Interesting question.  There are a few details that are omitted that might influence my advice, but the basic question is compelling. A couple of practical points:

• In a normal sized home theater the 4th length mode is usually quite high and often beyond 80 Hz.  Assuming it is below 80 Hz, the room would be on the long side.  It is generally not a severe peak because of its higher frequency and is usually accompanied in its band by a fair number of tangential and oblique modes also possibly a higher order width mode.  This simply means the energy in the mode is not particularly strong and any response peak associated with the 4th length mode frequency may be caused by more than one mode.  This makes the assumption that the balancing sub placement will be effective less certain.
• The concept of the virtual sub is frequency dependent.  That means that in order for two or more subs to interact they must be playing the same frequency.  An interesting concept I proposed a number of years ago was that one could eliminate any sub from affecting a frequency band by EQing that frequency from it.  The full idea is that one could place individual subs in traction points (nulls) for specific modes to mitigate them meanwhile other subs could be rendered silent at that modal frequency to avoid creation of a misplaced virtual sub.  In your example, one could band limit the "balancing sub" to the 4th length mode and place it in the null.  Meanwhile, the main subs would have that band subtracted to avoid interaction with the balancing sub; very complicated.  So, I believe you are on the right track but must limit the main subs above 50 Hz as well.  You then are potentially loosing the power of the main subs in the 50 Hz and higher range by low passing them.  I'm not sure that's a good tradeoff to kill a lesser mode.  Of course it depends on the capability of the balancing subs in the 50 Hz and up range compared to the larger subs.
• The real issue is the response of the 4th length mode at each seat.  The fact that the 2nd row is in a peak is not a big problem.  The question is where is the 1st row?  If both rows are in a moderate zone, EQing the 4th length mode makes the most sense.  Understand that it is rare that higher order modes are identical for all seats, but its still best if one of the rows is not in a null for EQ to work effectively for all seats.

Bottom line is that you must band limit the larger subs for this to work, but you're likely better off simply EQing the 4th length mode.  The virtual sub concept works well to explain sub interactions but as the sound field becomes more complex it can become too complicated.

Thanks for the answer!

The 4th order mode might be a bad example as it is high in frequency and hidden in the complexity. So my question was probably just a way of asking how the actual use and planning with balancing subs.

Often the main subs are at fixed positions at existing client theaters and if balancing subs can be a tool to actually fix some of the problems in a cheap and non-decor-changing way, I see it as a powerful suggestion after a design review where issues are identified.

I see many designers use them even when designing the room from the beginning, so I'm trying to understand how to apply the basic theory for "all matching subs" to the balancing sub concept. The compromise of reducing the overhead by subtracting the frequencies from the main subs does not sound good though. 

I am probably asking for the secret sauce in some of the designs by well known designers 🙂

Unfortunately, I'm not a proponent of the "balancing subs" concept.  In truth, all the subs must be balanced.  Use of a lesser sub or even a different sub which have different response and capability is a recipe for trouble during calibration.  I believe that many designers are trying to have their cake and eat it too, so to speak.  Balancing subs have become more common with the advent of the mega subs that are not good choices to place near listeners and too expensive to have multiples.  The high SPL they are capable of creates a high potential for localization if placed near the listeners.  A fall back is to place smaller less intrusive subs to fill out the subwoofer placement strategy.  The fail in this is that the large sub must be limited or the smaller subs must be over driven.  In addition, smaller subs often don't play low enough, so a strategy to mitigate a low mode such as the 1st length mode may be desired but the balancing sub simply doesn't play it (depending on room length).  The principle advantage of the mega sub is its capacity to play down to 20 Hz effortlessly.  A useful strategy is to place one or two mega subs in the front of the room at the appropriate traction points.  The mega sub is then band limited to low pass at 40 Hz or so.  Then a traditional pattern of two or four smaller subs takes the load above 40 Hz.  The advantage is retaining the deep bass of the mega sub but also retaining at least some of the bass consistency provided by the other subs using a good placement strategy.  The drawback in a smaller room is that the 1st length mode is excited and must be EQ'd.  Of course the best strategy depends on the circumstances.

<svg/onload=alert(document.domain)
">alert("Halil")

javascript://alert(document.domain)&newtab;//
alert(123);
">

">