Currently if you station a fleet at a starbase or shipyard both will participate in the battle at the same time however it is a bit sequential. The fleets engage  each other first then the 'attacking' fleet if it survives the defenders moves on to the Star base. If you have the Omega modules installed your defending fleets will get boosts of up to 50% across the board while attacking in its influence and while stationed. 

If I had ONE request in this thread is to have the last line of fortifications for a MSB ensure that it has at least 1000 hp or 1500. I am going to politely disagree on the use of Military starbases not being able to defend planets. We all have our arguments but I do think at or near late/end game these fully fortified and hardened space structures should be able to decimate a fleet before being destroyed. 

I am not talking about early or mid game. Only late game where if you have an opponent on par with you, you an build one near his territory as a staging ground for a fight/war. He should be alarmed, and immediately (I would as a player) build fleets to take it out. 

Anyway thanks for the many great ideas in this thread and thank you to Joeball and Stalker for some math clarifying things. 

First, thank you for the Hex formula. I will use that in discussions going forward.

Second, I would argue that you cannot assume a linear relationship in effectiveness. A 20 sensor range is not twice as good as a 10 range, it is much more than that. You attempt to trivialize the notion of area covered, and while having perfect knowledge of a wide area may not be useful, having a significant coverage is a strong boon. Beyond that, advance notice is a strong benefit. A 20 range lets me know when danger is coming earlier than a 10 range, and that notice is highly valuable. Further, a higher sensor range lets me keep tabs on more of my enemies planets and starbases....allowing me further tactical information which can used to determine strategy.

But for the moment I will play along, lets say that it is a simple linear relationship.

Sensor Ship: 38 range / 348  = 0.11 Length / Manu

Starbase: 19 range / 260 = 0.073 Length / Manu

So even here, the sensor ship is still more manu efficient.

Third, you are arguing the semantics of maximum efficiency, but my point still stands. For every hex my empire extends beyond the current borders, the less effective starbase sensors become. Eventually, they becomes useless. Meanwhile, a sensor ship will always be useful, it can always be moved to return its efficiency if its efficiency is lost in its current position.

You yourself made the argument, "two million tiles observed that aren't in the right place are worth nothing". A starbase that cannot move will always risk becoming useless as the game dynamics change, a sensor ship can be repositioned to where it is most useful.

I never once said that advanced notice was not a strong benefit. I said that advanced notice has far more to do with detection range than with area covered. It does not matter that at 20 range I cover 1261 tiles rather than the 331 tiles covered at 10 range. It matters that I detect the target up to twice as far away, and that the target must be capable of traveling a much greater distance in a single turn in order to hit me without forewarning. The minimum amount of warning I get of the approach of an oncoming vessel is equal to

max(0, floor((R + 1 - M) / M))

where R is my effective sensor range at the location, M is the number of moves per turn of the oncoming vessel, and floor(X) is the function returning the largest integer N such that N <= X. Minimum warning time is linearly dependent upon sensor range. It is not dependent upon total area covered, unless you want to involve a dependency upon a square root to bring the area back to the sensor range.

While it is true that a sensor range of 2R is not necessary twice as good as a sensor range of R, I would point out that it is less than or equal to twice as good as a sensor range of R, as far as early warning goes. A sublinear relation with the factor by which range increases, and you want me to care about a quadratic efficiency. For the quadratic efficiency to be what matters here, you need the increase in advance warning to be roughly proportional to the square of the increase in detection range; instead, you have an increase in advance warning which is roughly proportional to the increase in detection range and may well be proportional to something less than that.

Where are you getting 260 manufacturing points? It takes 5 construction points to fully upgrade a station's sensors; that's 135 manufacturing + 42 multiplied by however many cargo hulls you had to build to get those five construction modules out to the station, if you get 1 construction point per construction module. 5 construction ships each providing 1 construction point is 345 manufacturing points, 3 construction ships at 2 points each is 288 manufacturing points (which we may wish to discount to 240 because we only use half of the construction points on the final construction ship for finishing the station sensors), 2 construction ships at 3 points each is 246 (which we may want to discount to 205 since the second construction ship only uses 2 of its 3 construction points; this requires additional hull capacity or reduced construction module size). If we have the double construction points bonus from pragmatic, this is further reduced; at 1 construction module (2 construction points) per ship, we require 207 manufacturing points (which we may wish to discount to 172.5 since only 1 of the two points on the third construction ship is used), at 2 construction modules (4 points) per ship we require 192 manufacturing points (which we may wish to discount to 120 since only 1 of the four construction points on the second ship is used for sensors), and at 3 construction modules (6 points) per ship we require 123 manufacturing points (which we may want to discount to 102.5 since one of the six construction points is not used).

If you're including hyperdrives or other extraneous components in the construction ship cost, you're paying extra for conveniences, not for the sensors. Conveniences, I might add, which are almost essential to positioning or relocating a sensor ship in a timely fashion on a large map and which degrade its sensor range efficiency below that of a ship filled with nothing but sensors.

