In financial market investing, the term Alpha refers to profits that are generated through genuine investing skill.
Imagine you give your money to an investment manager who buys shares on your behalf. At the end of a couple of years your money has grown by 10%. You’d be happy, right?
But what if the whole stock market had grown by 15% in the same time? Suddenly the 10% profit your guy made you doesn’t seem so great. You would probably have done better if you had just picked a bunch of stocks yourself at random. Or gone with something that tracks the market, like an S&P 500 fund.
Alpha is the difference between what an investment makes and what a generic investment vehicle like a market tracker would have made. It is a way to identify genuine skill in investing, controlling for macro market conditions that are out-with any individual’s control. It is the excess return over a fair benchmark, stripping out the effect of market conditions.
Can the same principle be applied to the people who play, and bet on sports? Yes, we think it can.
Take a bookmaker. How does a bookmaker make money? Ask a load of people to answer that question and you’d probably hear answers like ‘bookmakers always know who is going to win’. But that’s not true at all. Bookmakers can’t know for sure who is going to win a horse race, football or tennis match any more than the punters really.
The reason they win, and the punters lose long-term, is that bookmakers have a margin in their favour when they offer odds. Often known as the ‘over-round’ or ‘vigorish’, bookmakers add on this theoretical profit margin to their best guess of the true 100% price for any outcome, much as a baker adds his profit margin onto his costs of making a loaf of bread.
So when a bookmaker makes a profit from its customers, on prices that include an over-round, it is just doing what it is supposed to do. They don’t need to know who is going to win, they just need to do a decent job of estimating how likely things are to happen, and remember to add on their margin. Statistical probability and the law of big numbers will do the rest.
So how would we measure the Alpha of a bookmaker, or more specifically its compilers/traders who set its odds? Well, you could measure the difference between the theoretical margin that they bake into their prices, and the actual profit they make at the end of a reasonable period of time (like a year).
If an odds compiler adds on 10% over-round to his prices but only makes 5% profit from the bets placed on them, then you could say he had negative Alpha. He still made a profit, so he deserves his salary. But probably not a bonus.
At investment banks and hedge funds the culture of paying exorbitant bonuses on such negative Alpha performance was a large contributor to the global economic crisis of a few years back. You might think they should have known better – but it is a common human failing that we tend to focus on headlines, and ignore the underlying truth contained in the smaller print below. Glorying in superficially impressive looking results (like profits on trading sub-prime mortgages), but failing to control for the wider context is prevalent in virtually all walks of life. Even among companies who look at profit %s for a living.
Professional sports has more than its fair share of hyperbolic headlines, and unwarranted claims toward greatness. At OddsModel, one of the benefits of our analytical and dispassionate ways of looking at sport is that we can see past the superficial, to the genuine Alpha of sporting performances.
Different sports need different techniques and controlling adjustments to peg performance to a fair benchmark; some fairly minor, and others a whole lot bigger.
In Formula One for instance, what is the Alpha of a driver who drives a Mercedes and gets 350+ championship points in a season versus a driver sitting in a Sauber and only getting half a dozen points? Is the Mercedes driver really 50 times better than the guy in the Sauber? Such is the importance of the car in F1, probably the only decent way of measuring racing driver Alpha is to compare qualifying and race performances of drivers in the same team.
In a sport like tennis the equipment the players use isn’t nearly as important. At least, they all get supplied with top-of-the-range rackets and shoes. But the seeding system of tennis tournaments means that, paradoxically, the best players get to play easier opponents on average in the early rounds than the weaker players. So you could argue that Alpha is greater for a tennis player breaking into the top 16 than it is for staying there.
What sports are set up to be pure tests of sporting Alpha? Golf is good. There’s no seeding, everyone starts each tournament on level par. There’s little or no equipment advantage for anyone. Individual athletics events are pretty pure too. Whether it’s the 100m or the marathon everyone wears pretty much the same shoes and kit, and the race is a pure test of who is the fastest. Unless somebody had been cheating by taking drugs of course, but that’s a story for another day.
In US Sports like baseball and basketball it is common now to talk about a player’s VOR: ‘Value Over Replacement’ (sometimes also Wins Above Replacement). This is a metric that attempts to measure how much better/worse a team plays (or how many more more wins/losses it has) when a given player is playing, versus a benchmark ‘standard’ alternative. VOR is effectively a measure of sporting Alpha.
In team sports like football which are fluid and involve greater numbers of players than a game like basketball, measuring the Alpha of individual players is a tricky business indeed. This is especially the case with midfielders and defenders whose role and worth to the team is less easily quantified with stats than goal-scorers and providers.
Nobody who has ever seen him play would dispute that Lionel Messi has Alpha. Lots of it. But how much? How do you measure it, and put a numerical value on his contribution over a benchmark replacement?
At OddsModel we work with ratings systems that put number on player contributions, but we are aware that they will always be imperfect measurements due to the massively complex interactions that go into making up team dynamics in a fluid invasion sport like football. Does Messi have such an incredible goal-scoring record only because he plays with guys like Suarez and Neymar? Would his Alpha and his goal-scoring be more or less if he played for a mid-table team?
You would think he would score far fewer goals without the dominance of possession, and the multitude of chances that his Barcelona team provides for him. But perhaps if he played for a lesser team he would be even more the focus of his team’s attacking game, and would score something like 80% of an ordinary team’s goals, rather than about 30% of Barca’s. Which one would equate to more goals at the end of the season? Which contribution has greater Alpha?
Measuring the Alpha of football managers is a lot more straightforward. You can get a really reliable benchmark of performance for any professional team by looking at how much money it spends on wages compared to its rivals. For a football manager, winning trophies doesn’t necessarily equate to having Alpha.
For a manager, his club’s wage budget is the equivalent of the power of the engine in the car for an F1 driver. Winning a few trophies in the first couple of seasons with a team who spends more on wages than your competitors, but then overseeing everything falling apart in your third season is the equivalent of a driver winning a few Grand Prix in a Mercedes, but ultimately finishing 100+ point behind your team-mate in the championship. Or an investment manager making 10% on the stock market when everybody else is making 20%.
The Alpha of a football manager needs to be measured over a medium-to-long term (a 38 game league season is short-to-medium) and adjusted for wage budget. The headlines will always hail the trophy-winners, but the managers with the biggest Alpha can often be found further down the table, shaping a team that is greater than the sum of its individual parts.
A single season games is too few to be sure that a football manager or his team has genuine Alpha. The inherent randomness of a low-scoring sport like football can create crazy patterns of results. True underlying value takes a bit longer to be revealed, so Alpha can only be properly measured when results have been given a chance to revert to their mean.
So, despite their astonishing and headline-making win in the English Premier League last season, the true extent of the Alpha of Leicester City (I.E. how much better they are than their wage budget suggests they should be) and their manager can be seen now in a better context, with them currently much closer to relegation than the top of the table this time around.
Claudio Ranieri is a nice man and probably a decent coach who was in the right place at the right time last season, when his team enjoyed the benefit of several chunks of good fortune on the way to winning the league. But realistically he has little or no Alpha. But then, football manager Alpha is almost certainly much rarer than we imagine.
Back in 2012 Newcastle United finished 5th in the Premier League. Their owner and board mistook a high league position as a signal of Alpha in manager Alan Pardew and chief scout Graham Carr and gave them both eight year contracts. But league positions are noisy and unreliable data, and a bad metric to use to judge the Alpha of a football team, as that club’s relegation just four seasons later attests. A much better way to get at the underlying true value of a team’s level is to use ratings such as the ones we produce at OddsModel. Or look at how the betting markets rate a team, or even just to look to Goal Difference.
That season Newcastle finished with +5 GD which ranked them 8th by that measure – a more accurate reflection of their true level than the loftier league spot, and about bang in line with what they were spending on wages. Measuring true underlying worth, and separating luck from real merit in a team is something data-driven operations like ours are generally much better at doing than even those who know and follow a team or league closely.
The benefit of hindsight can often shine a less flattering light on performances several years after they took place. In cycling on track and road, Bradley Wiggins and Chris Hoy won multiple Olympic golds and a yellow jersey in the Tour de France with performances that suggested massive Alpha achievements, and earned them both knighthoods. Those achievements are still considerable, but to be strictly analytical about it, their Alpha now seems much smaller since Britons Chris Froome, Jason Kenny and Laura Trott have matched and even surpassed their achievements in the last few years. The advantage given to riders by being within the British Cycling/Team Sky environment isn’t as huge as an F1 driver gets driving for Mercedes, but what is apparent now is that it’s certainly worth a lot.
The idea of controlling for context and external factors to measure Alpha could usefully be applied to politics too. A President or Prime Minister who enjoys great success, or crushing failure will either be lauded or vilified. But probably a smarter way to measure political careers would be to sift out the factors over which the politician had no control, from the stuff they could directly influence.
And ask the question ‘how much better or worse did this guy do than an average hypothetical replacement would have done?’ In other words; ‘what is his VOR?’. Political tribalism means that this way of looking at politicians rarely if ever happens, but it probably should.
The world is still coming to terms with the reality of a Donald Trump Presidency. His appeal was partly based on his reputation as a brilliant billionaire businessman. But what is his Alpha?
It appears that his Dad lent him around $14m to start his business back in the 70’s/80’s, and then left him $100m or so more when he died. If Trump had simply put that money into an S&P 500 tracker fund he would be considerably richer than he is today, without the mess of all those chapter 11 bankruptcies.
So America has just elected a guy with negative Alpha. Good luck with that. To be fair, having negative business Alpha doesn’t guarantee he won’t make a good President. But anyone who voted for him because of his business acumen has been taken in by a con trick.