Scorecasting Economists

Premier League 5-7 November 2021

 Premier League

Premier League football returns tonight after the midweek European matches.

Southampton v Aston Villa

Southampton play Aston Villa on Friday night and neither team has gotten off to a great start to the season. Villa have lost their last four games, whilst Southampton’s form has taken a turn in the opposite direction with two wins and draw in the last three. Our model has this as a 1-1 draw and favours Southampton, with bookmakers mean odds only just over 1% higher than our 45% confidence.

Chelsea v Burnley

Chelsea are very likely to win this (78% in our model). Again there is only just over 1% between our favoured winner and bookmakers mean odds. The most likely score here is 2-0 to Chelsea. Chelsea have won their last 4 league games with Burnley managing just one in the league all season, although that was their last match.

Everton v Tottenham

One game where there is a difference between our model and bookmakers mean odds is this match. We favour Everton at 46% (and with over 1.5 expected goals).

Everton have actually lost one more game in a row than Tottenham recently, have scored 7 more goals over the season so far and are at home. Bookmakers slightly favour Tottenham (and probably rightly so). This may well be for factors that our model doesn’t currently take into account – like a manager change.

Tottenham were somewhat fortunate yesterday against Vitesse Arnhem, with 3 sending’s off in the game but with new manager Conte’s first Premier League game, it does seem reasonable to expect some changes to Tottenham’s form in the coming weeks. Our model predicts a 1-1 draw.

Man Utd v Man City

Man Utd managed a win last weekend but prior to that were beginning to fall away in the title race. Our model favours Man City at 47% with bookmakers more confident with their mean odds at 56%. Can Ronaldo rescue Man Utd, yet again, with Varane out Man Utd’s defence will potentially be a bit shaky? We have a 1-1 draw as the most likely result but conditional on a win occurring we have Man City winning 2-1 away.

English Championship Midweek 2 November 2021

English Championship Midweek 2 November 2021

As we start November, the Championship midweek games continue tonight and tomorrow.

We have a bottom of the table clash with Barnsley v Derby County. Barnsley were one of the surprise teams last season, supposedly using data to maximise performance from players and it also was supposed to matter less whom the manager was, as part of their overall strategy.

Derby will fancy their chances, however, as Barnsley have just sacked the latest manager, not least because Barnsley last got a point 7 games ago. Our model says Derby’s chances aren’t that great though (27.5%, compared to bookmakers mean odds having them at 21.5%). The most likely score is a 1-1 draw at 12.1% likely, with a 1-0 Barnsley win at 11.5, so home advantage and a different manager could be key.

Elsewhere in the league it would be hard to go against Bournemouth winning against Preston (nearly 60% chance of a Bournemouth win).

West Brom also look pretty likely to win against a Hull City side that are in the relegation places, they have a 64% chance of a win (the highest odds from our model), with a 1-0 exact score 12.5% likely and a 2-0 home win also at 12.2%. Bookmakers pretty much agree with us within 1-3% on these two matches.

Millwall probably fancy their changes at home to Reading who have lost 3 games in a row and our model predicts misery for Reading being not far from 50% likely (48.2) with bookmakers again within around 3% of those odds.

One game that might be quite close is Blackburn v Fulham, Ben Brereton-Diaz caught the eye in the Summer at the Copa America for Chile and is currently Blackburn top scorer with 10 goals in 11 games, against a Fulham side that are currently outscoring everyone in the Championship. Fulham have 10 more goals than the next-highest scorers, Bournemouth. Fulham’s higher scoring is reflected in their higher expected goals 1.66 to 1.27 and although bookmakers mean odds are very similar to our most likely winner, a 1-1 draw is also the most likely score here.

Premier League 30 10 2021

Premier League – last weekend in October

From the weekend’s fixtures, one game stands out for me in particular and that’s Tottenham v Man Utd.

Both teams are underperforming this season. Tottenham are sixth in the table but actually have a -4 goal difference, having scored just 9 goals all season in the league.

Only Brighton, one place above them and Wolves have the same poor scoring record in the top half of the table.

