I don't think the three weapon types should initially have different ranges.

Different weapon ranges is, in fact, historically accurate ;)  It's just that B4.2 damage rates must all be tweaked to achieve a playable game balance.  Using energy-per-damage as the unifying concept:

• long range + fast transit (beams) ==> low damage rate
• short range + slow transit (kinetic mass) ==> high damage rate?
• This postulates that the most efficient energy delivery is kinetic.  That's true in 21st-century Earth.
• medium range + fast(er) transit + warheads (missiles) ==> ... inscrutable.
• Generally, a missile delivers less total impact than an equivalent mass of dumb bullets at equivalent velocity.  Each missile (including naval torpedoes) is also about 3-5 orders of magnitude larger than a dumb bullet/shell ==> slower fire rate and reload, much higher cost per shot.  On the plus side, missiles can track/home, and deliver a warhead (which we can abstract away as a WarZip compression of energy for shipping).  If the warhead is big enough (nuclear or antimatter), then the energy of the missile itself is insignificant.  Tracking overcomes the problem that your opponent will dodge/hide from a dumb bullet, and so one precious guided missile might really have a more efficient hit-rate than 4,000 blind flak shells with altimeter fuses, and actually work out to be cheaper.

Maybe what we want is a kind of Kepler's Second (Gaming) Law: all weapons of same tech level sweep out equal total damage per (area * time) (ergo, longer range = much larger area = much lower rate).  Currently, beam weapons have long range and equal damage rates (their full beam rating per shot), which is just a Beta 4 ba-roken-ness.  Fear not; I'm sure Beta 5 will evolve Combat Viewer (and all combat numbers) by another huge leap.

Flip this around: Consider a fixed set of weapons with different ranges, damage rates, and (um) reload penalties.  Fight wars.  Every nation will converge toward a mix of weaponry that optimizes winning.  This drives a great deal of weapons development.  Such leapfrogging should plausibly continue into the future, including GC3's tech tree over the span of a long game.

• Greeks: phalanx (4 tight ranks of very long spears).  Loses to: anything, in bad terrain.
• Romans: Javelin (one-shot, ~15m) + short sword.  Loses to: bad leadership (and lead in the water?)
• Mongol: horse archers w/short composite bows.  Loses to: they won everything and got bored.
• Knights: lances with stirrups, then sword + shield.  Loses to: longbows.
• Prussia: flintlocks + 2 ranks of pikes, then sabers.  For rate of fire, they had 4 ranks of muskets, where they pass fresh muskets forward, and pass the empties back to be reloaded.
• WW1: bolt-action semiautomatic rifles, no armor.  In close quarters, pistols and bayonets.  Loses to: machine guns, artillery.
• WW2: fully semiautomatic rifles, 1-man submachine guns, 2-man machine guns.  Loses to: hand grenades.
• late-20th: fully automatic rifles.  We still use bayonets.
• early-21st: sniper rifles.  The sniper rules the non-traditional battlefield.
• mid-21st: laser flak, railguns, hypersonic drones, ...

The overall lesson seems to be: you can always pay for longer range/higher rate than average, but by definition, they are not your average-case weapons.  Hence they come with restrictions: finite usage (no more ammo), slower rate, higher cost.  So preparation-for-combat becomes a grand optimization problem, where you buy the best bang for your (inefficiently allocated) buck.

Projecting to GC3, it could be argued that beams are an exceptional case, and should not be Simply Superior to non-beam weaponry.  An all-beams fleet/player could be analogous to an all-longbow expeditionary force with no close-in fighting capability (and the all-kinetic player would be the mounted knights with no force-projection beyond the tips of their lances).  Building trade-offs into the numbers enables interesting in-game decisions.  I'm confident that Stardock already has this on their list.

I think there needs to be variety in the starting weapons. If you make the ranges all the same but keep the damage different, then missiles will be unbalancing. If you make all the ranges and damages the same, then there is no reason to have 3 different weapons. Beams need to be balanced, but not by taking away what makes them distinct.

Kinetic damage should be increased dramatically and early beams should be nerfed. Early battles should be ruled by kinetic and missiles.

I think Gilmoy nailed the balance here. Beams 'should' be weaker in damage but make up for it by having longer range thus getting in more hits. Missiles will be the base line for the other good range, good damage, offset by either rate of fire or accuracy. Mass drivers being the 'shorter' (but not by much) in ranges but doing (slightly) more  damage than the other two, again also offset by either accuracy or rate of fire. Both of those variables can raise up or lower the 'effective' damage of said weapon.

