Manchester City face Inter Milan in the 2022-2023 season Champions League final on Saturday, June 10, 2023. Manchester City is a heavy favorite and has become a dominant force in world football/soccer since the arrival of their star manager, Pep Guardiola.
The Champions League is currently a combination of six group-stage games home and away in eight groups of four teams, with the top two from each progressing to the ‘knockout stage.’ There the teams play home and away against their knockout opponents, with aggregate scores from the two games determining who advances through to the quarterfinals, semifinals, and then the final. The final is a single game, with this year’s being held in Turkey.
While called the ‘Champions League’ field each season has been diluted to include many non-champions from the various domestic leagues, with four from the English Premier League (EPL), for example. Manchester City has won the EPL trophy five of the past six seasons - North American-centric fans can think of this as the ‘regular season,’ whereas the Champions League knockout rounds are closer to the ‘playoffs.’
Inter Milan in contrast, finished second in last season’s Italian Serie A and third this season, yet have made their way into the finals. With football being a relatively low-scoring sport, despite being a heavy favorite, it should not be a surprise if Inter Milan is able to win despite Manchester City being a juggernaut.
Despite being dominant the past six seasons in what many consider to be the highest-quality league in the world, Guardiola’s Manchester City side has lost in the Champions League in the quarterfinals three successive seasons, then the finals and semifinals the past two seasons. Will this finally be the season in which Pep gets them over the last hurdle?
After being acquired by an Abu Dhabi-controlled entity, Manchester City has gone from a financial minnow in England to a top-three rank in wage bill. In the last few years, the controlling company has invested heavily in modernizing operations and its analytics staff. In the modern sporting world, similar to the migration for Wall Street in the 1990s, STEM PhD’s have taken over those responsibilities at most of the best-run clubs. For example, City now has a PhD in computational astrophysics heading up their move into artificial intelligence/machine learning.
Despite this project to get ‘smarter', Guardiola has developed a reputation for making some erratic team selection and tactical decisions in some of these highly pressurized knockout games over this period.
Overall, football has lagged behind North American sports with the adoption of ‘big data’ and more advanced statistical analysis. The era was obviously ushered into the awareness of the broader public via the book and movie, Moneyball. Notably, the team profiled in the movie, the Oakland Athletics, has failed to progress through the MLB playoffs in any subsequent season.
Whether it is a machine learning model or Peter Brand looking at a spreadsheet, the idea was and is to look for competitive advantages through the lens of arbitrage and/or optimization. Perhaps the baseball club that has been most successful at this endeavor has been the Tampa Rays (about $76 million), who has contended in MLB’s most difficult division most seasons, despite playing in a small market in a mostly empty stadium, and with a payroll a fraction of the New York Yankees, Boston Red Sox, and Toronto Blue Jays.
The Rays have not quite gotten to the promised land either, though they have been close by reaching the World Series. Perhaps a better comparison to Guardiola’s Manchester City has been the Houston Astros. The Astros have combined a club with significant financial resources with being run as one of the ‘smartest’ run organizations in all of professional sports. The two clubs also share various allegations of ‘cheating.’
Now that I have lost most readers, both of you who have made it this far are probably thinking, “What the hell does any of this have to do with the usual nonsense that comes from Kayfabe Capital Towers?”
Much of modern data-centric economic and financial market analysis suffers from similar issues as Man City, Rays, and Astros. Can this create a false sense of ‘smart’ when in fact it may be a delusion? Does the data being used to feed models reflect reality? Are all or even most factors being captured? What about intangibles - both are attempting to model the behavior of humans. We see this issue with current limitations in AI tools, as the general intelligence of homo sapiens is not the same as an advanced optimization algorithm.
I played baseball as a weekend warrior from the age of eight to 44, with much of my life running parallel to the data and analytics revolution in the sport. The 2019 World Series offers an excellent case study of the challenges presented through the lens of analytics relative to balancing longer-term optimization with the here-and-now, IMO.
Like with any single football match, a playoff series or single game within something like the World Series introduces far more variance. What may be a huge statistical advantage over a 38 or 162-game season may not be representative within the specific circumstances of a game.
One of the ‘bible’ analytically-driven tendencies many baseball teams have adopted is based upon historical evidence of batters experiencing far better success in a game once they get to face a pitcher for a third time. I believe this was likely at the core of the decision to remove Zack Greinke in game seven of the 2019 World Series by the Houston Astros.
Greinke had been dominant, conceding only a single hit and no runs through six innings. In the top of the 7th, Greinke surrendered a solo home run to make the game 2-1, still in the Astros’ favor. He was then removed after walking the next batter, who was the Nationals’ best hitter, Juan Soto. Soto has been the only player to have a hit prior to the home run, but almost no hard contact had been made against Greinke.
Greinke was an older pitcher by then, known as a heterodox and cerebral pitcher that no longer threw very hard. Basically, he had evolved into what is known as a ‘junker baller’ who uses changes of speed and varying movement on his pitches to keep batters off balance and induce weak or no contact.
For a standard game in June, in just 1 of 162 regular season games, ‘playing the odds’ could be contextualized as part of a broader and effective decision-making framework. Even if all the variables related to that specific game, including Greinke’s specific characteristics, over 162 games the discipline of sticking with the framework may be optimal.
Unfortunately for the Astros, game 7 of the World Series is not the same. The pressure and stress are on a different planet. What may be an x% win probability decision using a big data and machine learning model was/is kayfabe, IMO. That is a false sense of being ‘smart,’ as the dynamics are so…..human.
Some people wilt under such conditions, while others do not. One can go through all sorts of training and conditioning, such as soldiers, yet the specific responses of a small group once in the trenches may be wildly variable versus decades of ‘big data.’
Having played the game for many years, I know what it is like to face a ‘junker baller’ who is dominating, let alone within that sort of pressure cooker, which I cannot imagine. Obviously, the benefit of hindsight offers an unfair context, as the removal of Greinke ended up as a mess.
I was screaming at the TV at the time that it was a terrible decision - not because of some ‘model,’ but because of my human intuition as to the very specific situation at that time. Greinke was/is the sort of quixotic personality for whom ‘pressure’ is less likely to impact. As game 7 in a World Series moves along, I could imagine the grip on the bat getting a bit tighter, with the prospects of facing a junker baller on a role being down 2-0 as daunting.
Of course, my own biases as a former cerebral junker balling pitcher may have driven what some would see as an irrational decision. But this is at the root of all decisions and the probabilistic modeling that people and institutions engage in to try and reduce risks and navigate what Yogi Bera famously said:
It’s tough to make predictions, especially about the future.
The use of advanced statistical models is ubiquitous in economics and on Wall Street. I see historical studies citing x% of the time when this or that happened meant this or that. Breadth thrusts, seasonality, presidential cycles, moving average crosses, etc. - it is an orgy of varying levels in quality of analysis for this specific point in time, with many closer to astrology than being efficacious about predicting the future, in my opinion.
Markets and business cycles do not involve reliably stable statistical relationships. In that specific way, every cycle is actually ‘different this time.’ I view that home run off of Greinke as having been a ‘bear market rally,’ regardless of what 3rd-time-through-the-order probabilities suggest.
I remember reading an analysis in January 2001 declaring that the US stock market had never been down 6 or 12 months after the Federal Reserve had cut rates x # of times. I remember 2008 and the belief that home prices do not go down nationwide.
As the likes of Benoit Mandelbrot and Nassim Taleb have written, the real world does not conform to the statistical models still widely used across economics and Wall Street.
‘Black swans’ have a way of defying backtested statistical models. With Guardiola’s team around a 75% favorite for the final, let’s see if any darkened waterfowl descend upon Turkey.
Unfortunately, for those investors currently enamored with various backtests, I suspect they may be facing a Taleb-Hitchcock collaboration in the coming months.
That feeling when the sports blog post you're reading unexpectedly turns into a finance blog post...great writing, it really came together nicely. The Yogi Berra quote was just classic too.
This was exceptional. We both share a more than casual interest in America’s Game & The Beautiful Game. As a long suffering Spurs fan, I watch City’s dominance with respect for their dogged commitment to a process that demands progress & scorn for my crosstown rivals, gunners, for the brilliant side they have built from the ground up with Pep’s understudy. Loved the piece but the match, while competitive, was a bit of a slog. That said, even with loss of KDB, City stayed committed to the plan & we’re just good enough, winning a match that they would have lost in past. Thanks again. Woooo!