Frédéric L Note, sensors don't measure a "rotational" effect.
I know that... You did not follow my argument (maybe because I did not explain myself well...). As CG mentioned above, these membrane switches don't measure absolute time but difference between successive activation of two switches: when the key is pressed, it first activates the back switch and then the front one (and also mid sensor, in 3 sensor actions), and these are usually only a few mms apart. If the pivot is short, the key bottoms at a steep angle (what I called rotational effect); if the pivot is long, the angle is shallow. In a theoretical infinite pivot (same displacement at the back and front of the key) you simply cannot use successive sensor strategy because the sensors are all activated simultaneously. Thus, membrane switches are a bad strategy if you have optimal pivot length, unless you place one switch close to the fulcrum and the other close to the key front.