In the end, each weapon needs to be great with its end game effectiveness to be desirable, no one type should be 'best' all the time. Having an all beam fleet is ok but SHOULD be handily defeated by a fleet with mixed weapon types. 

On that note support modules MUST be lower mass to be effective. Mass driver ships MUST have thrusters to boost quickly into a battle and not at the expense of removing weapons or baseline defenses. 

 Edit to add.

What is hard is the balance. All types should allow a player to succeed if he wants to go that route. If you want an all beam fleet fine, The trick here is not the weapons but the defenses. As I mentioned before, BUFF the defenses A LOT and make the strategic resources (invaluable) to each defense type. This would do two things:

1) Early weapons would be awesome fast fights but as you progress and learn to over come Beams with Elerium based shield technology, then fights drag out and the need to retool for other weapon types or 'mixed' loads would be needed. 

2) The player can plan his strategy and choose to rush those now valuable resources to be used in command ships against some types of weapons. Got Elerium? Great Now you can design mass driver ships that can finally get in close and shred those beam fleets because now you have invulnerability shields. Lets say adding a resource will at 100 to 200% effectiveness to that defense, I would RUN to get said resources. 

My strong impression of the game as it now, pre-balancing, is that beams are ridiculous.   Facing an AI with beam/shield ships, I built missile and mass driver boats that had strong shields.    Went into battle and got wiped out.   Ships blown apart before they could fire.   Did far better just building beam/shield ships myself, which doesn't make any sense (by the way I might add that free upgrades to ships allows one to experiment with this stuff, so very useful at this stage).    But I expect that to be addressed.  

The general (rough) principle should be that longer range stuff should have a weaker impact, as others have stated.

In the GC2 community update they are experimenting with weapon cycling (effectiveness of weapon types cycles slowly over time).   It's a very interesting idea, don't know if it is really workable though.   I think I would prefer weapon cycling that happens naturally as a result of tech development, such as Larsenex is suggesting.

I think something that might be interesting is that if a ew tat working like specialized mass for reactor output were added. Right now, there really isn't any Eason to diversify much.. But if beams had high energy requirements  nd range but mid range damage, missiles low energy /middling range/slower reload, ballistic short range, good damage , and minimal nergh requirements... It would mean that you would want a mix with the mix serving different needs

I dont know if a secondary mass type Sta can be nodded in with the XML files and  ill try to check tomorrow if its something that looks doable, but it could be very interesting

It seems good balancing will be  the key here.

Beams and carriers will be nerved.

Missiles and Kinetic buffe, pretty sure about that. 

Only thing that makes sense ! 

All weapon ranges set to the same level it makes differences senseless...

I just mean initially.  Then techs could be researched to increase range.

Since balancing is yet to occur, here's my take on weapons. I evaluate purely
intuitively on several parameters. These are all very rough assessments that will
change based on tech level and so forth.

1. Tech difficulty. (How expensive/difficult (in terms of placement in the tech
tree):
a. Mass Driver. To me a mass driver is like artillery. "Mass Drivers"
were used in the Stone Age - pick up a rock and throw it. In ancient times
peltasts (slingers) were important to battles. Rock throwers. Point, this tech
should be super cheap. Of course, fire control, such as the Bismarck sinking the
Hood from about 16 miles away, granted, is a bit more trouble.
b. Missiles. Quite a bit harder to tech. Guidance, munitions, fuel,
etc.
c. Beam weapons (Lasers et al). The US Navy is just now testing laser
weapons for use on ships. How they will pan out is unknown. This is
sophisticated physics, which is much harder than "rocket science" i.e. missile
science.

2. Space requirements: Roughly the same. Beams might actually win (slightly)
here because they don't require ammunition storage. Of course I don't see any
attention to ammunition requirements in GalCiv3 and don't expect them, but
whatever. One can imagine they are folded into mass requirements.

3. Cost (in BC). Roughly in the same order as tech difficulty, although because
of the ammunition thing, missiles will be closer in cost to beams than to mass
drivers. But mass drivers should be cheap while the others are quite expensive to
build and maintain.

