Just take station combat as a very good example. The viwer and auto calculator produce very different results.

 

So does also my unarmoured Guardians with missiles. Ships with Guardian role hang back behind the line and get to fire their missiles and generally be unharmed. With the auto calculator this is not that case.

 

The role of the ship matter and it is clear to me that it's not generated before hand, the result are way too different.

 

Save the game before a battle and run them several times using both methods and you will see they clearly work different. You need to have a few ships with different roles to experience it though.

 

I also fail to see what the battle log has anything to do with it... its not like you see what happens in the log before it happens... or I might be wrong in that?

 

As I said... the discrepancy  between the auto resolver and battle viewer are most clearly seen in battles with stations involved.

 

And... please... don't believe my words... just try it for your self and you will find out.

You actually do. First the hit is listed, then the shot is fired visually. I reported that as something they could improve maybe a week ago. ^^

yes... I know... the animation of a missile or shot is disconnected from each other. Every shot is basically a hit versus dodge roll and then damage is applied if it is a hit. A ship only fire ONCE per weapon type and do so depending on their cool-down. That is what you see in the log and why you clearly see missile attack results before the missile impact a ship. 

 

The animation is just for show, that is true.

 

That does not mean the actual shot is by any means pre determined in any way at the start of a battle, there are no proof of that in this model.

The Wiki, which is also checked by the Devs afaik, states:

http://galciv3.gamepedia.com/Ship_combat#Combat_Viewer

To me it basically say that you may auto resolve the battle or use the battle viewer and the result are the same, which is you may not directly influence the battle directly,

 

Just try some station assaults with a fleet that are fairly matched with both the station and its defenders. With auto resolve the battle will be even in the viewer you defeat them both in detail. The fleet charge out and is destroyed after which you attack the base and destroy it easily. You have to experience it to believe it I guess...

 

Use the campaign to test it quickly...

 

It's the same as "Deadliest Warrior" Other than a Hit happening; it doesn't matter what the ships do. They could all spin in circles. All that matters in the process is hits. These are all calculated before. The ships moving and dancing about are just graphical fillers while the sequential hit data is represented.

 

Actually... I think you are right and my previous experiences wrong...

 

I tried a few battles and in one I managed to get the exact same result no matter what I did down to the exact hit point every time. Even if I did attack different stacks in a different order. The big ship you get in the first campaign mission is so powerful it can take on enemies very reliably. An in one instance it attacked three targets and two of which it was never scratched and against the third it always ended with 278 hit point left. This was with or without the battle viewer. The worst part is that the outcome is the same no matter what target (stack) the ship attack first, so it does not even appear to be random even... that is even more disconcerting to me if that is the case. I have a real time finding this even a remote possibility to be a coincidence, will investigate this further.

If the HP between different tries didn't even change slightly, I guess something like HitChance isn't calculated with pseudo-random factors but procedurally. This means that there might be chances like 70%, but the seed the randomizer uses is constant upon reload.

So what you might want to check out is if something changes if you change the seed. We don't know what the seed is. But for example it might be the turn you are on. So even if you always get the same result upon reload, you might get a different result if you reaload, then skip a turn, and try again. You might also try to attack the opponent from a different direction. Maybe the location on the map acts as seen for the "randomness".

Yes, what actually scared me was that it did not change depending on which combat I resolved first... having a ship with 750 HP get to exactly 278 HP in the same combat against the same stack although it had been into one, two or no other combat first in the same turn seem quite worrisome to me... as I said... I will keep testing this. This really has to be a very unlikely coincident.

Yes, the results are calculated beforehand - the viewer isn't 'live', it's a 'replay'.

However, given that weapons have ranges and ships have combat speeds, I don't think it's only hits that matter. The relative positions of the ships must be part of the combat simulation, otherwise why have those stats?

Yes, the position and role of the ship do matter... I know that for sure... but here are some more observation running rampart with the crusader in the campaign.

 

In the battles I mentioned above no matter in which order or even between turns did the outcome of the battle change at all.... Every time I engaged the bigger stack of Drengine ships the Crusader ended up the winner with 278 HP. Every time I attacked the weaker stack second the crusader regenerated to 377 HP which means it took the crusader exactly the same amount of turns to kill the enemy ships.

 

So... no matter if it happens in the same turn or in which order these particular battles ALWAYS have the same result. Would be interesting to use a smaller size example, but what are the likelihood that I end up EXACTLY the same in HP every time even across turns if there are no real randomization going on in the background. This means we can calculate the formula done in the background with some effort and then use that to always know the result by simply plug in the numbers, or at least get a very rough estimation and find the best use of the roles since they seem very deterministic.

 

If this is the case I would suggest to through in some randomization on the targeting and to numbers... but this requires further investigation.

Ok... did some battle against pirates with smaller more simple ships...

 

The result are that damage is randomized, have managed to get a slight difference, it's just it matters very little given how deterministic ships target other ships.

