I've decided to completely try to unravel the mysteries of Strategic Combat.
I'll post any new information I find out here.
1. The spell power influence of books of a realm a node can counter are completely ignored for the attacker. In other words, it is assumed that all such spells will always get countered.
2. Nodes have no effect on defender spell power at all. In other words it is assumed that nodes never counter for the defender.
3. If neither the attacker, nor the defender is the human player, all non-neutral AI players receive a +50% stats boost to Total army Attack, Defense, Ranged and all spell power ratings. In other words, all AI players have a huge advantage against neutrals. (50% might not sound like a lot but it's applied to everything so the effect multiplies)
4. Spell power is divided to 3 categories. One is directly added to defense. One is directly added to attack. One is a percentage bonus to the attack sum of the army.
5. Casting skill/Mana is spent in chucks of 1/3 of the total available. So it'll only influence the first 3 turns of combat.
6. During the first 3 turns of combat, only ranged attack power is used in calculations.
7. The weights of realms contributing to each spell power :
Nature = 50 Attack, 10 Percentage Attack, 25 Defense.
Sorcery = 75 Attack, 25 Defense
Chaos = 100 Attack, 30 Percentage Attack
Life = 15 Percentage Attack, 40 Defense
Death = 75 Attack, 15 Defense.
These numbers are per book.
8. Actual Spell Power[attack] is sum of the above for all books/30
Actual Spell Power[percentage] is sum of the above for all books/20
Actual Spell Power[defense] is sum of the above for all books/20
9. For each 5 mana spent, Spell Power[attack] is added to the army total attack rating, for the turn it was spent only.
For each 5 mana spent, Spell Power[percentage]*attack rating of army is added to the army total attack rating, for the turn it was spent only.
For each 5 mana spent, Spell Power[defense] is added to the army total defense rating, and this extra amount will carry over to future turns.
For example, a skill of 50 and 10 Chaos books means (33*16)/5=105 attack power per turn for the cost of 16 MP per turn. For reference, a unit of 6 melee and 6 figures is roughly 720 attack power, so this is roughly as good as 1/6th of the attack power of a single halberdier. Unless my calculations are wrong somewhere, the effect of magic is very poor in this department, which, being the highest part of most realms, is alarming.
The same books and skill would mean (15*16)/5 = 48% attack boost for the entire army each turn from the percentage multiplier which is far more reasonable.
Assuming the same in Life books, (20*16)/5= 64 defense power is added to the army total. For reference, a unit of 2 health, 6 figures and 4 shields is 384 defensive power.
10. Damage is subtracted from the attack power to simulate how killed units don't get to fight in later rounds. Nice, something is done correctly at least! Damage done in the ranged phase is subtracted from both the melee and ranged attacks. Damage is also subtracted from "defense" which contains/works as health.
Conclusions :
3 is just unfair. The AI already had a huge advantage in these battles (neutrals don't have spells). Furthermore, the bonus is applied AFTER the health is saved for end of combat damage distribution for survivors. This means the extra 50% health all non-neutral sides gain is extra army health that, if damaged, will not translate to actual damage on units. Meaning in AI vs AI combat, a significant portion of the damage is automatically prevented on the winner.
5+6 means spell power will only ever affect the ranged phase of the combat (unless both parties have 0 ranged). Meaning melee armies benefit somewhat less from the "percentage attack" part of spell power.
7 is just stupid. Spell power should be increased by researched combat spells, not books. As is, a wizard with no spells known but 10 books is as powerful in battle as one who has all the spells.
I'll post any new information I find out here.
1. The spell power influence of books of a realm a node can counter are completely ignored for the attacker. In other words, it is assumed that all such spells will always get countered.
2. Nodes have no effect on defender spell power at all. In other words it is assumed that nodes never counter for the defender.
3. If neither the attacker, nor the defender is the human player, all non-neutral AI players receive a +50% stats boost to Total army Attack, Defense, Ranged and all spell power ratings. In other words, all AI players have a huge advantage against neutrals. (50% might not sound like a lot but it's applied to everything so the effect multiplies)
4. Spell power is divided to 3 categories. One is directly added to defense. One is directly added to attack. One is a percentage bonus to the attack sum of the army.
5. Casting skill/Mana is spent in chucks of 1/3 of the total available. So it'll only influence the first 3 turns of combat.
6. During the first 3 turns of combat, only ranged attack power is used in calculations.
7. The weights of realms contributing to each spell power :
Nature = 50 Attack, 10 Percentage Attack, 25 Defense.
Sorcery = 75 Attack, 25 Defense
Chaos = 100 Attack, 30 Percentage Attack
Life = 15 Percentage Attack, 40 Defense
Death = 75 Attack, 15 Defense.
These numbers are per book.
8. Actual Spell Power[attack] is sum of the above for all books/30
Actual Spell Power[percentage] is sum of the above for all books/20
Actual Spell Power[defense] is sum of the above for all books/20
9. For each 5 mana spent, Spell Power[attack] is added to the army total attack rating, for the turn it was spent only.
For each 5 mana spent, Spell Power[percentage]*attack rating of army is added to the army total attack rating, for the turn it was spent only.
For each 5 mana spent, Spell Power[defense] is added to the army total defense rating, and this extra amount will carry over to future turns.
For example, a skill of 50 and 10 Chaos books means (33*16)/5=105 attack power per turn for the cost of 16 MP per turn. For reference, a unit of 6 melee and 6 figures is roughly 720 attack power, so this is roughly as good as 1/6th of the attack power of a single halberdier. Unless my calculations are wrong somewhere, the effect of magic is very poor in this department, which, being the highest part of most realms, is alarming.
The same books and skill would mean (15*16)/5 = 48% attack boost for the entire army each turn from the percentage multiplier which is far more reasonable.
Assuming the same in Life books, (20*16)/5= 64 defense power is added to the army total. For reference, a unit of 2 health, 6 figures and 4 shields is 384 defensive power.
10. Damage is subtracted from the attack power to simulate how killed units don't get to fight in later rounds. Nice, something is done correctly at least! Damage done in the ranged phase is subtracted from both the melee and ranged attacks. Damage is also subtracted from "defense" which contains/works as health.
Conclusions :
3 is just unfair. The AI already had a huge advantage in these battles (neutrals don't have spells). Furthermore, the bonus is applied AFTER the health is saved for end of combat damage distribution for survivors. This means the extra 50% health all non-neutral sides gain is extra army health that, if damaged, will not translate to actual damage on units. Meaning in AI vs AI combat, a significant portion of the damage is automatically prevented on the winner.
5+6 means spell power will only ever affect the ranged phase of the combat (unless both parties have 0 ranged). Meaning melee armies benefit somewhat less from the "percentage attack" part of spell power.
7 is just stupid. Spell power should be increased by researched combat spells, not books. As is, a wizard with no spells known but 10 books is as powerful in battle as one who has all the spells.
