I think that the RSLs on the robots are good indicators here. They can clearly show us when the robot is in the "teleop disabled" mode, and when it transitions to "teleop enabled" mode.
It appears that when the timer reads "135" on the audience display, this is acting as a placeholder for the 'teleop disabled' period. One would think that the only time that the timer would be valid at "135" would be at time 135.00000, meaning that the timer is never really valid at 135.
Using that logic, we would expect that the robots would be enabled as the timer counter is changing to "134" (1:44 YouTube time) - however, using the video you've provided as evidence, we can see based on the RSL status that the robots are clearly enabled at the beginning of the "135" counter, or at 1:43, as you've suggested. We can see (not in the video that you've linked due to camera angles, but others, such as this one
https://www.youtube.com/watch?v=fI8r...st-Vt0&index=9 ) that the robots are enabled through the "0" second. This would lead me to believe that matches are actually 2:16, not 2:15.
*** DISCLAIMER - the audience display is provided only for convenience of the audience, the field timers are the only valid clock on the field.
What would be interesting would be to look at some DS logs of matches, and compare those to DS logs of the "Practice Mode" function of the DS software.