Does Form Exist

Team	"Form" games played	Avg "Form" Pts	"Normal" games played	Avg "Normal" Pts
Arsenal	17	2.18	54	2.02
Newcastle	5	1.80	67	1.12
Man City	17	2.29	55	2.07
Aston Villa			72	0.97
Tottenham	4	1.00	68	1.81
Liverpool	16	2.19	56	1.91
QPR			34	0.79
Everton	7	1.86	65	1.58
West Brom			72	1.01
Burnley			34	0.76
Hull			72	0.99
Norwich			38	0.87
Leicester	2	1.50	32	0.88
Stoke	2	0.00	70	1.39
Swansea	1	0.00	71	1.30
West Ham	4	1.25	68	1.16
Crystal Palace	5	1.80	67	1.16
Sunderland	2	1.50	69	0.94
Fulham			38	0.84
Southampton	6	1.33	66	1.59
Man United	11	2.09	61	1.74
Chelsea	13	1.38	59	2.44
Cardiff			38	0.79

table: Average points of teams on a "hot" streak vs normal

There is something to be said about form. When a team is on a hot streak are they more likely to win? Or is a match decided on the day with past matches having no influence?

To investigate this I looked back on all the games played by the Premier League teams over the last two seasons. I categorised each match as either in "form" or "normal". A team was in "form" simply if a team had won their last three matches. The table shows how teams performed in "form" compared to their "normal" performance.

The immediate conclusion would seem to show that the chances of winning (or accumulating more points) are correlated with the teams very recent results - which was a surprise to me. I would have believed that a match is decided in the 90 minutes on the day, and independent of previous results.

There are some limitations of this simple analysis. Clearly 3 wins in a row may constitute good form for a mid table team like Everton, but perhaps for a top table team 3 wins in a row may be expected. Or for a relegation team any back to back positive result might signify a run of good form. The definition that I used doesn't take this into account. 

Also note that many teams have not strung 3 victories together in the last 2 seasons, or they have but only once or twice. For these teams the sample size is not large enough to draw any conclusions.

The one big anomaly is Chelsea - whose performances appear significantly inferior when they are in good form. They only average 1.38 points when their previous three matches were victories. Is there some sort of complacency at play here?