<Click here for part 1 of this series>

<Click here for part 2 of this series>

Having determined a number of point thresholds, our next step in creating an accurate forecast is to match the 20 teams to the respective positions.

To do this, there are an almost infinite number of things that could be considered, but some of the most important are:

• Outliers and anomalies from 2010 performances;
• Personnel changes, development or deterioration; and
• Management changes; development or deterioration

I am going to look at the first of these points today and leave the personnel/management assessment for Part 4 of this series.

So what are we looking for when assessing possible outliers?

Well, what you are trying to identify is teams with point totals that did not correlate to their level of performance.  Such teams are often under/overvalued by the market in the following season and are possibly mispriced.

This is an area which the statistics can only take you so far; they are a tool but not a solution.  An element of this analysis has to be subjective.  That said, there are definitely some numbers that can help guide us here.

And surprise surprise, it’s my old favourite goal difference again.  I can hear the groaning at the back.

Joking aside, goal difference and points are very closely correlated and over the course of a season it can be a useful indicator.

2009-10 EPL points & goal difference

Team	Points	GD
Chelsea	86	+71
Man United	85	+58
Arsenal	75	+42
Tottenham	70	+26
Man City	67	+28
Aston Villa	64	+13
Liverpool	63	+26
Everton	61	+11
Blackburn	50	-9
Birmingham	50	-14
Stoke	47	-14
Fulham	46	-7
Sunderland	44	-8
Bolton	39	-25
Wolves	38	-24
Wigan	36	-42
West Ham	35	-19
Burnley	30	-40
Hull	30	-41
Pompey	28	-19
Total	1044	-

Observations

As you can see, team’s finishing positions roughly correlate to their respective goal differences.

Does anything jump out at you that could indicate an under/overvalued team?

Well, as I have highlighted, there are certainly a few numbers that stick out for me:

Liverpool’s +26 goal difference is quite staggering for a 7^th place team and I believe that such a points/goal difference ratio highlights a number of things:

• The points total required for a top 6 place has increased;
• Liverpool were, at times, probably unlucky last season;
• Rafa Benitez failed to maximise the capabilities of his squad; and
• With the appointment of Roy Hodgson, Liverpool should be back in the mix for a Champions League place this season, irrespective of any changes to their squad.

To put this goal difference / place ratio into perspective, prior to this season only one team has amassed a goal difference of +24 or greater in the 20 team Premier League era and not finished in the top 4 – the inconsistent Chelsea side of 2001-02.

Just above Liverpool, you will see Aston Villa.  They did very well last season to amass 64 points from a +13 goal differential and even under normal circumstances may have been looking at some downward regression.  I won’t linger on this one as their situation will be discussed in more detail in part 4 of this series… and I don’t want to pour salt into open wounds.

In the bottom half of the table, the number that jumps off the page is Wigan’s alarming -42 goal difference.  Prior to last season, no team had ever managed to avoid relegation with a goal difference of -30 or worse.

However, as we all know, last year’s Wigan side were no ordinary beast.  They were a statistical anomaly.  They lost two games by 8 goal margins (Tottenham and Chelsea) and lost a combined 0-10 over their two tete et tete’s with Manchester United over .  That is -26 in only 4 games.  Over the other 34 games they were only a -16 goal side.

Wigan are probably one of the hardest teams to predict going into this season and I can promise you there won’t be any BOTW cash going within a country mile of them anytime soon!

They could kick on and finish 9^th as easily as they could implode and finish 19^th.  After all, they lost 7 games last season by 4 or more goals, which is almost as many as the three relegated teams combined.

But on the other hand, Tottenham were the only side who managed to beat Chelsea, Liverpool and Arsenal last season.  Bizarre stuff.

The most valuable point I would take from Wigan’s performances last year actually has nothing to do with them.  The aforementioned Tottenham pipped Manchester City to the final Champions League spot with a goal difference of +26 – but 32% of this differential came from the freak 9-1 mauling of Roberto Martinez’s men at White Hart Lane.

They will need to improve considerably this year to be playing at the elite level again in 2012 given the likely improvement of Manchester City and the return of the real Liverpool.

As an aside, a special mention here for Paul Hart and Avram Grant’s Pompey side.  They really weren’t that bad.  In fact, I would go as far to say that with the 9 point deduction they might even have managed to survive in what were very difficult conditions.

They only took a couple of really heavy beatings throughout the entire campaign and were very unlucky in the seasons early months.

They played with a surprising amount of passion for a team in such dire straits, and I for one hope that the club manages to stabilise itself and that we see them back in the top flight in a few years time – although it will be nice to not have to see the crazy naked guy on MOTD every single week.

I would also note the large gap in goal difference between Everton (+9) and the best goal difference from the “middle 5”, which in this case happens to be Fulham (-7).  To me this further indicates the increasing disparity between the two halves of the league.

This worrying trend for the future of UK football is shown in the graph below, which profiles the mean Premier League goal differences by position over the last 15 years, along with maximum and minimum error bars.  The orange markers indicate the goal differences from the 2009-10 season.

EPL goal differences by position – 1995 – 2010

As you can see, there were goal difference highs at 6 of the top 8 positions last season as well as a number of goal difference lows in the bottom half of the table.

For all the teams not discussed above, I would suggest they finished roughly in line with where you would expect for their goal differences.  And this in itself could actually prove to be the most valuable bit of information here.

