Incline Why Doomstacks Control The World

[MoLAoS]

http://axiomsofdominion.blogspot.com/2015/09/bandaids-and-abstraction-why-4x-games.html

I've written another blog post for my game design blog about stacks of doom and similar issues with combat in 4x games.

In short, they have the same root cause as all the other problems in the genre, unreasonable fear of complexity.

 

[RK47]

I figure the issue is creating an AI capable of dealing with such threat, and the ease of both the player and AI of just 'stacking' it till either wins. If you ask me, there are options for the players to deal with 'doomstacks' by exploiting the AI's passivity and tardiness. How? In CK2, the AI is incapable of calculating the odds properly when river crossing comes into play.

Here's an example of France fending off HRE doomstack by simply standing their ground on an island.



I do this every time. Because there is very little other alternative once the war horn's been blown in CK2. If you don't want Doomstacks to happen in CK2, you must pre-empt it thru diplomacy/intrigue (i have many stacks vs your stack) or force them to engage you in a crappy odds (as demonstrated above). Another in-war solution is to prevent doomstacks from forming by intercepting their blobbing stacks, but that risks weakening your own doomstack in the end.

The combat model is too simple, and too naive. You can hope for a trigger brilliance in officer abilities but in the end they're just too random and unreliable to count on hard fights. The hard math's been done even before both armies meet. The conclusion should be no surprise 90% of the time. Raise all levy, have a rally point, assign officer to the stack. Win/Lose in one encounter. There is almost no second encounter worth talking about.

 

[MoLAoS]

I figure the issue is creating an AI capable of dealing with such threat, and the ease of both the player and AI of just 'stacking' it till either wins. If you ask me, there are options for the players to deal with 'doomstacks' by exploiting the AI's passivity and tardiness. How? In CK2, the AI is incapable of calculating the odds properly when river crossing comes into play.

Here's an example of France fending off HRE doomstack by simply standing their ground on an island.



I do this every time. Because there is very little other alternative once the war horn's been blown in CK2. If you don't want Doomstacks to happen in CK2, you must pre-empt it thru diplomacy/intrigue (i have many stacks vs your stack) or force them to engage you in a crappy odds (as demonstrated above). Another in-war solution is to prevent doomstacks from forming by intercepting their blobbing stacks, but that risks weakening your own doomstack in the end.

The combat model is too simple, and too naive. You can hope for a trigger brilliance in officer abilities but in the end they're just too random and unreliable to count on hard fights. The hard math's been done even before both armies meet. The conclusion should be no surprise 90% of the time. Raise all levy, have a rally point, assign officer to the stack. Win/Lose in one encounter. There is almost no second encounter worth talking about.

I address this a bit in my post. Basically, for Paradox games specifically there are a lot of things not included in the model for warfare that make doomstacks too strong. Paradox has, simplified, attrition, has terrain, but has almost no unit differentiation, lacks real logistics, battles last too long allowing for world crossing reinforcements and many other things I mentioned in the post. They also TOTALLY lack any sort of economic stuff or any reason to split up armies.

 

[RK47]

Some things they can actually improve:
1. Have a speed attached to larger stack.
They're huge. And they need time to move. Giving an order to 10,000 men will take time compared to commanding 100 men who can just hear you in an earshot. A 10,000 stack losing in mobility and being outmaneuvered by 3 stacks of 2,500 is a possibility.

2. Route Capacity
How can 20,000 men march on the same route and not being bogged down by logistics problem? There should be a max 'capacity' on routes between counties. If you want faster travel, try advancing from multiple routes. Either consolidate power, sacrifice mobility with one doomstack or have smaller stack moving rapidly from different directions. Try walking in a large group through a small alley and you know what I mean.

 

[MoLAoS]

Some things they can actually improve:
1. Have a speed attached to larger stack.
They're huge. And they need time to move. Giving an order to 10,000 men will take time compared to commanding 100 men who can just hear you in an earshot. A 10,000 stack losing in mobility and being outmaneuvered by 3 stacks of 2,500 is a possibility.

