Introduction

This article gives an explanation of various things to consider for troop compositions for attacking players in Guns of Glory, including a few suggested troop compositions.

Scope of this Article

The inner workings of the Guns of Glory combat engine are not known, so don't see this as the definite guide on making troop compositions providing you with the perfect solution, but read it and for both relative inexperienced and the more experienced players there will be useful information. I had been playing Guns of Glory and/or King of Avalon) for over two years and still learned a lot of new stuff when starting to investigate this topic based on feedback I got on the video that I made.

I will start by sharing two posts I got on my YouTube channel from someone far more experienced in this area, Labvem Kah/LivE EviL from K133, which in my eyes contains very useful information for all players. This article was written for King of Avalon, I have changed it where necessary for Guns of Glory, which were just name changes, the combat engine principles are the same. After that I will share my own considerations and give few possible troop compositions.

Post 1 from Labvem Kah/LivE EviL from K133 King of Avalon

The sole reason that everyone uses at least a few of each tier troops is because of the fact that the damage algorithm is based on the law of diminishing marginal utility – this means that the more number of a single troop type (meaning same class , i.e. infantry/cavalry/distance/artillery, as well as same tier level) you use, the less subsequent damage to the enemy per extra troop added . This applies to both PvP and PvE. This means that you need a good mix of troop tiers (meaning your highest tier all the way down to T1 troops, although having less troops per tier as you go down the tier level) and troop types (infantry, cavalry, distance and even artillery – having a few thousand or few hundred siege could help you because of the high initial utility of siege troops. But if the crippled march speed really bothers you, you can discard artillery).

I am not sure if the damage received algorithm is also based on diminishing marginal utility (which would indicate "increasing marginal troop damage received" in this case since the notion of diminishing marginal utility basically means that it gets worse per subsequent addition, and for damage received, "worse" is when a troop takes more damage), or if damage received is simply a flat, additive function in this game. This has yet to be confirmed. But if the former case is true, it gives everyone another reason to mix troop types in their troop formations. Even though you might suffer more losses in terms of troop count if you use a mixed tier march, your actual troop power loss will be less. For instance: 5000 T2 lost is better than 1000 T10 lost.

No one knows the exact formula for each troop type's marginal utility and the extent of how it changes with subsequent additions, but the combat boosts (aka "stats") and the default base stats of each troop type defined by the game (which could be seen by clicking the blue "i" icon above the troop type of interest within each troop training building in your city) are definitely factors that affect troop utility and marginal troop utility. Until then, the best way to figure out the "optimal" troop formation is by experimenting. Yes – it has to be done the hard way, especially because no one knows the secret formula that the developers do not seek to disclose, and because everyone has different stats. The spire is a great place to test this out, and from that it allowed me to figure out how to knock off people with higher march capacity and around 100% higher stats – just by rearranging my troop formation.

In fact, there is no such thing as a "perfect" PvP formation. Although one might think that perfect formations could be achieved through finding out the perfect ratio of troop types to achieve equi-marginal utility, it has to be taken into account that the enemy's stats and troops also affect battle results. For this reason, I do not fill up my entire march capacity when making my default/preset troop formations in my drill ground inside my city. What I do is leave around 50k troops unfilled, and depending on the enemy that I am facing, I will add different troop types to my preset when attacking them. Example of this may include: if the enemy has a lot of Pikemen infantry (T2, T4, T6, T8, T10 or T12), I add more heavy cavalry (T1, T3, T4, T6, T8 or T11) to my march when I attack them. (I know you have covered this fact in your video, but I need to stress my point that you have to take advantage of the different strengths and weaknesses of different troops depending on the situation ). Although I do not know the extent of this increased damage (i.e. how many percent of extra damage), this has to be taken into account when sending the marches to different enemies.