Moving on from the dreaded goal differences, below is another of my favourite pieces of analysis when retrospectively analysing a team’s performance in a league competition.

I have broken up the league into four mini leagues.  Importantly, this is not based on subjective opinions of who is a “big team”, but simply done on the finishing league positions, given the reliability of the points thresholds we established yesterday.

After all, allowing yourself to be influenced by team’s names makes you predict stupid things such as that a weakened Leeds team, that have been struggling for half a season, will amass 7 more points than the unfashionable Millwall side this season.

2009-10 EPL points analysis

Team	Top 4	Euro 4	Mid 5	Bottom 7	Total
Chelsea	15	10	26	35	86
Man United	12	13	21	39	85
Arsenal	3	15	22	35	75
Tottenham	6	15	20	29	70
Man City	10	6	22	29	67
Aston Villa	10	6	20	28	64
Liverpool	6	11	14	32	63
Everton	9	8	17	27	61
Blackburn	5	7	8	30	50
Birmingham	4	5	13	28	50
Stoke	3	5	15	24	47
Fulham	4	8	12	22	46
Sunderland	7	6	9	22	44
Bolton	1	4	12	22	39
Wolves	6	5	10	17	38
Wigan	6	7	4	19	36
West Ham	2	6	10	17	35
Burnley	7	5	8	10	30
Hull	2	8	8	12	30
Pompey	-	3	3	22	28
Total	118	153	274	499	1044

I have highlighted a number of things in the table above, but I don’t plan to go through them all in detail.  Some are more important than others, but you may find it useful to have a scan over them and see what jumps out to you personally.  It may help you to remove any false perceptions or biases you have for or against particular teams.

My key points to take from this would be:

1. Arsenal are a far better side than people give them credit for.  If they were not facing one of the worst injury crisis’ in the history of the Premier League going into the season, I was going to take a large position on them to outperform Manchester City at 11/10.  However, with so many players unfit or unprepared, I cannot justify it at current.

   They were unfortunate to only pick up 3 points from their 6 games against their fellow top 4 sides and in a normal season you would expect them to amass 4 -7 more points from these games.  And given that these games are “6 pointers”, they can have a huge impact on the league standings.

   Both Arsenal and Manchester United totalled more points than Chelsea last season against the 16 sides that filled the 5^th – 20^th places.

   Chelsea played very well in the big games.. and had the odd slice of luck along the way too.  I expect to see regression on both sides for Chelsea in this area.  They will be hard pressed to match a 15 point haul versus their rivals but will also likely pick up more point against the rest of the league.

   
   

2. Bolton and West Ham were potentially unfortunate against the top 4 sides, amassing only 1 and 2 points respectively.  It is likely that they will return more points from these games next season, which could be a driver which sees them move up the standings.

   I would note that last season’s Bolton side were actually two different assets and should be treated as such.  They are now playing with a different style and approach and with different management and personnel, which could obviously be a positive or a negative.

   
   

3. Tottenham amassed a crazy amount of points against the 5^th to 8^th place “Euro 4” teams.

   The fact that they chalked up more points against this group than Manchester United and Chelsea is an indicator of over performance and possible luck, given that those sides were +71 and +58 goal teams.

   It is likely that Spurs will fail to repeat this and they may have to find 4-5 points elsewhere next season.  If I was Harry Redknapp, the first place I would go looking for them would be against Wolves.

   The other side of this equation is Manchester City.  Given their performances against the other three mini leagues, they were unfortunate to garner only 6 points against their main rivals.  Their loss at home to Spurs at the back end of last season really was a cup final for them.  I expect them to pick up more points from these games this season which will push them over the 70 point mark.

   As I stated above, I don’t think the current squad or managerial regime has got what it takes to push them up to the 75-77 point range I feel they will need to finish above Arsenal, but only time will tell.

   
   

4. I won’t linger on Liverpool long as they were discussed earlier.  But it is safe to say they had some random results looking at these mini leagues.  As Sporting Index have forecast, I think a regression to around 70 points in very likely.

   And with the addition of Christian Poulson and the subtraction of the most overrated footballer on the planet (Ed – subjective bias from BOTW?!) they could even challenge the 75 point mark.  If El Ninio’s hamstrings allow it.

   
   

5. Blackburn totalled 30 points from a maximum of 42 against the bottom 7 sides.  This is quite an achievement and they are likely to see some regression in this figure.  However, they also struggled within their own mini league, the “Mid 5”, and I expect these to variations to be much of a wash.

   I am expecting slight regression here from last year’s 50 point haul, but not as much as the 7 points Sporting Index are forecasting.

   
   

6. True to form, this analysis highlights Wigan as a real mixed bag last term.  They recorded as many points against the “top 8” sides as any of the “bottom 12”, but then failed miserably against the mid table sides.

   
   

   Just plain unpredictable and they should carry a health warning.  The new Newcastle United!

Any teams without any numbers highlighted do not appear to have amassed an irregular number of points against any particular mini league against historic averages.

In my eyes, this reduces the chances of getting dramatic shifts in their point tally next season, before you consider the impact of management/personnel changes/development/deterioration which will obviously have a huge impact on the team’s performances in 2010-11.

And after all, when betting on sport, finding low variance selections can often be almost as important as finding high margin ones.

Check back in the morning when I will be wrapping this years preview up by reviewing a few individual squads and making a couple of predictions for the season ahead – and finally putting my money where my mouth is!

<Click here for part 4 of this series>