Scorecasting Economists

Premier League Week 3

Premier League Week 3

 

Man City v Arsenal

The first match this weekend is Man City v Arsenal and although our model has been somewhat kind to Arsenal in the first two weeks it now shows no mercy. City are on at least 2 ‘expected goals’ and at 68% to win. Bookmakers ‘Mean odds’ are even more confident than that, putting them at over 76% to win and Arsenal on just 8% and it really doesn’t feel like Arsenal are equipped to withstand City but let’s see.

Most likely score: 2-0

West Ham v Crystal Palace

The other game with a similar odds margin for us is West Ham v Crystal Palace. West Ham are somewhat provisionally top of the league after two games, with Palace only managing one point from their first two. We have West Ham with a 2-0 home win as the most likely score. We are even closer to bookmakers Mean Odds with this prediction (just 3.2%).

Most likely score: 2-0

Other games

Elsewhere in the league, we favour Man Utd slightly less than bookmakers (which is possibly because Wolves get home advantage. Personally can’t see Wolves drawing with Man Utd, although in our model it is the most likely score (1-1) at 12.4%). We also have a Brighton win against Everton at exactly the same odds (39.9%) but a 1-1 as most likely score. Tottenham look likely to beat Watford and are close to 2 on expected goals (I’d expect that to go up in the next few weeks as Kane starts scoring), again not much between our model and bookmakers on the likely outcome.

Liverpool v Chelsea

We also favour Liverpool more heavily than bookmakers do in the Liverpool v Chelsea game although it does feel like a 1-1 in that game is the most likely outcome. It does feel pretty likely that both teams will score.

Most likely score: 1-1

 

 

 

Premier League Week 2

Premier League Week 2

 

The weekend starts with with Liverpool and Burnley in the lunchtime kick-off. Liverpool got off to a good start last week against Norwich so this could quite a one-sided game. It’s only been a week but already our scorecasts are giving some clear predictions. Our model thinks Liverpool are likely to win (surprise!) at 73% (see table below). 2-0 is the most likely score (at 12.3% likely).

The Man City v Norwich game also comes in at 2-0 as the most likely score (assuming Man City can manage to score at all…). Man City are over 80% likely to win. Bookmaker mean odds have them at a massive 87% to win. Giving Norwich less than 4% (a 1-0 Norwich win is at 2.1% likely in our most likely scores).

 

Other results are fairly predictable but in a different way.

As both Brighton and Watford will be looking for 3 points against each other you’d think a draw was quite likely (our model naturally predicts a 1-1 draw but favours Brighton, as do bookmakers).

Similarly, if you are a Crystal Palace or Brentford fan you’d be hoping for a win but this match looks incredibly close at 38%-35% in our model or 37%-33 with bookmakers mean odds.

On Sunday Arsenal are at home to Chelsea. It’ll be interesting to see if any of the new signings make much difference for either side. It’s hard to see past a Chelsea win, after Arsenal’s start to the season last week. Both teams are remarkably at 1.22 on expected goals in our model though. This is because although it factors in ELO ratings and past results, it also gives home advantage. In our ‘Outcome Probs’ indicating the most likely winner, each team is on exactly 36.3%. Bookmakers of course have Chelsea on 54% to win.

 

English Championship Week 4

English Championship Week 4

 

Bristol City play Swansea tonight, kicking off the weekend’s Championship fixtures.

Bristol City managed to win their midweek game away to Reading but Swansea have gotten off to a pretty dreadful league start managing to lose two and draw just one game. The odds however, are not in Bristol’s favour in our model (we favour Swansea at 46% to win) and I suspect this is in part based on last season’s results. Bookmakers though have Bristol City as slight favourites at 40% – 30% (Swansea).

Six teams are on seven points in the league and two of those are playing each other, West Brom (top on goal difference) are away to Blackburn in sixth. The most likely result is a 1-1 draw but our model favours West Brom at 43.6% which is very close to bookmakers at 42.6%.

At the other end of the early season table, neither Preston nor Nottingham Forest have picked up a point. Things don’t exactly look promising for Forest away to unbeaten Stoke (26.4% for a Forest win in our model, 23% bookmakers mean odds) although Preston may fancy at least a point against newly promoted Peterborough (Preston are favourites at 48.4%, our model and 43.4% bookmakers).

Some other likely results of note are Coventry at home to Reading (both our model and bookmakers favour Coventry, although a 1-1 is the most likely score), and Bournemouth at home to Blackpool, where there is just 0.4% between our 58% likely Bournemouth win and bookmakers having them on 57.6%.

 

 

 

English Championship Midweek 17-18 Aug

English Championship Midweek 17-18 Aug

Midweek matches are back to normal times in the Championship with no more 5.30pm or 6pm kick offs.

Again though there isn’t that much between teams in our prediction model.

All the most likely results are 1-1, this was the same at the weekend where we got only 1 match exactly right (unlike the opening weekend where 5 matches ended 1-1).

This week there is one game where a 1-0 home win and 1-1 draw are almost exactly likely in our model and that is Swansea v Stoke (see table below under Home wins and Draws, both exact scores are at 12.1% likely).

If you look to the right of the table below, you can see that in our ‘Outcome Probs’ and Bookmaker ‘Mean Odds’ there really isn’t that much difference in terms of likely winners in each match. The Swansea v Stoke game is the biggest difference between our model and mean odds. We heavily favour Swansea as well as Barnsley, Hull and Reading. There is also a decent chance of a win for Bournemouth, Blackburn and Cardiff.

Bookmakers agree that Reading are favourites against Bristol City (Reading have the 2nd highest expected goals this week), West Brom favourites at home to Sheffield Utd and most other games where we have favoured a team.

However, Preston North End, despite losing both their matches so far, are ahead in our model by 10% away to Huddersfield but bookmakers have Huddersfield favoured at just 38%.

All this just shows that at the early stages there really isn’t that much in it and quite a few teams have a fair chance at some surprise results in the next few weeks.

One match does stand out more than most however and where bookmakers agree pretty closely on our odds – Hull City v Derby County, where we are 0.4% different on the likely winner odds (Hull at around 49%  and Derby to lose at between 27 and 24%).

 

 

 

 

 

English Championship Week 2

ENGLISH CHAMPIONSHIP WEEK 2

 

The Championship still looks pretty close this week and will do until we have a few more wins in place.

Our model slightly favours two away teams (Preston away to Reading, and Derby away to Peterborough) other than that we’ve predicted a 1-1 draw as the most likely score for all games (there were actually 5 1-1 draws last week!).

Compared to bookmakers mean odds, we favour West Brom at home to Luton but they have West Brom at 55% compared to our 41.6% (our model also includes home advantage, otherwise there would be less difference between West Brom and Luton).

 

The table below shows our expected goals, which team we favour (Outcome Probs), Most likely score, the odds of each exact score in each game and bookmaker mean odds.

There is a big difference in the odds of a Huddersfield or Fulham win, for instance, where bookmakers mean odds show Fulham at 50% likely to win, whereas we have them only only 31.7%. This may be still affected by ELO ratings and immediate past results, although Huddersfield’s won’t have been great, Fulham were obviously relegated last season from the Premier League.

The two teams with the highest expected goals in our model, Hull City (who scored 4 last week) and Blackburn (2) both won, and are on 1.5 expected goals. Both are favoured by our model at 43.4% and 42.8% to win. Neither team are favoured by bookmakers though so these will be a couple of games to watch out for.

 

 

 

 

Premier League New Season 2021/22

Premier League New Season Preview 2021/22

This week we have a preview of the Premier League and at the of that a preview of the weekend’s games.

So, who do Scorecasting Economists predict to win the Premier League this season?

We have two tables below, the first table shows as a heatmap, which teams are most likely to win the league, finish in the top 4 (Champions League spots) or get relegated. The redder it is the more likely it is to happen.

