A classic study by the psychologists Tversky, Gillovich and Valone (1985) on 'hot hands' in basketball is still as relevant to sports today at it was 30 years ago.
In the study, the researchers analyzed thousands of basketball matches and concluded that 'hot hands', whereby a player is said to be 'hot' if he sinks three or four baskets in a row, is just a cognitive illusion, a trick caused by our brain's compulsion to try to see patterns in everything and to attach causal factors to those patterns, even when there is no good reason for doing so. Psychologists call this the local representativeness heuristic.
In fact, the players didn't do anything differently in terms of movement or technique when they had 'hot hands' than when they didn't. Sinking four baskets in a row is just a random act of chance for a top player, there is no causal effect. Although some basketball players (and some teams) are clearly better than others, this can only be demonstrated over the long-term (I.e. only when a large statistical sample is available).
But what's all this got to do with the Premier League? Let me explain.
Say your favourite team achieves the following sequence of results.
WWWLLL
A typical pundit or fan would say that the team were on form for the first three games but not on form for the last three. They would put it down to some problem with the tactics, or the fact that a certain player is (or is not) playing. They may also go further and hail the manager as a genius if he wins three games in a row, and then condemn him as overrated, out of his depth, or someone who has 'taken the club as far as he can' if he loses three in a row.
But now consider this sequence:
WLLWLW
A typical fan or pundit would say that this team hasn't been either particularly in or out of form. The BBC pundit Mark Lawrenson would say that their 'form has been mixed'.
But the statistical probability of either sets of results happening is exactly the same. The spread of the results is completely random. A team can win or lose several consecutive matches completely by chance, by a mere statistical fluke. Similarly, a player who scores 20 goals in a season might score in six games in a row completely by chance; he's not on form but it's just the natural spread of random results. Therefore form does not really exist at all here. Our minds are simply tricking us by seeing statistical patterns and attaching causal factors to those patterns which are actually completely unjustified.
Fans and pundits following the Premier League fall for this trick far too often. They say that a striker who doesn't score for a while is experiencing 'a goal drought' rather than just acknowledging that this will happen sometimes in a natural spread of random results. Worse, if a team experiences a few bad results in a row we say the manager is overrated and should be fired, even though he achieved excellent results in the past (think Brendon Rodgers, Alan Pardew and the clowns who wanted them fired earlier this season). We really should be judging managers on a season-by-season basis (i.e. a sufficiently large statistical sample) not on an arbitrarily selected number if matches. Radomness plays a far greater role in sports than we like to acknowledge.
Postscript
The full hot hands study can be read here: