This is kind of an addendum to today’s football analytics post.
Couple of posts that caught my eye. One from Martin Eastwood, who has designed his own mathematical model to determine ideal team performance based on the relative skill of their opponent. The results are fascinating:
So what tactics should a manager choose? In the case of the underdog it surely makes sense to take the high risk strategy of attacking the match to increase their chances of winning all three points. The downside is of course that they are at more risk of losing by a heavier score line. But whether a team loses by one goal or by four goals, the net outcome in terms of points is the same – zero.
Over the course of a season it is much more beneficial to gain additional points at the risk of worse goal difference. One extra victory is worth more in league placement than having a superior goal difference to the teams around you. Plus, you need to hold on and scrape three draws to equal the benefit of getting that one extra win.
This is music to the football purists’ ears, and it certainly confirms my own personal bias regarding the relative, season long effectiveness of conservative play. My own favourite example? The United States defeating Spain in the 2009 Confederations Cup. An expansive, proactive approach, while risky, worked extraordinarily well. Compare to Germany’s tepid World Cup semifinal, in which Germany significantly altered their preceding gangbusters style and lost to Spain.
Another interesting post: Prozone’s Blake Wooster on leadership. He concludes:
Having studied and observed leadership in its many forms over the past few years, there is little doubt that no one style offers a holistic winning formula. Football would do well therefore to move away from the stereotype of a leader as only someone who shouts the loudest in the dressing room and sweats the most on the field. Maybe they’re sweating from having to recover from being out of position – something that Moore rarely ever needed to do owing to his composure, anticipation and game intelligence.
Can we measure true leadership in football? Any takers?
Interesting enough, I had a bit of an Allen Curve moment on way home to Toronto after SSAC. I met with Nicole Forrester, a former Olympian high jumper, who was completing her PhD in Sports Psychology. She was convinced that certain character traits aren’t merely intangible, but that a lot of work has been done in measuring for these attributes. I’ve since emailed her for papers on the subject and am waiting to hear back.
In any case, I don’t think this field is as flighty as some might think…
I think the state of soccer nowadays is such that if a lower/lesser team risks and goes for it against a stronger top 4 team for example and gets beaten 3-0, 4-0 or more….the fans start to turn on the manager and players and if you have 2 or 3 results in the span of 5-8 games go this way, the managment sacks you!!!
I am not saying this is right or wrong but ultimately management would rather be happy with the team grinding out draws than winning against a top team and then losing comprehensively against other teams…its a shame but most managment and owners are only people with deep pockets and dont understand the game and tactical analysis and other tangibles of the game other than the financials, TV revenue and staying safe/promotion
This is basic math, the problem is, as described above, having a chairman and fans that buy into the strategy.
I mean the strategy is made even more obvious in basketball, yet most coaches don’t grasp it …… if you have a vastly inferior team you want to increase the variance as much as possible – in basketball this means slowing the game to a grinding pace and shooting a high number of 3′s.
I don’t know what the correct style of play is for football but the thought pattern is the same – if you’re playing a top 6 side and you’re a bottom feeder you want to gun for 3 points and not be concerned if you get blasted 5-0 a few times in the process.
Wigan has been doing this for years – they taking some absolute hammerings but in return they always steal all the 3 points a few times a year …… much better than trying to grind 12-15 nil – nil’s.
The math extends to teams chasing Champions League spots – if you’re say expected to be a 4th to 8th place side coming into the year you need to be EXTREMELY aggressive when you play anyone “worse” than you and then also play really tight against teams that are at your place in the pecking order – this is why I’m not nearly as big of a David Moyes fan as the average person – his teams are way too negative/rigid in matches where they need to be gunning for 3 points ….. sure some years you’re going to run bad as heck and you might finish 10th – 13th in the table but eventually you will run good and those extra W’s could get you top 4 …… I think it’s much better than trying to grnd, grind, grind for 10 straight years and always inevitably coming up short. They simply tie way too many matches, they’re better off gunning for W’s in most of those.
I liked Easton’s article on risk, but I think assuming that more risk still results in a normally distribution of results just with a high variance is an inaccurate one. I have a feeling that by taking more risk you would see a more skewed distribution towards the losses rather than an normally distributed one. It’d be interesting for someone to measure win distributions for different formations, obviously there’d be a selection bias, but I think it could yield some meaningful results in terms of how risk pays off in terms of team shape.
^ I would agree that you would probably lose more than you would win, especially once teams catch on (until you adjust to their adjustment)
But you only need a 33% win rate to match 3 draws, so you have a lot of room. The real problem is not giving away the points to the teams in your area. So for example if you’re say an Everton/Liverpool caliber team you would sell out to win matches against say the top 2 sides in the league who are clearly better than you, and you would do the same for the bottom half of the table. Anyone from spots 3-6 you wouldn’t go kamikaze for fear of giving them the points.
Maybe in practice it would be more long the lines of:
if you’re at say the 70 minute mark in a tie game, how do you play the last 20 minutes?
Do you go out of your way to do things that are going to cause a goal for somebody (formation changes, pushing your line up, etc) .. This should depend on your relative place in the league and what your aims are
It just seems very weird to me that you have sides like Sunderland and Stoke who spend a lot of money only to play a low variance grind, grind, grind 0-0, 1-1 style against basically everyone ….. the way Stoke approaches away matches to big sides is epically stupid.