The second table shows how likely each team is to finish in each exact league position.

Quite reasonably both tables put Man City as the most likely to win (after all they won it last season). They, along with the others we predict to finish in the top 4, have no chance of being relegated.

Outside of the likely top 4 the next likely contenders are Arsenal, Tottenham and Leicester. Everton for instance have less than a 5% chance of getting into the top 4. None of these teams have much chance of getting relegated.

At the other end of the table, sadly, the newly promoted teams are currently the most likely to also be relegated.

Norwich have a 44.2% chance of this happening. Away from that, both Brighton and Burnley are between 32 & 33% to go down. Even the teams we have listed 12th and 13th (Southampton and Aston Villa) have a 17% chance of being relegated but less than 1% of the Champions League. In the next section though, when we look at the exact positions, we can see that Norwich also have a reasonable enough chance of finishing outside of relegation.

Where exactly do we think teams will finish?

In the table below, as per above Man City are 73.4% to finish top but there are some other interesting aspects. Liverpool are 30% to finish second. Chelsea 25% to finish 3rd for instance.

Everton, West Ham, Wolves and Leeds all have between 12.7 to 6.4% chance of finishing 7th and thus getting into Europe next season. One of those (West Ham) is already in Europe this season so the extra matches this year may actually reduce that chance as the season goes on.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

At the other end of the table, in terms of exact likely positions, it gets quite close once we get to the last 8 teams or so.

Even Norwich have a 5& chance of finishing 12th, and Crystal Palace have a 6.1% chance of finishing 11th.

Norwich also have an 18.7% chance (the highest) of finishing dead last though. Newcastle for instance have a 10% chance of finishing 18th place but only a 7% chance of finishing last or even 11th. So in essence the exact positions at the lower end of the table are much harder to predict this year.

 

WEEK ONE Preview

Newly promoted Brentford start the new season at home to Arsenal and because it’s Arsenal, the model favours Brentford (actually it’s not because it’s Arsenal but, Brentford’s ELO rating is pretty decent due to the amount of matches they won towards the end of last season and home advantage is factored into our model).

Our model has Brentford at 41% to win with Arsenal on just 32.4%. It also predicts a 1-1 draw as the most likely score.

Our model has actually predicted a 1-1 draw as the most likely score for every match this week although it is unlikely you will see the same thing in future weeks as we base it on more recent results (see Score Picks).

If you look at the section to the centre-right called, ‘Draws’ you can see that a 1-1 draw is between 11.2% likely (Tottenham v Man City) to 12.7% (Everton v Southampton). 

This week our predicted outcomes vary quite wildly from bookmakers mean odds. For instance they have Chelsea at over 75% to win against Crystal Palace and Liverpool winning away to Norwich (Liverpool are at over 72% with Norwich on just 11% to win).

One of the biggest matches taking place is Tottenham v Man City. Tottenham finished the season quite poorly and have a new manager. Harry Kane is officially still a Man City player at the time I’m typing this, so if he starts our 1-1 draw and the model favouring Tottenham’ are not actually totally unrealistic. Bookmakers disagree an have City on nearly 60% to win. Our model has home advantage factored in though which gives Tottenham’s chances a boost.

The other match where I suspect our model may turn out to be more accurate than bookmakers is Man Utd v Leeds. With Man Utd missing Cavani, I think will hinder them and Leeds will prove difficult opponents. Our model favours Man Utd slightly 40%- 34.5% but bookmakers have them at 62% to win.

 

 

 

 

 

English Championship 2021/22 Preview and week one

English Championship 2021/22 Season Preview

 

Week One

 

Football is back, so we’re back and predicting things again.

This week we’ve a double-header as we have a table predicting the most likely teams to be Promoted, Play-off contenders and Relegation candidates, as well as the usual weekly likely scores.