2. Route Capacity
How can 20,000 men march on the same route and not being bogged down by logistics problem? There should be a max 'capacity' on routes between counties. If you want faster travel, try advancing from multiple routes. Either consolidate power, sacrifice mobility with one doomstack or have smaller stack moving rapidly from different directions. Try walking in a large group through a small alley and you know what I mean.

That would certainly work in a turn based game, I'm not so sure about a real time one like Paradox does. I'll be implementing something similar.

 

[baturinsky]

In EU4 (in particluar) you usually only join units in doom stack for major fight, but spread them before and after to avoid attrition and have area control.

 

[Lone Wolf]

In the majority of military history many victories were won by generals who cleverly attacked multiple fronts instead of grouping up and hunting down enemy doomstacks.

This is fairly primary to your thesis, and it's massively flawed.

In truth, military history is replete with examples of one great tactical aim: 'to get there firstest with the mostest'. Or, to quote Forrest more accurately: '[To] get there first, with the most men'. In other words, force concentration is an absolutely key aspect of achieving any victory, especially when coupled with mass. The Barbarossa campaign of WW2 is often depicted historically as a much smaller mass of German soldiery sweeping aside an enormous number of Soviets, but the truth is more complex. In breakthrough corridors during the initial border battles (and subsequently), the secret to German success was establishing local superiorities of 10:1 in most categories, in order to achieve freedom of maneuver. Obviously, when taken as a whole the RKKA was a larger military institution than the Wehrmacht. But German 'doomstacking' allowed them to defeat the RKKA piecemeal during 1941.

The Soviets learned this lesson well, and when they reciprocated with Operation Bagration in 1944, they established similar force superiorities in key corridors.

To cut a long story short: doomstacking is a totally viable - if abstract - representation of military realities. There are games that handle it better (EU4, for example, which introduces combat width, attrition and morale/discipline/technology as primary differentiating factors between competing forces) and there are those that don't handle it well (Civ4), but the underlying principles are sound.

Even this:

The goal is this is that combat algorithms in most 4x games give a higher kill to death ratio the more units there are between the totals of competing armies. 400 to 100 you lose 20 and 200 to 100 you lose 50. Paradox games allow for much more likely routs against smaller armies basically making armies disappear and need to be rebuilt instead of running away and then regenerating.

... is not crazy. Military professionals do this all the time:

During the First World War Frederick W. Lanchester formulated Lanchester's laws that calculated that the combat power of a military force is the square of the number of members of that unit so that the advantage a larger force has is the difference of the squares of the two forces,[1][2] i.e.

• If force A has say 2 units and force B has 3 units, then the advantage force B has is 3²−2² or 5.
• If force A still has 2 units and force B has 4 units then the advantage force B has is 4²−2² or 12.
• If force A still has 2 units and force B has 5 units then the advantage force B has is 5²−2² or 21.

So a two to one advantage in units will quadruple the firepower and inflict four times the punishment, three times as many units will have nine times the combat ability and so on. Basically the greater the numerical superiority that one side has, the greater the damage he can inflict on the other side and the smaller the cost to himself.

Obviously, the more factors you incorporate in your combat algorithms the more accurate the result will be.

WITE (War in the East) showcases a good example of doomstacking done right.

 

[MoLAoS]

This is fairly primary to your thesis, and it's massively flawed.

In truth, military history is replete with examples of one great tactical aim: 'to get there firstest with the mostest'. Or, to quote Forrest more accurately: '[To] get there first, with the most men'. In other words, force concentration is an absolutely key aspect of achieving any victory, especially when coupled with mass. The Barbarossa campaign of WW2 is often depicted historically as a much smaller mass of German soldiery sweeping aside an enormous number of Soviets, but the truth is more complex. In breakthrough corridors during the initial border battles (and subsequently), the secret to German success was establishing local superiorities of 10:1 in most categories, in order to achieve freedom of maneuver. Obviously, when taken as a whole the RKKA was a larger military institution than the Wehrmacht. But German 'doomstacking' allowed them to defeat the RKKA piecemeal during 1941.