Brighton, only one place ahead can be said to be overperforming compared to expectations though. Man Utd on the other hand, for a team that was realistically able to outspend or at least compete with the league’s highest spenders are clearly not doing very well in 7th. How long either manager has left to correct things remains to be seen but for both teams the poor run of form can’t go on for too long.

At the start of the season virtually every prediction model and pundit had a top four of Man City, Chelsea, Liverpool and Man Utd (not necessarily in that order). The first three are are the top three with Man Utd the only serious underperformer.

In the table below you can see that our model favours Tottenham (we have home advantage included in our model), at 43% to win. With Man Utd on just 31%. The most likely score by far though is a 1-1 draw at 12.2% as an exact score.

Bookmakers Mean Odds (in the far column) show Man Utd as marginally favoured on 40.5% so almost the opposite of our model. All in all, with two out of form teams playing each other a draw does seem a realistic outcome.

The other game that looks quite interesting is the lunchtime kick off with Leicester v Arsenal. This also looks close. Our model slightly favours Leicester at 40%, with bookmakers having just 1.5% less confidence than that. Both teams have shown signs of improvement in recent weeks.

In fact overall, there is really not a great deal of difference in our model and bookmakers mean odds this week. With matches like Liverpool v Brighton and Man City v Palace favouring the home teams at over 75% minimum likelihood, it’s not surprising.

English Championship 29 10 2021

English Championship – last weekend of October

The English Championship returns tonight with a game that when I was a child was a regular fixture in the old Division One and early Premier League. QPR v Nottingham Forest kicks off at 7.45 on Friday night and we think QPR are favourites, at just over 50%.

A 1-1 draw is the most likely score however as both teams have an expected goals of over 1.

Bookmakers are more cautious about QPR having them at 45% to win.

The rest of the Championship fixtures take place tomorrow, there is a big lunchtime game between two of last season’s Premier League teams, Fulham and West Brom.

It’s big not just because they were both in the Premier League last season but because they are also second and third in the league this season. This also is most likely to be a 1-1 draw but favours Fulham at 47%, with boomakers more or less agreeing at 45.4%.

This round of matches finishes with a Saturday night kick off between Reading and Bournemouth. Reading’s form has been patchy this season, although still perhaps better than some fans had hoped for at the start of the season when they were financially restricted in terms of buying players.

Bournemouth on the other hand are top of the league, are unbeaten and have won their last four matches. Perhaps surprisingly, our model gives Reading a reasonable chance (29%) of a win and Bournemouth less than 50% (44%), but bookmakers are not wildly different (Reading – 20.3% and B’mouth at 53.5%) in terms of their mean odds.

A 1-1 draw is the most likely result with Bournemouth, as said before, favoured.

English Championship 23 10 21

The is the first week for a while, in the Championship, that we’ve had a team on two expected goals. In the table below you can see West Brom hit that milestone this week at home to Bristol City.

West Brom have actually lost two of their last three games but both our model and bookmakers Mean Odds (in table above) seem pretty confident of a West Brom win against Bristol City (64.2% our model; 65.2% bookmakers Mean Odds). According to our model there is an  equal 11.5% chance of a 1-0 or of a 2-0 win as the most likely exact score).

Another game where there is a high likelihood of a particular outcome is, Bournemouth winning against Huddersfield. This is a bit less likely than West Brom winning but still at 60% according to our model. Bournemouth have maintained an unbeaten record so far and although Huddersfield are 6th in the league bookmakers have Bournemouth as likely winners as well at 57.2%.

Our model predicts a 1-1 draw as the most likely score in the Birmingham v Swansea game, this is the game with the biggest difference in likely winner when comparing our model to bookmakers. We favour Swansea at 41% but bookmakers odds have Birmingham on 39.6%.

Premier League 22 10 21

Premier League 22 10 21

Of the games I’ll be looking out for this weekend, I think they’ll probably be Man Utd v Liverpool – can Utd’s defence stand up to Salah, Mane and Firmino? And does Solskjaer have it in him to outwit Liverpool and get the best out of the attacking part of his side?

In terms of upsets, Brentford v Leicester looks interesting. Leicester had an exciting mid-week in Europe, winning 4-3 against Spartak and Daka scoring 4 goals. They have also had some patchy league form (despite winning last weekend) with Brentford outperforming everyone’s expectations.

