Do Pre-Match Favourites Maintain Their Advantage In A Penalty Shootout?
NFL and football fan. I've seen my two favourite sides, Stoke and the San Diego Chargers play at the new Wembley....and both lost.
Who is more likely to win a penalty shootout? The pre-match favourites or the underdogs? Today on the blog, Mark Taylor breaks down the data to the answer the question, do pre-match favourites hold their advantage when it comes to penalties?
The longer the duration of a sporting contest, the more likely it becomes that the best team or individual will emerge as the winner. This is tacitly acknowledged in such sports as snooker, where the number of frames played round by round increases as the effect of the seeding declines. A seeded player breaks off against an unseeded one over the best of 19 frames in the first round, but this has extended to the best of 35 by the final. Grand Slam tennis events similarly require the men to win a best of five set game rather than a best of three for the more mundane tour matches and the longest race at Royal Ascot, the 2 mile four furlong Gold Cup has been predominantly won by the favourite over the last decade.
Sport needs competition to thrive, but equally it should not appear overtly random and the length of a sporting contest appears tailored so that the most talented also enjoyed the opportunity to make their skills count, especially at the highest level.
Talent Plus Time
A 90 minute football match would appear to be close to the optimum length to fulfil the previous requirements. Spread over 38 games, the best teams always fill the top four or five places by early May, but contained within the season are occasional surprise victories, such as Blackburn's win at Old Trafford and markedly inferior sides are well capable of remaining competitive for significant periods of a game, especially at home. The balance between excitement and rightful reward largely prevails.
Better sides will be more accomplished at performing the actions that go towards winning a football match, such as passes, crosses, tackles, saves and shots and with greater frequency of these actions, the more likely it is that their superiority will be accurately reflected on the scoreboard.
We can demonstrate how match length relates to a team's chances of winning by plotting the respective win probabilities for two typical teams from the EPL as time elapses in a scoreless match. I've made the home side about 4 tenths of a goal superior to their visitors and recalculated the win/draw and loss probabilities as time ticks away and that initial 4 tenths of a goal superiority declines.
At kick off the home side has around a 45% chance of winning compared to just below 30% for their opponents. By half time the home side's win probability has reduced to about 38%, similar in level to the chances of the game ending stalemated. The shape of each plot demonstrates the constant interplay between goal expectancy and time remaining and slightly more goals are scored in the second half on average compared to the first as caution gives way to tiredness and increased adventure.
The win and draw probabilities from the 60th minute onwards give a reasonable representation of the extra time period in knockout competitions. Time available is now just a third of that available at kick off and even allowing for the likelihood of more frequent scoring, it is the draw that very quickly becomes the most dominant outcome.
The pregame favourite is much less likely to win under this reduced time frame, but they still remain more likely to win than do their opponents right up until the moment the referee blows for full time.
In short, favoured teams retain that advantage whatever the length of the contest, but the size of their advantage is dependent upon contest duration. Therefore it seems legitimate to consider how win probabilities change when we witness the shortest contest two sides can compete in, a penalty shootout.
We speculated that better teams assert their long term advantage by a steady accumulation of slightly superior actions on the field of play. During a penalty shootout a team will touch the ball around ten times (five for the shooters and an equal number for their keeper, if he's good enough). This number seems insignificant when measured against the number of passes, touches, headers and shots that occur during a game. However, many possession events that occur during a game are probing precursors to a possible shot on target.
Last gasp Champions in 2011/12, Manchester City, for example turned an average of 750 touches per game into 6.5 shots on target. Seen in these terms the outcome of a penalty shootout should at least resemble the outcome of a regular contest, stripped of the preamble. The gap in abilities will be narrowed by equality of opportunity, but the better team will retain a slight edge through superior keeping and shooting ability.
Considering Penalties & Penalty Shootouts
One way into this problem is to look at how efficient sides are at scoring regular penalties throughout the seasons. I therefore compared the failure rates from the penalty spot for teams with multiple EPL seasons since 2002/03 with their cumulative team success rate for wins and draws. Success rate is merely the number of wins plus half the number of draws divided by games played. A top 4 side will have a success rate of around 0.75 compared to 0.42 for relegation fodder and they are good indicators of performance levels over a period of games.
Penalties are relatively rare events, averaging between four and five a season per team, but the correlation is reasonably clear and in the expected direction for penalty kicks to be considered a talent. In general the higher a team's cumulative success rate over the last ten seasons, the lower their failure rate was from the spot. WBA for example managed to miss almost half of the 17 penalties they were awarded, while alternating between Championship and Premiership, but Chelsea missed just 13% of theirs, while winning the league, FA Cup and UEFA Champions League.
Penalty shootouts are of course most commonly played out during international tournaments. So to test our assertion that better teams take better penalties (and therefore should win penalty shootouts more frequently), I compared the results of 54 such contests from the World Cup, Euros, African Cup of Nations, Asian Cup and the Copa America. FIFA rankings, despite their poor press and occasional tweaks are an excellent proxy for team quality. So I recorded how often since the rankings were first formulated in 1993 the higher ranked side prevailed in the shootout and 36 of the contests saw the higher ranked side successful for an overall strike rate of nearly 67%.
Hopefully we have demonstrated how the chances of the favoured side decreases with shortening contest length and the logical conclusion that the favoured side's advantage should remain, albeit diminished in size even during the curtailed contest of a shootout is backed up by evidence from the EPL and various global tournaments.
However, the 67% strike rate should be treated cautiously. A 36 and 18 record could arise without triggering any statistical alarm bells if the average success rate of the higher ranked FIFA sides was around 53%. Nevertheless, even this extremely cautious conclusion is a useful residual advantage to take into the death throes of an international knockout tie.
Read more of Mark's work on his The Power Of Goals blog
And follow Mark of Twitter: @MarkTaylor0
You must be logged in to post a comment! Sign up + or log in in the top right corner.