4. Accuracy.
a. Mass Drivers "total accuracy" is poor and falls off badly with
distance. Obviously, it isn't hard to shoot in a straight line, but I say "total
accuracy" because mass objects will be easier to evade (slowest velocity). At
very short range, however, mass drivers could be devastating.
b. Missiles. Guidance systems mean missiles should be very accurate,
(i.e. cruise missiles) but this ignores defense. Chaff and point defenses should
be quite effective, while armor vs. mass drivers less so because armor doesn't
prevent a hit. Sort of like the silly asteroid disaster movies. If you blow up a
big asteroid, what have you accomplished? The same mass is still headed for a
collision with Earth. Ouch.
c. Beams. Because they propagate at the speed of light, accuracy will be
very high at reasonable ranges. Evasion will be much less effective. However,
sophisticated shields (read tech and BC very expensive) should be useful.

5. Range
a. Here we assess "effective range". Clearly a bullet in space could go
forever until it experience a gravitational field. But so what. Effective range
is very short.
b. Beams dissipate with distance, but have much longer range than mass
drivers.
c. Missiles (equipped like cruise missiles with internal guidance and with
sophisticated fuel usage) could be exceptionally long range. Just might be a
while before impact.

6. Defenses
a. Armor is cheap and maintenance is very low. Useful every time you take
a hit.
b. Chaff, Point Defenses are expensive to use and maintain and are not
effective on every attack like armor.
c. Shields are very tech and BC expensive. I see them as effective on each
attack that lands, but their effectiveness should degrade if they are hit.

7. Damage
Mass Driver < Beams < Missiles (missiles can carry nukes, e.g.)

Summary of how this might play out in a battle.

Early ships use mass drivers, then later in the tech tree more expensive
missile tech becomes available, and finally, last of all, beam weapons.

In an actual battle with all three types available, missiles fire first at
long range,but approach the targets relatively slowly. Beams are next to engage
at medium range, but hit almost immediately. At very low range (if that is
achieved), mass drivers engage.

Well, that's my 2c. It is definitely not rock-paper-scissors. In fact, the
result will be that the best weapons will be mass drivers then missiles the beams,
although both missiles and mass drivers will have effectiveness in certain
situations. For example, with small, very fast ships (tactical speed) mass
drivers might be effective even at very late stages of tech development. Evasive
technology on missiles might make them harder to intercept, etc.

Caveat:

All this is just thinking on paper as we are dealing with a game that needs to be
fun, not a realistic similation. Please trash this as you see fit, fellow
testers.

Kentucky Fried 4X does tactical combat!

• Well ... Mounted!
• Long ... Range!
• Enormous ... Warhead!

And if you get past those, then you can invade.

More seriously, today's prototype next-gen beam (laser) and kinetic (electromagnetic railgun) weapons still run into an energy wall.  To deliver a certain number of MJ (megajoules) to a target, you must generate that energy somehow, which means you still must carry around at least that energy / c^2 in frozen form.  Eliminating low-density chemical propellants and lumpy bullets is not a simple panacea.

• Modern lasers typically use massive quantities of two chemicals reacting.  One argument vs. Star Wars anti-ICBM lasers in space was that each one would need tons of chemicals as the fuel for their warshots.  In fact, this made it prohibitively expensive, and such lasers were never flown.  We probably still couldn't build them even today.
• The Navy's new prototype ship laser vs. drones uses (I think) the ship's normal electricity supply (which is ultimately generated by running the ship's gas turbine).  Ergo, it's so low-power that a drone is about the limit of what it can affect, and it still requires several seconds of uninterrupted time-on-target.
• The Navy's prototype railgun uses electromagnets to fire a goofy-shaped rod at ridiculous speeds, Mach ~15 (which they think will enable 100+ mile direct fire ranges).  Each shot consumes so much energy that the ship must recharge the capacitors for minutes per shot.

I'm sure humans will just learn to adapt new tactics to the new capabilities.  Nonetheless, the ship doesn't get (much) lighter just because you took away its magazine of shells and powder.  The mass requirement just changes shape, and becomes maybe huge coils, capacitors, engines to pump them, and ultimately a fuel (which is simply frozen energy).

Consider now that all of this thinking (including all of GC3's tech!) is, perhaps, stuck at the bottom of a bowl-shaped space of possibilities.  We're taking for granted a large assumption:

• Weapons destroy ships by delivering energy.  Energy performs work, which rearranges ship atomic structures into more (um) diffuse structures ... like vapor.

