After gaining some experience in the game you may find out that in the opening phase the pieces are naturally attracted to form fragments of certain pattern: the shield wall.
It’s advantageous in Divoshi to push the opponent pieces back towards their side, crowding them in defensive positions and decreasing their ability to retreat. Having your own pieces side by side allows you to guard the fronline, leaving them free space for retreat at the same time. So the continuous horizontal front may arise where all friendly pieces are effectively covering each other. There is nothing easier for the opponent than to create the opposing shield wall in the subsequent row.
But the board has an odd number of rows. So inevitably one of the sides has an advantage of space and may call themselves the attacker. It seems to me that the more passive side needs to avoid forming the double shield wall, because it makes them even more passive, giving out a lot of space unnecessarily.
See the diagram which shows a model, ‘ideal’ situation. The white is the defender now. If it’s his turn, many of his moves lead to a loss or to an easy enemy breakthrough. For example: 1.Rh1-h2 g45xh3 2.g3xh3 f45xh2 (losing the rider). Or 1.Ag1-f2 f45xf3 2.e3xf3 h45xg3 3.h3xg3 g45:g3 (shattering the wall). And the white doesn’t want the enemy to break through easy, because he sacrificed his space in order to reach a stable defensive position.
To be honest, practically every move of the white – even those looking best, like 1.Wf1-de2 – now allows an easy breakthrough. So the black may either break through now, or prepare it first, moving the pieces from the 8th to the 7th row. He may even move some of his pieces on the 6th row since it usually doesn’t lead to a material loss (as you can make sure by yourself).
See the next diagram where the black have prepared the attack while the white was waiting and moving his Warrior to de2 and back home.
1…g45:g3+, sacrifying the savage, starts the melee.
After 2. h3xg3 f45:g3+ 3.g2:g3 Ac7xg3+ 4.Ag1xg3 Rb8xg3+ 5.h2:g3 g45:g3+ 6.f3:g3 de45xg2 the black side gains both material and positional advantage.
There is also a similar possibility of 1…h45:h3 breakthrough.
So it seems better to prevent forming the wall rather than accept creating it while being on the more passive side. But keep in mind that we are talking about an “ideal” situation where the wall is complete. In the most game situations the wall is either absent or incomplete, and so other possibilities come into play (e.g. attacking the wall from the side). The shield wall rarely grows to a full length because one or the other side often prefers another formation.
In any case, the wall defender can (and sometimes must) release the tension by attacking the edge of the board. On the first diagram, something like this is possible:
1.g3xh45 g45xh45 2.h3:h45 f45xh45 3.Rh1xh45+ h6:h45 and the shield wall becomes somewhat disrupted, while still keeping the black at bay.
When some of the savages get captured, heavy pieces can replace them on the wall if needed, standing side by side with their weaker comrades. But it makes the wall “inconsistent”, targetable by enemy savages.