Scorecasting Economists

Premier League 19 March 2021

Premier League 19 March 2021

 

The Premier League clashes with the FA Cup this weekend, so there are only four Premier League matches taking place. Our model below has somewhat struggled with this, so shows a lot of the matches that have subsequently been re-arranged.

Fulham v Leeds

As with the West Ham v Arsenal match our model is very close to bookmakers mean odds for these two teams. Both teams have very similar expected goals for this match with the expectation that Fulham’s defence effectively cancels out Leeds attacking options.

Most likely score: 1-1

Brighton v Newcastle

This is a pretty important relegation battle with 16th v 17th. Brighton have at least won a game in the last five, unlike Newcastle. One place separates the teams but 10 goals are the gap in goal difference with Brighton scoring more and conceding less than Newcastle. We favour Brighton at 47%, whereas bookmakers have them at 57%.

Most likely score 1-0 (our model shows this at 12.7% likely. It also shows a 1-1 draw as 12.6%!)

West Ham v Arsenal

West Ham are the highest placed of the teams playing this weekend. This game is 5th v 10th. Both teams have won 3 games in the last 5.

Given all that it’s hardly surprising that the most likely score in our model is a 1-1 draw.  The odds in favour of an Arsenal win or a West Ham one are pretty much the same in our model as with bookmakers. Neither team gets above 40% likelihood for a win and Arsenal just edge it on expected goals.

Aston Villa v Spurs

After Tottenham’s performance in midweek this is a match Tottenham need a win from, now they are out of the Uefa League. If the sort of Tottenham that turned up in midweek shows up for the rest of the season it will be no Europe next year for them.

Our model is completely out of sync this week due to the FA Cup so there is no prediction for this match, however I think Villa will win 1-0 and if they do Mourinho is in serious trouble.

 

 

 

Championship midweek 16 March 2021

Championship midweek 16 March 2021

This week our model favours the same teams as bookmaker mean odds for all matches (see the table below for our ‘Outcome Probs’ and bookmakers ‘Mean odds’).

I can’t remember that happening hardly at all this season.

The emphasis however is different, i.e. there are, in some cases, differences of up to 15% in terms of the extent that a team is favoured to win.

Blackburn v Bristol City

The biggest difference in our model and bookmakers odds is with the Blackburn v Bristol City game tomorrow.

We favour Blackburn at 39.5% but bookmakers have them at 53%. Based on that, our model predicts a draw as the most likely (a 1-1 draw is showing 12.8%).

Norwich v Nottingham Forest

Norwich are surprisingly predicted to draw with Nottingham Forest in our model (1-1 being 13% as the most likely scoreline) but they have won their last 8 games, hence the fact there is a 24% difference in the likely winner with bookmakers (24% Forest to 48% Norwich).

Forest haven’t won for 4 matches but our model takes into account home advantage (despite Covid-19 meaning the biggest home advantage, fans attending has somewhat nullified that effect).

Results between the teams show Norwich have won ten more games (28-18) 

Cardiff v Stoke

Our model heavily favours Cardiff at 57.5% with a 1-0 home win predicted. Bookmakers however only have Cardiff at 44% likely winners.

Cardiff have almost an extra goal in terms of expected goals at 1.8 to 0.95 for Stoke.

Only 3 places separate the two teams but Stoke have lost four of their last five and this is reflected in that 1-0 most likely score.

 

Premier League 12 March 2021

Premier League 12 March 2021

I covered last night’s Newcastle v Aston Villa game earlier in the week but there are plenty of other games this weekend.

Arsenal v Tottenham

This weekend’s London derby is naturally pretty close.

Both teams have looked interesting at times this season, with Spurs possibly edging it slightly this week in terms of form, winning their last three league games.

Of course it all depends whether Spurs can carry on their most recent form, where the reintroduction of Ali and the return to form of Bale have clearly helped.

Arsenal’s play led to a ridiculous goal last weekend and if the Spurs forward line turns up and presses them, you feel like Arsenal could be in trouble.

That said, although Arsenal are stuck in mid-table, bookmakers slightly favour them at 37.4% to Tottenham’s 34.5. Our model slightly favours Tottenham at 38% to 36.5% for Arsenal.

Most likely score: 1-1 

Leeds v Chelsea

This is today’s 12.30pm game.

Historically speaking, it’s not really that surprising that they are reasonably well matched over time, with Leeds winning 39 times to Chelsea’s 35

Chelsea have been in excellent form since getting a new manager and are back in the top four and pushing injury ravaged Leicester for third place. Leeds are one place below Arsenal, in 11th.

Given all this, our model predicts a Chelsea win at 54.2% with bookmakers mean odds being almost identical at 54.3%.

