There is a "short stay" option, where you return on the same synod. This requires a higher-energy transfer, involving Venus on at least one leg, to get there early enough to hit the return window and also have enough time to actually do something on the surface. Apparently you can get up to 90 days in some situations, but 30 is more likely, out of a total elapsed mission time of 400-650 days.

https://nssdc.gsfc.nasa.gov/planetary/mars/marsprof.html

Of course, this in in the "resource-constrained" mode of thinking. If you allow for giant vehicles with on-orbit refilling, SpaceX-style, you just spend a bunch of energy to get there fast instead of monkeying around with Venus.

There are faster transfer opportunities, but not 3 days. But I think he way saying 3 days of surface ops. Which is still an exaggeration. I think the NASA reference mission is 30 days?

I had a peek at the Mars Design Reference Architecture 5.0 [1], and under "trajectory options" they talk about two classes of missions.

"Opposition class missions" stay on the surface for 30 to 90 days. "Conjunction class missions" stay for 500 days or more.

That confirms the 30 days that you mentioned as a reasonable minimum.

[1] https://www.nasa.gov/pdf/373665main_NASA-SP-2009-566.pdf