I gave specific quotes in the followup here http://bg.battletech.com/forums/index.php?topic=49502.msg1144989#msg1144989 . To convince me that the StratOps rules say otherwise, you'll need to be similarly specific. (Yes, I understand that there is some contradictory Q&A now http://bg.battletech.com/forums/index.php?topic=49502.msg1145341#msg1145341)
W.r.t. time to train you just need to connect the dots: 8 hours per maintenance/repair cycle + one m/r per scenario + the number of scenarios needed to advance a level.
StratOps (2nd printing), p. 187.
No mention is made of hours. It's strictly stated that the scenario count is what matters. The only time factor is the time in between scenarios. So take two teams, A and B, both starting at Green. Both need 5 scenarios to reach Regular. Team A is attached to a unit that does quickie raids (one per month or less), & the team is kept safe back at the DropShip during actual combat. The soonest they can reasonably expect to level up is after 5 raids, so maybe by the 5th month they'll level up.
Team B, OTOH, is involved in a running campaign, like you see in some of the official sourcebooks, where the unit has a day or two in between scenarios (if they're lucky). Within a week or two of in-game time, they could easily reach the 5 scenario count they need to level up... assuming they aren't killed off (requiring a brand-new team) or replace their team casualties (which restarts the scenario count).
Just as stated in the examples in the book, it's the scenarios they survive without replacing casualties that tracks their skill improvement, not how many hours they spend working on the job.
The only issue that needs determined is if 2 or more linked scenarios count as separate scenarios every time, or if the team needs time to attempt repairs between scenarios for them to be considered 'separate'. I lean towards the latter, but also feel it's up to the GM & players in each gaming group.