Beyond that, what alternatives could there be?  How could you discombobulate a ship('s atomic structure) without delivering energy to brute-force move atoms around?  We can sketch out some wack-o ideas ...

• Borrow from the zero-point energy field.  By Heisenberg, any amount of energy E can simply pop into existence from nothing, as long as it vanishes within t = h-bar/E.  Hence vacuum seethes and teems with particle/anti-particle pairs that happily flit in and out, self-annihilating before nature notices them.  N.B. this underlies Hawking radiation, where the pair straddles the event horizon of a black hole, and one escapes (and becomes a real boy!!) but the other one doesn't.  Somehow do something that requires vast energy, but you borrow it somehow, and then ... the Laws of the Universe come looking for it like a debt collector, and repo half the enemy ship for you.
• Create a field that ... induces a gravitational slope.  Then atoms just slide down the slope.  (Stephen R. Donaldson's Gap singularity grenades; Star Trek's red matter creating a black hole; Star Wars Interdictor star destroyers)
• Fold space to connect two points, thereby connecting his interior atmosphere to a vacuum.
• Project a field that -- changes a universal constant, e.g. the strength of the strong nuclear force, or the stickiness of the Higgs boson (which would change mass in weird ways).  (Orson Scott Card's "little doctor" (mass dispersion, or MD))  You deliver no energy at all; the atoms (or quarks) provide their own "energy" to seek out the new stable configurations.
• Gently deliver a quantity of tranquil, slow-moving magnetic monopoles (each the mass of a proton, with a negative charge).  These fall inward through electron shells and steal the nucleus away, jilting the electrons, and producing monopole-atoms of radius much smaller than the electron's shell.  The matter shrinks by several orders of magnitude, and an equal number of electrons wander off, no longer electrically bound.  You don't need to add energy, since the mere existence of such a monopole + normal physics is already sufficiently disruptive (to normal matter).  (Then we learn, to our chagrin, that the Zor are composed entirely of monopole-matter ... so all of the ships simply shrink by 5 orders of magnitude and defect to the Zor, and they're still fighting you with weapons just as massive as before :))
  
• This generalizes to other kinds of bizarre elementary particles (strings, etc).  Postulate that the Extended Standard Model of Physics has a new tier of configurations-of-constituents, which greatly prefers QQ-combos, only we never see them because there are no QQs.  Then have your weapon deliver some QQs, and stir.

This is far beyond most SF authors, and GC*.  Interestingly, a tiny niche of SF authors have dared to consider vast wars on time/space scales that beggar human thought, and some of these ideas do pop up there.

• Stephen Baxter's Xeelee Sequence is his life's-millieu, recounting a war between humanity and the Xeelee, a billions-years-old Type IV civ Battles span galaxy clusters and tens of thousands of years, i.e. entire human civilizations can run for billions of years and then get snuffed out by a single galaxy-heating weapon (whose one explosion takes the tens of thousands of years).  The Xeelee have time travel and edit their own evolutionary history, which apparently allows them to extract their grumpiness from their own DNA and engrave it into the physical laws of the universe.  Even weirder: the Xeelee are getting their own butts kicked by an even grumpier and older race -- and eventually they lose.  (But then, humans get time travel, too, and the series gets ... just bleak and weird.)  That's some of the weirdest SF ever.  (Does it even make grammatical sense to say "and then we got time travel"?  The concept fits poorly into English ...)
• For a 12-page short story, he postulates a listening outpost of humans, held in a ball suspended over a star, 60 thousand years old, with a stable breeding population of humans -- 1/3 the evolutionary history of modern humans, disjoint and alone, doing nothing but breed, recycle (themselves), listen, and wait.  At the end of this short story, the ball is sacrificed, dropped into the star, finis.  It's no big loss (so it is implied in 1/2 a sentence of dialog), there are millions of other listening posts throughout the galaxy, they will continue the evolutionary history of humans elsewhere.  Sheesh ... it is very far away from being heroic.

I think that would be the 5X of Galactic Supercluster Empires, where you start with a Galaxy and some Ships, and you explore the nearby strand of galaxies.  It would have to be a 5X -- no way could you micro that ...

I am a fan of mass drivers/rail guns and I think that at the highest research point they should be the strongest weapon available but the research should be expensive.

