I can't rec this enough! This is the point I'm trying to make to the devs. If they add sliders and cheats to the AI, I will optimize my own race better and play with a better over-all strategy, end result,.. a less fun game as I have to slog my way through even more defenceless ships... No they can't and the solution was to give them a bonus in 1.04. They were already twice as powerful as Battleship, but because they were getting chewed up by the player, they received additional points to their hull... End result, it takes a few more shots to kill one, but the worse is what happens if the player fields one. If the AI can't handle my Cruisers, which are nothing special by the way, how will it handle one of my Titans? Answer, it won't And this Is so critical and basic, I'm going to quote it twice @McUH Are you shock from Steam? Either way, nice to have you here.
No, I'm "Mc" there, but I did post in the Shock's thread. Thanks. I mostly registered here because this AI issue is critical in my opinion and should not be that difficult to fix (assuming AI has some bonuses so it can afford to research duplicated tech. to keep up with armor/shields/targeting and enough production to upgrade ships). Aside from this ISG is actually very well designed game.
Yes it is and it brings quite a lot of good things to the table Ok then, that’s where I ran into you. So I guess I may as well hijack my own thread for a bit and talk about Ai. One of the games I play is hearts of Iron, darkest hour, it’s my favourite of the series. For anyone that knows the game, you can start as any nation and select the start year as well. A skilled player could take a small minor nation and conquer the world. In one AAR I read, another player started as Germany in the “Endsieg” scenario. That means the scenario starts in December of 1944 where the objective is to see if you can last longer than Germany did IRL. Well this one player not only lasted, he actually beat back the Allies and started winning the war. Thankfully the devs did not respond by further tweaks to the AI as that player is in the extreme minority. A good game needs to be accessible to the casual player, and a challenge to the average veteran. But there will always be someone that can beat it no matter what, there is nothing wrong with that. Another game I played a lot was Ageod’s civil war. In 1.06 I could choose the confederates and defeat the superior union by 1864. Well those devs started listening too much to the complaints. The result was that by the last version before civil war 2. (1.17), I could beat the AI one year sooner. Every action has visible and not so visible consequences. ISG, at the hardest difficulty, gives the AI two extra infrastructure at the HW, extra starting population, 6 free techs, a robotic mine, an economic capitol, and very good worlds nearby. This means that if the player chooses emperor, the AI starts with about double the production capability, extra research capability, extra income generation, less tech to research and quicker expansion compared to the player. These are not insignificant bonuses, in fact, they are quite powerful, they offer benefits on turn one, and if used correctly insure that the AI can outpace the player throughout the game. Adding additional bonuses does not thrill me if the existing bonuses are not being used well. Play the game long enough and you know how the AI expands and behaves, even at the hardest levels. It will seek to expand and exploit, focusing on civilian aspects, and only start building a military if it is at war (but not with the intensity it should), or it has expanded to the point that further expansion is difficult and a few hundred turns have elapsed. The problem has multiple aspects to it The AI neglects its military in the early game The AI neglects building a true "fleet" The AI relies too much on single ships and not small task forces The AI ships are designed poorly The AI is predictable in the tactics it uses in combat The AI tactics in combat are not diverse enough (see above) and inadequate to handle even the most basic play style In the end, adding some production and research cheats may help the type of player who took Germany at the end of 44 and fought on to victory… but only if the military aspects of the game are addressed.
Continuing the war against the AIs, I saw some puzzling behavior On turn 217 I destroyed a single Lemurian FF at Zrops. The FF was equipped with a combat jump and appeared right next to my Destroyer. The result was obvious, it got hit at point blank range by my beam weapons and my missile salvo finished it off quickly as there was no distance to travel. This special, “combat jump”, can be a very powerful tool when used correctly… this was not it. The following turn I destroyed a Kaek Cruiser that Raided Sofiela using a single Destroyer. I also destroyed a Kaek colony that had been rebuilt at Acriu. In the middle of such a conflict, I do not see the wisdom of rebuilding colonies in a hotly contested zone and leaving them un-defended. Finally on turn 219 I chased away a Lemurian cruiser at Zrops and destroyed, (get this) a Sulak Battleship at Chemoria. This too makes little sense, The Sulak are in a bad way and should be conserving forces, sending a Battleship, by itself, into a far away Nebula was not ideal in my book. (I made a mental note to exterminate them entirely soon). I also sent one of my advanced Destroyers to intercept a small Kaek task force at Hodorth. Here are the results.
