Technically, none of the default map sizes are a couple thousand hexes across. Even on Insane, the maximum straight-line distance is only 761 tiles.

Why is there this claim that sensor stacking removes exploration from the game? Sensor stacking at its current levels does not remove exploration from the game except on map sizes where exploration is essentially nonexistent anyways. Even on Tiny maps, your basic turn-1 maximum-sensors ship still needs ~84 actions to fully explore the map using a perfect exploration pattern that wastes none of the ship's tiles revealed per turn, and this on a ship design that has a base of 1 action per turn. I'm sorry, but if the exploration phase isn't essentially over ~40 turns into the game on a map where the maximum straight-line path length is only 61 tiles, then I think there's something wrong or you invested far too little into exploration. There's a case to be made that Interstellar Sensors are too much of an improvement so early on (a sensors-only cargo ship with no capacity bonuses or non-hull non-component sensor range bonuses goes from 13 to 22 sensor range by switching from Navigational to Interstellar Sensors), but this is more of an issue with component progression than component stacking, and that Interstellar Sensors are available early on is mostly the fault of the value of the +1 movement at no cost other than research offered by Interstellar Travel.

I dislike both of these suggestions. Your suggestions remove most of the costs of getting high sensor range and high numbers of actions per turn. Sure, it puts easily-controlled limits on these, and yes, if you put a cap on the number of one type of component you can have you do sort of need to have a cap on the number of the other type of component, but now there's little real tradeoff between actions per turn and ship capabilities. You don't have specialized exploration vessels (which have a balance of speed, range, and sensor range), you don't have specialized surveillance ships (which, depending on objectives, might mostly focus on sensor range at the expense of most other things), you don't have much of a tradeoff between lots of actions per turn and ship durability/firepower.

Actually, people like games to be fun. "Balance" is only a concern when the issue is significant enough to prevent the game from being fun.

Beyond that, this

is a question of game pacing, not game balance.

This

is a balance argument, though it's also one whose validity is questionable. Assuming no capacity bonuses, a turn-1 max-sensors cargo ship will have 1 base move and a base sensor range of 13, and costs 185 manufacturing. A turn-1 max-sensors tiny ship under the same assumption will have 1 base move and a base sensor range of  5 while costing 50 manufacturing. The max-sensors cargo ship will be able to explore up to 27 tiles per action; the three max-sensors tiny ships you can build for about the same cost will be exploring up to 11 tiles per action each, and moreover can explore in several directions whereas the cargo ship can only explore in a single direction. The initial sensors-only cargo ship is not so superior to the initial sensors-only tiny hull as to be a no-brainer choice for exploration. Nor, for that matter, is the sensors-only cargo hull so superior to a sensors-and-engines cargo hull as to be a no-brainer choice there. Switching to Interstellar Sensors allows the cargo ship, under the same assumptions given earlier, to explore up to 45 tiles per action, while the 4 full-sensors tiny ships you can build for the same manufacturing cost and using Interstellar Sensors each explore 13 tiles per action. The only place where the cargo hull really has an advantage for exploration is the degree to which you can alter the balance between sensors, engines, and life support components to get the performance you want.

If you sacrifice sensor components to fit a hyperdrive, then you're getting a a base of up to 7 tiles explored per action and 2 actions per turn on each of 3 tiny ships or a base of 25 tiles per action and two actions per turn on 1 cargo ship using Navigational Sensors, a Hyperdrive, and the base hull capacities and component requirements.

3 tiny ships can explore in three directions at once whereas 1 cargo ship can only explore in 1 direction at once. 3 tiny ships cost roughly the same amount of manufacturing as 1 cargo ship. 3 tiny ships with 1 hyperdrive and 1 sensor explore tiles at roughly the same rate as 1 cargo ship with 1 hyperdrive and the rest filled with sensors. 3 tiny ships filled completely with sensors explore at roughly the same rate as 1 cargo ship completely filled with sensors. As far as picking exploration ships goes, this does not seem to be the completely trivial decision that you've made it out to be. It may well be the case that a cargo ship with a balance of hyperdrives and sensors will be greatly superior to the equivalent cost in tiny ships for exploration, but I would not consider a cargo ship with a balance of hyperdrives and sensors to be a 'sensor barge;' that designation implies to me a vessel which has sacrificed most or all of its other capabilities to enhance its sensors.

As the guy above said, it's about pacing.