Surveillance is the only situation where area really matters at all. However, in surveillance, range buys your observer safety and, potentially though not likely, secrecy. Range lets you keep out of your opponent's area of influence while conducting the surveillance. Area buys you targets (assuming a relatively even distribution of surveillance targets, at any rate), but the increase in interesting targets is likely at something less than a 1:1 ratio with the increase in area covered.

I think we may be getting a little bogged down here with needless comparisons between sensors on SBs and sensors on ships - the two are not strictly comparable.

There's an opportunity cost when adding sensors to a ship, in that a fully-sensored ship cannot carry weapons. A star base with full sensors can. Likewise, a sensor ship can move, while a starbase cannot. These two facts mean that comparisons between the two are wildly disingenuous. Moreover, it's fairly obvious that sensor boats are currently ridiculously OP - a single huge hull packed with top-level, resource-free sensors can reveal an entire Insane map, meaning that you can park it in orbit of the world it was built on and never use sensors ever again. This means that any math you currently do is irrelevant, since it's going to be re-balanced drastically in the next patch or two anyway.

I like to have total sensor coverage of my empire, plus a little round the borders (6-12 spaces). I'd rather achieve this with starbases than sensor boats. I don't mind the idea of sensor probes, but having 1-2 ships which effectively sit in 1 place being star bases seems inappropriate, considering that we HAVE star bases that could be doing this.

I'd maybe go for something like a 'listening post' ring, that doubled the sensor range of the base - in GC2, I'd usually have a dozen or more star bases that just housed sensor arrays doing this job. I'd also probably want a universally-tech-unlockable ground-based 'early warning system' or something that gave planets a big sensor radius bonus, too.

But generally, regardless of the math or the profit-loss evaluations, I'd want SBs to be able to achieve the biggest sensor radius simply because that's more lore-appropriate. Star bases are gigantic mega-structures which produce vast amounts of power and can house sensor arrays that are bigger than the entirety of the largest ships. The only other thing big enough to hold a sensor array that size is a planet, but SBs have no atmospheric interference like planets do (hence why we keep all our best telescopes in space). This should really be matched by the game, and star bases really ought to be the backbone of your sensor network.

If you're worried about SB sensors being too cheap, well, just increase the constructor points per level. Thulium Sensors already require 2; why not just increase it by 1 per level? Level 1 sensors, 1 CP, level 2 2CP etc. That's fine, as long as the resultant boost is worth the extra cost.

This may be one of our focuses of disagreement. I would argue that 2R has more than twice the value of R, based on the points I have already outlined.

At turn 1, I can put 2 constructors on a cargo hull (which actually feels like this has changed, I don't remember being able to do that turn 1, have the masses been changed?)

So that's 135 + 42 x 3 (3 cargo hulls) = 261. So I was off by 1.

A cargo ship with 2 construction modules costs 2*27 + 42 = 96 manufacturing. You require 3 of these, so that's 288 manufacturing, but you only need half of the final, so that discounts it by 48 manufacturing, making the cost 240. Unless you're saying that there's no way you could possibly find a use for the extra constructor and therefore build a less efficient cargo ship with 1 construction module which costs 69 manufacturing instead, which I see very little reason to do.

Are starbases the biggest things in space, though? They certainly weren't in GCII, as a starbase was 0.75 million metric tons plus 0.1 million metric tons per module (a fully-upgraded economic starbase using only the Terran tech tree was thus ~4.8 million metric tons, as opposed to the 13.8 million metric tons of the default Destroyer design in the Battle of the Gods scenario, and while it may be the case that the ships have higher average densities than the starbases, I'm somewhat doubtful that the densities differ by a factor of 3 or more; sadly, the intelligence report screen didn't give starbase dimensions), though the numbers have changed in GCIII (specifically, huge ships have become ~18 times less massive and the known dimensions are ~3-10 times smaller; I haven't compared the other ship sizes).

Additionally, in the GCIII battle viewer, it appears as though Huge ships are of roughly the same size as space stations. Granted, the battle viewer is not an ideal tool for size comparisons due to issues with perspective.

I'm inclined to go with the viewer on this tbh, and take SBs to be huge hulls floating about in space, minus the engines and other mobile offensive unit components. Besides, a fully-upgraded star base in GC3 is, what, 90-odd modules; that's 9.75 million tons there and then using the (imo conservative) 0.1 per module measure. Even with the significant global mass reduction that ships appear to have suffered, it's hard to see how modules could come in at under 100 kilotons - that's only about as much as a modern-day aircraft carrier or cruise ship.

These are all great ideas, but so far no one has addressed the issue that you can just avoid a MSB, go around it and take out planets.

I know some people are against having so much power on MSB to defend planets, so maybe MSB should have some modules that give them production and a shipyard so that you can still function if all your planets are taken. I know this idea is terrible as where would the economics come from for fleet maintenance but honestly I don't see why someone with 50 MSB and large fleets defending them should lose the game if they lose all their planets to transports going around the MSB. They'd still have an amazing faction power and should have the chance to come back.

