The lower limit for the pull-up resistor is based on how much current the sensor will sink, which I think is ~30mA, so 5V/0.03 = 167Ω.
The upper limit is based on either the input current into the PIC port pin while HIGH plus the Off leakage back into the sensor output pin while HIGH, or how long it will take the pull-up to charge the capacitance of the wire and both pins when the sensor's output transistor turns off.
When the wire is at 4V (which is a clean HIGH, >Vih), there is 1V across my 10K pull-up, meaning it can supply (5-4)/10K = 100uA, which is well in excess of the leakages, which are likely ~1uA each.
A WAG at the capacitance on that wire node is 15p (pin) + 15pF(pin) + 30pF(wire) = 60pF. The time-constant when the pull-up is charging the node is 10e3*60e-12 = 600e-9 = 600ns. I dont think there will many pulses you will miss....
10K is a good compromise; does the job, doesn't produce much heat while the sensor output pin is low...