A ballistic shell hurled at near light speed would have range, pretty good accuracy, and it would be impervious to almost any armor 

Thinking further on it, I suppose at the highest point they should all be nearly invincible, doom rays, anti-matter missiles etc. I still see kinetic as the more expensive energy intense research and missile being the least expensive

I also thpught you could hurl a bullet,  or missile at near light speeds then galactic civilizations two called a particle beam a beam weapon. Your idea is iconian technology. I never realised that lasers and ships could have different fuel requirements for better effectiveness. Definately the game coulf bring in s nrw mechanic power supply. 

Bullets and Missles when shot in space for all intents and purposes have near unlimited range.  There is no air to cause friction which slows them down.   In the current battles missles are shown as going round and round which does not make sense.  They should go in straight line in the absence of a gravity field which would alter their trajectory.  When using a guidance system (for missles), the trajectory would get ajusted with side thrusters to adjust the trajectory.

The difficulty when using bullets and missles over large distance is not range but calculating where your target will be when the bullet / missle gets within vincinity.  That is in the absence of a guidance system.   A kinetic firing system would need to spray the most likely area that ship may find itself by the time the bullets get there.

Also, firing kinetics and missles should also push the ship back from the recoil if it is not equiped with thrusters to counter the force of firing.  

Missiles:

One key aspect of missiles overlooked in GCIII is missle inventory. Small ships never run out of missiles.

In real life, Large ships can have more space for more inventory, Supply ships can restock combat ships.

1. Range Techs boost missile range
2. Missile Engine Techs - boost the speed
3. Miniaturization - Reduces its size, allowing ships to stock more missiles
4. Ship Design Techs - boosts shipboard inventory - think external or internal missile pods
5. Missile guidance systems

Counter Missiles include:

1. Counter missiles
2. ECM
3. Close in defense lasers to stop missiles that get by the counter missiles

Beams

Two weaknesses of beams are power requirements (more powerful beams require more power, may have a longer recharge) and shield defenses. Unlike missile defense systems, shields are 100% effective until breached. Shields require power. Larger ships - battleships, dreadnaughts - can be equipped with more powerful shields and the most powerful beam weapons. Small ships with their small power plants have comparatively weaker shields and weaker beam weapons.

For Beams, they could be stronger if a battle is taking place near the sun. (Solar Panels).

Alternatively, bullets and Missles could make use of a gravity to extend their range if near planets.  Somewhat like they did in Star Control melee fights.

A fleet could be required to re-visit a star base to re-charge // restock after a fight.   Unless freighters are used to re-supply fleets from planets or star bases.

Wait, isn't a huge part of the plot of Galciv based on humans discovering feasible fusion power that can fit on a ship ? Thus making solar panels a really dumb choice and practically solving all of the energy requirements?

With that out of the way..

Here is a railgun

 https://www.youtube.com/watch?v=wa_vuX5_oAk

This is a mass driver weapon. I actually see no benefit of pure missiles(in combat) to railguns except that they are cheaper and more compact. In a figter drone where available space is limited, missiles might make sense since you can have more missiles and therefore more potantial damage than if you used that space for the gun + ammo + energy source.

For any larger ship where space is more available, i think that railgun-launched missile would be superior to pure missiles in every concievable way...  (or light sail assisted missile, assisted by ships lasers, same idea, basically why not launch a missile from a mass driver weapon, you get more damage, speed AND control...)

All in all, i think that missiles should be low cost low volume low tech weapons intended for small ships while 'big boys' fight with either hybrid systems or ditch missiles altogether.

I apologize for not contributing to the actual discussion about game balance, but I had to rant about pure missiles

Having to go all the way back to a starbase/planet after EVERY. SINGLE. FIGHT.  would be Unfun.  It would introduce even more micromanagement hell and encourage the worst sort of constructor spam (plopping starbases all over the place simply to facilitate battles [which would, in turn, cause people to call for starbase restrictions {which would in turn, well you get the idea}]).

Maybe games with smaller scope/different mechanics can get away with it.  But one where there will be literally hundreds of ships in the late game?  No way.  This sort of thing needs to be thought of and implemented at the very beginning of the game design so it can be taken into account when other game play factors are decided.

If you need an immersion explanation, consider restocking weapons part of the range restriction.  One can already upgrade ships in the middle of nowhere, so it seems it is implicit that range includes not-seen supply ships and what not.  Call them commercial contractors if one must.

So just presume that these supply ships are also restocking weapon caches.  Simples.

I am thinking the whole thing doesn't make much sense.

Lasers are a beam. The light disperses with distance. So beams should be more powerful the CLOSER the target is. The rate of fire should be slowed or tied to engine capacity for reloading. They will be accurate though.

Kinetics should be the least accurate, but most powerful, agreed. However, this is space, there is no real drag as there is no air resistance. To me damage should be consistent (or closely anyway) at any range.

Missiles also deliver a steady damage, less than perhaps both, but missiles take time to gather speed. In space, I would imagine this to be very slow acceleration unless it was launched from a tube. Launching them from under the wing of a fighter would only have the speed of the fighter and a slow acceleration because there isn't much for the rocket to react against. I would imagine a missile in space would be fairly easy to dodge as it will be much harder to turn in space vs an atmosphere. The Tactical speed of the target should really come into play and the missile itself should have a rating for maneuver. A large missile will have a slow maneuver rating but do more damage. A small missile has higher maneuver and less damage. Also missile range, really missiles are guided/unguided mini ships unto themselves. 

So in short

Lasers - accurate, damage drops over distance, slower reload Really there should be large and small lasers. Large lasers would be great against capital ships, much less effective against smaller targets because of slow reload. Small lasers would have less damage, but faster reloads. 

Kinetic - least accurate at medium/long range, but pretty accurate at close range (could be buffed with targeting computers) damage consistent. Have large rail guns for bigger targets and smaller guns for smaller ships and anti-missile.

Missiles - can lock on target at medium to long range, but easier to dodge at a distance. Missiles at short range, so many variables... I feel they would mostly be used as capital ship killers.

Suggestion - large/medium/small of all three types. Large weapons not good for small targets, medium are kind of jack of all trades, small poor vs large targets and good against small.

Side Note:

Although I can't help but think light beams, which can go one until absorbed or dispersed. Kinetics with little to slow them down until they hit something would be horrible IRL because this would great LOTS of collateral damage. Missiles could be programmed to explode at a distance or run out of fuel. However, this is a game, so no worries.

That said, if I was attacking a planet, I would fly my ships between the planet and the enemy ships. Every enemy shot of kinetic and laser would hit their own planet! Especially kinetic with gravity pulling the ordinance at extreme speed. The laser would likely disperse unless it had little or no atmosphere.

Actually missiles in space would perform almost exactly the opposite of this.  They do not need an atmosphere to push against to gain speed.  Their acceleration is from the combustion reaction pushing against the missile itself, that physics thing about for every reaction there is an equal and opposite reaction.  The atmosphere (and gravity) environment of our planet actually slows missiles down a lot.  In space a missile launched from under a wing would accelerate incredibly fast, and theoretically keep accelerating until it approached the speed of light (or it ran out of fuel ).  You are kind of right about the turning thing though.  It would be incredibly easy for the nose of the missile to turn very very fast with mini-thrusters in the nose, but the momentum of the missile would still be pushing it in the direction it was initially going since there is no atmosphere pushing against the missile to help change the direction of it's momentum.  Steering a missile in space would actually require much different methods than we use to steer them here on Earth.

Sorry I was bored eating cake for breakfast, and just felt a need to clear up some physics misconceptions

As for the light dispersing over a distance, it has to be quite a distance after all we are seeing light from stars and this light took several million years to get here.   So a close combat with ships laser dissipation should not be a factor.  There is no atmosphere do dissipate the light.

Actually, until it approaches the exhaust speed of its propellant.  A rocket can't go any faster than that (because then its exhaust particles would end up with net forward drift, which would require some exhaust-exhaust for the exhaust to have pushed against).