The first table below (with the shades of Orange and Yellow) gives the most likely odds of each of those outcomes for each team. It’s split into ‘Promotion, Play-Offs and Relegation’. On the left hand-side we have the teams listed by abbreviation (SHU = Sheffield United, WBA = West Bromwich Albion)

As you can see, unsurprisingly the three relegated teams are the most likely to go back up. All three benefit from ‘parachute payments’ which probably give them an unfair advantage for the next couple of seasons. A lot of the media assessments of the Championship seem to think that at least two of these three teams will go back up.

There are however also some interesting teams in amongst the more likely playoff contenders – Bournemouth, who were strong for most of last season and certainly don’t seem to have been massively weakened over the Summer, Cardiff and Barnsley (Cardiff who I think made a decent late run and Barnsley who emerged through clever use of data and building a ‘system’ – that will be properly tested with a new manager as their previous one has gone to West Brom) are also contenders.

HOW WILL MY TEAM DO THIS SEASON?

Although in our model Swansea are likely play-off contenders again (quite similar odds to the top 5 in our model at 37.5), the factors not in our model, such as Swansea only appointing a new manager only a few days before the start of the season suggests it may be quite a bit more difficult to repeat last season.

Reading, another team that did well for part of last season (they were in the play-off positions for most of the season, only falling out in the latter part of the second half after a dismal end to the season) are now under severe financial restrictions so are more likely to be in a relegation fight than promotion battle (they are nearly 9% to get relegated in our model).

At the other end of the table we have Huddersfield in last place giving them next to no change of promotion (0.1%) and with a 45.7% likely relegation looming. Birmingham don’t fare much better on 40.7% and Bristol City and Derby County are also on 26% to get relegated. The other interesting thing is we’ve included a ‘promoted’ variable which has possibly given Hull, Peterborough and Blackpool a better chance of staying up than might normally be expected, that said most pundits seem to also give them a decent chance of staying up, in large part because teams like Derby are expected to struggle.

First Weekend of Football

On to this weekend’s games and you can see our likely winners in the ‘Home and Away’ section of ‘Outcome Probs’ below in the next table.  This is then compared to ‘Mean odds’ on the right of the table below. ‘Mean odds’ are the average bookmaker odds for this weekend. There is a decent amount of similarity between our two models but there are some differences which I will highlight below.

One or two, as usual stand out. With Derby v Huddersfield, bookmakers heavily favour Huddersfield at 54% (the highest odds they give for any team winning this weekend). Our model favours Derby but that’s because it takes into account past results and ELO ratings and not the number of fully contracted players Derby have!

HOW WILL MY TEAM DO THIS WEEK?

As you can also see – all of the matches bar one game have a 1-1 score as the most likely outcome (all within 10.8%-12.4% likely to happen as the most likely score).

This is in part because there are no recent matches and although the table is partly based on ELO ratings, once the season gets underway in form teams will start to heavily influence the likely future outcomes.

The exception to the 1-1 scoreline is Sheffield United who also top our likely Promoted teams, they are at nearly 60% to win in our model, with bookmakers having them on 52%.

 

 

 

 

 

 

 

 

 

 

Premier League end of season

Premier League – last day of the season

Premier League – last day of the season

Today is the final day of the Premier League season.

We end the season with some almost exact odds as bookmakers in our model’s predictions:

Liverpool v Crystal Palace – both our model and bookmakers have Liverpool on over 80% to win. Liverpool have really picked up in the last few weeks and after 4 wins in a row are actually back in 4th. Their goal difference is pretty close to Leicester’s (+4) but Leicester have a tougher match against Tottenham (they are still favoured in our model but only at 43.6% – with bookmakers having them on 48.4% to win).

There isn’t much to play for in the Leeds v West Brom game but our model has them at 66.6% to win, with bookmakers having only 0.3% difference in their mean odds. So a Leeds 2-0 win seems a reasonable prediction.

Man Utd don’t look so good against Wolves, where our model has them at just under 50% to win but bookmakers actually have Wolves edging it on 41%. This sounds like bookmakers are taking into account Utd’s UEFA League final in midweek.