So your next challenge is to play against the optimized AI races on highest difficulty with this race: I'd give you -75% to ship attack and defense and ground combat if it would let me. An additional rule being you can hire no leaders and cannot use missiles or torpedoes.
We suck??? You are being far too kind I don't mind an asymmetrical challenge, but this is too much for my taste, I must decline. A game should be challenging but winnable, even at the hardest settings, what you propose is too close to a kind of cheat that turns me off entirely as I know I can't beat it no matter what I try. Perhaps you could soften this somewhat and I will consider it
You perhaps sell yourself short, sir. But I think it's the maluses to the home world and low-G and Lava preference that would hurt the most--all the per pop and combat maluses aren't really *that* significant. Perhaps just this, then: the ship defense and attack and research maluses; no special abilities positive or negative; no bonuses whatsoever; but you can't use missiles, any weapon mods, or any ship leaders.
Before I forget this amid all our AI discussion which is absolutely necessary IMHO ... The best strategies often involve creative elements. And with a game like this, these may revolve around minimizing micromanagement, like restricting yourself to limited tech or fleetsizes. Anyway, I recommend to make your savegame from that turn available for download for everyone to check out. It should become our touchstone to measure AI progress in the future. If the AI cannot win in this situation, with the allied Kaek and Lemurians vastly outnumbering you in every respect, the game will remain worthless as a strategy game and remain a pure sandbox to try out a couple of times. This savegame is also of interest to all players, who would like to see how the AI matches up against their skills. Don't be afraid. Simply be resourceful. You see it is not so difficult.
Perhaps, but to be honest I am getting a little burnt out with ISG and more so with the AARs. The later demand that I break and document actions, which at the later stages of the game mean I break immersion on every turn. So If it is all the same, I pass. It's time for me to take a break, perhaps when 1.05 comes out I will return. I haven't forgotten about your WIKI, I will work on the topic you requested over the next couple of days and contact you by pm Yes, that is true, but by strange I meant more than this. Two things here. First I think it's obvious that the AIs will lose, even if I were to take them all on at the same time. The second thing is that you may be a bit harsh. The game is easy for a group of players, this includes you and I, however, some players complain that it's too hard on normal. Worthless depends entirely on who is playing it. We also will see 1.05 come out soon, I will remain neutral and reserve judgement until that time. As I am not one of the devs, I am not privy to all they have planed, so let's take a wait and see approach. I've already given up on this AAR, and the only save available is 219, if anyone wants it, contact me by pm... though as I said, it's over, I have 3 dozen ships now, was set to start producing Battleships at a time when one of my Destroyers can beat any AI cruiser and most Battleships by itself, what's the point of continuing?
By all means take a break if you're getting burnt out. And I wasn't asking for an AAR--just offering suggestions for limits you could place on yourself to have a more challenging game. For myself the challenge of 4x's is to get absurdly advanced in tech, back from the days when my goad in Civ 1 was to see how early I could get rail roads. My record was just after 2000 BC. The combat parts never interested me as much--most 4x's don't have great combat systems. In a game like this with customizable ships, my goal would end up being building one perfect ship that by itself could conquer the galaxy.
Ideal would possibly be turn 215, but 219 should work as well. Could you post it as an attachment here? To be more precise, I think it is an ideal touchstone to measure new ISG versions. No need to hope for a good map and nurture your empire for 200 turns. Simply load it with a new version and have a go. If it offers a challenge it is fine, otherwise you know it doesn't work. It depends. Most people think, hey other 4X games (looking at you, Civ V! ) are hardly better, but as I am working on an AI that is very, very similar to the AI required here, I know it is not just wishful thinking but doable. It is a mountain of work and needs an implementation that reflects the recursive nature of the decision making process that is typcial for strategy games. Strategic goals, strategic planning, leveraging of threats, lookahead, variant prediction for actions and counteractions and more. Tons of work but doable.
Well that's just it. I handicapped my self here on purpose. On turn 115 my base production was 660 and my research was slow. Just a small adjustment to my race and the choices I made could have easily resulted in this, twenty turns earlier. That's the same Cruiser type under construction that is still effective in Terra dark after turn 200. There are many roads I can take to ramp up my game, both in production and research. Playing this game as a 4x with role playing elements is quite fun, as a strategy game it needs some work
I tried new game over weekend now on impossible/emperor, playing as Nova Vida (which fared worst of all AI's in my first game). There are several layers which bring AI down: *** economic *** Sometimes around turn 60-100 I erased AI start advantage and was keeping first in most categories. 1.05 patch might help by giving AI continual bonus on top of start bonus (which, once erased, is gone for good). *** warfare *** Aside from AI neglecting defences (armor, shields), the biggest problem is attack mechanics. The hit chances are too low and especially early the fleets just exchange missing shots and hitting frigates is almost impossible. Human can cope with it (leaders, crew experience, continuous beams or missiles). AI struggles, can't hit anything and so loses without destroying any ship. IMO hit chances should be greatly improved, especially when ships get to close range. To compensate maybe some mid/late game defence module can be introduced (like inertial stabiliser/nullifier in MoO2). *** planet defense *** In MoO2 it was tough to proceed early. Battlestation was threat, and it was backed up by missile base and sometimes fighter garrison. You needed real power to take fortified worlds. ISG - there is no planet defence like missile bases. Battle station is easy to hit and it itself can't hit anything (no leader, crew experience, no custom ship design leaves it with very low attack). Because of this bomber ships can easily move from planet to planet, bomb them to dust not getting any attrition in return. Planet defences should be improved so that you need at least something like 2-3 battleships to reliably win against fortified worlds. You can still blockade with weaker fleet, but taking down fortified world should be costly unless one employs big fleet. Human can adjust using fleets for defence and "dance" with them intelligently. AI is not capable of it and really needs improved static defences.
Well that's a problem. At maximum advantage in 1.04, the AI starts out with over a 4 to 1 advantage in production... think about that for a moment... A player can certainly optimise their own race, but losing a 4 to 1 advantage so quickly is troubling. I think it's a bit more complex than that. Nice suggestion True, In ISG you need air defence buildings and ships in orbit as a solid defence, a Battlestation alone won't cut it. The AI however builds too few units early on, there should be a better balance, a ratio of forces that need to be respected relevant to how many outposts/colonies the AI has. Right now it tend to exploit the space around it, and then move on to building a fleet, leaving itself vulnerable because it doesn't have enough forces to defend well, especially early on It might, but some players may have a real issue with this. For example, I do not consider giving the AI a starting bonus a cheat, as it is something I can replicate myself in the game. I do consider a flat production bonus a cheat as it cannot be replicated by me as a player. Besides, as I stated earlier, at the hardest levels it has a 4 to 1 advantage in production and significant advantages in research and cash generation, how much is enough? 8 to 1? 16 to 1? How does this make the AI look if it needs such perks to be competitive? As you say, it may help and be welcome by a percentage of players... it may also be viewed negatively by others
Sure, it will be optional and I think turned off by default. But it is common practice since the CIV1 (if not before). I view the difference as follows: Static bonus (e.g. current start advantage) - this is stronger on paper because of snowballing and exponential growth in 4x. However, you must be able to use it correctly, and this is difficult. Continual bonus (%) - this is weaker, but easier to use (you never lose it). Which is why most games prefer it. And lately even experiment with dynamic bonus (AI % bonus increases as the game progresses). Good thing is that you can avoid early rush by AI (as AI does not really have some big early advantage) and it might keep the pace as the game progresses despite being inefficient. I don't consider it cheat if it is transparent (clearly stated) and AI still needs to use game mechanics (e.g. is not getting free units/buildings out of nowhere). I consider AI cheating when it is avoiding game mechanics as that makes the game mechanics bit pointless (e.g. no morale penalty from number of cities in CIV 5 for AI, no need to pay maintenance in Stellaris, no need to pay for developing colonies and never runs out of fuel in Sots etc.). Of course this is mostly about economic part, it won't solve ship battles.
Yes @McUH there are many ways to look at it. I concede that with a game as complex as ISG the devs are in a hard spot. Difficulty levels are not influenced only by the setting a player will choose, but also by that same player's skill level and how she/he may optimise their own race. As such, while I personally do not care for production bonuses (it's one of the reasons I stoped patronising Civ games), I do understand why it should be offered as an option. As far as ship battles, I partially agree. A good production bonus will see the AI field more units early on, a tech bonus means they will be better equipped. For some players this may be enough. However, if the production bonuses were introduced in conjunction with some alterations in how the AI designs ships, how it fights tactically, and how it deploys operationally...then we would be looking at something far more challenging, even for the old time veterans