Also for nostalgia reasons, Everton v Watford is one I’ll be interested in largely because it was the first full F.A. Cup final I actually watched (I watched the Man Utd v Brighton replay, the year before, when they still had those). I’ve covered all of these games below the table.

Man Utd v Liverpool

This is one of those matches where our model stands up to bookmakers mean odds. We have Utd on 33% to win, they have them on 32% – we have Liverpool to win at 42% with bookmakers mean odds they have 0.7% less. All in all this suggests a very close game. Liverpool edge it on expected goals being 0.2 (so 1/5 of a goal more!) higher.

Most likely score: 1-1 (exact score 11.5% likely according to our model). – Personally I think Liverpool will win.

Brentford v Leicester City

Brentford have had a great start to the season and although Leicester managed to win in the Uefa League and a win last weekend, their form has been a bit patchy. This is reflected in our model. Home advantage probably gives Brentford an edge at just under 41% to win. They also are 0.14 higher on expected goals. Bookmakers slightly disagree with us with their mean odds as they have almost nothing at all between the two teams (0.2%!)

Most likely score: 1-1

Everton v Watford

The likely outcomes in this game are much clearer. Watford have a new manager but looked utterly hopeless last week against Liverpool. Although it was a very much in form Liverpool. Everton are clearly not Liverpool but are 8th and have only lost two games all season compared to Watford’s five. Everton are very close to 2 expected goals and this is reflected in our exact scores section (Home wins with 11.4% for 2-0 as the most likely score and 11.7% for 1-0). We have Everton at 63% to win.

Most likely score: 1-0

English Championship 19 10 21 – 20 10 21

English Championship 19 10 21

The Champions League may dominate the headlines tonight but as usual the English Championship looks like it will produce some much closer games.

Weekend review (How did we do?)

The most likely score in any Championship match in the last 18 months or so (and probably for many years) is usually 1-1 or 1-0, or unsurprisingly 0-1, at least it is when we are trying to predict the most likely exact score.

Our model for instance will often predict more 1-1 draws than anything else in a given week or weekend.

Looking back at the weekend’s fixtures though there was just one 0-0 (Preston and Derby – we had this as 1-0 as the most likely score), two 1-0’s (the first being, Reading v Barnsley – we had this as 1-1, favouring the home side by 2%).

Of those 1-0 home victory’s we got one exact score, West Brom v Birmingham City. There were no 1-1 draws with most games producing at least one team with more than 2 goals.

Across the rest of the matches that took place our model was broadly correct in terms of which team we favoured to win, although it predicted a 1-1 as the most likely score, with one exception – Millwall v Luton, where we had both teams as fairly even, favoring Millwall by around 8% and bookmakers mean odds favoring them 1% higher, Luton actually won 2-0.

Tonight and Tomorrow

Tonight however, it is a similar story to the weekend, for our model, with no teams yet regularly on over two for expected goals.

The closest we get to that is Fulham at home to Cardiff. Our model predicts a 1-0 home win as the most likely score (11.8% as an exact score, 11% for a 2-0 home win).

This feels pretty obvious as Fulham are third in the league but Cardiff have also had an utterly dismal run, losing their last six games (we have Fulham to win at over 60% likely, with bookmakers Mean odds (see table below) very close to that at 63%).

Elsewhere the 1-1 strikes again for a number of games, about half of which we strongly favour the home team (Reading, Sheffield Utd, Preston, Hull), one of these, Preston v Coventry, although we favour Preston at 46.5%, bookmakers actually just about favour Coventry instead by 0.3%).

Then there are the rest where for the most part there is only a maximum of 4-5% points between how heavily our model favours them (for instance, Bristol City, Derby, Stoke). It also looks like our model has under valued West Brom’s odds with bookmakers mean odds favouring them to win at 47.5%

Premier League Mid-October 2021

Premier League Mid-October 2021

Leicester v Man Utd

Leicester have not had a great time in their last four games in the league, drawing the last two and losing two before that. They have home advantage against a United team that has lost then drawn in their last two games.

