I just completed my first successful flight to Mars following one of the TransX tutorials. Even the tutorial I followed just basically said "it's up to you to decide when to do your corrections". I made the first one after I had completely escaped Earth's gravity (G=0.0 in the Orbit MFD with Ref set to Earth, you will also know when you're at this point when TransX drops the initial Earth escape stage of the flight plan). Other than the final orbit insertion burn, this first correction was the biggest one with a Dv of around 270. After that I just waited until a couple of months had passed to make my next one. Using the 1/2 distance guide given above is probably better but I just "eyeballed it". The subsequent corrections were really small, like Dv of less than 2 each time. I used my RCS thrusters for them since a short main engine burst could easily be too much.
Here's a couple of things I noticed that might help you. When you go into maneuver mode and set up your variables, be careful not to change the "man date" to a time before the current sim MJD or the time to maneuver will be negative when you go into the target view - basically telling you that you're already past the maneuver time that you set up. Also, the Cl. Approach value represents the radius from the planet's center, NOT distance above the surface. I failed to recognize this on my first approach since I usually have the Orbit MFD PeA/ApA/Alt set to display as altitude from the surface. So instead of setting up a nice tight low Mars orbit I had set a course for deep inside the core of the planet! Set your Cl Approach to be higher than the Maj Rad value shown in the Encounter View, and get your estimated altitude above the surface from the Min Alt value shown on the same screen. If that value is negative, make a correction burn to raise the Cl Approach! I made most of these adjustments well outside of Mars' sphere of influence (G value in Orbit MFD was still 0), and the required Dv to make significant changes to my approach was ridiculously small, like I said - only required small linear RCS bursts.
I'm sure on my next try I could do it better, and I have some other questions that I'll probably make a separate post for, but most importantly I made it and TransX makes more sense to me!