OK - variable reluctance (2-wire) crank sensor operation.. :-
As the target metal (flywheel tooth in this case) approaches the sensor, the output signal will start to rise due to magnetic flux change, until the sensor face is covered by the target when the signal will then fall to zero until the target starts moving away, when the signal will rise again in the opposite (negative) direction. If the target is of a similar width to the sensor, the transition from rising signal to falling signal will be continuous, passing through zero on the way.
By convention the sensor will be wired so the signal goes positive as metal approaches, and falls negative as the metal passes the sensor. The trigger circuit will be 'armed' if you like as the signal rises, then detects the centre of the target as it passes through zero ('zero-crossing detector' in electronic terms) . This provides a reliable mechanism to detect the precise position of the crank as the target will be centred on the sensor when this zero crossing condition occurs.
The convention to use the positive rising signal and subsequent fall through zero is due to there being no need for a negative supply voltage in the ECU to detect the negative signal, it can essentially 'ignore' the negative transitions as they are irrelevant.

Now if you wire the sensor in reverse, this sequence of events doesn't happen in the same way.. the target approaches and the signal goes negative, followed by a rising signal through zero as the target passes, the signal falls again as the gap between targets passes the sensor.
As you can see, the ECU now detects the falling edge sometime after the target passes..

sorry , bit long winded, I couldn't find any handy pictures to demonstrate..