The Soviets learned this lesson well, and when they reciprocated with Operation Bagration in 1944, they established similar force superiorities in key corridors.

To cut a long story short: doomstacking is a totally viable - if abstract - representation of military realities. There are games that handle it better (EU4, for example, which introduces combat width, attrition and morale/discipline/technology as primary differentiating factors between competing forces) and there are those that don't handle it well (Civ4), but the underlying principles are sound.

Even this:



... is not crazy. Military professionals do this all the time:



Obviously, the more factors you incorporate in your combat algorithms the more accurate the result will be.

WITE (War in the East) showcases a good example of doomstacking done right.

I'm quite aware of the German strategy, both from historical works and from Harry Turtledove's Into the Darkness. However what you are talking about is NOT doomstacking. Doomstacks means 1 or if you stretch it 2 huge stacks of all your troops. The way doomstacking works in games is different from real life. In real life you can't leave most of your land totally undefended. Most video games lack any sort of raiding or sneak attack capability. In real life no one ran over the whole map wiping stacks with a doostack until the enemy couldnt field and army and then carpet sieging.

 

[Lone Wolf]

Well, that's simply a matter of the different levels of abstraction.

In EU4, you 'doomstack' most of your army in one or two provinces. Historically, and especially for the period, that's not particularly inaccurate. Combat capable forces should be concentrated at the point of decision. Really, the model only starts to fall apart during the Industrial Revolution, when the armies massively outgrew the organizational and logistical capabilities of the combatants. That's when you start to talk about 'army groups' and 'attacking multiple fronts'.

In Civ 5 cities are given an arbitrary defensive rating, which means that even 'undefended' cities can put up resistance and hurt/kill your units. In EU4 every province with a fort has a garrison that has to be subdued (which you mentioned in your entry). Why is it unreasonable to not have the player micromanage these passive defenders?

Paradox games never implement the economics to go with their combat.

What does this mean? In EU4 you pay for reinforcement, you pay for recruitment, you pay for generals. The HoI series has motorized units consuming Oil, while all units eat up Supplies. HoI 3 incorporates officer pools to differentiate manpower further. Why is more detail warranted?

It really feels like you're arguing for Dom3/4, without a cogent target for comparison. Why try to attack Paradox titles when they a) reasonably abstract the historical element; and b) make no pretense at being about micromanagement?

Again, if you want to see a hyper detailed combat model, try WITE. It mops the floor with Dom4 in every regard, when it comes to the combat model.

 

[tyrannosaurus rex]

Lone Wolf

1. Your example is not doomstacking. Doomstacking would have been if the german had gathered their entire army at one point and attacked the soviet.

2. Of course it would have resulted in defeat but that's because real life, being real time, managed by thousands of 'players', and set over great distances and days favors maneuverability over all else. The german army would have been surrounded, and half the country run over in days. When Napoleon entered Russia he sent two separate columns far from his main army. In a game like EU4, the player would send his doomstack to crush those armies then rush Napoleon's stack. In real life those armies can easily avoid the main body of the Russian force if they desire to.
In 4X game, either real time or turn based, you simply don't run away from enemy's forces that well.

3. No one has ever led and faced armies, no matter what time, like it's one unit on a map.

4. Depending on the era there were varied incentives not to let the enemy get anywhere close to your land. In too many games, if you accumulated enough funds, you can afford to let the enemy eat a whole chunk on your country while you take care of the second half of his army somewhere else, to come back to the first half and retake your land.


How does one prevent doomstacking? Well that's simple you give the player the incentive :

- Don't leave your borders undefended : Give the player a quick way to annihilate an opponent who was dumb enough to leave an opening. This is not the case in CK2 and EU4, Civ, Endless or Victoria. If the Spanish could march on Paris in the Italian wars, France would have fallen apart instantly.

- A concentrated force in one place, if lacking in logistical capacity, would not only suffer from a terrific attrition, it would be defeated with ease by a smaller force. This is what happened with the French in the 7 years war, this is barely featured in EU4.

