Montza's formula provides an excellent approximation, but it is not suitable for accurate calculation.
3-4 swordsmen will die... See the damage in the status bar and you will find out not the range, but the exact number.
For example, we attack 10 hydras, let the damage be 151 HP. This means that 2 hydras will die. We write the number 10-2=8 in the large window.
To avoid misunderstandings, I will add that the program finds the number of warriors in the attacking squad itself, as well as the minimum and maximum damage, and the attack. The parameters of the attacked (except for defense) are needed to predict morale. Since the number of changes in R in Heroes I depends not only on the number of responding units, but also, strangely enough, on the damage! Moreover, this number is not limited to ten, as in "Heroes III".
P. S.
Minimum damage (for one djinn or for 3?). I set 24 (for 1 djinn)*;
Maximum damage (for one djinn or for 3?). I set 36 (for 1 djinn)*;
The problem is probably here. Instead of 24 and 36, try specifying 4 and 6 (I do not recommend testing on djinns due to the chance of halving the enemy).