Pure random targeting in ground combat?

[vorpal+5 (OP)]

From the manual (circa 2020):
Once a front line formation (or a light bombardment element in the Support position) has been matched
against a hostile formation, each friendly individual unit (a soldier or vehicle) in that formation engages a
random element in the hostile formation, with the randomisation based on the relative size of the hostile
formation elements. The targeting on an individual unit level represents that the different elements in a front
line formation will generally be attacking in conjunction (infantry supporting tanks, etc.).

Is this still how it works? Pure randomness (aside from weight being factored in) when one ground element selects another ground element? I understand the general principle: you can't specifically target the optimal target for your weapon class or the biggest threat to your unit, but pure randomness doesn't accurately reflect real-life scenarios. In reality, elements aim to make tactical choices to the best of their ability, considering various limitations (one being that the enemy doesn't cooperate much), with a degree of tactical savvy. Imagine a Soviet infantry battalion charging with T-34 tanks against a position of German Pak with infantrymen. Clearly, the anti-tank units will prioritize targeting enemy tanks, not just because tanks are larger than men. They would count on the German foot soldiers to shield them, protect them from enemy infantry, while they attempt to take out the enemy armor.

[nuclearslurpee]

This is a common inquiry which has been raised many times in the past. Each time, two common factors come up:

On one hand, Aurora is not a tactical simulator where ground combat is concerned, it only models the operational and strategic scale really. Of course, if an anti-tank gun has the choice between shooting at a tank or an infantryman, it will choose the tank nearly all of the time. However, that is a tactical consideration which Aurora does not attempt to model. In operational terms, if an anti-tank gun is faced against an infantry platoon, it will have to fire at the targets available. On the large scale, having more anti-tank guns in your army will mean you kill tanks more effectively, which is sufficient for what Aurora seeks to represent.

On the other hand, it has been shown that the mechanics do not support any kind of targeting but random without leading to a collapse-to-optimum in the ground combat mechanics. If an anti-tank gun is mechanically inclined to target armored units over infantry, for example, then a combined-arms formation becomes strictly inferior to monotype infantry or armor formations (assuming random targeting of formations). This is because preferential targeting of units within a formation leads to higher loss rates compared to random targeting, so the only way to counteract that is to make every formation of the same unit type so that only random targeting on the formation level applies. Since one of Aurora's key principles for ground combat is that almost any roleplay setting can lead to reasonably effective ground forces (if your roleplay setting calls for exclusive use of AA guns, not so much, hence almost any), this would be a serious problem. You can devise additional mechanics to try and balance this out, but then you're just adding a bunch of extra mechanical complexity that doesn't translate to decision-making depth - since, again, Aurora does not attempt to model the tactical layer for gameplay.

So while it does offend the sensibilities of some folks who would prefer greater tactical fidelity, ultimately purely random targeting works best in Aurora to accomplish the actual goals of the ground combat system. That's not to say that different mechanics would be worse, by any means, only that different mechanics are better-suited to a different game with different overriding objectives.

[Garfunkel]

The combat model would have to be remade almost completely to accommodate such gameplay. Basically, for tactical gameplay elements to be relevant, the game absolutely needs a map and formation movement(s). And since Aurora ground combat ranges from few individuals fighting aboard small space ships to millions of troops battling it out across giant planets, having a random map generator that makes maps with suitable scale and that are sufficiently accurate for our needs is no simple task. And then Steve has to create movement rules and now we would need to include all possible propulsion types - how much is an infantryman marching slowed down by 1.2 G gravity or how much faster they are in a 0.5 G gravity? Wheeled light vehicle should be faster than wheeled heavy vehicle, a jet engine cannot work without a sufficiently thick atmosphere, and so on. Steam engine, internal combustion engine, all futuristic engine types - what about ships? Do we need to model sail power? It is making a comeback now, after all, in a high-tech version. Then we need to think about how to add actual tactics and how to model them in the game - after all, Napoleonic marching columns of riflemen are not the same as modern fire-and-movement infantry squads advancing in a staggered, spread-out formation, even though both have their place - just on very different battlefields. There needs to be some rules about command and control - how many NCOs and officers are needed at each formation level to maintain unit cohesion, how many radios/phones/etc are needed to maintain basic or improved level of communications. Should there be additional communication techs? Because again, a modern networked military formation operates on a decision-making cycle that is an order of magnitude faster than their WW2 counterparts. And all of this is just modelling humans fighting other humans in space - what about aliens?

Now, it is easy to scoff that this is all unnecessary detail but if we want a proper tactical level combat model, it isn't. Just adding preferential targeting, as nuclearslurpee said, is not going to solve this problem, it will only add further problems. I would love to have all of this but it would be a huge task for Steve. The current model works 'well enough' for our purposes.

