How Large Leads Can Alter The Dynamics Of A Sporting Contest
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.
How do football teams change their approach when leading by a large margin in a match? Today on the blog Mark Taylor takes a look at the tendancies of teams enjoying comfortable margins in both Association Football & the NFL.
The quality and quantity of data that is now available to use to deconstruct a football match has never been greater. Goals have been supplemented with shots, passes, tackles and interceptions to name just a few, all time stamped to allow a better understanding of the subtle interactions that occur within the ninety plus minutes that make up a game.
Football is a relatively low scoring sport. Premiership games yield an average of about 2.5 total goals per match and trends that occur as a result of the current score may be difficult to spot in the bulk of unexceptional scorelines. It is therefore often a good idea to take a long look at games where exceptionally unusual scoring events have occurred and shifts in playing style by either team may be more pronounced and easier to spot. If such shifts of playing emphasis can be seen in these type of games, it may indicate that they are present in all games, but to a lesser degree.
An Example: Wigan vs Chelsea
The opening week of the Premiership season produced a couple of high scoring wins for both Fulham and Swansea, an uncomfortable bow for Brendan Rogers at Liverpool and a second consecutive 3-2, come from behind win for Manchester City at the Etihad Stadium.
Chelsea also won 2-0 at Wigan, on the surface an unsurprising outcome, but both of their goals were scored within the first seven minutes and the visitors, who were already strong favourites to win at kick off very quickly built up a virtually unassailable lead. Teams such as Chelsea visiting the likes of Wigan are likely to find themselves 2-0 up in such short order around one time in 200 attempts.
Inside The Numbers
So how did the frantic and rare opening to the contest impact on the rest of the match?
Chelsea, the Champions of Europe, is naturally one of England's finest teams. Last season they averaged 15 shots per game away from Stamford Bridge and around half of those efforts were on target, conceding only 12 with 7 on target.
However, having taken an early grip on Sunday's game they didn't come close to matching last season's average for goal attempts on the road and in the remaining 83 minutes they fired just four additional shots at goal. In stark contrast, Wigan was much more adventurous, firing 15 shots towards the Chelsea goal, four of which were on target.
So did Chelsea sit back or were they being pinned back by Wigan's urgent need to score, resulting in a last 83 minutes that was possibly untypical of a Chelsea Wigan game.
A Meaningful Comparison
It's easy to use data from a single match to create a narrative that is persuasive, but unrepresentative of a larger group of games. So I took a larger sample of Premiership games where one team was leading by at least three goals at the half and compared the average supremacy enjoyed by the leading team in the second half to how they would have been expected to perform post half time by pre-game estimates.
As you would expect teams that enjoyed such a comfortable half time team talk were on average much superior to their trailing rivals and pre-game they had around a 63% chance of winning the match.
In goal terms this is the equivalent of winning a large number of repetitions of such a match-up by an average of 1.1 goals. Goals are more prevalent in the second half and after allowing for this phenomenon, a team who was 1.1 goals superior to their opponent would expect to have an average second half supremacy of 0.64 goals. However, in my sample, such teams when leading by three or more goals at half time "won" the second half by just 0.5 of a goal. Well shy of the pre-game estimation of over six tenths of a goal.
It appears that as with Chelsea at Wigan, teams in command of the scoreboard partly through choice, partly through necessity, drew back and allowed their badly trailing opponents to create more, but lower quality chances than they would usually allow and made do with many fewer, if better quality chances for themselves.
A Study: The NFL
The object of all team sports is the same, to perform tasks that result in credit on the scoreboard and ensure that you are ahead of your opponent at the end of the game. If teams try or are forced to achieve this aim through shifting the emphasis of attack and defence, might it not be present also in sports other than association football ?
The NFL is currently in preseason and in terms of available data has been well ahead of football for many years. Consequently it has been relatively easy to track whether an American Football side is adopting an adventurous, higher risk passing strategy aimed at putting more points on the board or is merely trying to close out an already secured win by using a lower risk, less reward, running approach.
Betting lines in the NFL are just as reliable and accurate as those for football. If you group together every team over a stretch of time that was predicted to win by six points, you'll find that's pretty much what happened. We can therefore use the average betting lines to see if teams who were leading by large margins at the end of the second quarter keep going at comparable rates or whether they overtly adopt a more defensive strategy similar to that used by their round ball cousins to try to ensure victory.
Playcalling On 1st Down
Since the start of the nineties over 800 teams have found themselves leading at home by a margin of two or more touchdowns and on average these teams were favoured pre-game to win by just under six points. Points scoring is split more evenly between the halves in the NFL compared to football and slightly less points come after the interval, mainly as a result of leading teams running out the clock inside the final two minutes even if they are well placed to add to their score.
Therefore, pre-game estimates from universally accurate betting lines would suggest that such six point favourites who through a combination of good play and good fortune find themselves up by two or more scores at the half would "win" a second half by an average of about 2.5 points. But as with football, NFL teams carrying big leads into the locker room only "win" the second half by a fraction of the pre-game expected margin. In this case they win by only half a point.
As with Chelsea’s shooting statistics at Wigan, we can demonstrate this shift of emphasis in the NFL if we look at play calling tendencies in relation to the scoreboard. 1st and ten is a fairly neutral position for a team, it isn't an obvious passing play such as 3rd and 13 for example. So how a team's proportions of passes and runs change on 1st and ten will be largely a function of the scoreboard.
When leading by a large margin, a team runs more. This won't lead to frequent, rapid scoring, but it will virtually guarantee possession is maintained and just as importantly, the clock will remain running hastening the end of the game.
By contrast the trailers adopt a higher risk, but potentially higher scoring, pass heavy approach and as a result they get closer to their superior opponents during the second half than they would have done with a less extreme scoreboard, but not close enough to often threaten their large lead.
We've seen two sports where large leads appear to result in the trailing side partly through desperation, partly through their opponents reverting to a less attacking mode being able to play slightly above expectations for the remainder of the match. The better teams, as measured by pre-game estimates unsurprisingly remain superior once they have jumped to a large lead, but their superiority is diminished. Football teams who lead by three or more goals at the half are about 20% less potent in the second half than you would expect them to have been had the game still been tied.
In running is a complex time where many competing factors are subject to subtle change, but uncertainty also brings the opportunity to make informed judgement calls. As a general rule trailing teams become slightly more likely to score in a match and winning teams become slightly less likely than they were prior to the change in scoreline. It's a point worth remembering even when the season opens with a couple of 5-0 wins.
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.