Eyes on **World Cup 2022 in Qatar**, one of the most controversial editions of the FIFA World Cup and after almost a week of matches, bettors from all over the world are trying to predict which teams could be the favorites for the final victory. According to the bookmakers, Brazil is the most favored team for victory, but this is not the only feasible prediction and according to **Anders Nysteen**Senior Quantitative Analyst at BG SAXO, there may be some surprises here.

## The Elo rating

In order to quantify the possible scenarios it is necessary **estimate the probabilities for the different outcomes of a match** and this on the basis of some metrics related to each single team. To carry out this statistical exercise, says the BG Saxo analyst, “we have to classify the various national teams on the basis of their Elo score, which is a value that emerges by comparing the skill level of football teams / players” or in general of any sport . In fact, this metric has proven to be effective for predicting the results of a match, attributing a **higher probability of victory for the team with the highest Elo rating**.

### The most probable scenarios

Based on Elo ratings and assuming that the team with the highest probability of winning will also be the World Cup winner, Brazil emerges as the winner of the tournament, a result in line with the results of the various bookmakers.

In the chart below we have the teams that are most likely to be finalists in the quarters, semis and finals and how we see the **Brazil** has a high chance of winning.

### Find the least likely scenarios with bootstrapping

To find the least likely scenarios we have to use the** bootstrapping**, a statistical technique of resampling widely used both in the financial sector and in the sciences, which allows to approximate a sampling distribution of a statistical survey. Through this statistical system we can in fact artificially simulate that the **World Cup is played 10,000 times** and we use the odds of winning to determine the outcome of each individual game. Using this method we obtain several possible scenarios and some of these see outsiders as national winners such as**Australia**.

But why would Australia be the least favored team to win? As mentioned, with the bootstrapping method we are able to simulate the possible results and out of 10,000 possible scenarios **on only 2 occasions has Australia emerged as winners** of the competition.

As we can see from the percentages shown in the adjacent graph, Australia will need one or more sensational events, as well as a good dose of luck, to beat the Netherlands, since the Australians have an estimated 15 chance of winning against it. %.

### Who are the favorites at the 2022 World Cup in Qatar

By combining all the 10,000 scenarios obtained through the bootstrapping method, the probabilities that the different countries have of winning the World Cup emerge. As we can see, the results show that the three favorite nations for victory are: the **Brazil** with almost 24% probability, the**Argentina** with 22.5% and**Holland** which turns out to be more detached with an 8% chance of winning.

Conversely, the least favored countries to lift the coveted World Cup would be Senegal with 0.03% probability, followed by **Tunisia and Australia** who have only a 0.02% chance of winning.

But that’s not all, another peculiarity emerges from this overview: there are some lower-level teams that have a fair chance of winning such as **Denmark** which has an estimated winning probability of almost 4%, but also the**Uruguay** with an estimate of 2.1%.

### Risk management

Can all this reasoning just made also apply to investment risk management? Of course. Let’s say we invested five years ago in the S&P 500 index and suppose we reinvest the dividends, we get a 66% return. This would seem a more than interesting yield but as Anders Nysteen points out, you cannot look at yields without considering the risk associated with the investment and therefore a question arises: what would have happened if the macro events of the last five years had been different? Would the investment in question have been as good?

As we have seen before, **one way to quantify risk is to apply bootstrapping**, thus estimating the possibilities of realization of all possible scenarios and not just the most probable one. Since five years ago, the return of the S&P 500 index has been in most weeks around 0.5%, while in half of the weeks the return of the index has been between -1.1% and 1, 7%.

As we can see from the adjacent graph, however, there are some weeks in which the American index has made very important weekly gains or losses and these events can be considered exactly as the scenarios in which Australia wins the World Cup.

Here is that when considerations are made on investments, the so-called **queue scenarios** they can have quite a large impact on investment performance and therefore cannot be ignored despite the small odds.

### Measure queue scenarios

A standard way to quantify the impact of these low probability events is therefore through the bootstrapping method. We do this by simulating 5 years of performance for 10,000 different scenarios and for each scenario we create a **artificial path of 5-year performance** by plotting the weekly return from the histogram above. As we can see from the graph below some of the scenarios will contain many positive weeks, other events will be very negative.

The graph below shows the 10,000 simulated scenarios that could occur with equal probability, including the performance that actually occurred indicated with the black line.

So if we compare the events that (according to this model) have the same probability of occurring as the probability that Australia wins the World Cup, i.e. 2 out of 10,000, the lucky investor would have earned 632%. And the hapless investor would have lost over 72% of the value invested in the worst 2 out of 10,000 scenarios.

When it comes to asset management, tail events can have a major impact on performance.

There are several ways to minimize the impact of these queue events, such as having **diversified portfolios** where the different activities have low correlations.