34

When a rapid transit ticket machine in Germany refuses to accept a coin, I've often seen the ticket-buyer scratch the same coin on the machine's metal surface before re-trying it. Some scratch with the coin's edge, some with a flat face.

It doesn't always work, making me suspect confirmation bias. Over time, it also causes a large round scratched patch adjacent to the coin slot, which could be enabling a bandwagon effect, reinforced by word-of-mouth.

"Removing static" is the only theory I've heard for why the technique might work, but I'm skeptical as metal coins are conductive.

Can a machine's coin-approval mechanism be influenced this way?


Significance support: This blog article contains a photo of a ticket machine in Frankfurt badly scratched by the practice, and a text explanation. Here's a photo of a metallic "coin scratch pad" reportedly installed on a ticket machine in Norwich, UK. The associated Reddit thread mentions the "removing static"-speculation, as do these two independent forum threads.

  • 3
    I've seen this loads of times, in multiple countries, and done it myself quite a few times. Would love to know if it actually works. – P_S Sep 7 '14 at 18:53
  • 6
    This might be an extension of the practice of rubbing paper money to flatten it out when it is rejected. In the US at least, bill accepter devices tend to be twitchy and the theory is that flattening the bill (stretching it on the edge of the machine) will make it more likely to be accepted. I can't tell if that's real or not either, but it at least has a better story than the coins. – Michael Kohne Sep 7 '14 at 22:18
  • 1
    Definitely true for 5 franc coins on Swiss railways (scratched patches abound). – Benjol Sep 8 '14 at 7:47
  • 2
    may be related to the coin being slightly too heavy and scratching off some rust will let it be accepted – ratchet freak Sep 8 '14 at 9:25
  • 1
    My first thought was scratching off crud that's hanging it up in the machine. As for bills--I've never tried flattening them on the edge of the machine but I have frequently found that doing what you can to flatten a crease does get a rejected bill to be accepted. – Loren Pechtel Sep 9 '14 at 3:45
12

The short answer is probably not.

See this article for a short description of how these things are done.

For coins, the machine measures diameter, thickness and number of ridges (on part of the circumference - not the whole edge) so you may think that grinding the coins would affect this.

However, because coins are treated pretty harshly - banging against each other in our pockets and purses, the machines have reasonably high tolerance bands and a coin has to be substantially damaged before it will be rejected in the first place. The whole point is to recognise the denomination and that the coin is domestic. Most coins worldwide are different enough that even with wide tolerance bands foreign coins get screened out. Australian and New Zealand 20c coins being a notable exception.

A more likely explanation is that the coins are fed in too fast for the machine to process and they will shunt the excess to the reject chute. Rubbing the coin will simply give the machine time to clear itself. So would counting to ten!

Edit

The answer above has, quite rightly, been criticized so I leave it here as a lesson to myself on what not to do. I will now approach this more rigorously.

How does a coin detector work

I have been unable to determine the exact model of detector used in rapid transit ticket machines in Germany. This indicates that there are 5 separate U-bahn operators and a single S-bahn operator - it is possible they all use the same detector, or maybe they don't, or maybe there are different detectors in use within each system.

This is a modern coin detector. According to the literature it tests "40 measurements of coin parameters", for obvious reasons it doesn't enumerate them. Other products on the site check fewer parameters.

Notwithstanding, from this general overview, we know that some of the parameters measured include:

  • weight
  • optical characteristics (diameter, thickness, ridges, head & tail embossing)
  • alloy detection (through use of the Hall Effect)

The Hypothesis

Coins that were initially rejected are, after being scratched on a metallic surface (either on the face or edge), more likely to be accepted on subsequent presentation than coins that were not scratched.

Evidence in support of the Hypothesis

None (anecdotes are not evidence).

Possible reasons why it may work

Given the way coin detectors work, none.

Scratching a coin will not appreciably alter its mass, size and shape or the alloy it is made of.

It may remove some surface dirt but probably not enough to affect its optical signature.

It may damage the coin but this is more likely to increase rejection rates.

It may remove or add static electricity, but none of the detection methods would be affected by this.

Conclusion

Given the lack of evidence for the phenomenon and the lack of any physical mechanism by which it can occur we reject the hypothesis in favor of the alternative hypothesis - it doesn't do anything.

That is not to say that people don't do the coin rubbing thing, at least in Europe - the blog article the OP cites is evidence of this. However, here in Australia I have neither heard of or witnessed the practice - a statement that you should disregard as mere anecdote.

This means that the alternative hypothesis comes with a corollary - people are idiots.

  • Might they also test mass by using some kind of inertial test? I've sometimes noticed that pushing coins down slower/faster seems to make a difference. – Benjol Sep 9 '14 at 4:48
  • 1
    You need to base your answer on evidence, but your answer is speculating of what should happen in theory. Why this could happen is interesting, but certainly not an answer to does it happen? – Sklivvz Sep 9 '14 at 6:24
  • 1
    Designers of these machines tend to be naturally wary of explaining the exact details of what their coin slot checker looks for when identifying a coin, for obvious reasons. – Shadur Sep 10 '14 at 4:42
  • 1
    Much better. It's not as good as finding a study on the matter (someone that actually tried it to see if it works), but it's not making drastic conclusions either. As a suggestion, you should include as quotes any relevant passages from your sources to prevent link rot (as you can imagine, link rot is a bit of a problem here) – Sklivvz Sep 10 '14 at 8:50
  • 2
    Regarding the static electricity reason, unless there are some coins made of a non-conductive metal, I doubt that that is possible. Most coins would almost instantly (think microseconds) discharge any static electricity when they touched any grounded piece of metal on the machine. In general, this would probably happen as soon as the coin first touches the machine. Certainly, rubbing is not necessary to remove static electricity from a conductive object and rubbing two conductive objects together (especially when one is grounded) will not generate a potential difference between them. – reirab Nov 29 '14 at 18:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .