Tuesday, 3 August 2010




我費了不少唇舌解釋 sample size 越大,結果理應越接近50:50,他似懂非懂這個數學上的概念,但卻明白了事實有時候是會給感覺或 impression override 的

每次經過紅燈,我們因為要停下來,所以印象特別深刻。但綠燈呢?經過時連望也不用望一眼,完全 being taken for granted,就像船過水無痕,又怎會放在心上?



基本上是沒有人有必要丶沒有責任或義務對我們好,但每一天,我們也不停地在接收,那些 being taken for granted 的“綠燈”。




Anonymous said...

The ratio shall not be 50:50 because green light lits for longer time than red light during a day. Your 50/50 assumption was wrong - not to mention amber light exists and it is ignored in your model as well. M was empirically deriving the probability ratio. If he continues to collect more observation results (samples), his estimation will get better and it tends to be closer to the actual mean value of the probability ratio. Of course, he can use moments to characterize his estimation better....

Anonymous said...

If red light to green light ratio is 50/50, I would also take red light for granted as well bcoz they are equally probable as well.

SKII said...

Yes, you are totally right from a statistical modelling point of view. Amber is being ignored and so as "red + amber" (we have this in HK and it's the signal before red goes to green). But it's too technical to explain to M at the moment.
Also, from a 'feeling' perspective, the impact of the above on people's behaviour is not much different from red and green. That said, people will simply take "red + amber" as green and go ahead without stopping anyway. For amber, either stop (i.e. treat as red) or go (treat as green).
It would be great if we can take red light for granted like the green light. But it takes really good 修為 to face others 缺點 without any bad feelings....
Thanks for raising this, it makes me think deeper....

Anonymous said...

他似懂非懂這個數學上的概念,但卻明白了事實有時候是會給感覺或 impression override 的。

M's confusion is correct. At least, his observation (or empirical model) does not match with what your model suggests. Don't confuse him no more with "impression override" idea. M can continue to fine tune his empirical model so he can estimate when the next red light will be.

Btw, amber light is not crucial because the fraction of time that it lit is relatively smaller than that of green light and red light. The total time for amber light that lit can be ignored. But, for a given day, the total time for green light "on" is much longer than that for red light "on". So, we should see "green" light more often than "red" light by default.

I don't know why the traffic designer used "red" to represent "stop" and "green" to represent "go". I bet it has something to do with human psychology. However, Chinese communist party may find it otherwise --- "red" goes first. :-)

But the concern seems to be the "change" of color. We, human being, are good at detecting changes, especially with our eyes. I think it reckon to your theme "take it for granted". That is, having "green" light as the normal condition and "red" being as the outlier or "change" that is significantly different than the normal color. Our human being will be easily catching such change of condition and we will response to it accordingly.

Perhaps, we should ask (or encourage) M to think of something to counter-argue your 50:50 idea and/or how to make his empirical model to do prediction better. For example, taking "amber" light into consideration should be a trivial way to invalidate such idea.

Related Posts with Thumbnails