One annoying feature of the MTR trains in Hong Kong is the beeps before the doors close. Unfortunately they're going to get worse, supposedly for a good reason. From today's SCMP:
How do you stop impatient passengers charging onto trains as the doors are closing and getting themselves or their belongings trapped? Simply increase the number of beeps.
When the number of beeps is increased from 9 to 20, travellers who are timing their run for the doors now arrive before the doors close. The newspaper article reports that since this new policy was introduced, "the number of cases of people or objects caught in the doors dropped by 10 per cent on the Kwun Tong Line in the two months after the change" and 1 to 2 percent on the Hong Kong Island Line.
I am highly sceptical. First, there's no mention of whether these reductions in objects getting caught in doors is statistically significant (1-2%? Give me a break!), and second, basic economics would predict that any benefits will be only temporary. Let's outline this second argument in more detail.
Economics views humans as rational agents, who choose their behaviour to maximise their happiness. Now we can quibble about whether this is truly what humans do, but as an empirical model of observed behaviour, it's not a bad starting point. Their behaviour will depend on the environment around them. Change that environment, and they change their behaviour- at least once they have learned about the new environment. Passengers who run for the train once they hear the beeps do so because they think they have 9 beeps to get on the train, and they value being on that train.... not the next one, even if the wait is only 3 minutes.
Now increase the number of beeps. In the short run, passengers rush for the train, and arrive in time to hear another 11 beeps before the doors close. The first time, they may be surprised by this. But eventually, if the MTR continue with their policy, they'll start to notice a pattern. They're rational, utility-maximising consumers, after all! With time, hurried passengers will start charging for the train later in the cycle of beeps, and eventually, when everyone has adjusted to the new regime, I predict that there will be just as many blocked doors as before.
Of course the MTR, in their great wisdom, could take the experiment one step further. If increasing the beeps from 9 to 20 increased for a while, why not increase it from 20 to 30? And after that wears off, from 30 to 40, and so on? If consumers are rational, even increasing the number of beeps in a predictable pattern shouldn't work eventually- as people will notice the pattern of beep increases, and start to predict this when determining their behaviour. The MTR could move up a derivative: why not increase the rate of growth of the number of beeps? Even then, rational people will eventually notice the pattern.
I plan to buy some earplugs....