# How do you prove s r

## Why the video evidence isn't evidence ...

• headlines Alex Feuerherdt in Collinas Erben on Friday, May 25th, 2018.

In summary, you can also describe this post with a quote from Stefan at our user meeting, which I would like to give the best here:

Quote from Stefan (analogous user meeting 2018)

There are things that can be divided into black and white and there will continue to be things that will always remain a gray area.

Also interesting is the link to some statistical evaluations, which in the end brought me the following findings.

The video evidence is not evidence.

At least not in the sense in which the couch potato referee would like to imagine. In the run-up to the introduction of the VAR, the proponents always argued with the so-called "smoking gun scenes". However, the basis of any evidence procedure is that the underlying claim or the fact to be ascertained can also be proven. Unfortunately, this black-and-white proof, desired by many, only actually leads to a TRUE-FALSE decision template in mathematics or logic.

The fact remains that in many cases someone has to make an assessment of the evidence. Which ultimately leads to the fact that the fingerprints at the crime scene do not prove the crime, but only the presence of the accused at the crime scene.

Very few false perceptions.

I do not want to use the word "decision" on purpose, because strictly following the rules of logic, a logically correct derivation of a decision based on a wrong perception can only ever be correct.

Quote from An everyday example of decision logic

If I have a wrong map, the decision to follow a certain path is logically correct, even if this ultimately does not lead to the goal. This sounds paradoxical at first. However, strictly following the logic of decision-making processes, it is not.

The fact that the decision is based on a wrong assumption does not change the fact that the actual decision based on this assumption is nevertheless correct, because according to the information available the path taken should lead to the goal.

So it is not the decision that does not lead to the goal, but the insufficient information on the map. Any other decision than to follow the route marked on the map would be illogical with the information available.

What nobody - including our colleagues in our guild - is really aware of is that a soccer referee makes at least 500 to 600 decisions in a game. This is only one every 10 seconds on average. That may sound like a lot, but I think there are a lot more decisions.

• Continuing to play is also a decision, not every decision leads to a whistle.
• Virtually every touch of the ball is checked to see whether it conformed to the rules
• Virtually every duel is checked for compliance with the rules.
• The position of the ball is constantly checked
• The position of the player is constantly checked
• The "look at the clock" is also used to make a decision.
• Etc.

If you don't believe it yet, just write a decision algorithm for a robot referee.

76 times in the last BuLi season a decision was changed by the intervention of the VAR. With 34 game days with 9 games of 90 minutes each, that makes a decision change every 362 minutes (i.e. about every 6 hours of play or one in every 4 game)

It gets interesting when you look at the VAR reviews, which may be crucial to the game. (Tied score or 1 goal difference and penalty goal or FaD) Here we only come to a game-decisive VAR bet every 451 minutes (7.5 hours of play or in every 5th game).

Conclusion:

Strictly speaking, the VAR is business nonsense. No company in the world would invest such a disproportionate amount of effort with such a low error rate.

• FC Bayern would have become German champions even without VideRef
• Eintracht even without a VideoRef cup winner
• HSV and Cologne would have been relegated without VideoRef
• Aue remains in the 2nd division despite the lack of VideoRef
• Wolfsburg would stay in the 1st BuLi even without VideoRef
• Maybe you can still argue about the Europe tickets from Hoffenheim, Dortmund, Leipzig and Stuttgart.

What remains is a quote from Felix Brych at the beginning of the "VAR season"

Quote from Felix Brych (analogous)

The good thing about VideoRef is that we referees are no longer the idiot of the nation in the week after the game

But I also have to put that into perspective, Felix. At least for Bayern, your namesake is the "dork of the cup final" if you look at it through the white-blue hooness diamond glasses that you can buy in every FC Bayern fan shop.

• Reactions
385
place of residence
Regional association
Hessian Football Association

Especially since it is not proof, but only the possibility to look at certain scenes again and additionally from different perspectives. Given the cup final, the statistical analysis would be interesting to see how many video views led to a change in the decision - because in the end, only such cases are really interesting.

Football is the most beautiful Minor matter of the world.

• how many video views led to a change in the decision

I wrote to Manfred:

A total of 76 (but not all with video sequences in the VAR zone) You can find the evaluations here:

Especially since it is not proof, but only the possibility to look at certain scenes again and additionally from different perspectives

But that is precisely the linguistic problem. The couch potato referee speaks of the evidence and means it that way, while technically it is actually "ONLY" another SRA that supports the SR with electronically extended perception.

• Reactions
64
Regional association
Lower Saxony Football Association

With the 500 decisions per game, "keep playing" will already be included, right? And basically - in order to calculate the error rate and its correction - you would only have to rely on the decisions that are VAR-capable. Of course, it doesn't really make sense to say that the referee makes the decision every second not to issue a penalty or expulsion.

Edit:

For example like this: Since we do not know how often referees decide on "no penalty kick", you can take the penalty kick decisions.

There were 88 penalties last season. Including after the ntv-Link 26, which were only given by the VAR (there are not those that were only confirmed by the VAR, right?). There are 62 penalty decisions left by the referee, plus the 13 that were revoked by the VAR deployment. So makes 13 mistakes out of 75 and thus an error rate of 17.33%.

Since there were twice as many penalties that were given by the VAR than those that were denied by the VAR, one can come to the statistical expectation (not finding!) That the error rate there is also twice as high (which yes is also somehow clear, because referees only decide on a penalty if they are very sure). The quota then of course only refers to the "tricky" or deliberately judged scenes (i.e. not to not giving a penalty with every step or duel).

35% error rate would be pretty severe. In my opinion, it is these unwritten rules of doubt that ensure such high error rates, such as "I only give a penalty if I am absolutely sure". We had a similar discussion at the Aue gate; you should only decide to score if you are absolutely sure. Or with offside, where the actual written rule of doubt is even reversed by the referee tactical recommendations (because the shouting is less if you decide offside right away, before a possibly irregular hit can even be made).

• Reactions
385
place of residence
Regional association
Hessian Football Association

Regardless of how I think Felix Zwayer's justification, this is a really great example:
Who can judge how many of the penalties that were awarded or denied after the video "evidence" were in the "gray area", so they could (not) have been given, but the referee decided on the basis of the television images has chosen the "more convenient" route?
We must not forget that it can have a psychological influence to know that the images are not only "dissected" on television, but also by a VAR - and if that comes to a different conclusion ...

Football is the most beautiful Minor matter of the world.

• So makes 13 mistakes out of 75 and thus an error rate of 17.33%.

Ahhh, this interpretation is just as wrong as we compare it to my 500-600 decisions. The fact is that the many CORRECT decisions are missing that are not mentioned here at all and never led to a VAR check.

That's why I simply reduced everything to the time factor and put it in relation to the final result in the table after 34 match days.

Maybe I'll summarize one of the main results of my introductory contribution.

A huge effort is made to actually revise a decision in every 5th game that may have been decisive for the game. However, only the fan soul is calmed down because these decisions would not have had any real impact on the essential factors of the final table.

However, it is noticeable that significantly more penalties were given after VAR verification than were withdrawn

• Can it be that this is because of being 100% sure that it was a penalty or
• Can it be that a wait and (VAR) see leads to more "preliminary decisions to continue playing" because an interruption by (possibly wrong) whistle can no longer be undone but the 11s that were not given can still be given afterwards?

• Reactions
64
Regional association
Lower Saxony Football Association

Nothing is missing at all. Again, there were 88 penalties taken last season. According to ntv-Link, 26 of these were only recognized with the help of the VAR, leaving 62 that the referee correctly recognized without the VAR. In addition, there are 13 that the referees recognized but were revoked with the help of the VAR (and therefore never carried out).

This means that a referee decided on penalty kicks 75 times out of the game, and 13 of these decisions were so clearly wrong that the VAR intervened. From the game, 17.33% of the penalty decisions were clearly wrong. I think that's a lot. Especially if you take into account that a) referees actually only give a penalty if they are very sure and b) this effect is reinforced by the VAR because the referees rely on the VAR (and of course prefer to whisper a penalty to themselves rather than cancel).

• This means that a referee decided on penalty kicks 75 times out of the game, and 13 of these decisions were so clearly wrong that the VAR intervened. From the game, 17.33% of the penalty decisions were clearly wrong.

And there are no decisions in which the penalty was NOT awarded and this decision was correct. Don't tell me about statistics

Just to complete:

• Penalty Decided - Penalty Wrong (decision was wrong)
• Penalty decided - Penalty correct (decision was correct)
• Penalty not decided - should have been a penalty (decision was wrong)
• Penalty not decided - neither was it. (Decision was right)

Only in this way is it complete.

Unfortunately, the last point is not recorded statistically. This is why the comparison is incomplete because an essential part of making the right decisions is missing.

You have to consider all decisions in order to calculate the CLEAR wrong ones as a relative statement. But maybe we have a statistician here who can explain that a little more precisely.

I only deal with such things professionally on the sidelines and have not studied them.

