Mulligan decisions may actually be the single most difficult aspect of competitive Magic to learn and constantly improve on. I've personally struggled with it throughout my career, and it's problematic trying to take lessons with you since the standards for which you keep or mulligan a hand change drastically based on your deck type and the matchup you face.
Last week , I declared my intention to win Pro Tour 25th Anniversary with Reid Duke and William Jensen. We've played as a team many times in Grand Prix and many more in practice matches not even counting how long we've been on the same Pro Tour testing team. We don't really have any set rules for how much we do or don't communicate during our matches, but we do appreciate the fact that anytime you say something aloud you could be giving away information at the table and you're interrupting someone's game when you decide to volunteer information. The level of confidence you need to be introducing your advice to the game of someone you respect deeply needs to be strong, and we try to be careful.
One thing we don't skimp on, though, is advice with regards to opening hands. There's so much to learn and it's the type of mistake that will have long-term ramifications on the rest of the game, so you want to be sure you're giving yourself the best chance to win. As a rule of thumb, I like to ask myself:
How lucky do I have to get to have a competitive hand?
Another way to interpret this statement is that I want my opening hand to clearly be advantaged against someone without enough lands. I haven't studied my games to know exactly what percentage of the time I win as a result of mana screw, but I have played Magic for long enough to realize that it's a significant loss in equity to lose a game when my opponent has one of their hands that would fall in the bottom 5%. You could even go one step further and instead of quantifying what it means to have "not enough lands" you could instead consider many hands to be nonfunctional.