Our model favours Leicester at 41% (bookmakers disagree having Utd on 45%) to win.  We don’t take into account player status but if we did, we’d probably consider that Utd could be playing with a second-string defence if both Maguire and Varane are injured. It’s also going to be a test of whom Solskjaer picks as his forwards as their have been accusations that rotation has cost them recently.

Most likely score : 1-1

Norwich v Brentford

The two promoted teams here have had wildly different fortunes. Brentford have probably surprised most people by being in 7th, whereas Norwich managed their first point in their last game. Norwich have home advantage and our model slightly favours them at 40%.

This is one of three games where there is either a difference in outcomes in our model and bookmakers mean odds or a sizeable difference in the confidence levels. Here bookmakers favour Brentford at 45%, so this is a decent test for our model.

Most likely score 1-1

Watford v Liverpool

Our model gives Watford over 22% to win this which seems unlikely as Liverpool are hardly struggling. If there is a win then our model gives Liverpool a 55% chance of winning with a 2-1 away win the most likely winning score. Watford have just over 1 on expected goals with Liverpool approaching but not quite at 2. Bookmakers have Liverpool at 70% to win though.

Most likely score 1-1

Elsewhere, Burnley are given next to no chance against Man City (7% our model, 4.4% bookmakers mean odds). Also both our model and bookmakers agree on Arsenal winning against Palace (58% (with 0.2% between our Arsenal likely win and bookmakers mean odds).

 

Championship Mid-October 2021

 Championship Mid-October 2021

Club football returns this weekend in England for the top two divisions, after the International Break.

Our scorecasting model has thrown out a few interesting results this week.

Huddersfield v Hull

Our model has a 1-1 draw as 13% likely as the exact score here, the highest probability for an exact score this week. Both teams are as you’d expect, on almost 36% to win (see table below).

Bookmakers disagree however, favouring Huddersfield as home winners at 47.7%. Huddersfield are actually 7th in the league, compared to Hull in 21st, although at this stage of the season there is only 8 points between the two of them so there isn’t that much in it.

Blackburn v Coventry

The other match where our predictions are different to bookmakers is Blackburn v Coventry, where we have Blackburn at 46.6% to win but bookmakers marginally favour Coventry. Our model gives a 1-1 draw as most likely.

Elsewhere, we have Preston winning 1-0 and West Brom also with the highest odds for winners at 63%.

 

Premier League 2 10 2021

Premier League 2 10 2021

Man Utd v Everton

Man Utd and Everton kick off the weekend games.

It’s only six games in but Everton under Benitez have performed pretty well and above expectations. They are fifth. Man Utd are fourth. The two teams are both on 13 points having each lost one game.

Man Utd have home advantage and with this factored in and the fact they have a higher ELO rating. The two teams have played each other 206 times with Utd winning 90 to Everton’s 70 wins. With all this taken into account, we have Man Utd winning 1-0, although Everton’s expected goals is at 0.99 just 0.1 shy of ‘1’ that would take them into a likely 1-1. The most likely exact score is at 11.4% likely (1-0) with a 1-1 draw at 11.3%. We favour Utd at 54.9% with bookmakers having a Utd win at 64%.

*Although having just noticed at 11.30am that Man Utd are starting without Ronaldo and Pogba, I do wonder if this now gives Everton a decent chance to win.

Most likely score: 1-0

 

Liverpool v Man City

Pre-season, most pundits seemed to fancy either Chelsea or Man City to win the league.

Liverpool are currently the only unbeaten team, one point ahead of the rest of the top six. This is going to be a close game in terms of determining a winner but should see some goals. It is also the only game across the league this weekend where there is more than a 10% difference in the likely outcomes from our scorecasts and bookmakers mean odds.

That Liverpool have Trent Alexander-Arnold injured and that may have some impact, especially against such a good City side. Individual player absences or selection isn’t included in our model.

This game is closer than Man Utd v Everton in terms of odds, in both our model and bookmakers. We favour Liverpool at 44%. The most likely score is 1-1 but conditional on a win happening then 2-1 is most likely. Bookmakers favour Man City at 39%.

Most likely score: 1-1 

Skip to toolbar