Evaluating surprising results

SoccerSTATS.com - updated on: 30 Jul 2020

As a low-scoring game, football can produce match results that can seem surprising. The strongest team on-paper may lose, even the strongest team on the pitch may lose. Each football weekend carries its load of surprises, big or small. Let's try to quantify how much of a surprise a given match outcome could represent.

The final whistle blows and the fans realize what has just happen. This is it. That last minute goal. That home defeat that nobody seems to have seen coming, and certainly not at the league leaders' stadium. This is a known feeling though, it happened before. But this time this feels like a bigger surprise. Rightfully so? Last year's defeat was against the team ranked 4th in the league. This time it was against the 8th. Yet the winners may be ranked in 8th position in the league, they also have an impressive record on the road this season, so maybe this could be less of a surprise after all.

Evaluating match outcomes, as opposed to specific scorelines

Of all the surprising results - and in this article, we'll be discussing "outcomes" (like a win or a draw) as opposed to results with a specific scoreline - some can be deemed more surprising than others. If we were to calculate how much of a surprise a given match outcome could represent, we'll have to come up with a set of criteria.

Choosing the criteria

Let's say we want to come up with a value representing the "surprise-level" of a particular outcome. To calculate such value, we'll be taking into account the same type of indicators covered in other articles in this blog related to team performance evaluation: home and away Points Per Game (PPG).

The goal is not to obtain a value that would absolutely define with a number what we call a "surprise", as a "surprise" is also mainly a feeling and will always be subjective to some degree.

To a fan of the winning team, the "surprising" outcome may represent less of a surprise than to a fan of the high-profile team that has been defeated. The idea is to compare matches with each other, so that we can calculate values by which we could say something like: "According to these criteria, the outcome of this match is more surprising than the outcome of that match."

Outlining a formula

In line with the Run-in analysis, where we use the opponents' Points Per Game value to evaluate how "valuable", in terms of performance, a win could be, we will determine the outcome surprise-level for a given match based on the teams Points Per Game. We will be using two separate formulas: one to evaluate a win (used for both home wins and away wins), and one specifically for draws. In both formulas, we will only be using the home PPG of the local team and the away PPG of the visiting team.

Outcome surprise-level (win):
(Losing team PPG / Total PPG of both teams) x 100



Outcome surprise-level (draw):
(PPG difference between both teams / Total PPG of both teams) x 100

Reasons for using such formulas

There could certainly be a lot of different ways to evaluate such type of "surprise" values. The formulas above have the advantage of allowing the following conditions to be considered:
- the higher the PPG of the losing team (relative to the PPG of the winning team), the higher the surprise-level value
- the higher the difference in PPG between the two teams, the higher the surprise-level value for the match having ended in a draw

Example

You can view an example of the Outcome suprise-level table for matches of the Serie B or the Bundesliga.