• Hence we have a 21st-century problem using chemical rockets, all of which max out around 3-5 km/s (of the 11 km/s you need to escape Earth orbit).  The rocket equation overcomes this by exploiting mass fraction, where you split your rocket's mass into stages, and drop off empty stages to shed 50-90%+ of your mass, which changes the equation for whatever's left.
• The cost of a rocket launch just presumes that you throw away / let burn up a 747 or two per flight.  Hence each flight is tens to hundreds of M$, which is right around the cost of a jumbo jet that could have served you for 25-40 years.
• We know it's a waste.  Shuttle re-used its solid-rocket boosters that were fished out of the drink after every flight, but the main tank falls from too high up, and just burns up to vapor.  Automated recovery of boosters is an old idea; SpaceX is pretty close to making it commercially viable.  Just give the boosters some brains, gyros, and a fuel reserve, have them fall back upright, and levitate themselves to a soft landing.  In 20 years, it'll be as ho-hum-old-hat as watching a 747 land.  It follows that these soft-landing boosters must weigh a bit more than a disposable booster (and/or they jettison a bit earlier, before running dry), so they'll contribute less oomph to the launch.  But if they last 20 years each, your total launch costs would drop to just about the cost of the fuel you burn, and then orbit might cost as little as air-freighting an equal mass from JFK to New Zealand.
• The mythical single-stage-to-orbit (SSO) tries to achieve orbit (and back, if it's a space plane) without mass fraction.  You can do that (just barely) by having the 95% of your mass be fuel, which you burn up.  Such designs are known (for decades), but they have payload capability far below modern commercial requirements to be viable.  Also, it's structurally flaky to imagine any hull or shape that takes off full, lands while nigh-empty, and is good in both modes.  It'd be as fragile as an eggshell on the way up, and as loose as a kite on the way back.
• The ion drive, which the Dawn mission has used very successfully to hunt down the big asteroids Vesta and Ceres (yaay!), is a game-changing rocket tech: it accelerates xenon(?) ions through an electromagnetic field, which gives utterly ridiculous exhaust speeds (of 30+ km/s? -- way beyond solar, and possibly galactic, escape velocity), for the drawback of carrying only a few kg total of reaction mass, for a relatively tiny total delta-V it can "buy" over its lifetime.  That trade-off actually worked out well: it can make tiny but significant orbit-sculpting maneuvers for years, whereas an equal amount of gasoline would have lasted maybe 30 seconds and achieved nil.  That's how Dawn can scuplt its orbit down to a close circle around Vesta, and still have the power to boost itself to escape again, and then chase down Ceres.  And after Ceres, they'll extend its mission and keep doing that, until humans catch up to it in 2050 and we retrofit it in situ.  (Then we shall do the same for Voyagers 1 and 2.)
• A Bussard ramjet could have exhaust speed approaching c.  If it accelerates at 1g for 1 year, it would be moving at nearly c itself -- surely a sign that humans are destined to colonize the galaxy!!  Niven wrote a charming/unsettling short story about two Bussard ramjets chasing each on autopilot for 30,000 years, and thereafter forever.  (He later admitted that he started with the idea of the chase, and then hacked the story to fabricate a plausible rationale for it.)

Starlight is harmless, though.  Laser light is collimated (photons of identical wavelength and phase), which is what makes it powerful.  Quantum effects cause collimated photons to be "jealous": they tend to disperse, and thus decollimate.  After enough decollimation, even a laser is as weak and ignorable as starlight.  So light as a weapon does has a finite range.

• We're pretty sure that laser light pressure can remain significant over light-years of distance, e.g. batteries of launch lasers to drive solar sails.  It's no longer a weapon, but if you cooperate by surfing the beam, it can give you milli-g's of acceleration.  Niven/Pournelle's Moties use that to colonize nearby stars.
• Robert L. Forward used it to send humans to Barnard's Star, and wrote a detailed technical description of the entire mechanism (which is his forte; he makes up for it by being a bit weak at human conversations .  The sail is in the form of a huge parachute, ~3,000 km in diameter (not kidding).  The laser beams must bounce through a massive Fresnel lens in orbit around Jupiter (and later, they chain it through a 2nd Fresnel in orbit around Saturn).  For 1/2 the flight, it simply pushes the entire sail.  At half-way, the sail detaches into two parts (1) an annular ring, with autopilot, and (2) a smaller inner sail, still attached to the crew pod.  The pod + inner sail now turn around and point away from Earth (toward Barnard's Star).  The launch beam now does a majestic Z-bank in space, from Mercury through Jupiter/Saturn, off the annular sail, back to the pod sail.  This accelerates the annular ring like crazy, but the bounceback slows down the pod, yeah baby.  "All" you need is enough raw lasers to generate the power, and sufficient aiming precision to hit a moving lightsail smaller than the Moon at 20+ light years distance, with no FTL feedback loop.
• Finally, the ring zooms off into interstellar space, and the pod + sail remnant drops tranquilly into Barnard's system, where it lives out its crews' lifetimes as a fully-functional solar sailboat, tacking and puffing its way around the planets.  (Forward isn't maudlin: of the crew of ~20, exactly one survives, and he takes wicked glee in devising spectacular and sad endings for everybody else.  They knew it was coming, they all volunteered.)  Punch line: Humans do catch up to him with FTL jumpships, and give him a medal

Oh right, thanks for correcting that .  I was thinking of a theoretical nuclear drive I heard about, using particles from fission and/or fusion reactions that were travelling near the speed of light.  Something similar to the ion drive.  Obviously that is something much different from standard chemical reaction missiles.  It hadn't even occurred to me the exhaust speed of the propellant would be the limiting factor.

