This is basic watch knowledge 101, so bear with me while I give some details...
The theoretical power reserve of a specific movement is designed into the movement. It is based on the length of the mainspring, and the ratio between the number of turns on the mainspring barrel, and the turns of the center wheel typically. As the spring unwinds, the barrel turns (in most modern watches) and the teeth on that barrel turn the center wheel, which then drives the hands.
Manufacturers base the power reserve they advertise on this calculation, so assuming a new watch in good condition that is fully wound, it should run as long as they state. Most companies understate the power reserve, so you can often see a new or freshly serviced watch that will exceed the stated reserve.
Of course the condition of the watch can affect if the watch runs for it's full power reserve or not. In any watch, when the torque delivered from the mainspring is sufficient to overcome the friction in the movement, the watch will run. When it can no longer overcome that friction, it stops. This always happens before the spring is fully unwound, and when you add a higher load (from dirt, dried oils, running the chronograph, etc.) the point where the friction is high enough to stop the watch comes closer to the point of full wind, so the watch will stop earlier.
This same idea of friction can make the run time of even a new watch vary, due to small variances in how upright the wheels are, finish on the wheels, pinions, and pivots, etc. - all typical manufacturing tolerances like this can have a cumulative effect. This is why most manufacturers are conservative with the power reserve numbers they supply.
Hope this helps.
Cheers, Al