In our model, Chelsea are favourites against Villa at 52% whereas bookmakers gave them on 65% to win.

West Ham are favourites against Southampton and again our model has their likely win at pretty close odds to bookmakers – 53.7% to 55.4.

So if our predictions come right, West Ham will finish 6th; if Liverpool win by more than Leicester then they will claim a Champions League spot but in 7th, a Tottenham draw, Everton defeat to Man City, could lead to Arsenal sneaking in on the last day – they are predicted to win 2-0.

 

Premier League Mid-Week 18-19 May 2021

Premier League Mid-Week 18-19 May 2021

There are only European places now to play for in the Premier League (and whatever increase or decrease in prize money you get for finishing one or two places higher or lower in the table).

We have a full set of games though crammed into two days in midweek, with another full set of final matches playing out on Sunday.

One of the curious things is that as the table narrows and we have a full set of results to measure across the whole season, football becomes sort of a bit more predictable (apart from the odd shock). There is only one game where our predicted most likely outcome differs from bookmakers mean odds – that is Southampton v Leeds, where our model favours Southampton at 39% whereas bookmakers actually favour Leeds by almost the exact same amount.

Apart from that, there is one game where there is a significant difference in the extent of the odds that might effect the overall outcome:

For the West Brom v West Ham game we slightly favour West Ham at 40% to win, with West Brom on 34%. Bookmakers have West Ham on almost 20% higher odds at 59% to win.

Both teams are also pretty close on expected goals at 1.25 (West Brom) to 1.38 for West Ham. Our model predicts a 1-1 draw as the most likely score but if there is a win it’s more likely to be West Ham (see Score Picks, Most & Conditional).

Elsewhere, Man City are on close to 3 expected goals against Fulham, with a 79% chance of winning.

The repeat of the FA Cup final has Chelsea winning 1-0 and getting some consolation against Leicester.

Our model has Liverpool winning 1-0 away to Burnley but only giving them a 62% chance to win compared to bookmakers who think Liverpool will walk it at 76.5%. This would give Liverpool a last day of the season chance to get in the Champions League.

 

Premier League weekend 14/5/21 – 16/5/21

Premier League weekend 14/5/21 – 16/5/21

It’s the FA Cup final on Saturday afternoon so four teams are not playing in the Premier League this weekend (Arsenal, Chelsea, Leicester and Man Utd).

This seems to have played havoc with our data collection so the only two matches we can cover at the moment are tonight’s game between Newcastle and Man City and the FA Cup final itself, largely because it is being played out again on Tuesday.

The title is won and relegation places are finalised so the remaining weeks are really about the jostling for European places.

Liverpool have made a late surge and if they can win away to West Brom they are then only one place off a Champions League place behind Chelsea.

I think West Ham are probably too far off for them to claim a Champions League spot now but an away win against Brighton would keep them in contention. Both teams are relying on Chelsea to drop points though (our model has them winning on Tuesday so this probably sorts that question).

In tonight’s game though, Newcastle don’t have much hope against league winners Man City despite the fact they won their last game.

We have City on 8.5% less than bookmakers to win at 65%, whereas they have them on 73.5% to win.

Now they don’t have to win it will be interesting to see who they actually pick and whether a few lesser known players get to play. The most likely score is a 2-0 Man City win at 11.3%, with a 1-0 City win at 11%.

FA Cup final

On Tuesday in the league Saturday’s FA Cup final is replicated between Chelsea and Leicester and we have that included in the model.

The model favours Chelsea as it factors in home odds and has Chelsea winning 1-0 with a 57% likelihood to win with bookmakers having them at 53%. Leicester are only on 19% to win in our model and 21% with bookmakers.

Even if you take away the factor of home advantage which Chelsea won’t really have at Wembley, it still looks like a Chelsea win is most likely.

*We will likely update the remaining fixtures on Saturday

Skip to toolbar