Most likely score: 0-1 (this is at 14.4% a pretty high figure for an individual scoreline, although Man Utd and Leicester have higher percentages for their most likely results this week).

Southampton v Brighton

Our model favours Southampton in this game and whilst Brighton have lost their last three games, Southampton haven’t done that much better lately, although have managed a win in their last five.

Our model gives Southampton at 47.5% as likely winners. Bookmakers on the other hand only have them 31.5% and actually favour Brighton at 37.3%.

Brighton don’t score as many as Southampton but have let in 14 less goals.

Historically Southampton have won 38 times to Brighton’s 20 with Brighton last winning in 2012

Most likely score: 1-1 (this is showing at 11.9% likely in our model but with a 1-0 home win showing at 10.5%).

English Championship 12 March 2021

English Championship 12 March 2021

Blackburn Rovers v Brentford

Tonight’s game sees 14th v 4th, with Brentford having slightly lost their way a bit since their 21 match unbeaten run.

They have still won two of their last five games whereas Blackburn have won only 1 of their last five, drawing two.

The teams haven’t played each other too many times, Blackburn have won 19 to Brentford’s 13

Our model slightly favours Brentford at 38% but bookmakers mean odds are higher at 45% (see table below). The teams are also pretty close on expected goals with only 0.08% between them.

Most likely score: 1-1

Nottingham Forest v Reading

Reading have won their last three games and Forest have lost their last two. Reading have also scored 21 more goals than Forest although they have let in three more.

Our model slightly favours Forest but Reading have a higher ELO score  and somewhat surprisingly have a decent historic record against Forest winning 17 times, losing 20 but in the last four games have won 2 and drawn 2. The last time Forest won was in 2018.

Bookmakers have very similar mean odds to our scorecast with less than 1/2 a percent between our model and bookmakers odds for a Reading win (33.2% to 33.6%), however, I still think Reading will win.

Most likely score: 1-1

Sheffield Wednesday v Norwich

Sheffield Wednesday haven’t won in ages and are 23rd in the league against a Norwich team that is ten points clear at the top of the league so although this is the only Sunday match for the Championship, it certainly doesn’t seem like a very fair match up.

Naturally our model favours Norwich. There is less than 1% between our model and bookmakers mean odds, in terms of the likelihood of Norwich winning. Norwich are showing at 1.82 on expected goals with Wednesday on 0.88.

Most likely score: 0-1   (0-1 is 12.3% likely; 0-2 is 11.2% likely).

 

Elsewhere in the league it is noticeable still how close the league generally is. Norwich are the highest on Expected Goals (1.82) this week but the nearest to them are QPR on 1.48 and this is only because they are playing Huddersfield who have the second worst defence in the league.

 

 

Midweek Games 8th and 9th March 2021 (Premier League and Championship)

Premier League

Only one Premier League match in midweek this week (there is also Champions League). There is also a Friday night game.

Man City v Southampton

It felt like Man City were unstoppable and defeating every team in their path until playing Man Utd at the weekend. That said they will still look very difficult to beat with any modelling or odds based analysis . They have a 75% chance of winning tomorrow in our model and nearly 80% with bookmakers mean odds.

Southampton are still just about vulnerable to relegation but won at the weekend and although it was against Sheffield Utd, as far as football cliches go, “you have to win those” and they did. This probably gives Southampton a better chance that at any other point this year at least but it still looks like a City win.

Most likely score: 2-0

Newcastle v Aston Villa

Also showing in the model below is Friday’s game.

Our model gives Newcastle a better chance than Villa of winning (why???) but is corrected by bookmakers mean odds giving Villa a 48% chance of winning. This is likely to be due to the model including home advantage and that in the last 2 games although Newcastle have drawn both, Villa have lost and drawn.

Most likely score: 1- 0

 

Championship 

Although the table below shows four games this midweek, the Luton v Rotherham game is postponed, leaving two games tonight and one tomorrow.

Barnsley v Derby

Barnsley are now in the top six and just keep winning. This match has slightly less of a predictable feel about it than the other games this midweek (all of the rest are much closer when comparing our model to bookmakers mean odds). We have Barnsley on 56% to win whereas bookmakers only have them on 49%,

Most likely score: 1-0

Blackburn v Swansea

Tonight’s first game is the closest in odds with almost nothing between the two teams and very similar likely outcomes. Bookmakers very slightly favour Swansea, whereas we favour Blackburn. One of the reasons for this is that despite Swansea being third, Blackburn have the best goal difference of the bottom half of the table and also have scored the most goals.

Most likely score 1-1

QPR v Wycombe

Our model is again almost identical to bookmakers odds giving Wycombe only 17% chance of a win. QPR although mid-table have the 2nd best goal difference in the lower half of the table so our model thinks they will win

Most likely score: 1-0

Premier League 6 March 2021

Premier League 6 March 2021

Manchester City v Manchester United

Are Man City unstoppable?

This season, in terms of winning the Premier League, yes

If United win tomorrow, will it change the title race, such as it is?

No, because even if they do win, City have enough to rotate and will win more than enough of their remaining games.

This is first v second but there are already 14 points between the two and Man Utd have drawn four of their last five games. United actually look reasonably vulnerable of not being in the Champions League if two or three teams below them put together good end of season runs and they carry on drawing.

They also have the worst defence of the current top four and the second worst out of current top six.

Man Utd do score goals though and are in fact second in the league only to Man City on that measure.

Man Utd have won quite a few more of the games played between the two clubs so there is some hope, especially as they have won 2 and drawn one of the last 4 match ups between the two.

Man City are showing at 1.88 on expected goals to Man Utd’s 0.7 and are 65% likely to win in our model. Bookmakers have them on similar odds at 61%.

Our model predicts: 1-0 Man City win.

Elsewhere in the league:

For a lot of the games this weekend our scorecasting model is very close to bookmaker mean odds (see table below).

There are one or two exceptions. There is not much between Brighton and Leicester (with each model favouring a different team). Brighton have the joint best defence with Fulham in the lower half of the table and apart from Leeds, the best goal difference in that half of the table.

Our model favours Sheffield Utd over Southampton whereas bookmakers heavily favour Southampton despite their dismal recent form (45% for a Southampton win).

In many weeks of the season, our model has slightly underestimated Everton. This week is no exception, as a draw is predicted with West Brom but bookmakers favour them at 50% compared to just 23% for West Brom.

It’s also one of the few weeks this season we’ve had where there are no teams on above 2 for expected goals.  The game with the biggest difference between two teams this week is between Spurs and Palace.

 

English Championship 5 March 2021

English Championship 5 March 2021

Huddersfield v Cardiff

Cardiff manager Mick McCarthy has just signed a new contract and Cardiff are one of the teams just outside the playoffs. They are putting together the sort of run of form that often leads to promotion or a playoff final at the least.

Cardiff also are the second highest scorers in the league now (53 goals for, only Brentford, in 2nd place have more with 61) and have the joint 4th best goal difference.

Huddersfield on the other hand are only 5 points away from the relegation places and have the second worst defence in the league (they’ve let in 51 goals and only bottom placed Wycombe have more with 55).

All of that surely suggests a Cardiff win tonight which will increase the pressure on the teams in the playoff places…..

However, our model predicts: 1-1 but looking at the details below, this is very marginal – an exact 1-1 draw is 10.7% likely. However, conditional on a win, Cardiff win and away results are all three of the next most likely results:

0-1 (10.4%),

1-2 (9.9%)

0-2 (9.6%)

Cardiff are also 56% to win in our model (with bookmakers more cautious at 44%).

Reading v Sheffield Wednesday

Finally a match where Reading are outright predicted to win both in our model and bookmakers mean odds (see table below for both). This has probably cursed them, so sorry about that.

Reading have more than half a goal extra in expected goals than Sheff. Weds (1.56 to 0.9).

Reading are 53% to win in our model and 52% with bookmakers. This is in part because Sheffield have lost their last five league games (their last win was also against Wycombe). Reading have won their last two so hopefully this scorecast will herald good news.

Our model predicts: 1-0

 

 

English Championship Midweek, 2nd Mar 2021

English Championship Midweek, 2nd Mar 2021

Our scorecasting model for the English Championship is currently producing some fairly similar predictions when compared to bookmakers mean odds at the moment (see table below). There are one or two exceptions each week but not many. There are occasional teams that have broken this slightly though – one is Barnsley who have come to serious form in recent weeks and Reading who often defy predictions of defeat or a draw. This week there is also a difference in the Coventry v Middlesbrough game.

Bristol City v Bournemouth

This is one of the few games where there is a large difference in the odds for a win for either team when comparing our model to bookmakers. Bookmakers have Bournemouth as overwhelming favourites, we have the two teams within 1% of each other in terms of who wins.

Most likely score: 1-1

QPR v Barnsley

Barnsley have now won 5 games in a row and I suspect our model is slightly underestimating them.

Our model slightly favours them at 37%, against 35.4% for QPR with bookmakers also placing Barnsley at 37% but QPR 2% lower. There certainly looks like there is space in the playoffs or even promotion so wouldn’t be surprised if the current layout of the table changes quite a bit before the end of the season.

Most likely score: 1-1

Reading v Blackburn

Reading managed to halt a two match losing run and win at the weekend. Blackburn managed a draw against ten men, after losing 5 in a row. Reading are at 47% to win in our model (bookmakers have them at only 40%) with Blackburn only 26%.