NO NO NO NO NO NO.

It makes absolutely no difference if you have a rocket engine with an exhaust speed of of 4 km/s or  a monkey throwing balls at the same speed.

Taking that analogy, if YOU can throw a ball at 20 m/s, and you are standing in a train going 40m/s and throw a ball, it will not come back when you throw it... It is still pushing you. For you, it is absolutely the same if you are standing still or moving at 5000 kilometers per second, the ball you throw will push you just the same. Sure, to an outside observer it is moving in the same direction as you are, but faster/slower depending on direction, and conserving momentum, making you move slower/faster respectively.

edit: Sorry, i think i have misunderstood you. I thought you were saying that a rocket cant go faster than its fuel exhaust velocity, when you were actually saying it cannot go faster than exxhaust velocity because it will run out of fuel before that.

Thanks for the math-check!  I refreshed my understanding, and you're right: rockets do go faster than their exhaust.  I'm (partially) right, for the right reason: it costs inverse-exponentially more fuel (as a % of your starting mass) to exceed your exhaust speed, which is why we drop off dead tanks and engines to shed weight.  The simple reason is as I described above: your fuel exhaust ends up with forward drift (from the perspective of a stationary outside observer), and so you do pay a "luxury tax" of even more fuel to have essentially accelerated your exhaust "forward".

The details, with math, are given by the Tsiolkovsky rocket equation , which reasons from Newton's 2nd law and conservation of linear momentum to obtain a nice equation:

• Fuel mass Mf = 1 - e^(- total change in speed dv / exhaust speed ve)

Let's call (dv / ve) your velocity ratio, in terms of your fuel's exhaust speed.  So dv / ve = 1.0 means you will exactly reach your exhaust speed, 2.0 means you'll go twice as fast as your exhaust speed, and so on.  (Each kind of rocket fuel has a characteristic exhaust speed, which roughly measures the energy released by that particular chemical reaction.  Orion, which uses thermonuclear detonations, would have a ginormous exhaust speed = blast wave speed, since it's a nuclear reaction, not a chemical one.)  Tabulate the equation for some velocity ratios, in terms of fuel mass you must burn to achieve it:

• velocity ratio 0.1: e^(-0.1) = 0.905, Mf =   9.51%
• velocity ratio 0.5: e^(-0.5) = 0.607, Mf = 39.35%
• velocity ratio 1.0: e^(-1.0) = 0.368, Mf = 63.21%
• velocity ratio 2.0: e^(-2.0) = 0.135, Mf = 86.47%
• velocity ratio 3.0: e^(-3.0) = 0.050, Mf = 95.02%

Interestingly, the type of fuel you use is irrelevant; you can ignore everything about it except its exhaust speed.  No matter what kind of rocket engine you use, you must burn off 63% of your starting mass as fuel to achieve a speed equal to its exhaust speed.  The more powerful the fuel, the faster that speed will be, but your cost of 63% mass doesn't change.  As you exceed your exhaust speed, your fuel mass asymptotically approaches 1.0, and so your payload mass plummets toward 0.  Put another way, suppose we hold payload mass constant, and make the rest of the rocket bigger.  Then to achieve ratio 1.0, you're at roughly 2:1 rocket-to-payload, but to achieve ratio 3.0, you're at roughly 19:1, i.e. you need a 9.5x(!) bigger rocket.  That's the difference between Saturn 1B to low earth orbit and Saturn V to the moon!  (N.B. low earth orbit is around ratio 2.2 with the best chemical rockets today, but a moon/Mars shot requires Earth escape velocity, which is around ratio 3.0.  That's why we need a Saturn V to lift off, and only that tiny command module comes back.)

I think my recollection of exhaust speed as the velocity limit applies to air-breathing engines, including turbines and hypersonic scramjets, where most of the reaction mass is borrowed from the outside world en passant.  If you're flying through air, and your wake ends up with net forward drift, then it must leech your momentum away.