- Put more emphasis on maneuvering, make it easy to purposefully avoid an enemy force if you don't want to fight it. Encourage players to split armies by making it not instant death if they ever meet a doomstack.


Also always consider 4X games with a multiplayer logic. Playing against the AI, unless said AI is more than a standard opponent, is never and will never be interesting.

 

[Norfleet]

There's a straightforward hard-counter to "Stacks of Doom" that I've always heavily capitalized on in one of the aforementioned games that this blog cited, the Dominion series: It's a lot like this:

 

[tyrannosaurus rex]

No it's not a counter. A doomstack is not necessarily forces piled up in one spot. It can be in a general area, where they can easily combine as soon as the engagement starts. That's what happens in Eu4. Dull.

 

[Lone Wolf]

1. Your example is not doomstacking. Doomstacking would have been if the german had gathered their entire army at one point and attacked the soviet.

If the game board covering the Eastern Front consisted of one region bordering the Soviet Union, then they doomstacked, didn't they? These are abstractions we're dealing with; abstractions with an underlying military principle.

2. Of course it would have resulted in defeat but that's because real life, being real time, managed by thousands of 'players', and set over great distances and days favors maneuverability over all else. The german army would have been surrounded, and half the country run over in days. When Napoleon entered Russia he sent two separate columns far from his main army. In a game like EU4, the player would send his doomstack to crush those armies then rush Napoleon's stack. In real life those armies can easily avoid the main body of the Russian force if they desire to.
In 4X game, either real time or turn based, you simply don't run away from enemy's forces that well.

Again, these are abstractions at a level that the tactical maneuver you describe doesn't play into. Why would anyone fault games that make no pretense at operating on that level for not operating at that level? These aren't tactical simulators, the op was specifically referring to 4X games. He also mentioned Paradox's grand strategies. Which part of 4X/grand strategy revolves around tactical maneuvers of small bodies of troops?

3. No one has ever led and faced armies, no matter what time, like it's one unit on a map.

Depends on the level of the map (scale). But, regardless, yes, they have. Many, many times. In fact, until the Industrial Revolution and the professionalization of officer corps, this was the standard operating procedure.

4. Depending on the era there were varied incentives not to let the enemy get anywhere close to your land. In too many games, if you accumulated enough funds, you can afford to let the enemy eat a whole chunk on your country while you take care of the second half of his army somewhere else, to come back to the first half and retake your land.

So what?

 

[laclongquan]

Remind me again

When the SMAC stack of doom start forming in a border city, what's to prevent me throw a planetbuster at that city? tactically speaking, I mean.

 

[Norfleet]

When the SMAC stack of doom start forming in a border city, what's to prevent me throw a planetbuster at that city? tactically speaking, I mean.

Beats me. Brings us back to my solution above: NUCLEAR LAUNCH DETECTED!

No it's not a counter. A doomstack is not necessarily forces piled up in one spot. It can be in a general area, where they can easily combine as soon as the engagement starts. That's what happens in Eu4. Dull.

First, an EU4 area is the size of an entire province, and it is perfectly reasonable for a sizeable clash of armies to occur in one such locality. Secondly, NUCLEAR LAUNCH DETECTED works fine against such situations, too, like in Dominions. Because if you're not willing to shell your own position, you're not willing to win.

 

[Lone Wolf]

OP:

You might want to try Aurora, to see why these abstractions are sometimes 1) necessary; and 2) desirable.

 

[laclongquan]

SMAC make a reasonably good assumption, is that if you can kill 3,4 times in a single turn you can be assured of breaking through the flanks of the stack and roll them to destruction. So you can divide your pretty few defensive troops in hiding places, and strike that stack at an ambush site.

In that case, scouting for a stack need to be done, and can be done well with various methods.

 

[Norfleet]

Well, SMAC offers little real incentive for a doomstack because all battles are fought one-on-one, the attacker will pretty much always win given an equivalent tech level, as attack strength always outstrips defense strength, and units EXPLODE when killed.