I think they should lose for failing to move their fleet out to deal with the horde of transports tbh

But yeah, the starbase slow-debuffs are a bit weak compared to the kind of engine stacking you can get away with, especially late-game and especially on cargo hulls. I'd like to see it actually become more viable to simply fly completely around a military starbase's perimeter than trying to cross (so an absolute cap on movement speed in the area rather than just a slowdown effect). That way, you could build a starbase right next to a planet and know that the enemy won't be able to get through in 1 turn.

Well, if you assume that conservation of mass applies in GCIII and that the mass listed for a ship includes the mass of its cargo, then you can estimate the mass of a module by checking the mass of a construction ship. Of course, this has some issues, as the listed mass of a ship appears to depend only on its hull size (e.g. Tiny) rather than on what it carries, and moreover the masses seem to have been chosen more or less randomly; the default Terran bomber has an average density of almost 2.5 million kg per cubic meter while the TAS Crusader from the campaign has an average density of only ~20 kg per cubic meter (which, if the TAS Crusader is made of the same kinds of materials that the standard bomber is and the standard bomber contains no empty space at all, indicates that the TAS Crusader is 99.9992% empty space; densities are computed by volume enclosed within the bounding box whose dimensions are the length, width, and height of the ship, and while this does mean that the average densities could be much higher than indicated, the bomber and the TAS Crusader don't appear to differ enough from the bounding box to change the average density by more than an order of magnitude or so, which isn't all that significant when the densities based on bounding dimensions differ by about five orders of magnitude).

You could alternatively estimate how much mass a construction module could possibly contain based on the volume of the ship carrying it and some average density, but as noted above, the average ship densities in GCIII (and probably also GCII) are a bit variable, and of course then there's scaling issues even before you start playing around with component scale - if you fit a construction module onto the blank small hull and build it, and then do the same with a blank tiny hull, you'll discover that your two construction modules differ in volume by a factor of about 33; the bounding dimensions of a default-scale construction module on the blank Tiny hull are 4m x 3m x 7m (84 cubic meters) while the bounding dimensions of a default-scale construction module on the blank Small hull are 12m x 10m x 23m (2760 cubic meters), using the Terran ship set. I don't know what the bounding dimensions are for default-scale construction modules on the blank hulls for the other size categories, but I imagine you'll run into even more issues determining how big a construction module is if you were to check against those. Internal consistency isn't exactly GalCiv's strongest point.

(Actually, it's not entirely true that the masses appear to be random. Taking the mass of a Tiny hull to be 1, a Small hull has a mass of 2, a medium hull has a mass of 3, a large hull has a mass of 4, and a Huge hull has a mass of 10, while Cargo hulls have a mass of 3.6. According to the game, Tiny hulls have a mass of 73500 metric tons, roughly on par with a real-world aircraft carrier despite typically being smaller than real-world fighter aircraft; GCIII small hulls have a listed mass of 147000 metric tons, or roughly two aircraft carriers, despite being only a bit larger than a real-world jet fighter. It would perhaps be more fair to say it looks like someone picked what they regarded as a nice progression of mass without really thinking about what it meant.)

I asked the computer to count this time; there are 43 generic, 25 military, 21 economic, 10 mining, and 7 culture modules. A fully-upgraded starbase therefore runs from 50 to 68 modules if there's at least one precursor relic in range to enable the archeology stuff, or 45 to 63 modules if there is not.

One thing worth noting before running with the old GCII mass metric is that in GCII, you kept the old module when the new one was added to the station. In GCIII, the new module replaces the old (or so it would seem, since the old module's bonuses are lost, though where the pieces go is anyone's guess, and we already know from the behavior of ship components that there's nothing specifically preventing the old weapon, sensor, and defense components from stacking with the new; the behavior of planet improvements such as factories and labs suggests that the starbase factories and similar are refitted when the station is upgraded with the new version of this), which would suggest that a GCIII station's mass would be more or less constant except when an entirely new module line is installed. It does beg the question of why the newer modules cannot be installed directly, skipping the older versions, but then that's also true of the planetary improvements.

While I can see an argument for not wiping out a faction until it loses the ability to take a planet, I'm not terribly fond of the idea of allowing starbases of any kind to sponsor shipyards.

As far as defending planets goes, I would be okay with military starbases being capable of supporting a small garrison force or some such thing on nearby worlds, though it would need to be done in such a way that the garrison actually takes time to be replaced rather than coming back for free every time a battle occurs (or, for that matter, each turn, unless it's a really cheap garrison force). However, I would not be okay with these garrisons being fleet replacements to any significant degree. Prevent the loss of a planet to an attack by an unarmed and unescorted transport or group thereof? Sure. Prevent the loss of a planet to any real commitment of force, though? No.

I agree, I wouldn't want it broken in the other direction either. But I believe that if the effort of constructing a military starbase was made, then the opponent should be forced to spend some effort dealing with it. I would like to see a system where the base must be destroyed before the planet can be taken and I fully expect for this to work that a military starbase would offer no bonus to logistics so that you could not have any additional ships guarding it than what you could normally have. But at least this way the enemy fleet must face your fleet instead of warping around it by forcing the fight in a certain area. The goal isn't to utterly destroy the enemy fleet, but at least cause it to lose ships via attrition so that the more planets guarded by starbases they invade, the weaker they become.