With the release of update 2.2.0 of King of Avalon a tweak to the Battle Logic was announced. This guide explains the tweak and shares some ideas about how the Combat Engine could work.
The tweak to the Battle Logic per update 2.2.0 is that Dragon Skills, Lord Equipment, Research, Talent points etc of all participants rallying or reinforcing will contribute to the overall result! (In previous versions only the battle benefits of the war leader were counted.)
In the old situation whenever someone joined a rally or reinforced, they did not have to worry about their boosts/benefits at all, for example your talent points and equipment could be set for resource production. I you joined a rally, you would get the boosts/benefits from the rally leader, if you reinforced, you would get the boosts/benefits from the player you are reinforcing.
In the new situation the announcement indicates that your boosts/benefits "will contribute to the overall result" when you join a rally or reinforce, but it is not specific about how. I have done some experiments, but the reporting does not seem to have changed, it is currently still the case that both rally leaders and rally fillers see the same boosts/benefits the rally leader would have if he did a solo attack.
So it is unclear how exactly the boosts/benefits from the rally fillers contribute, but they contribute, for this reason it is now important that rally fillers also make sure their lord is wearing the right equipment, the dragon is included and skills are set, and talent points are set for combat.
I have done some analysis of a situation where a group of players rally a high level castle (in this case the target was too strong and the attackers lost more than they killed). I want to share some findings so others can benefit from this. This was done before the battle logic changed, but since the recent change was introduced as a optimization of the existing battle logic, I believe the bulk of the way it works has not changed.
Both losses and kills were distributed evenly based on total troop amounts of each type. An example to illustrate. In a rally there were 3 people contributing different amounts of Double Crossbowmen: 12784, 2923, 415. Each of these players killed roughly 0.32 troops for each Double Crossbowmen they contributed, and they each lost 64% of their Double Crossbowmen.
Considering all troops types, either all troops of a type were killed, or all troops survived, and only for one troop type part survived and part was killed. This suggests there is a certain order of troops types. In this situation we attacked with troops of tier 7 and lower. All cavalry and infantry troops were killed, and lower tier bowmen and siege were also killed completely. The tier 6 and tier 7 troops survived, as well as the tier 6 bowmen. The tier 5 bowmen, the Double Crossbowmen, were the only troop type that partly survived. What the exact order is in which troop types and tiers die is not yet clear to me, but it is clear it does not (just) depend on one of the key stats like HP or Defense. Interesting is that the Tier 7 Bowmen troops died before the Tier 5 and 6 Bowmen.
Additional note: I analyzed a few more battle reports and the one I did the analysis on for this guide is atypical in the sense that we as attackers were overwhelmed by the defenders and that caused some types of troops to be killed of completely. Looking at another rally, in which we suffered 26000 losses and had 60000 kills, I noticed that for each troop type (infantry, bowmen, cavalry, siege) there was one troop tier that got partially killed, the rest either survived all or were lost all. This suggests that for each troop type there is an order in which the troop tiers die (potentially depending on the troop composition of the defender).
Concerning the kills, in general the Bowmen have the highest kill ratio (number of kills per number of troops), followed by Cavalry, followed by Infantry. For Siege the kill ratio is very different for the different tiers.
These kills and losses are likely influenced by the enemy troop compositions to some extend, so don't see it as hard evidence to do one way or another.
No matter the recent change in Battle Logic it is still best to have your strongest players (highest level Hall of War, best research, best equipment etc) as rally leader, as his contribution to the rally will still be more significant than the fillers.
Infantry is a buffer. Although they themselves do very little kills, without infantry your other troops are wiped out a lot faster. It is best to have a combination of all troops, and let the amount of Infantry depend on the strength between you as attacker and defender. The stronger you are as your target, the fewer Infantry troops you need.
In case a defender has a lot of traps, you might consider having Rams or Towers to absorb the damage from the traps, or Distance Siege for increased attack against traps.
Similarly you might also let your troop composition depend on the troop setup of your opponent, e.g. if the defender has a lot of spearmen, you might put extra Crossbowmen that do increased attack against spearmen, and also take into consideration that Spearmen have an increased defense against Light Cavalry.
You will likely lose most Infantry in combat and you want to prevent from being forced into situations in which you need to attack, but you don't have enough Infantry to function as buffer, so make sure you train plenty of Infantry troops.
If different rally participants have different troop compositions, they will have different losses. It is good to keep this in mind and keep things fair. Likely it will lead to unhappy situations if one person provides all infantry and cavalry, and has heavy losses with little kills, while another has just bowmen and has little losses and huge amount of kills. It is good to discuss rally compositions upfront and potentially let players use their Army Formation presets.
I did not fully understand when troops would end up in the Alliance Hospital. I mistakenly assumed that attacking a City that invaded our Alliance Territory would still cause my troops to end up in the Alliance Hospital, but this was clearly wrong. The Alliance Hospital is purely an extension of your own hospitals. If your own hospital capacity is full, your troops will go to the Alliance Hospital when lost in battle. Troops do not go to the hospital if lost while attacking a city, but they do go to the hospital when lost during defending (including reinforcing), attacking monsters, or attacking enemies on a tile. This is a good reason to be extra careful when attacking instead of defending, typically strong players can teleport to hive, but they limit their attack to much weaker players and hope that players closer to their strength will attack them, possible even by a rally, in which case they have the benefit of the large number of troops and the use of hospitals.
This is just a start, but I hope this gives a bit more insight in the battle logic and will give some tips. Hopefully others can share their findings in the comments so that together we will come to a better understanding.
Tool(s): Troop Training Calculator
Find this website useful? Please consider providing me with some much needed energy to keep going!