Fact check: 45th / 90th minute momentum

Share

All FIFA players recognize the frustration of conceding goals in the 45th and 90th minute. But do EA actually script goals or trigger momentum swings in the final minutes of either half? With some new evidence at hand, we decided to drive the final nail into the coffin…

The claim

As always with scripting / momentum / handicapping, it is somewhat difficult to get your hands around what the exact claim is. What seems to be the common ground in regard to late goals can probably be summarized as per below:

1. Goals scored in the 45th or 90th minute happen very often …

2. … and more often than they should.

3. Late equalizers and winning goals happen more often than they should.

4. These late goals happen either to create drama or to make matches more even.

So, what is up and what is down? In the following sub sections, we will walk through the five claims above one by one.

Fact checking

1. Are goals in the 45th and 90th minute very frequent?

Let’s start out by establishing that different players playing different FIFA versions should experience varying numbers of late goals. Some people concede – and score – more goals than others. Some FIFA versions allow the players to score more goals than others.

But for the purpose of this article, we aren’t interested in the exact numbers. We merely want to know whether the broad statement that goals scored in the 45th or 90th minute happen very often is mostly true or false.

How often do these late goals happen to ordinary players?

According to our own survey from 2014, two-thirds of the respondents experienced decisive stoppage time goals in minimum every other match.

According to the data collected by tedditor Arlington69, 45 % of his matches contained 45th or the 90th minute goals. He noted down  the timing of all goals in 761 of his own matches.

With the above, it seems fair to conclude that 45th and 90th minute do occur very frequently.

2. Do late goals happen more often than they should?

According to claim #2, goals not only occur often in the 45th / 90th minute: They also occur more often than they should under natural circumstances.

But how often should goals occur in the 90th minute under natural circumstances?

The natural starting point would probably be that they should occur with the same frequency throughout the match. And when we say frequency, we mean number of goals per effective minute. Obviously, goals won’t be scored while the ball is out of play. And when we talk about minutes here, we mean “in game minutes”, not real time minutes.

So, do goals actually occur with the same frequency in the 45th and 90th minutes as in the rest of the match?

Below, we plotted the occurrence of goals per minute based on Arlington69’s data. At a glance, it definitely would seem that goals occur much more often – at a 4 times higher frequency – in the 45th and 90th minute as in the rest of the match.

There is however a problem: The “45th” and “90th” minutes aren’t necessarily one minute long.

If you look carefully a the post match report, you will notice that FIFA records all goals scored from 44:00 / 89:00 and onward as 45th / 90th minute goals.

And it gets even more complicated: The fact that stoppage time is added to a match implies two things:

  • That the 45th / 90th minute contains more than 60 seconds of effective playing time.
  • That the previous minutes on average contained less than 60 seconds of effective playing time per minute.

Keep in mind that stoppage time is added due to stoppages. And obviously, teams can’t score during stoppages, even if the clock keeps ticking. So, what actually happens when stoppage time is added is that effective playing time is moved from the first 44 minutes of the half and into the “45th minute”, cf. the illustration below.

Additionally, please note that when a referee adds 5 minutes to the first half, the “45th minute” in fact will be 1 + 5 minutes long. The 45th minute starts at 44:00 and ends when stoppage time runs out. But at the same time, the fact that the referee added 5 minutes means that the previous 44 clock minutes only contained 39 minutes of effective football.

There are some very natural reasons why we should expect goals to be reported more often as scored in the 45th or 90th minute rather than for example the 44th and 89th minute.

Variance

Further, the assumption that goals under normal circumstances should happen with the same frequency throughout the match clearly needs some fine tuning. Below, we divided Arlington69’s fairly big sample into 3 minute intervals.

You will notice that some 3-minute intervals contain more goals than others despite this being a fairly large sample. Not surprisingly, we see that the first three minutes contain fewer goals due to kickoff. Additionally, we see some fluctuation which most likely is the product of to statistical variance.

In Arlington69’s matches, the interval that contains fewest goals is the 39-41 minutes interval (.053 goals/min. or 2.7 % of all goals), whilst the 42-44 minutes interval contains most goals (.077 goals/min. or 3.8 % of all goals). If we zoom down to the earlier mentioned 1 minute intervals, the goal frequency varies even more, as we have observations fluctuating between .037 and .087 goals per minute.

Nothing indicates that this variation is unnatural in anyway. Despite people claiming that goals are particularly common in … just about every minute of the match, no evidence suggests that the variance we see above is anything but natural.

The fact that there is considerable natural variation in goal frequency throughout the match of course implies that we can’t assume that the goal frequency in stoppage time should equal the average goal frequency during the rest of the match.

So, it is more correct to say that the goal frequency should be the same plus / minus natural variation in all parts of the match – with the exception of the first 3-4 minutes.

The duration of stoppage time

We want to know whether the stoppage time goal frequency exceeds what can be explained by natural causes. But without knowing the amount of added time, we can’t determine the actual goal frequency in stoppage time. Instead we will have to attack the problem from a different angle.