The teams have played each other remarkably few times. Apart from one FA Cup match in 1913, it was not until 1979 that they actually played each other. Reading won 4-2 earlier in the season and the last 2 games between the clubs have produced 13 goals!

Most likely score: 1-1 (although a 1-1 win is at 12.6% likely to happen, a 1-0 home win is at 12.5%!)

 

Premier League 2nd March 2021, Midweek

Premier League 2nd March 2021, Midweek

Liverpool v Chelsea 

This is a pretty intriguing game as Liverpool’s defence continues to struggle and will face what is now a better organised Chelsea team.

Our model favours Liverpool, who are actually one place lower than Chelsea but our model takes into account teams ELO rating, so with Liverpool being 2 ELO places ahead of Chelsea among the English teams and 7 across Europe the outcomes showing in our model make more sense (47% Liverpool win v 25% Chelsea).

Unsurprisingly given the longer period of success in Liverpool’s history the record between the teams favours Liverpool. Liverpool have won 82 times to Chelsea’s 64. The last time Chelsea beat Liverpool in the league was nearly 3 years ago.

Liverpool also won their last match (despite losing the last 4) and although it was against Sheffield United who looked almost relegated, Chelsea also have drawn their last 2 matches.

Most likely score from our model: 1-0

Burnley v Leicester

I strongly suspect there are going to be academic studies into how it appears that players are picking up more injuries, across this particular season. Considering a large amount of the same players are expecting to play in Euro 2021, I wonder what state squads will be in for the start of next season as well?

Teams outside the top 6 don’t really have the squads to play the same first team in 3 games in ten days but are having to with the increase in mid-week games. Leicester are now out of the UEFA League but it looks like it has already taken it’s toll on their players with the amount of injuries they’ve had. This gives Burnley a chance tomorrow night I think.

Caution should be applied though, Leicester have only lost won game after a good run but whilst Burnley may not be winning that many games they are grinding out the points or were until Tottenham finally got their attack to work at the weekend. Bookmakers are slightly clearer out the outcome than our model, giving Leicester a 47% chance to win.

Most likely score: 1:1

Elsewhere in the league, our model continues to under-score Everton, where it actually favours West Brom, who despite having the worst defensive record in the league still show as slight favourites. Bookmakers overwhelmingly favour Everton though at 51.3% to 22.5% for West Brom.

Sheffield Utd v Aston Villa is another match where our model has undermarked a team and in this case it’s Villa. Our model slightly favours Sheffield United who are all but relegated. Bookmakers correct this by favouring Villa at 45%.

 

Premier League 27 February 2021

Premier League 27 February 2021

Chelsea v Man Utd

This is 5th v 2nd with Chelsea now looking much more consistent than at the start of the season.

Our model favours Chelsea, as do expected goals and bookmakers mean odds. Historically Man Utd have come out winners far more often (81-55)

I think this is going to be a real test for an in-form Chelsea. I still think Man Utd’s defence is too weak and Chelsea have a much better chance of winning that they did with Lampard.

Our model gives a 45% chance for a Chelsea win, with 14.5% for a 1-0 and 13.1% for 1-1.

Our model predicts: 1-0

Man City v West Ham

1st v 4th should perhaps produce better odds for the 4th placed teams but not this fixture.

That West Ham have surprised people this season is fairly widely acknowledged. However Man City look very difficult to beat and are regularly scoring and will surely try to dominate this game, crucially they are winning and have key players returning from injury as well.

Man City have an 80% chance to win in our model, giving West Ham just a 6% chance. A 2-0 win for Man City has more than twice the chance of happening at 14% than West Ham winning at all. City don’t just have the best defence (only 14 goals let in), they have scored more than everyone else (50). I actually think there is a strong chance City will get more than 2 goals today.

Our model predicts: 2-0

Tottenham v Burnley

Which Tottenham will turn up?

In midweek they put four past Wolfsberger and this does feel like the sort of match most managers would be looking to play all four of Kane, Ali, Bale and Son from the start – it’s got to be better than whatever else has been going on for them in the league recently.

Feels hard to justify not doing that but then there is Mourinho, so who knows?

Tottenham are still likely to win, regardless of that but will surely have to attack Burnley.

Our model has Tottenham with a 66% chance to win with bookmakers only 4% less. Tottenham also lead on expected goals (over 2), hence the 2-0 prediction below.

Our model predicts 2-0

Other matches

Elsewhere in the league there are two matches where our model’s likely outcomes differ significantly from bookmakers mean odds.

Both games are at the lower end of the table.

Palace v Fulham and West Brom v Brighton. Our model favours both home teams whilst bookmakers favour the away team.