Wembley beckons & Gambling strategies
Analysing a routine semi-final win for Tuchel's side where VAR struck thrice and more updates from the betting strategies devised by David Sumpter in his famous book Soccermatics
Credits: CHELSEA TV, FBREF, The Athletic, David Sumpter, FourFourTwo, Liam Twomey, Mark Carey, StatsBomb, Jan Van Haaren, McKay Johns
Contents:
Match of the week: This week we have another match to analyze and this time we take a look at the underlying stats at an aggregated team-level rather than an individual player
Reading updates: Continued to read Soccermatics and some interesting updates below from Chapter 12 “Putting my money where my mouth is”
Community Briefing: Few interesting bits from Twitter and a longer blog post currently viral on the social media circles that force you to put your thinking hats on!
Self-learning Progress: Updates of changes pushed to the GitHub repo and we take a look at some references from YouTube and lastly a quick summary of this week’s Podcast that I listened to
1 Match of the week
EFL Cup: TOT v CHE (2nd leg) | Date: 12/01/2022
Formation: Interchangeable 5-3-2 and 4-2-2-2 and their variations | Result: 1-0
Match notes dump:
Rudiger scores (yet again) at the Tottenham Hotspur Stadium. He scored at the same ground back in September in the 3-0 win and scored from a set piece again. He now has 3 goals to his name this season. Speaking of set pieces, Chelsea have now scored 9 goals (20%) from set pieces and including penalties that number is more than 1/3rd (36%) of all goals scored this season.
Werner-Lukaku link-up play - This was quite evident from the number of passes and key passes they played to each other. Playing 2 up front in attack the pairing worked in a similar fashion as to how Havertz-Lukaku had operated in the 1st leg of this tie.
CHO-Mount flexibility enabling Chelsea to play in 2 different setups - (4-2-2-2) in attack where they form the attacking midfield pairing behind the two strikers as clear from the 1st image below (taken from inside the first few minutes) and transitioning into a (5-3-2) in defence where CHO slots in at LWB and which pushes Sarr into the three-man defence alongside Rudiger and Christensen. Mount too drops back and supports Jorginho and Kovacic in the middle block of three.
Opinion:
Another feature of this rare 4-2-2-2 setup was the way Sarr was overlapping down the left flank when Chelsea had the ball as clear from the 2nd image above. Though not entirely productive given his limited abilities in attack, it (Sarr drawing a Spurs defender out of position) was used with great effect to allow CHO and Werner to make dangerous runs in behind.
Despite the amount of possession in the final third, creativity and inventive play was largely lacking from the Chelsea front four. On numerous occasions poor decision-making meant Chelsea never really managed to put the game to bed.
Transitioning between a back 5 and a back 4 (out of possession and in possession respectively) came with a tradeoff of having to let go the solidarity at the back resulting into the team being susceptible to counter attacks as visible in the last 2 images above. Lack of coordination between the back 3 and spaces to exploit between the centre backs and the wing backs lead to dangerous Spurs’ counter attacks suggesting some serious disorganisation.
Effect of which was notable VAR influence: 2 penalties over-ruled, 1 goal disallowed for offside. What might have been a Harry Kane hat-trick on another night - would only be settled with Andre Marriner losing friends on the night!
Wembley beckons! Overall across 180+ minutes of football, Chelsea were clearly the better side and deserved to progress to the final. It’s one more week until they learn their opponents for the final but until then this semifinal win over their London rivals being managed by a past manager will be enough to keep the spirits high. Fun fact: Thomas Tuchel is yet to lose a national cup final spanning across his 3 recent spells at Dortmund (twice finalists, once a winner), PSG (once finalist and three times winner) and Chelsea (2nd final in row).
Why is Tuchel having to experiment?
How ironic is it to be speaking of this experiment given the name of this newsletter! In this particular article on The Athletic, Liam Twomey (Chelsea correspondent for The Athletic) looks at the prospect of reusing the 4-2-2-2 setup and the pros of which can be summarised as follows:
Last week, Tuchel’s response to the Lukaku Sky Italia interview fiasco was: “"If you put a little bit of work in, in our systems, you find not a lot of systemic changes, if you find any”. So far in his Chelsea tenure, Tuchel has deployed 3 variations of the three at the back: 3-4-2-1 (most used), 3-4-1-2 (variation where split strikers in front of No 10) and 3-5-2 (three central midfielders to provide more cover at the back used usually when defending a slender lead).
3 draws, 1 defeat and 1 win in the last 5 PL matches suggest something that was working so well at the start of the season isn’t working anymore reasons of which will need further looking into than just tactical setups and personnel in the team. Injuries to Ben Chilwell and Reece James have meant a variety of players have had to fill in for their slots at left and right wing backs respectively. Alonso and Azpilicueta are the most obvious replacements but they may not be the most effective in those roles. And hence in order to provide more flair going forwards, Pulisic, CHO and Ziyech have all started in those positions recently.
Maximising Lukaku’s strengths and providing a consistent striking partnership to divide attention of defenders in central areas. This seems the most significant reason for the experiment given the timing of it. Lukaku was part of a 3-5-2 setup at Antonio Conte’s Inter where he paired up impressively with Lautaro Martínez last season. It would be interesting to see if Tuchel continues to deploy this high reward - high risk strategy in the matches to come.
