This study is very interesting. Reminds me of a statistics professor who made a point about humans and their tendency to average/balance results. He gave us a simple exercise. Take your laptop, open a word document and fill the page with zeros and ones, print it. Open another word document, use your statistics software to generate the zeros and ones and fill the page, print it. The professor then asked us if we could see any differences between the 2 printed pages. No one could tell!! As he walked around the classroom, he was able to point which were generated from the computer compared to the ones we did ourselves almost every single time. Why? Because the computer generated pages always had much much longer sequences of zeros in a row and ones in a row. Most people, after 6 or 7 zeros in a row, will alternate.
but regarding the rat, an important question still remains... how long would it take him to adapt if the odds were to flip? (4 out of 5 times, it's RED, instead of GREEN)