“All models are wrong; some are useful.” – George E. P. Box

 

The quote above comes from George Box. He was a brilliant statistician and professor, who thought long and hard about the use and misuse of statistics.

I was reminded of Box this weekend while watching the thrilling World Cup final between Germany and Argentina. (If you didn’t find Germany’s 1-0 win thrilling, that simply means you don’t understand soccer). From Goldman Sachs to fivethirtyeight, just about every major modeler with the temerity to forecast the outcome of the Cup got it wrong. Not merely wrong, but wildly so. Give credit to Macquarie for choosing Germany to win (me too!), but getting almost everything else wrong. (I did even worse).

There are trillions of dollars invested based on models. Many of the world’s biggest hedge funds, pension funds and foundations are highly dependent upon some form of modeling to put their capital to work. What does it say about the world of investing that nearly all of these folks in the business of modeling markets and economies were so far afield when they tried to predict the outcome of the Word Cup? I believe it is more than just a matter of sports being different from investing.

Continues here

 

Category: Investing, Mathematics, Sports

Please use the comments to demonstrate your own ignorance, unfamiliarity with empirical data and lack of respect for scientific knowledge. Be sure to create straw men and argue against things I have neither said nor implied. If you could repeat previously discredited memes or steer the conversation into irrelevant, off topic discussions, it would be appreciated. Lastly, kindly forgo all civility in your discourse . . . you are, after all, anonymous.

7 Responses to “What the World Cup Tells Us About Investing Models”

  1. NMR says:

    I never for a moment doubted victory for the Fatherland. Part emotion but part observation. Argentina played a great defensive game in the final but technically the Germans had it throughout the entire tournament.

  2. [...] What the World Cup Can Teach Us About Investing (The Big Picture) [...]

  3. NoKidding says:

    “There are trillions of dollars invested based on models… What does it say about the world of investing that nearly all of these folks in the business of modeling markets and economies were so far afield… I believe it is more than just a matter of sports being different from investing.”

    Now apply to the climate discussion.

  4. orsogrigio says:

    Dear Mr. Ritholtz, you are describing ‘models’ but, really, you are speaking mathematically of extrapolation, simple, blind extrapolation. A model requires the description of the LAWS driving the phenomenon. You can (and must) use PAST experimental points to look for the laws, not just to aptly fit data (this is visual cosmetics, really). F=MA is a model, simple and enormously powerful. A model is timeless, in the sense that the operation of the model is independent from time, a model works by definition, exactly as well in (any) past, present and (any) future, because a model rests on physical laws, and physical laws are inherent PROPERTIES of Universe. Now, what are the rules driving markets ? One must define such rules in order to have the initial framework for a model. Once you try this step you simply fall into a myriad of ‘talking heads’ ‘tricks’ ‘chatter’ … and you cant’ move any further, Extrapolations, taken as ‘model’ are based on things like ‘bubbles’ ‘bearish/bullish standing’ ‘risk aversion’ … that is hot air, simply statements describing things that are likely not even existing, let’s imagine having a definite, measurable, value. But outside the realm of measure you do not have data (you have a salad of numbers, big as you like, but just numbers). Is there any way out of this ? I simply do not know. May be one should work more with a framework based on induction laws, but it’s just a faint idea.

  5. Livermore Shimervore says:

    Germany were entirely beatable, Ghana scored twice on them and their efficient offense managed only a narrow one goal win over USA playing without Jozy Altidore. I’m not sure Germany would have gotten past Colombia if they first had to face the highest scoring offense (before Brazil’s Fukushima meltdown). Brazil seemed to have expended all of their energy (via mostly fouls) holding of wunderkind James Rodriguez. And this was of course without Colombia having Radamel Falcao who is perhaps the most lethal offensive player in Europe scoring 100 goals in only his first three seasons of play in Europe. He went down to injury in January an accounted for fully 1/3 of all of Colombia’s qualifying goals, yet the 2nd best Colombian player still manged to steal the coveted Golden Boot from Klose, Muller, Messi and Van Persie. The widely panned, foul-a-minute win by Brazil over Colombia on July 4th left Germany in the sweet position of finishing off what was left of Brazil’s energies to go with its very lackluster offense. Even with Neymar Jr., he could not score for Brazil up to the 70th minute of that match, so he would not have done much against Germany in the following match who have a stronger defense than Colombia. Germany were also aching to face a weak goalie for the first time which is exactly what they got with Brazil. Mexico, Costa Rica or USA’s goalies would not have allowed such a romp, particularly the first goal at 10 minutes and the last two goals Brazil gave up, even after they were doused with cold water by Brazil’s coach at the half. Germany then faced a very low scoring Argentina offense who’ve had big problems finishing plays all during this WC and continued this mediocre play in the final. Messi dragged that team all the way, but as a play maker, he needs a strong midfield, or at least a consistent one, to shine, a luxury he has at Barcelona but not on his national team. Not the toughest path for Germany to traverse but a win is a win.

    As for models, any that prove accurate is pure luck. As soon as you have a single injury or a team member playing with one, all bets are off. Which is what completely changed USA’s chances of improving their results from 2010 once Altidore was out.

    • einarben says:

      It wasn’t the easiest path for Germany to traverse either though. Don’t forget that Germany faced a lot of challenges throughout the tournament, most notably losing Khedira minutes before the final and Kramer (his only natural replacement) 30 minutes later.
      They also lost their best player (Marco Reus) before the tournament even started, echoing the point you make about Colombia losing Falcao. I have no doubt that with Reus in the lineup the ride would have been less choppy. Additionally, they faced what you could call tactical challenges throughout the tournament, and a lot of the early, less impressive games (Ghana and Algeria) were due to Jogi quite frankly getting his tactics wrong (fair play to him for making the right adjustments as the tournament wore on).
      Without a question, the best team in the tournament won.
      Also, the only thing narrow about the 1-0 win against the US was the result.

  6. faulkner says:

    A great start on a very important topic.

    In contrast with “the accurate and precise world of physics” model is the idea that our (mental) models are everyday, ordinary and ubiquitous (as well as those in specialized, technical and sometimes scientific fields). That these everyday models are (nearly) invisible does not render them any less important. In fact, they are the basis of all the rest of our thinking – influencing the form of content, processes and methods of interpretation.

    All models are abstractions (and often idealizations) of what is modeled. Their simplicity means there is less messy, ambiguous, noisy information to deal with, and this very simplicity of representation makes them seem more compelling and exact. It’s an example of a processing paradox with “clear and vivid images.” As Daniel Kahneman points out, “What You See Is All There Is (YSIATI),” and with repeated exposures, we come to think it must be right and true.

    Models, including graphs of all kinds, are most easily (and thoughtlessly) interpreted in terms of their visual resemblance to other models/graphs regardless of differences of scales, ratios, types and/or operations. Similar names and analogies often get confused with each other. Different levels of detail are applied without consideration for the context. And most efforts go into attempting to make a fit than seeking out better matched models.

    Meanwhile, unconscious conventions in the design of models/graphs orient and activate our attention. The vertical axis indicates increasing amounts – often with the associations that “More is Better” “Up is Better.” Another association with graphs, sometimes explicitly named, is “peaks and valleys” with all the entailments (and dangers) of that mountain metaphor.

    These everyday applications of models/graphs are a long way from “a numerical depiction of a tiny slice of a complex universe.” They are instead the way people make sense of what is bigger, more complex, invisible, and beyond themselves. The feelings they get from these clear and vivid images, whether it is single indicator or a goldbug’s conviction, is that they understand what is going on – which while significant, is a far cry from any mathematical correlation much less causation.