Some key stats to put the Tuchel era under the microscope:
Furthermore in the same article, Liam has compared some of the prominent stats from Lampard’s Chelsea (his last 19 matches in the PL) vs the present Chelsea under Tuchel (21 league matches this season). Here is a different version of those Pizza charts to give us a clear picture of how the team has transitioned since Frank left and I leave it up to you to form your own verdict:
*Stats’ definitions (as defined on FBREF):
Field tilt: Measuring the share of possession a team has in a game but only in terms of touches or passes in the attacking third — and it is a useful measure of how often a team was able to take the attacking initiative in a game, rather than just holding onto possession for the sake of it.
xG: xG (or expected goals) is the probability that a shot will result in a goal based on the characteristics of that shot and the events leading up to it. If you want to have a detailed read at the expected goals model currently used widely across clubs, leagues and media, this explanation on FBREF is brilliant!
PPDA: This is a relatively new measure and can be broken down as being equal to (Number of Passes made by Attacking Team) / (Number of Defensive Actions)
A smaller PPDA value signifies a greater level of defensive intensity, as in essence, the defence has allowed a smaller ratio of uncontested passes to be made.
2 Reading updates
Continuing from last week, I completed reading Chapter 12 titled “Put my money where my mouth is”. To give a gist of what’s covered in the chapter, the author tries to put a portion of his hard earned money (from writing this book itself) into a pre-calculated gambling strategy and hence this title of the chapter. Throughout the chapter there are simple explanations provided to complex concepts within the betting industry:
How betting odds are different from probabilities?
In order to incorporate a built-in advantage for the bookmaker’s, it is easy yet not obvious to observe that the odds listed against a match outcome almost never reflect the expected probabilities of that outcome.
For example, if we look at the pre-kickoff odds (European) for Manchester City vs Chelsea on the top online betting platforms, they are averaged at 1.69 (City win), 3.97 (draw) and 5.26 (Chelsea win). We can convert these into probability equivalent values by simply finding their reciprocals. So we get these values to be 59.17% (City win), 25.19% (draw) and 19.01% (Chelsea win) suggesting that the bookmaker’s think City are likely to win the game as much as more than 3 times compared to Chelsea. And as you might have already realised, the probabilities add up to 103.37% indicative of the 3.37% cut for the bookmaker’s to take home as their profit when averaged out over the entire crowd. Thus, it is mandatory for us to adjust the betting odds so that they add up to 100% before using them in our calculations.When to bet and on what outcome?
David recommends doing a simple check before you gamble away with your friends the next time. Leaving aside the social aspect to betting and looking at it from a more objective and practical standpoint, he suggests we should only consider placing a bet on a match result if and only if p > (1/o); where ‘p’ is the expected probability of an outcome (win/draw/lose) and ‘o’ is the European odds’ value for that match outcome. With plentiful practical knowledge on offer, a dedicated article to follow later this week where we’ll look at how these betting strategies can be worked out using a few quick google searches, basic probability know-how and an Excel sheet to answer the questions imposed here!
3 Community briefing
Mezzala Football: Stumbled upon this Twitter account over the weekend. It has cool examples of using visualization techniques to analyze players. Given we are in the middle of January, you can explore some of the recent posts to get a glimpse of these techniques used to provide a dashboard that can then be used to compare similar players for scouting and recruitment purposes.
Statsbomb article: “PL 2021-22 - the story so far”
https://statsbomb.com/2022/01/premier-league-2021-22-the-story-so-far/
Mid-season review from StatsBomb again featuring series of intriguing and debate sparking visualizations. One of the charts can be seen in the image below which looks at the xG difference [ (xG or the expected goals for) - (xGA or the expected goals against) ] for all the 20 teams. The ranking isn’t entirely different from the league table at this stage but there are few notable movements (Crystal Palace at 6th and Man United down in 10th the most prominent ones) enough to raise a few eyebrows!
4 Self-learning progress
Check out my latest set of updates to my main GitHub repo. I push frequent updates on a weekly basis and the code/resources pushed there should complement the content presented in the weekly blogs.
Manually add shots on the pitch and plot using R
Another YouTube video that I encountered during the week. It walks you through the simple and intuitive process of creating a shots’ map by manually annotating on a free online tool and then later translating it over to a chart created in R.
The Athletic Tactics pod notes
Episode name: “Pep's Innovation, Lukaku's System & Ralf's False Start: Premier League Notebook”
You can listen to the Podcast on almost all major audio streaming platforms (Spotify, Apple Music, Amazon, The Athletic to name a few).
City unstoppable already? 12 different goal scorers with 2+ goals and the highest goalscorers (B. Silva and Sterling) on just 7 goals apiece suggest Guardiola’s rotation policy is in full swing and the lack of reliance on an out and out striker also supported by the selling of Ferran Torres in this window.
Chelsea v Liverpool (2-2) the best 1st half of the PL so far this season? Chelsea tactics and Chalobah/Left Wing Back/Striker issues touched upon.
Arsenal passing sequences: (shots/goals) statistic looking good for Arsenal and has significantly improved of late ever since the fallout of Aubameyang with Arteta. More direct attacks and better link-up play in the final third might be signs of better things to follow.
Man United: Looking at the PPDA stat again, all the talk around Ralf Rangnick’s high pressing tactics doesn’t quite make sense. United have been uninventive to say the least in the last few matches and Rangnick might have more than tactics to sort out.
4-way relegation battle with {Burnley, Watford, Newcastle, Norwich} discussed in depth. From what I could make out, Norwich look doomed already. Watford and Newcastle trying their best to recruit some fresh crop in January to scramble for safety. Burnley have the history backing them and from their recent experiences of playing with the relegation spots might be crucial in avoiding the cut.
Quote of the week:
“The hallmark of successful people is that they are always stretching themselves to learn new things.” - Carol Dweck