AnFi - so we come to yet another inaccuracy that you hadn't appreciated in your arrogance.
There is an accepted way of calculating the speed of an object from a video - you could try googling it but I will help you out and it doesn't involve the use of a stopwatch.
You need to know the frame rate of the video and then use software (VLC for example) to measure the movement of the object between each frame. Then after doing that for 50 frames, you will have an accurate time scale (of 2 secs if the frame rate is 25 fps) and, depending on how you measure the distance, and accurate one from which you can accurately distill the speed.
I say accurately measure because the accepted way to do this is to either overlay a scale - you could do this using the known length of the Apache, or use accurate and precise reference points in the video (as dCl did).
None of the accepted methods of ACCURATELY assessing speed from a video have the camera moving (like it is in the apache video) because it adds another level of error.
Can you finally admit that your assessment of the speed of the apache is flawed and cannot be relied upon in any way, shape or form to support any of your hypotheses?