 

In the test I had one tiny ship with 24 shields designated as "Assault" and one tiny ship with 6 laser strength designated as "guardians" and no modifications. The pirates had two small ships with a total of 8 laser strength. I managed to win this fight every time without taking a single hit point loss except for one time when I took 4 damage.

 

The other were with two tiny ships that had 4 laser strength and 8 shields both designated as "Guardians" against the same pirates. Now I lost one ship but still won every time with some damage to the last one.

 

Possition do matter but it is in my opinion to deterministic and there seem to be no regard to a ships threat level. Once they target a ship they will keep fire on it until it die. This can so be abused by putting defenses on assault ships and just weapons on your guardians. The enemy will target the first ship it get the opportunity to fire at... I have not tested this with all types yet, but it will be abusable by those the choose to do so.

 

So... load up your designated "Guardians" with only missiles, no need for defenses. Give your Assault and Escort only defenses and then Capital with a mix of Beam and Kinetics and no direct defenses. You will win pretty much any fight no matter what... at least I presume you will as long as the targeting behavior is like it is. But I will have to test this further to be sure...

Or just stick with just Guardian and Assault designations, that might be enough for the most part.

Ash, I asked this very same question just a few weeks ago.  I was specifically requesting a dedicated ship role that would allow missile ships to stay at range and leverage their range advantage.  And I agree with some sentiments expressed earlier in this thread about the battle system in general.  I honestly think the dev had grander plans for it than what we got but it seems that a lot of the more tactical mechanics of combat that were hinted at initially weren't implemented or have been pushed back to dlcs.  Given Stardock history I feel that the dev will get around to this eventually but combat still needs some tlc  in my opinion .

My friend I knew exactly what to expect with GC3. If combat alone was a sticking point I wouldn't have stuck around after GC2. I put over a 1000 hours into GC2 and I've been a Stardockian since then. I didn't fall in love with the series for it's combat, which should be a given for any of us here as combat has never been the game's strong suit, but rather the whole 4X package. With a game like Civ: Beyond Earth though depending on how you view it, you could say even it has controllable battles because you're just clicking on your troops and then clicking on what you want them to attack. More or less same here.

As for the larger maps and how many battles you have to fight, with them being both controllable Or auto-calculated it's a moot point and solely the players "opinion" on how many is too many to do your-self. The same could be said for any war strategy game on a large scale. In Total War or even in old school MoO2 I'd often do 10 or 20 battles per turn and not bat an eye. A long term player isn't going to do every little battle or a battle they know for sure they'll win manually anyway, but they'll do every "important" battle manually.

Even for those of us who want controllable battles in Gal Civ with knowing what Gal Civ is, none of us would be wanting or expecting anything as in depth as Total War anyway. Hell I'd be happy with just being able to click on a ship in the battle viewer, then click on it's target and watch it go. It doesn't need to be any more in depth than that as the rest of the depth is handled in how you build your ships and fleet composition and ship roles/classes. I don't think any of us wanting manual battles would expect a total overhaul of the battle system, nor want one. I don't think any true GC fan is here just for the battles. We're here for design choices. For strategic choices. For in depth research and planning how to grow your empire every time you start a new game. Will you be a war monger this time or will you go for that galactic peace victory? That's the beauty of it. Not the battles. The depth is everywhere in this game and it's the whole package that completes the big picture. As long as the auto-calculate is always there, how many battles the player choses to play through and how many they decide to auto is always left to the player. No-one is saying take out auto-battles or force the player to do every one of them manually. Then the player would be Required to do every battle and that would get monotonous in any large strategic game, not just here.

Personally I fell in love with GC because of the total package and what I could do with it, the customization. I've never finished a story campaign in GC. Don't get me wrong I like the game's normal races, they are well designed, fleshed out nicely, and they have great AI personalities I enjoy playing against. I know the lore but that's only from reading information screens and not so much because I care about the story. It's a strategy sandbox so technically you make the story up as you go along colonizing planets or going to war with one species or another, or making peace with one species or another. To me every map is a blank story I write as I play.

 

Range, tech, defense, support items. Everything is calculated and a hit sequence is enacted and all before combat. What we see means nothing.

Seeing how our fleets perform in combat means nothing?

BigBadB, when you try hard to misunderstand, you will.

 

Frankly I'd wish for a lot more variability in battles, but for that they would have to be calculated on the fly, not in front. If battles were calculated on the fly, ship behaviour would mean a whole lot more.

I honestly am not trying to misunderstand. I really don't get what you're saying.

Whether the battle is calculated while the viewer is running, or whether its calculated first and the viewer is a replay makes no difference - the result will be the same.

The viewer allows me to see how the result was achieved, and thus adjust my designs, roles and fleet composition to do better in future battles.

You don't even need the battle viewer for that. The rules that govern ship behavior are very deterministic as you can see from my test results above.

 

Just put defenses on ships such as Assault and Escorts and weapons on Capitals and Guardians and you are good to go (In MP it will be different though)...

 

As for what Voyager were trying to say I would agree that a real battle viewer with real combat would be nice, I actually though it was that at first. This would make combat much more complex and you would actually need to build really balanced and interesting ship designs... perhaps GalCiv4 will get that...

Just because the results are deterministic doesn't mean that seeing how they were determined isn't very useful when designing ships and putting fleets together.

Again, my apologies if I am being obtuse, I'm really not trying to be.

 

This is what I don't get - the battle viewer is 'real' combat, at least as far as I am aware. It's a replay of the battle simulation, but that doesn't make it less 'real'. As far as I can see, it's doing what you seem to want, or at least I don't understand what it isn't doing that you want it to be doing?

That is the part you don't seem to understand... ...the combat simulator is very simple. Which mean that understand what happens are pretty easy once you understand how ship behave in the simulator.

 

By real combat that means that ship range and maneuver really matter, that is why you sometimes see lull in the combat viewer when ships are out of position and need to catch up to what the simulator already decided, more or less.

 

The battle viewer are basically a view of an advanced form of rock, paper, scissor simulator... that is not ships firing from it's true position against a target within its zone of control or ships actually intercept or defend other ships based on whats going on in the current battle.

 

To simplify the simulator put all ships on a two dimensional line at different ranges... then each weapon has a range and ship has a speed with which it can close with the enemy battle line to engage. Each ship then have a primary and secondary target type and will use them once they are in range. Once they target a ship they will fire on it until it is dead.

 

A ship will target its primary target whether that ship is a real threat or not, so the system can really be abused that way.

 

This is why the battle viewer is NOT "real" in the logical sense... it is a real representation of the result of the simulator... nothing else.

So what you're saying is that the viewer is not an accurate representation of the combat simulation? If that's the case, then I agree that it sucks, as it means I have no way of seeing how my fleets actually perform in combat.

I was under the impression that it was displaying the battle simulation accurately. If it's simply eye candy then I'm very disappointed. Do we have any confirmation from the devs either way?

It seems to be like this shouldn't be overly hard to fix. In fact a combination of 2 changes may indeed improve the combat model immensely. Best of all, I actually modelled the below system for a JS/PHP based web game I created years ago with simple AI and similar to GC2/GC3 weapon types. I can confirm it definitely works! I also had a "flank" ship type in who would drive up the sides and take out enemy flanking ships then drive straight at their capitals. Nudge nudge, Stardock. Because firing arcs and speed were important, they were pretty useful.

First, enable firing arcs. A simple check to see if the target is within a set arc (say, 180* - that's easy using basic trig, it's an extra 3-4 lines of code). If it isn't, prevent firing. Or maybe allow this to be modded via ship extras? A turret mod or some such. Then running missile ships couldn't run and gun as easily.

Then, if that is or isn't applied, there is still the second option which may actually be modifiable - slow down ship movement in combat, increase combat ranges for weapons, increase the actual battle area and most importantly, massively increase the time it takes for ships to turn. It would let missile ships hammer at an enemy from range but it would take them significant time to turn and run. In order to stack this and make the run and gun viable they would give up vital space used on weaponry over to thrusters, a handicap the opponent wouldn't have to deal with and could load up on more weapons & PD.

This would allow missile ships to fill a valuable combat role, while heightening the importance of combined ship roles (Guardian, interceptor) and allow some proper combined arms approaches to fleet combat.

I think this would open up a fairly hefty can of worms for both the combat simulation and ship design. It could very easily end up being a massive pain for both the devs and players without much gameplay benefit.

This sounds more practical, imho. Although I am far less convinced that missiles need a buff than you appear to be.

Ah but it's not really a buff. There's a trade off. Load missiles on, stand back and hope they go down because if they don't, you're not getting out of there any time soon due to slow turns. Maybe momentum would be a good idea too, though that seems pretty covered. Less thrusters = shallower turns, further increasing turn times. 

If someone loads up on slow missile boats, shove a thruster in there and load up on PD because you're gonna take him down easy. The strength I think perhaps here is in order to protect his missile boats,  he will need appropriate guardian class vessels, interceptors. Plus he has to contend with *your* missile boats. It forces a combined arms approach, a very positive change I would say.

Making it designable (IE changing fire arcs), maybe yeah. However, I just tried to write a function to test my hypothesis (In PHP, no less. It's not really even a real programming language compared to some) and I can calc the angle between 2 points (using a static point as a guide) in 4 lines - 2 lines to calculate the angle between ships (1 line for the angle and 1 line for direction of movement to give a reference for facing) and 2 lines to echo a result if they're in LOS. This is for a 180 degree test from the ships center.

Instead of just flying in at full speed face first and firing in a 360x360 arc.