We saw above that the goal frequency varies for natural reasons. On average, the goal frequency was .067 goals / minute, but in some 3-minute intervals, it went as high as .077. In some 1-minute intervals, it went up to .087 goals per minute.

So, how much stoppage time would be needed in order for us to conclude that the goal frequency in stoppage time is perfectly normal?

We can calculate the amount of necessary added time in order for goal frequency in stoppage time to stay on par with the known natural level observed in the sample:

Goal frequency Added time
Average for entire sample (.067 goals per minute) 3:01
Maximum, 5 minute intervals (.072 goals per minute) 2:43
Maximum, 3 minute intervals (.077 goals per minute) 2:29
Maximum, 2 minute intervals (.077 goals per minute) 2:30
Maximum, 1 minute intervals (.087 goals per minute) 2:06
Calculation:
GF = Goal Frequency 44:00 – end of stoppage time                                                   (input parameter)
G = Average goals 44:00 – end of stoppage time per match                                      709 / 761 = .54
M1 = Avg duration 44:00 – end of stoppage time per match                                     G / GF
M2 = Avg duration 45:00 – end of stoppage time per match per half                     M1 / 2 – 1

The above first and foremost tells us that unless Arlington69 and his 761 opponents experienced less than 2 minutes of added per half, the goal frequency in stoppage time would be considered normal.

Now, as already stated, we don’t know with certainty how much time was added or normally is added. But based on our experiences, it is fair to assume that having 2 or more added minutes is very common. According to this reddit post, the author experienced 2 added minutes in almost every match. Based on the responses, most readers seemed to agree, although some added that they had seen 3 and 4 added minutes. In another post, the author claimed that 5 minutes was added every time.

Maybe more importantly, we didn’t manage to find any complaints from people who thought that two little time is added. And on top of the above, we have to remember that “minimum 3 added minutes” doesn’t mean that stoppage time can’t be extended.

As we don’t know with certainty how much time is added on average, the status of claim #2 is “not proven”. But based on the considerations above, it is fair to add that it seems very unlikely that late goals occur at a higher frequency than other goals.

3. Do late equalizers or late winning goals happen more often than they should?

The evidence does not support the assertion that stoppage time goals happen more often than they should. But what about late equalizers or late goals that give either team the lead, and which thereby have a potential to change the outcome? We often hear people claim that goals that change the outcome happen particularly often in the 45th and 90th minute. But is that actually true?

Arlington69’s 761 matches contained 409 45th / 90th minute goals and 4576 goals in total. In the table, we counted the percentage of goals which either (1) were equalizers, (2) gave either team a lead or (3) increased an existing lead.

All goals 45/90th min. goals 45th min. goals 90th min. goals
Total 4576 409 206 203
Equalizing goals 17 % (776) 14 % (59) 14 % (29) 15 % (31)
Goals giving either team a lead 32 % (1449) 22 % (92) 30 % (61) 15 % (30)
Goals increasing a lead 51 % (2351) 63 % (258) 56 % (116) 70 % (142)

Notably, the percentage of equalizers and goals giving either team a lead is significantly lower in the 45th and 90th minute than in the rest of the match. So, as opposed to what people claim, the chance of conceding an equalizer or a winning goal is lower in the 45th and 90th minute, not higher.

And when we look at the 45th and 90th minute separately, we may an idea about why that is the case. We see that the percentage of outcome-changing goals is much lower in the 90th minute than in the 45th minute. This is due to a natural trend, which we have illustrated below. Goals being scored early in the match – i.e. where few other goals are scored – have a larger probability of causing a lead rather than being equalizers.

As for claim #3, it is definitely false.

4. Late goals are put in to create drama or to make matches more even

Claim #4 is about EA’s alleged motive for “scripting” goals into our matches in the 45th and 90th minute. Supposedly, EA either wants our matches to become more even or they want the opponent to win.

But when we look across Arlington69’s full deck of matches, it is notable how rare outcome-changing 90th minute goals are. In total, his 761 matches contained 4576 goals. Among them, only 30 (.7 %) were equalizers scored in the 90th minute and only 31 (.7 %) were goals taking either team into the lead in the 90th minute.

If EA has an opinion about the results of our matches, they how come they only interfere in less than 1 % of them? Don’t they care about the other 99 % of our results?

Claim #4 is already heavily challenged by the findings presented above. EA hasn’t put in any extra goals in the 45th and 90th minute, and the goals that do happen there do not create drama or make matches even to a larger extent than they should.

Claim #4 is clearly false.

Conclusion

We looked into four different claims related to the 45th / 90th minute momentum theory.

In our earlier post on late goals, we demonstrated that there are natural reasons why goals will happen more often in those minutes. In this article, we have added some additional weight to our argument by demonstrating that goals don’t happen more often than they should in the dying minutes, and in particularly not decisive goals.

No matter how much we hate conceding a decisive goal late in the match, EA isn’t the reason why it happens.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: