New perks for ship levels and other things

I’ve read through the Dark Avatar posts looking for these thoughts; I could not find them, so please forgive me, if these things have been mentioned before. Here are just a few things that after a lot of thought, would seem to add to the game- These suggestions are listed by there theoretical ease to implement

1) Since it has been determined that the only advantage a ship gets from experience is more hit points- which by the mid game are worthless, because ships attack values far exceed a ships hit points - a bad defensive roll? Your ship is toast. So there needs to be another bonus for achieving higher levels, giving the higher level ships a real value in the mid and end games- Namely weapons efficiency. The easiest idea to implement this would be for every level you go up; your weapons and defense become 1% more effective. This would not add to the ships top damage, but instead add it to the bottom end- i.e. while a rookie ship with 100 attack would do 1-100 points base damage, a level 23 ship would then do 23-100 points damage. This idea should be easy to express mathematically
1a) Also, ships should get more experience from fighting a ship solo, then it should get for being part of a large group. Experience earned should be value of ships killed / by ships that survived

2) Add a new defensive feature: Evade, would basically allow each ship before it makes its defensive roll, to make an evade roll. If it makes its roll, the attack on the ship fails, throw in a few new graphics like the evade ship kicking its engines on full, and then rolls out of the way, and presto! Now we have some cool new things to watch on the combat view screen
Now, please don't misunderstand, the evade roll should not be a generic percentage simply added across the hull spectrum. Tiny hulls should get the biggest bonus, then Small, then Medium up the line. With the Huge hulls getting what would amount to a ceremonial evade chance. You could even throw in a top gun like wonder which would increase the chance to evade.

3) Improved AI. Let’s face it. The computer is the master when it comes to Mathematics. Even the best human in the world can't outperform it in speed or accuracy. And yet the AI on this game can't even build its planets right- Its not uncommon to find Factories on Research squares, and research buildings on Approval tiles all on the same planet. Or my favorite; when I capture a 5pq or less planet only to find all factories with no starport. This should not happen. There should be a set build policy that achieves the desirable outcome, such as percentage of Factories, to research, to Stock Markets, etc.
Or triggers on what to build if a planet has a low PQ score. All factories and no starport on a planet should be a no-no.

4) Improved the AI's defensive/offensive capacity. The AI should know how many of its ships can fall on a square at any given time. If such a system were in place, this would allow the AI to know, that when you end your turn with a fleet worth 27 logistic points into parsec 135,57, where it has 51 logistic points of ships instantly available to strike, it should then compute how much force it needs to win with minimal losses, and then dispatch those ships to do it.

I don't really know how hard 3 and 4 would be to implement- 3 should be the easier then 4 because a planet isn't going to have more then 40 squares to build on. While the size of a universe what? 1000+x1000+? The time for the computer to compute the value of each parsec maybe too time extensive- I don't know. Thou at the risk of offending Stardock and being ostracized, I'd simply suggest that they find some decent experienced Chess programmer who's brain you can pick. After all when it comes to putting values on squares that works, a chess program has gotta be tops. Thanks for reading my ideas.
10,214 views 13 replies
Reply #1 Top
I personally would prefer something similar to Civ 5's upgrade system.
Reply #2 Top
I personally would prefer something similar to Civ 5's upgrade system.


Those sneeky jerks! I can't beleive they managed to sneak another civ game past me.


Seriously, though I really do like the Civ IV upgrade system and I wouldn't mind seeing something like that. That said, this might get to be kind fo a pain if you tend to use smaller ship types. The UI would definitely need to be improved to make this work, since the current UI for ship stacks isn't really suited for this sort of thing. I also liked PolarPawn's evade idea, which would work quite well in this sort of upgrade system.


3) Improved AI. Let’s face it.


That's a lot easier said than done. Some of the things you mentioned seem like they should be simple (especially the factory heavy build without a starport). However, any sort of intelligent build system needs to take into account a lot of things, such as what your other planets have been building, the location of your planet (near enemy lines? Near other planets so I can use a starbase for the group?), and the race's abilities.

Thou at the risk of offending Stardock and being ostracized, I'd simply suggest that they find some decent experienced Chess programmer who's brain you can pick. After all when it comes to putting values on squares that works, a chess program has gotta be tops. Thanks for reading my ideas.


Again, not as simple as you seem to say. While there are chesslike elements, it is not nearly as simple. For one thing, you must consider that there can be a lot more sectors (much less actual spaces) in a galciv game than there are on a chessboard. Add in considerations like different weapon types, protecting dozens of fixed targets, and keeping in mind future research. Ideas that might be good in chess could be devastatingly slow in this setting. I'm not saying that it has nothing to offer, just that it is not nearly as simple as saying to look to chess for the answer.
Reply #3 Top
Civ 5? Is out??? And I wasn’t informed??? Man talk about being out of the loop…
Like I said, these things where listed in the order of theoretical ease to implement.
I never said 3 and 4 would be easy. Thou I have come up with a couple of answers for you.

Planet build orders would be easiest fixed thru set scripting, but from what I’ve read here Brad doesn’t want to do that. Defensive capacity of the planet is really a no factor in most cases. George S. Patton said something to the effect of “Great Defenses only prove the stupidity of man- Anything God can make, man can overcome.”
The primary goal of any world is to in-rich its owner. Either thru research, production or economic means. Lets face it, if a computer is getting its ships blown out of orbit around the planet, all a big population with the planet defense buildings are going to do is maybe buy it some time- until you show up with 3 transports with 3 billion soldiers a piece. First transport hits with mass drivers, and looses, getting the population down to a manageable size – if its not manageable after the first one, hit it again- the third ship well take the planet with mini soldiers. In a perfect world, the AI’s idea would be an active offense. (Not going to happen- I know)
There should be set percentage goals the AI has for its planets as a whole. Have a world with a low PQ? Make it a small research world if you need research- or an economic one…
As for idea 4- After picking my mind I realized that I could have programmed such a feature with Basic back on the Apple-(only problem would have been that systems computing speed – a game set to the galactic settings would still probably be in process today) So with today’s programming languages it should be much easier to do indepth. The basic string from the 80’s would look like this- here are the basic values for the rudimentary functions:
s(n) with n = ship production number and the S part meaning ship
sdata$(x,y) = X and Y would be your parsec location on the map. strings where the s(x) field is added, ending in something that looks like this when the sdata string is looked at: [sdata$(256,19)=”S(15), S(156), S(157), S(199)…” Which would mean that ships 15, 156, 157 and 199 could occupy parsec 256, 19 in one turn.
Now all we have to do is determine how to fill that string with the correct information- 20 years ago, it could have been done with this sub-routine: First step is to put the ship in its move corner:
Sbm(n)= ship base moves (ship number)/ x= ships number/sc= ships influence sphere/y=ships y location on map/x=ships x location on map/cx and cy are the x and y coordinates of the computations start points. (in this case, the South Western side of the move square)

10 sis =sbm(n)
20 bm=sbm(n)
lines 30 and 40 not needed after reread
50 cx=x-sis
line 60 not needed after reread
70 Rem *end of turn maintenance checking ships influence points*
80 I=bm: cy=y-sis
line 90 not needed after reread
100 Sdata$=sdata$+s(n)
110 S(n)(cx+I, cy)
120 I=I-1
130 If I>0 then 100
140 Cy=cy+1
150 If cy>y-sis+bm then 170
160 Goto 80

The only real problems I can see with this is that the time to compute all ships points may take to long.

Also the only reason I said look for chess for answers is because if I would where looking to improve the AI, I would hunt down and pick the brains of the people who have made the Best forms of AI for complex computations. (More or less, because I don’t do original thinking- I simply stand on the shoulders of Giants and then take credit for it) The Best, is hands down a chess program. The AI in chess gets its strength because of a few things. 1. Its massive data base it can pull form. 2. over time the mathematical computations for its calculations have been perfected, to the point where it can look at a 64 square board with 32 pieces locked in combat and see 10 moves deep in under 5 minutes, is nothing short of amazing.

The complexity of how every chess pieces moves far exceed that of how Galciv2 ships move. The only thing that adds to the complexity of Galciv2 is the board size. These ideas, (Especially 4) wouldn’t give the AI any kind of deep thinking; it would only give it a 1 move deep defensive/offensive option. The weak point could very well be thou, the average compute time for figuring the influence areas of all ships on the map, may take to long- Making the game unplayable

Note- Or you could look at a checker program too- Since Checker programs have held the title of “Best Checker Player in the World” for a few years now

Reply #4 Top
I personally would prefer something similar to Civ 5's upgrade system.


Wait a minute here... did I read that right? Civ V? When? How? Where? WTF!?! I just bought Civ IV/Expansion and haven't even played it yet because of this GC2 addiction. Please inform me, must know now!!!!!
Reply #5 Top
Morberis is just messing with us... Civ 5 well probably be out when Galciv3 is
Reply #6 Top
Your Chess analogy for the AI is so far off base. First problem, 1 oppenant, while GC has up to 8. In Chess waiting 5 minutes for the computer to plan its next move is okay. In GC waiting for 5 minutes for each AI to compute its next move, well that could be 40 minutes between turns. I would not want to play. Chess has 64 squares to compute, all visible at all times, with limited number of possible moves...ie pawns move 0ne space and one direction...bishops can only move on the diaganol... while each ship and each fleet can move any direction they want, and can move multipule spaces, revealing more of the map, and then having to recompute those recently revealed squares.

I applaud brad for not using scripts for the AI. For once a person figures out the script, the game becomes very easy to beat, and boring.

In my current game, I have tiny hulled fighters running around with 45 hit points, from experience, with ultimate minituratzation a fleet of these little wasps can take out a battle ship, and the battle ship gets one shot.

It is just like in Civ, one has to take care of those experienced ships, and upgrade them, and you will have one awesome attack force.
Reply #7 Top
It is just like in Civ, one has to take care of those experienced ships, and upgrade them, and you will have one awesome attack force.


I respectively disagree. A fleet of my Rookie Battleships from my end game would smoke your wasps. In the end, My Battleships normally have 200 space points for armament. Sporting sexy defensive scores of well over 1000, and an attack of well over 200. That 45 points would only make it one round. Besides I'd engage you with my small hull disposable "Defiant" Class ships before you even got the chance to engage my top of the line Battleships. Which do around 500+ points damage a round. Once again. that 45 points is academic.
And this game is ruled easily by adopting Henry Fords mantra, “Quantity is its own Quality”

Your Chess analogy for the AI is so far off base


The chess analogy is not of base. You’re getting too hung up on the forest to see the trees. We are talking about calculations- and what program is the best at calculating and implementing a brutal AI? What other programs do you know that can shift thru 50 to 100 billion positions within three minutes in a typical turn. (Source: Ivars Peterson’s Mathland) If you can think of one, let me know. I’d love to read about it.
with 10,000 ships, each having the average move rate of 10, gives us a total of only 10,000,000 calculations. Way below the 100,000,000,000 some chess programs have to do


Reply #8 Top
Your Battleship only gets to fire once, and takes out only one wasp regardless of how powerful it is. I get 28 shots to your one. Regardless of how powerful you are on each ship, I will take your ship out with a my wasps. I am not talking about tiny hulled ships with an attack of 1, but with attack of 24 each.

It is you who wants to set and wait for 40 minutes while the computer calculates all the AI moves...using your chess analogy. 5 minute wait in chess is acceptable, but not in stratagy games IMHO.

A chess so called AI has been being developed for over 40 years now. Again, a limited board space, a limited number of moves by each piece, and only 1 advesary. A much easier programing problem than 8 idividual AI's, dealing with 100's of ships, all having different capacity and capability, and calculating the best option against 8 other AI's. Chess is a very simple calculation in comparision.
Reply #9 Top
Just ran a test on my old athlon xp 2.0 (4 year old computer) It does 5,237 MIPS (Million Instructions Per Second). Or 5.237 BILLION caculations a second. My old computer could do the caluclations I have suggested in less then 2 seconds for all players. 2 second. I repeat: All players togather HUMAN & ALL AI's -2 second. So for the changes I suggest even the slowest 900 mhz Celeron could do that in no more then 10 seconds on a Gigantic Map/Abudant everything. If you doubt it, go crunch the numbers. You'll see that I'm right. Chess is far more complex then this game well ever be. This is in no-way a slight to Galciv2, Which is a fine game- but it is a testament to Chess. Here is a fact:

In 1949, information theorist Claude Shannon estimated that there are about 10^120 possible 40-move games. To give you a sense of how enormous this number is: It dwarfs even the most generous estimates of the number of atoms in the universe. If each atom were replaced by a supercomputer, it would still be impossible to complete all the evaluations in preparation for a perfect game's first move.

Also- I'm still not convinced on your high level small ships. I've had one of my Battleships last 200+ shoots on it from a horde of fighters (80-200 attacks) I'll agree that in the begining of the game when everyone has weak weapons high level small ships are the way to go. But in the mid and end game, when ships are dealing out 200-500 points a piece damage, the added hit points are worthless- Since when an attacker gets thru your defense the fight is over. What you want is space to house more stuff and the ability to produce those ships quickly. While you can build small ones fast they just don't give you enough room to house the stuff (never gotten a small up over 55)




Reply #10 Top
Perhaps it would be cool if you could give your ships standing orders, nothing complicated but just a selection from a list like "target smaller ships", "target capital ships", "distribute firepower" or "target transports" etc... and the higher your ship/fleet level you could have more options available.
Reply #11 Top
I understand the everyuthing but the "Distribute firepower", what where your thoughts?

That Target Transports would sureeeeeeeeer be nice for my border patrol ships
Reply #12 Top
Your Battleship only gets to fire once, and takes out only one wasp regardless of how powerful it is.


Sure about that? I've had my battleships rip through two ships in a single barrage before the return attack. I've had it go toe to toe alone against fleets of a half dozen lesser attack craft and come through with only a single shot fired in return.
Reply #13 Top
Sure about that? I've had my battleships rip through two ships in a single barrage before the return attack. I've had it go toe to toe alone against fleets of a half dozen lesser attack craft and come through with only a single shot fired in return.


"Return Attack"? You know that now every ship fires at the same time? This does make fighter wolfpacks less efficient, but having 6 ships wiped out with only one shooting is technically possible.

Polarpawn: It's all calculations. But as has been said, comparing galciv to chess is so ludicrous it barely bares thinking about.

Consider everything that a chess AI has to do. Times it by 9. Now make it consider the other 9 opponents, and the ship strategies it has to use, and the individual statistics of each ship, fleet, and planet, and what best to do with them, etc etc.

Your argument falls short instantly. A game like GC2 may have the same basic idea as chess (IE a strategy game) but is vastly more complex. The fact you think it's about ships moving shows you have no understanding of how the game works whatsoever.