Lastly, the amount of infantry you need in your march depends on the strength of your enemy. If they are weak, then you can send less infantry so long as they aren't able to penetrate your meatwall of infantry. By doing so, you can replace the quota with other troops to deal more damage and further minimize your losses. However, don't overdo this – once your infantry dies, the rest of your troops will get shredded easily. Once again, the amount of infantry you need to send is best tested with experience. As mentioned above, the spire is a great place. The royal arena can also be used for testing, though to a lesser extent because it only allows you to send 1 tier of troops.

In conclusion:

• Use as many troop types (i.e. troop classes AND troop tiers) as you can in your march to maximize damage and reduce losses
• Use the different strengths and weaknesses of different troops to your advantage, and adjust them depending on situation
• Use an appropriate amount of infantry. Too much will make your army deal less damage, while too little will make your army wiped out too soon.
• Keep experimenting. Use the spire to your advantage

Hope this helps, and sorry for the huge wall of text above. Formations are a complicated thing to master indeed.

Post 2 from Labvem Kah/LivE EviL from K133 King of Avalon

During a battle, the troops die in a specific order. As you have mentioned, each troop class (infantry, cavalry, distance and artillery) has 2 subclasses/subtypes each (infantry = Pikemen and Musketeers, cavalry = heavy and light cavalry, distance = Riflemen and Grenadiers, Artillery = Mortars and Cannons). For every single troop class, one subclass completely dies before the other starts to get casualties – Specifically: For infantry, Pikemen die before Musketeers. For cavalry, light cavalry die before heavy cavalry. For distance, Riflemen die before Grenadiers. And for Artiller, Cannon siege die before Mortars.

Within each subclass, the troops die in the order of ascending tier level – e.g. T4 Pikemen die before T6 Pikemen. Many people have the misconception that all infantry must die before your cavalry starts to receive damage, followed by your distance, then your artillery. This is wrong – it can be the case that you still have some musketeers survive, but your light cavalry and Riflemen are taking damage. Some might justify this by saying that they are lost because of enemy traps, but my point above can easily be proved when taking a defense where not all of your troops die in 1 hit. Yes, most of the time the infantry will suffer more casualties – this is because there are two troop subclasses, specifically heavy cavalry and grenadiers, specifically designed to target enemy infantry. These troops do not try to target other troop classes until all infantry are destroyed. On the other hand, for Riflemen and light cavalry, they will attempt to target the enemy's cavalry and distance layer respectively; however, not all of their "hits" land on the said target troops – the enemy infantry will protect them by taking a certain percent of the damage (you can test this in the royal arena).

One interesting thing to note though, the game says that crossbowmen have increased damage against enemy spearmen. However, despite that, the shots fired by crossbowmen will be taken by enemy warriors instead of spearmen unless if they run out/did not send any warriors. This is because the rule of the order of troop casualties apply in all circumstances. In this case, it can be thought as the warriors are protecting the spearmen from the shots of the enemy crossbowmen. For this very reason, I include lots of heavy cavalry in my march – they target the very first line of the enemy's defense.

With that in mind, the following can be concluded.

For infantry, the order of troop casualties from first to die to last to die are:
T2 => T4 => T6 => T8 => T10 => T12 (Pikemen)
=> T1 => T3 => T5 => T7 => T9 => T11 (Musketeers)

For cavalry:
2 => T5 => T7 => T9 => T10 => T12 (light cavalry)
=> T1 => T3 => T4 => T6 => T8 => T11 (heavy cavalry)

For bowmen:
T1 => T4 => T7 => T8 => T10 => T12 (Riflemen)
=> T2 => T3 => T5 => T6 => T9 => T11 (Grenadiers)

And for siege:
T1 => T3 => T5 => T8 => T10 => T12 (Cannons)
=> T2 => T4 => T6 => T7 => T9 => T11 (Mortars)

This set of orders is very important because it indicates what troops you need to replace most frequently if you attack a lot, and it can also affect your decision of how many of each troop to include in your march formation.

