! 👋