That is actually the desirable outcome to me - start of with weak sensors due to tech limitation, so you have to scout and explore and weigh up the value of researching sensors through research as opposed to all your other options. As it stands, I have *never* researched past 3-range sensors. Why bother? Ship capacity upgrades pretty quickly already, that'll do for more range...

I think it's important to remember that sensors don't sit in a vacuum; sensor range has a direct relationship with the maximum number of moves that a fleet can acquire. 

In a turn based system (TBS) game you start to run into serious mechanical trouble when units are able to move out of the FOW and then have pretty much free choice of targets to hit thereafter, particularly when the game doesn't include mechanics that allow you to cut enemy moves short mid-turn (by means of Overwatch, for example). In games where such things are allowed -- and GC3 is one such game -- you create a paradise for the attacker because the defender is utterly unable to station enough defense at all targets within range of an enemy fleet that has dozens, or even hundreds of moves.  To be blunt: in the mid-late game the TBS in GC3 pretty much falls apart.  (Play against a Human at this stage of the game to see how broken things are; it's nigh-on impossible to defend against as the attacker needs defense at least broadly inline with the attacking fleet power multiplied by the number of targets that you as defender don't want to lose next turn.  And I use the word 'broken' deliberately because the TBS in GC3 isn't equipped to deal with such things.  In single player the problem hides itself for a while longer, but it's still there; mostly because the AI is unable to exploit giant move ranges and because the AI gets confused when it's presented with vast amounts of information.)

Whilst it's true that you don't need to have a single sensor source that's capable of seeing far enough into the maximum likely moves range of any given unit (you can have a network of shorter ranged sensors that do the same job) it is important to recognise that changes to sensor ranges should also provoke thoughts about maximum movement range reductions.  Personally speaking, even if the sensors aren't changed (which it's looking like they will be) the maximum number of moves that a fleet is able to acquire needs to be significantly reduced; that, and/or the game needs some form of reliable movement interception mechanism.

Paul Stated he did not want to change sensor mass as a solution, as this would break many ships (IMO, who cares, how many sensor boats have people designed which take more than 5min to rebuild?)....but he does care, so instead of changing mass, he is looking into a mechanic where the more sensors you add, the less effect each sensor has.  Therefore, a nerf without a mass change that would break ships...unless you have the sensor range in the name of the ship

I would far sooner see them implement a change which breaks designs by rendering them invalid than a change which breaks ships in ways that aren't immediately apparent, and adding a stacking penalty of the form -X% sensor range per sensor is very much a change which carries a high risk of breaking ships in ways that aren't immediately apparent, even for low values of X. If X = 5, standard rounding is used, no other percentage sensor range bonuses are in play, and the base sensor range and the flat sensor range bonuses of sensor components are unchanged, then a ship achieves its maximum sensor range with only 6 navigational sensors. Even though designs with more than 6 navigational sensors are not invalid, I would still consider them to be broken as they would then be paying for components which do nothing for the ship; if you've managed to fit 13 or more navigational sensors onto the ship (which is possible for a turn-0 design using a faction with two levels of the hull capacity trait), those extra components will actually degrade the ship's performance. These designs will have been broken whether or not they were rendered invalid, but they will have been broken in a way which hides the issue from the player unless he or she bothers to check. This is not a good thing; far better to simply force the player to redesign ships because they're no longer valid than to conceal the fact that the changes broke the designs by leaving the design still technically valid.

Beyond that, I feel that this type of stacking penalty is generally a bad type of penalty and is confusing for the player. Generally speaking, you expect that adding more of a component will either help the ship's performance in a given role or at least not hurt its performance. This type of stacking penalty, however, has a number of components beyond which additional components are actively harmful rather than merely redundant. Even a flat -X% sensor range (rather than -X% sensor range per sensor) is somewhat superior; this too will run into the issue where adding additional components is actively harmful rather than beneficial, but the point at which this occurs will generally be much higher (though it depends on the values of X chosen in each case). The -X% sensor range per sensor model also makes it difficult to determine what exactly the benefits of getting +Y% sensor range from technologies happens to be if, as I expect, the technological bonus and the per-component bonus are additive. You would not, for example, expect that a 10% bonus from technology would increase the maximum sensor range using +3/-5% sensor range per component sensors from 16 to 19 without running the numbers, and yet this is the effect that a 10% sensor range bonus from technology would have on such components when the hull provides a base sensor range of 2.

this will effect every ship built with sensors from the stock surveyor to the armed ships to the sensor boats!! so yeah if and when it happens it will brake alot of ships!! where the FUN in that ???