Conclusion:

• In a battle, the order of troop casualties is one subclass before the other followed by ascending tier levels. This order applies in all circumstances.
• It is important to take the order of troop casualties into account when making troop formations or even when training troops in general.
• The belief that all infantry must die before cavalry before distance before artillery is wrong - they can all suffer casualties simultaneously, by following the order of troop deaths mentioned above.
• Heavy cavalry and grenadier's attacks land on enemy infantry 100% of the time, while the attacks of riflemen and light cavalry can potentially land on enemy cavalry and distance, but some of that damage will be "rescued" by enemy infantry.

Best PvP Troop Compositions

Above considerations were very useful for me to read, I hope for other readers as well.

Although I agree the best troop composition depends on various factors, including your own and your opponents stats, and your opponents troop compositiosn, I would still like to share what I see as a good general troop composition. Just use it like this, or see it as a starting point for your experiments to further improve upon it for your specific situation, including opponents.

My preset currently is 25% Infantry (15% T10, 10% T9), 30% Cavalry (20% T10, 10% T8!), 45% Distance (30% T10, 15% T9). Obviously I have not unlocked T11 troops yet.

The percentages Infantry/Cavalry/Distance were based on gutfeel and experiments. I started out with 20% 40% 40% but as my stats grew I got the experience that this composition in general worked better for me.

I do not (yet) have the conviction that lower tiers are a good choice for PvP combat, more about that in the next paragraph, but I do feel that using the different troop types is useful, as it can exploit potential vulnerabilities of opponents and makes you less vulnerable. The 2/3 - 1/3 ratio I use is based on gutfeel and experience as well. For Infantry and Distance T9 troops have a different subtype than t10, but Cavalry T10 and T9 are the same subtype, that is why I use T8 Cavalry in the mix.

If you are already convinced including lower tier troops are the way to go, you can still use the basic idea of my march composition, but just leave a certain amount aside for adding a couple of 100 of each of the lower tier troops and types.

Lower Tiers, Yes or No?

I might very well be missing something, because by far the most players I battle use small amounts of lower tiers, but in my own (limited) experiments I just haven't seen any indication yet that having lower tier troops is useful. On the contrary, when using it to the extreme, using my PvE troop composition (with thousands of each of the troop types and tiers) that works very well for Underworld and Kingdom Threat, inflicts less kills and takes more casualties than my PvP troop composition by far. I will likely do some more experiments and if I get convinced otherwise, I will for sure share it here.

There might very well be this diminishing marginal utility for PvP as I am aware there is for PvE, but if it is, it is not as straightforward, or easy to "use" as it is for PvE. Maybe it is implemented in a different way, but you also have to deal with the fact that in PvP, unlike PvE battles like Underworld Gates and Kingdom Threat, you take casualties as well, so it is not just about lower tier troops inflicting kills, it should outweigh the fact they die faster.

Another possible reason to use lower tiers is to use them as a buffer, to make sure your highest tier troops, which are harder to retrain/heal, don't die so quickly. I am not a fan of this tactic, because it also means your total health, defense, attack goes down.

As mentioned in the previous paragraph, I do currently only use the lower tier troops (the highest possible lower tier troops) when it benefits my mix of troop subtypes.

Difficulty with running Experiments - Randomness

When experimenting the best PvE troop composition you can easily use the Kingdom Threat and if your troop composition, march size and boosts are the same, you will do the same amount of damage, time after time. One of the complicating factors with doing experiments for PvP is that the result can be different. Just attack again and again the same opponent in your Spirit Mines revenge list, and although you might lose each time, if you look at the casualties you take and the kills you inflict, they differ. This makes it more cumbersome to experiment, you can't assume you have improved just based on one or a few trials.

I find the Spirit Mines, specifically the revenge list, the best place to do experiments.

Conclusion

This article should introduce you to many key concepts in determining your PvP troop composition, and give you some starting point for what troop composition might work as well.