[Jorgen_CAB]

You can still do sort of targeted combat in the current model with moving units from frontline to rear echelon or even to move them into space on ships and back to the planet.

If the enemy formation have mostly soft infantry then no point in even using your antitank formations... once most of the infantry is gone you bring in the anti tank.

Depending on the defensive values of your units and the overall anti armour capacity of the enemy you can use more or less armoured vehicles in your formations etc.

The only real difference in general is the amount of micro you are willing to endure for the sake of being efficient and saving your more vulnerable units and supply.

Likewise does the mechanic encourage mono type units, like the same weapons on each platform and grouping them in separate formations so you can optimise their use more efficiently such as armour or weapons profiles.

No matter how you twist and turn the mechanics you can always find loopholes and exploit them if you wish or you just role-play and let it be what it is. I'm not against a weighted targeting system as it removes some of the micro if you want to be efficient but I don't think it is necessary to enjoy the current model.

I do think though that there should be some hit modifier to reduce the squared law of overwhelming force odds. In reality it is way harder to apply a larger force more efficient than a smaller one, especially on a strategic or operational scale.

[xenoscepter]

 --- I have suggested the same thing before. I suggested that Recon and Concealment be factors. Your own Recon would increase the number of times a unit selected a preferred target while enemy Concealment would decrease it. And vice versa.

 --- I spent more time and letters than was sound, sane or wise arguing past people trying to make them see what I was saying rather than reading what they wanted to hear.

 --- Needless to say, unless Steve himself decides to add some form of mechanic to allow units to be directed in some way, there will be no such mechanic.

EDIT:

 --- http://aurora2.pentarch.org/index.php?topic=12687.msg156566#msg156566
 --- http://aurora2.pentarch.org/index.php?topic=13137.msg162888#msg162888

 --- If there was going to be an implementation of prefered targeting, then a recon / concealment AKA intel / stealth mechanic would be, IMO, the cleanest and simplest way to make that happen. And mixed units would remain just as viable as monotype, with only a change in WHEN and WHY.

[vorpal+5 (OP)]

So, the key point of the current implementation is that it lacks preferential targeting in order for the system to be highly tolerant of various formation compositions. Essentially, it's more about bringing a large quantity of troops (like half a million tons) and less about considering 'which troops to bring and whether my troops are well-organized with combined arms taken into account'.

"Quantity is a quality all its own."

[Ulzgoroth]

While how your troops are organized may not make a lot of difference (so long as you refrain from outright mistakes like not provisioning logistics effectively), quality is not irrelevant - while high-quality elements may not tend to punch above their production cost, they do take less transport tonnage per production cost.

I believe the suggested composition for an invasion force is not identical to the suggested composition for defense? Unless I'm badly misremembering.

On the other hand, it has been shown that the mechanics do not support any kind of targeting but random without leading to a collapse-to-optimum in the ground combat mechanics. If an anti-tank gun is mechanically inclined to target armored units over infantry, for example, then a combined-arms formation becomes strictly inferior to monotype infantry or armor formations (assuming random targeting of formations).

That parenthetical assumption seems like a bit of a suspect keystone though.

If targeting works by picking a random formation and then 'smart' selection of a target element inside the formation, yeah, that clearly creates a pressure to negate the smart part by not offering any choice within your formations. But, as with the new missile defense, it's possible to not have the formation structure provide that kind of influence...

[misanthropope]

number one, the farther the scenario gets from "attacker using exclusively PW equivalents and MAV; defender driving exclusively unarmored infy and armored medium tanks, all at equal tech level" the harder the logic gets.  if someone has sample logic that addresses all the permutations and is robust in the face of limited information, i would be highly impressed and considerably interested.

number two, its exactly as reasonable to say "the defender is trying to force the attacker to fire on anti-optimal targets" as it is to say the attacker wants good ones.  why shouldn't the defender get to reroll attacker target selection?

number three, this wouldn't even introduce any tactical element; combat would still be strictly "press 'GO'; await results".

i agree there is some kind of sin against verisimilitude going on, but it looks like a bad effort-to-effect frontier of the game's development.

[Jorgen_CAB]

While it is obvious that in reality you always try to match the weapon system at the proper target, such as firing an ATGM at a tank and not at a guy with a rifle. But in reality every type of equipment and unit type will have their advantages and disadvantages and uses in different environments. You can never win a war without infantry for example, no matter how much you try you will never win without having boots on the ground. Everything else just increase the likelihood that you can win in different ways.

The game does not model how different platforms function, the mechanics are just not tactical in nature at all so it has to be very simple. If you add the possibility to weight the targeting you will as explained get rather unwanted artefacts. The pure random is under the current model the simplest way with the least unwanted artefacts. Not to say that the system could somehow be improved in the future if Steve think it is worth the time.

The only really bad implementation currently in my opinion are air units, they are awkward and not very fun to work with and I basically ignore that part of the game.