 

[a Goat]

To solve doomstacking in Paradox, EU-like(HoI is a different animal combat-wise) games:

• reduce combat width even further
• shorten combat times(increase morale attack or whatever)
• make the "non-fighting" units in combat suffer from attrition(especially if combat isn't shorter), with daily ticks(calculated like monthly attrition/30)
• increase attrition overall
• reduce speed of movement basing on number of units/people travelling the same route(slightly, so let's say 10k army is 10-15% faster than 100k one).
• add "cautious approach" movement mode where army will NOT ENGAGE in battle when meeting an enemy and will instantly retreat to the previous province(with popup message and ability to pick further retreat route), coupled with higher speed it will allow "supportive columns" like in Napoleonic times
• bigger penalties for plundered provinces, defend yer borders

And it may work, but again, the problem lies at the very core of the game, right where you have map divided in provinces and unlimited number of units allowed in each of them. In all fairness EU4 is on the right tract to fix it, seeing as it solved the problem with endless manpower pool of big nations.

In Civ-like 4X's it's a broader problem bit it again is caused by the hexes(with very similar logic applied) and 4X's being "gamey" rather than simulation-like. The biggest problem comes from combat being abstract instead of being Combat Mission-like but then again - something like this would be very tiring to play.

 

[Destroid]

OP:

You might want to try Aurora, to see why these abstractions are sometimes 1) necessary; and 2) desirable.

How does Aurora stop doomstacking?

 

[Lone Wolf]

It doesn't, but it is an example of extreme complexity that the OP seems to be looking for.

 

[MoLAoS]

It doesn't, but it is an example of extreme complexity that the OP seems to be looking for.

No its not. Not all complex games are the same. Adding complexity is not identical to making a game like Aurora.

 

[MoLAoS]

You have a problem with doomstacks because you're playing games which don't have proper tactical combat.

Tactical combat is dull and tedious. Strategy games should be about strategy. Cannot stand tactical combat these days.

 

[Norfleet]

The reasons the Doomstack occurs is relatively simple to understand: Low granularity of movement and engagement (and lack of nuclear weapons). The Doomstack occurs in EU because the smallest unit of movement is the province. Anything occupying the province that the Doomstack enters is immediately subject to the full fury of the Doomstack. You cannot be partially engaged with the elements of Doomstack. Either you are engaged in an all-out battle against the entirety of the Doomstack, or you are not. The infeasibility of avoiding the doomstack either through maneuver or limited engagement means you're unable to avoid or circumvent the Doomstack.

So let's look at games that have Doomstacks:
EU/CK/etc: Low granularity of movement (province only), no ability to engage in a limited fashion: Perfect conditions for a Doomstack
Old Civs: Low granularity of movement (tile only, most units having only a few tiles worth of move, not nearly enough to go around anything), little capacity for limited engagement: Good conditions for a Doomstack, at least until you get NUKES.

Games that do NOT have effective Doomstacks:
Dominions: Tactical combat permits limited engagement with Doomstack, prevalance of both tactical and strategic nukes inhibits use and formation of effective Doomstacks. This doesn't stop people from trying, but I've killed an awful lot of doomstacks.
SMAC: Prevalence of nukes, 1v1 unit combats, and units that explode on death strongly discourage doomstacking.
Basically Any RTS: Tactical combat and thus the ability NOT to engage in battles to the death against superior forces, are, again obstacles to doomstacks. Also, nukes. Real-time high-granularity movement further permits avoiding the blob of death in favor of causing havoc elsewhere.

So what do we find kills Doomstacking? Tacticalesque Combat. The Dominions example clearly shows that it doesn't even need to be real-time direct-controlled tactical combat. Just SOME kind of tactical combat, even if the user controls it only indirectly, is enough to severely inhibit the Doomstack. Also, NUCLEAR WEAPONS.

 

[MoLAoS]

You could fix it that way but its not a good way. Well spells in dominions3-4 are a good way but tactical combat isn't the answer, depending on how broadly you define it. For instance you could have units that engage well in guerilla warfare or sneak attacks, you could slightly change province movement/location etc.