In your case I think panels in parallel will be the best solution. You are right that the shading of one panel in a series string will cause a serious reduction in output.
For small installations I doubt that mppt is worthwhile anyway. For a single 12v panel The matching most of the time is fairly good. The time when you may gain from mppt is on a bright sunny cold day with a low battery. If the panel is aiming for 20 plus volts and the battery is below 14 there is a fair mismatch. If the mppt device is sufficiently efficient then you should gain. Most of the time with a hot panel and a reasonably charged battery the match is so good that no mppt is going to gain you more than its own losses.
Mppt scores on larger installations with panel voltages higher than the battery voltage. You can then get a good match without having to use the correct number of cells to suit a given battery voltage.
Without mppt there will be very little difference with or without a charge controller ( a good controller should only add one fet drop and that should be very low. Unless you reach a state of over charge there is no need for a controller but it is a good idea in most cases unless it is constantly monitored.
You will gain with a mppt controller if the energy gained from the better match is not negated by the controller losses. Some of the basic mppt devices may have an efficiency of perhaps 90%, at this level you will hardly ever see an improvement on a single 12v set up, hot days will be worse than without it and only on the sunniest cold days will you be in with a chance. If you have one of the very best controllers with higher efficiency you may see better results except on very hot days.
With higher voltage strings of panels then even a 90% unit may prove worthwhile but you need a reasonable sized set up to get better value for money than buying another panel.
I never thought about this boat problem with inevitable shading but it certainly would seem virtually to rule out any likely gain from mppt.
Flux