Random variation: a need for skill and luck
The initial league format, followed by a two legged knockout phase and finally a single game to decide the winner, does mean that the successful side has to rely on both their own quality, but also trust random variation that is ever present, especially as the length of the contest steadily declines throughout the tournament.
Six matches, home and away in the initial four team group stage affords the best team in the group a better chance of progressing compared to a World Cup group where only three matches are played and a slip up is difficult to retrieve. A typical group favourite in the UCL will have upwards of an 80% chance of progressing with a top two finish, compared to around 70% if the truncated World Cup format is applied.
Therefore, quality in abundance is guaranteed in the knockout phase, although as Jose Mourinho ruefully acknowledged following Chelsea’s elimination by Atletico, ties can then swing on a single outstanding save at one end, quickly followed by a penalty decision at the other.
As the format moves towards a single winner takes all show down, the chances of the weaker side causing an upset increases.
Should Real be favourites?
Real Madrid arefavourites to lift the trophy – which is similar for the two La Liga matches between the two clubs this season. Real were, unsurprisingly strongly favoured to win at home (57%) and were also given around a 45% chance of winning the return, just over a month ago.
If we look at a comparison of the individual odds for all league matches played by both Real and Atletico over the course of the season to May 5th, the former has been consistently rated superior to the latter. Real has been a shorter price to win than Atletico in every fixture involving common opponents played so far.
Converting the win and draw odds for each game played by both teams into probabilities and then into expected points, Real have been projected to gain around 82 points from their 35 matches compared to only 74 for Atletico from their 36.
Real are currently on 83 points, largely performing to expectations, but Atletico are 14 points ahead of the curve with 88 points and lead the table by three points. In addition Atletico also hold the tie breaker having taken four points from their two derbies with Real.
So the oddsmaker’s consider Real Madrid superior, but actual results from this season appear to suggest that Atletico are at least their equal as the season moves towards a climax. This apparent contradiction can be partly explained if we accept that results, even over a 38 game season are the product of a side’s talent and also random variation.
If the assessment of the chances of both Madrid sides in their total league matches played is broadly accurate, their expected number of points is merely a random draw from many possible realities, some more likely than others.
Simulating points outcomes for the season to May 5th, Atletico accrue a range of possible points totals from the low 50’s to the low to mid 90’s. Similarly, Madrid’s points range lies from the high 50’s to occasionally breaching the 100 barrier.
Above we’ve plotted the outcomes of simulated seasons, expressed in terms of the point’s gap between the two sides in these virtual campaigns to date. The simulations have used the match odds over the season to date, which we have commented are more favourable to Real.
The most likely outcome from the simulations was a 9 point lead for Real, which neatly reflects the expected points from the match odds for each side. However, in 20% of the simulations, it is their neighbours whom hold the advantage and there is a cumulative 10% chance that Atletico hold at least the 5-point lead over Real that they actually enjoyed on May 5th.
The importance of head to head games is also borne out. As Mourinho said, individual games can be decided by single incidents and in simulations where Atletico mimicked 2013/14 by taking four points from the derby games with Real Madrid, their chances of heading Real in the table going into the final few games leaps to 35%.
In short, over a season there is a significant, minority chance that the better team can find itself below an inferior team in the table as the season draws to a close.
Are Atletico over-performing and could they regress?
Atletico are undoubtedly a fine side, improved under Diego Simeone, but it is perfectly possible that they have been good and fortunate in occupying a higher league position than their superior home city rivals.
Advocates of Real’s superiority can also point to Atletico’s inferior goal difference. A team has some control over how many goals they score or concede, but less so on when and where they do so.
Newcastle, famously recorded a goal difference of just plus 5 in the 2011/12 EPL season, typically good enough for a mid-table finish. But they achieved the heady heights of 5th spot and 65 points, partly by scoring in games when it mattered and conceding additional goals in matches that were already irretrievably lost. A more normal distribution of goals would have led to a lower finishing position.
Atletico’s goal difference and likely finishing position in 2013/14 is less extreme than the case of Newcastle, but it is more in keeping with a side accumulating around 2.2 points per game than their actual rate of over 2.4.
Overachievement by Atletico is also evident in the outcomes of their goal attempts so far in the UCL. Whereabouts on the pitch a goal attempt originates from and whether they are shots or headers are significant factors in whether they will result in goals.
Atletico have both outshot their opponents overall and managed to create efforts that have, on average been taken a couple of yards closer to goal than those they have allowed. So they are producing good shooting volume and chances with a higher goal expectation.
They have also fashioned over twice as many opportunities that carried a particularly high expectation of a goal than they have allowed. 17 attempts had at least a 20% chance of providing a goal compared to just 7 such efforts for the opposition.
Overall, based on shot location and type, Atletico’s 166 goal attempts might expect to yield 16 goals, but they have actually scored 25 and defensively they have also out performed this basic model, allowing 6 goals, where 10 might reasonably have been expected.
The reasons how a side can over perform against shooting models isn’t yet well understood. Finishing talent, defensive pressure or lack of it, caused possibly by a counter attacking style may be relevant, but random factors will also always be present. Once again, impressive as Atletico’s numbers are, they may also be buoyed by good fortune, which may eventually subside, although there is no guarantee that this will happen.
Real Madrid share some of these over performing traits, particularly in shooting, but their record is more prolonged and while Simeone’s appointment, along with a counter attacking strategy may account for some of Atletico’s improvement, we should not be blind to the possibility that they may also have been fortunate.
The bookmakers are siding with the team they consider to be the better side in Real, even at the stage of the tournament where we are down to a maximum of 120 minutes of playing time and the contest is at its shortest. Also the perception that more money will continue to flow towards Ronaldo and co may have pushed that stance even further. (Learn about market movement here)
But bettors backing Atletico eager to see value, should at least consider the possibility that part of their great run in 2013/14 – a two legged Copa Del Rey thrashing to the very same Real Madrid aside – is down to effects that may not repeat in the long run.