Each Premier League team has now played six league games, 16% of their season, with varying degrees of success and satisfaction.
Aston Villa are already over 25% of the way to likely Premier League safety. The primary target for the majority of EPL teams, especially one that has struggled in the lower half of the table over recent years and currently has a side of comparable quality to those of the recent past.
Chelsea top the table, having dropped only two points on their visit to title rivals, Manchester City. They also became the second, top ranked side in a week, following on from Arsenal, to expose Aston Villa’s hitherto mean defence in a comfortable 3-0 victory.
Prior to the visit of Arsenal, Villa had conceded just a single Premiership goal from their first four games, comprising 50 goal attempts. Only Hull had managed to score against Paul Lambert’s team and Villa’s four goals in reply was sufficient to earn them 10 points from four games and second place in the table.
The addition of Roy Keane was credited with bringing “a ferocious winning mentality” to the side’s newly discovered strong, defensive foundations, but there is a more mundane and less dramatic explanation for the apparent turnaround.
It is highly likely that Villa simply got lucky in the short term, particularly in terms of the actual outcomes of the attempts they conceded and the generous nature of their initial fixtures. It is important bettors don’t overvalue a teams performance based on results alone.
Streaks, either winning or losing occur all the time in sports, but they do not always indicate a permanent change of fortune. Conceding one goal or fewer over a series of four matches is a rare, but hardly a unique event.
If you take sequences of four games over the last eleven seasons in the EPL, conceding one goal or fewer happens around 6% of the time. Just over half of these occasions the stout defence belonged to one of the EPL’s top five teams, but that still left ample for the less talented teams to briefly flatter to deceive, including Aston Villa themselves, as recently as 2012/13, when they finished 15th.
Fresh data is both exciting and most relevant to the present side, so there is an understandable desire to use it to the exclusion of less recent information in evaluating a team. If bettors wish to do this they should try to temper their enthusiasm with some caveats.
Firstly, older data should not be entirely dismissed. The historical record of a side can impact on the present team. A team who usually finish in the lower half of the table will likely have a lower quality of squad players compared to the elite teams – making it more difficult to deal with the almost inevitable injuries and suspensions.
In addition, they will not be able to attract top quality reinforcements in the January window and may even see their most accomplished players subject to bids from more elite teams.
Secondly, small amounts of data are often extremely noisy. Luck can play a significant role in contests which are decided by an average of just two to three scoring events. Hence why the move away from actual scores to goal expectations based around the quality and quantity of the chances created in a game, is a more accurate way of extracting relevant information from a new season.
We already have over 1,500 attempts on goal from the 60 Premier League matches, so far. And while this number is still smaller than ideal, we can use the expected goal route to attempt to allow for the unforgiving consequence of a last minute shot from an excellent position cannoning off the bar, rather than nestling in the net for an equalizer, as happened to Newcastle on Monday night at a wet Britannia Stadium.
We can also attempt to correct for the differing challenges faced by sides in the limited fixture list up to the end of September.
Case Study: Aston Villa & Chelsea
Aston Villa’s defensive goal based stats now appear much less impressive than they did prior to the visit of Arsenal. They have currently faced 77 goal attempts and the quality suggests that a fair return for their opponents would by just under 7 goals.
This agrees with the seven goals they have now actually conceded, but we can attempt to add context to the performance which Villa’s defence has delivered so far by looking at the attacking, expected goals records of their six opponents in 2014/15.
Villa’s six opponents, to date have been Liverpool, Chelsea, Arsenal, Stoke, Hull and Newcastle. Intuitively they would appear to be an above average set of opponents, comprising as it does of three of the big five. As a group, these six teams have played a combined 36 matches against a range of Premier League teams.
During these 36 matches, Villa’s six 2014/15 opponents have created opportunities of sufficient quantity and quality that they would expect to score, as a group, an average of 1.4 goals per game. Villa have restricted these same opponents to an expected goals value of just below 7 from six matches or 1.16 goals per game.
So Villa has put in defensive performances that are around 20% better than their six opponents typically achieve from their attacking efforts.
In a similar vein, Villa’s attack has created chances consistent with scoring 0.75 expected goals per game, against opponents who have allowed chances that might expect to yield 0.86 goals per match. So Villa’s attack has performed 13% below this average.
In short, Villa’s chance creation profile in 2014/15 is consistent with a good, but not great defence - they are currently ranked 7th in the EPL by this method - and a below par attack (13th). Overall that’s good enough to just creep into the top ten of the EPL, compared to their current 6th spot.
Therefore bettors should be aware that they have been lucky so far this season, based on shot analysis, to be in their current position and should expect them to fall down the table the longer the season goes on.
We can apply the same methodology to their opponents from last week, Chelsea. The Blues have posted eye catching scoring feats, whilst allowing the entertainment to flow at the other end as well –something not renowned with Jose Mourinho teams. Overall they have been very good, but against largely limited opponents in Leicester, Burnley, Everton, Swansea and Villa – the game with Manchester City has been the sole exception.
Chelsea, of course can only play the teams put in front of them, but we can begin to measure their performance by combining their shot derived expected goals with their apparently weak strength of schedule.
Chelsea has created chances with the expected goals equivalent of 2.1 per game and allowed chances that might see them concede 0.86 expected goals per game. The six opponents they have played so far currently concede chances at a combined 1.35 and score at 1.00 expected goals per game. So Chelsea’s attack is over 50% better than this average and their defence is 14% better.
These are the best overall figures in the league, but only marginally ahead of their main title rivals, Manchester City.
These figures are based on 170 goal attempts from Chelsea’s six games and 912 opportunities in the 36 games involving their opponents, so we are still drawing conclusions from limited data. But we may have avoided making the extravagant claims that surrounded Villa’s early season exploits, based on relatively rare goal events.
|Home Team||Home Win %||Draw %||Away Win %||Away Team||Actual Score||Most Likely Score|
|Man Utd||28||32||40||West Ham||2-1||0-1|
EPL Week Six review – Strength of schedule
There were few causes for complaint in week six, where most sides who dominated the quality and quantity of chances received at least a point. Only West Ham appeared to be hard done by. But Wayne Rooney’s red card has skewed the game’s expected goals as the Hammers outshot their host by 5-0 following his petulant dismissal.
The table above showcases the data that is used in these performance evaluations. Expected goals replace actual goals to reward the team that is creating chances, but experiencing poor luck and lends an air of caution about teams who may be scoring or preventing goals at unsustainably high rates. Also as match numbers increase we can approximate and begin to account for strength of schedule.
We can then compare actual achievement, which may be heavily influenced by luck and will be reflected in the current league position to a team’s ability to create and prevent chances, which may be a better indication of future performances.
The above examples of Chelsea and Villa show how these numbers can be produced. Chelsea are currently the best team in the EPL but not by much. This is however consistent with their recent campaigns and they have the financial muscle and a tradition of strengthening the squad further if necessary. So “one of the two best in the league” is probably a fair assessment of their current abilities.
Villa on the other hand are producing stats typical of a mid-table side, better than their two most recent season, but inferior to 2011/12. So weighting these various indicators more generously for more recent events, they probably currently have the talent of a typically 12th placed team.
The value to betting is to recognise when the current table, an inevitably big influence when striking a bet, is at odds with these wider and more repeatable influences. Understanding this and factoring it in to your thought process will leave you more informed and therefore better placed to make a long-term profit.
Currently, Newcastle, Liverpool, Burnley, West Ham and Everton occupy league positions, which perhaps do not fully reflect their core talent. They may be better than their position implies.
By contrast Swansea, Manchester United, Crystal Palace, Hull, Leicester, Spurs and Villa appear flattered by their current league position from a purely shot based view.
The projections remain particularly enthusiastic about surprise pace setters, Southampton. They are flying higher than in any of their more recent EPL seasons, which may suggest they will fall back into the chasing pack, but their chance creation data remains
strong and a substantially re-jigged squad and a new manager only adds to the fascination and uncertainty around Saints’ future prospects.
Their achievements based on this term’s shot data and strength of schedule ranks Ronald Koeman’s team at just below the usual suspects of Chelsea, Manchester City, Liverpool and Arsenal. However, how well Southampton deal with the almost inevitable injuries and suspensions, could determine if they keep this level of performance going or not.