• Reactions
64
Regional association
Lower Saxony Football Association

Yes, I'll tell you about statistics. If I measure how many cars are driving too fast on the A road, you cannot say that the same cars could have driven slower or faster on the B road.

The numbers concern the given Penalty kicks. Not more but also not less. Both given Penalties were 17.33% wrong. It doesn't matter how many missed penalty kicks were right or wrong.

Since there is also no measurement when a penalty kick Not is given, and we do not agree on what one would actually measure (every step, every duel, every second, every gesture to continue playing?), one cannot calculate this statistically valid with the means available to us.

But there are two Clues (no more and no less) for that - if you get tricky and deliberately stopping decisions (everything else makes no sense if you want to calculate error rates) - the error rate for penalty kicks not given still significantly higher is to be expected:

1. With the help of the VAR, twice as many penalties were recognized as denied.

2. Referees only decide on a penalty if they are very sure. It is therefore more likely that an irregularity will go unpunished than that one that was not punished will be punished.

My estimate is that together (i.e. penalty kicks given and penalty kicks deliberately not given in tricky situations) you end up with an error rate of at least 25%.

• Reactions
142
Regional association
Football and Athletics Association Westphalia

In my opinion, statistical recording of penalty kicks that have not been given is feasible to a certain extent: Whenever at least one attacker requests a penalty.

• Reactions
385
place of residence
Regional association
Hessian Football Association

Whenever at least one attacker calls for a penalty.

Okay, then the SR's margin of error is out of the reasonably measurable range ...

Football is the most beautiful Minor matter of the world.

• Reactions
4
Regional association
Hessian Football Association

You can see from the video evidence, at the latest after the DFB Cup final, that there are also referees who do not adhere to video images

and stick with it despite the wrong decision.

So I don't know what the video evidence is supposed to do ... at least that's not how it should work.

The video referee should be able to overrule the referee, what else does he do if the referee on the pitch still makes the wrong decision?

Without a referee? Have fun it won't work

• Reactions
142
Regional association
Football and Athletics Association Westphalia

You cannot question the usefulness of the video "proof" in the few decisions where an alleged mistaken decision has not been changed.

Nor do I accept the argument that coincidentally in the first season the VAR apparently had no influence on the championship, CL participants and relegation. It can be a coincidence. It is also pure speculation how a game would actually have turned out without VAR changes (butterfly effect).

Therefore, the "true tables" or the argument that wrong decisions compensate each other in the course of a season are nonsense. I want as much as justice to be seen in one game and not over a long period of time.

The final decision currently has to be made by the field umpire based on his perception in real time and the images made available to him. In order to justify a change in the "field decision", the images must be unique. At the Bayern game there are enough arguments for the "corner kick" decision originally made.

Of course, a quick decision by the VAR without anyone noticing would be the perfect solution, but it will take a few more years if the project is given the chance. In my opinion, the first year went very well and we have to give this system the time to improve. In my opinion, there were far fewer blatant mistakes by the VAR than the other way around.

• Reactions
385
place of residence
Regional association
Hessian Football Association

The video referee should be able to overrule the referee, what else does he do if the referee on the pitch still makes the wrong decision?

No, no, and once again no! We should all know that there are situations in which there is no "right" decision.

My favorite example is the DFB's instructional video on the handball: From position 1 a clear "shot from a short distance, the arm follows the movement and was already there". Position 2 shows the same scene from an angle 45 degrees further - the face is already deformed into a question mark.Position 3 is another 45 (a total of 90) degrees further, there it is clearly deliberate handball, you even have to seriously think about a warning. Note: The same situation from 3 different perspectives - and there are still 270 degrees left. The VRA also has no images available from all directions, ergo it also has a limited field of vision - and now think of position 2 ...

The VRA should support the umpire and expressly not be a senior umpire - and that's a good thing, because besides the picture there are - we in the concrete league know that only too well - other impressions such as the noise (everyone has belongs to the hit) that are not available to the VRA. And every umpire also pays attention to things like facial expressions and more, which - keyword perspective - may not be available to the VRA or taken into account.

Don't we all know the gut feeling when someone lays a swallow down for us? It falls too precisely and too smoothly, the Bayern Rs are also very good at it. Of course there may have been contact, but mere contact is not necessarily a foul.

Therefore: The person who is on site must have the ultimate responsibility, it is not for nothing that the VRA is also called an assistant. In the end, he has to answer for the decision and nobody will lightly ignore a hint from the VRA.

On the subject of the cup finals: We have already discussed elsewhere whether the decision was right or wrong - and we did not come to a unified and clear result either, so why should the VRA have been better here than the SR?

Football is the most beautiful Minor matter of the world.