My AI tests have moved to the latest version 1.32. I'm using my own newest custom map, Sedition, which is a multistar. Game settings are unlocked FFA, no allied victors. I turned the pirates on because I feel they are usable now, while the Diplomatic Victory condition remains off until it is replaced, I can't see any future in the current version without drastic alterations.
On Unfair I played Vasari and was able to overwhelm the nearest AI, another Vasari, because it chose to build a dozen turrets at each of its planets like anyone teaching new players explains not to. The Fortifier AI evidently needs more work. Anyway, after this I simply ignored the diplomatic techs and destroyed each of the other AIs in turn, it was no contest. Two of the Advent AIs did manage to make a firm alliance but it should be remembered that this doesn't imply any cooperation between their fleets, and I was able to attack one without the other interfering, which suggests that the penalty for attacking the ally of another faction needs lokking at- possible automatic loss of ceasefire?
Playing at Unfair level was quite a dull exercise. I would recommend that the 1.31 levels of AI relationship bonus be restored to make the game more viable. I also noticed that the pirates 'attacked' a dead asteroid I owned but had nothing built on, weren't they supposed to go for high priority targets?
I boosted the AI to Cruel levels, took Advent and set up just one other Advent AI with three TEC and three Vasari, as the element of similar factions cooperating seems to work well. I'm not fond of Cruel and Vicious levels because they cheat so much that it just disguises poorly thought out gameplay, but whatever.
My start was similar, taking an ice and my roid, but just after I sent my colony cap to another roid the nearby Vasari jumped my ice with a Level 4 Skirantra, a Marauder, 2 carrier cruisers and some frigates. I only had one factory and some Disciples at this stage, so I had to send them in and use my godly battle tactics STOP EXPLODING YOU COWARDS. This failed and though I built a carrier cap, returned my colony cap and had some frigates, I considered it prudent to send the pirates after my opponent, which was expensive but effective in the short term.
So on Cruel level I was finally forced to investigate the relations tech tree, maxing the flat bonus and building 9 envoys. One of the TEC factions had taken a dead roid and by sieging the starbase there to complete missions, and resorting to outright bribery I was able to obtain a ceasefire with the Vasari and send the envoys in to make it stick.
The Advent in the other system quickly perished, this left 2 Vasari allied in the other system, an alliance between 2 TEC factions one from each system, and a third TEC faction attacked by me and my new Vasari ally. Even against cruel AIs I quickly gained an economic lead as they dont seem to comprehend the need to keep credit rates high or have the ability to make trade routes with starbases included, and I was able to profit the most from the destruction of the lone TEC.
There doesn't seem much left in this game as the cross system TEC allies have just surrendered. My Vasari ally is also allied with one of the paired Vasaris but I have ceasefires with all of them and the AI isn't equipped to recognise by how much my faction is ahead of any of theirs, so they should prove fairly easy to divide and conquer.
The Cruel game began fairly brightly but the AI just doesn't have the ability to play the middle and late games. It needs to recognise when the player is becoming a threat, especially if it is being cheated by a player who ensures that the profits of an alliance in terms of planets all go to him.
The AI relations factor that compares fleet strength actually rewards a powerful player, which makes for a poorer game. The intention may be to penalise players who go for a total economy strategy, but surely it is far better to simply have another separate factor which compares credit income, and penalise for that too? The penalty levels then need amping up.
Also the AI needs a sense of how many planets each faction has. The adjacency will suffice for this, and either it seems to be working better in 1.32 or my setup favoured however it works, however the penalty level needs a considerable boost.
Conclusion or TL;DR
The AI is poorer in 1.32 unless you play on the super-cheat levels. Even then, it is far too easy to have cease-fires with all surviving factions in the mid and late game, because there are insufficient means for the AI to develop animosity toward a leading faction, no matter how strong they are or how close to winning the game.
Stiffer penalties for proximity to another empire, as well as the leading factions being penalised rather than rewarded for stronger fleets, and a new penalty on the basis of compared credit income rates, would help to improve play later in the game.
Have to agree with DesConnor. Excellent analysis btw DesConnor
In regardless to the relations logic...
In 1.32 all you have to do in unlocked teams regardless of difficulty level is to research a little relations boosts in the diplomacy tree early on and you can effectively cease fire all the other opponents while you pick one off at a time meaning gg...
In 1.31 at cruel and vicious more than half the time the AIs form super alliances with often insane joint fleet support that could be both very challenging and very frustrating at the same time. Researching diplo boosts early became mandatory and even then were often not sufficient to prevent being classified as galactic pariah. Setting the diplomacy rate to normal did help this quite a bit based on the testing I posted in another thread pre 1.32.
In other words in 1.32 unlocked teams are way easier than FFA (which is still easy due to some of the AIs obvious glitches) and yet in 1.31 unlocked teams was usually way, way harder than FFA.
I will admit when I play I refuse to make SBs (well except maybe on vicious depending on the map) and I intentionally do not exploit the diplomacy aspect by spamming envoys. But now in 1.32 all one has to do is research the basic diplo boosts and you get to cease fires without virtually any other effort.
Something in between would probably be more enjoyable imo.
It is much too easy to get cease fires early on ... rather than a cease fire being a garuntee ... maybe it would be nice for an AI simply to just not WANT to have a cease fire with you, even though everything else matches up?
This could possibly be achieved by yet again another modifier that is seperate from race, location, etc and more related too a random value between X and Y with those x and y ranges modified slightly depending on AI type (i.e. Aggressor, Defender, Economy, etc ...)
I'm not sure that I agree, having had more experience with them the 1.32 changes might be good enough. Certainly with the AI boost at 2 max, that still represents quite a few completed missions. A great deal might depend on the starting positions. If you've got time, could you try this map 'Sedition', with unlocked Vicious AI? I imagine the positions are closer than is typical, which might help? Cruel AI might not work, though I suspect the problems lie elsewhere. Here are two examples of the best and worst of the AI from the game I'm playing:
The worst- Luopica Colonists are almost out. They had a roid beyond my roid next to the plasma storm, though they lost it due culture noone else could have it due to the starbase. Rasaeida Sect decided that they wanted it and sent a fleet there. Unfortunately for them, I want it too, to extend my trade route, so I have a colony ship waiting in the area. Unfortunately for me, this overestimates the capabilities of my ally. Their fleet is mostly frigates and the Luopica starbase begins to destroy these and the capital ship. Worse, they do have Drone Hosts, but these not only wander into the starbase range they have mines not strikecraft... truly rubbish. The day Drone Hosts are liberated from mines will indeed be a happy day for the AI- please put them on the Purge ships instead?? So my best allies were defeated by an entirely inactive opponent, and to gain the roid I had to send my own Halcyons and Drone Hosts.
The best- that plasma storm is cursed. This time I was determined to keep a starbase there, so I built it up to 2 attack 3 defence 2 meteor and the trade. However the plasma storm by its nature keeps the best ships in my fleet, the Halcyons and Drone Hosts, from protecting it. Adjun turned up with a 4 capital fleet with 20 overseers, and many heavy cruisers and lrf rather than carrier cruisers, and levelled it almost before my Illuminators could get there. It was the 2 colony caps that seemed to do the damage, my starbase was heavily down on armour and had a long list of red modifiers.. My Illuminators rapidly destroyed the 2 rather high level colony caps but still. I put this down to the AI having high level caps in the first place, which it rarely does, and also to target simplification, my starbases are typically surrounded with all kinds of junk that the AI randomly shoots at diluting its attack, in a storm this doesn't happen. Also my meteors ran out just before the starbase exploded so they are not endlessly spammable without more culture than I could afford, I suppose it makes a difference when you have a culture centre in direct support. (On another occasion I noted that Scramble Bombers still works in storms, though there were no Skirantras this time.) So much fighting in a plasma storm is not good for a human player, again this might be my clever map setup aimed at helping the AI a little....
I have been checking the relations situation in the Vicious game, something I never had to do in the Cruel game. I am finding the screen hard to interpret. You can quickly get both the numerical relations level us-them and them-us from the same screen, but to get the details of why you don't favour them you have to go their screen- and in 1.32 its typically your relations with them that are the issue. Also, I have issues with screens full of numbers it has too much background detail that should really be kept from the player- though this is very helpful and necessary for testing. For example-
Adjun Imperials are my main opponent now, so I want more information on their alliances. They seem to have fairly solid relations with the Hand of Illus, the Advent in their system, less solid relations with Lexmada, the TEC in their system, and their relations with Provians are very flimsy from the Provian side as the Provians have about a 3.5 liking for them, about the ceasefire level. Wouldn't it be more appropriate to have a mesage 'the Provians are uneasy about the ceasefire,' though, rather than an exact series of numbers? It would not be possible for me to obtain such detailed information.. how can I not see planets in the same system yet know that Provian public opinion is influenced by -0.5 due to the proximity of the alien Adjuns..? Also, though I have all these numbers my agents cannot be certain that there is indeed a ceasefire, without me having observed this on the map?
Rasaeida Sect are my firmest ally, but I also have solid ceasefires with Illus and Lexmada, Lexmada might need firming up from my side so I sent some envoys. It is very difficult to stop a war with a Vicious once started owing to the vast numbers of ships destroyed. My diplomatic options seem to be confined to making a breach between Provians and Adjuns, and even better making a breach between Lexmada and Adjuns. The question is how to accomplish this with the capabilities at my disposal? The one option seems to be to save a fortune and pour it into a mission for Lexmada to attack Adjuns. The mission menu is a novelty to me as up to this level of play I have never used it before. One big complaint about it is that it takes you far away from the battle screen, but I was able to make time to have a go, Forewarnings help. It seems that you can't offer at least low level missions on just anyone, for instance the Adjuns will only accept low level missions aimed at Raseida Sect. I hadn't enough credits floating to see beyond the lowest mision level as I had just increased supply, perhaps more lucre makes a difference? Any faction that hates you, like Adjuns and Provians, tends to reject missions 'It's laughable that you think we'd assist your agenda.' (Incidentally its more usual to say 'advance your agenda' but these are aliens, perhaps? If you want some variation in messages, I'm sure we could help...) So how do I ruin the Adjun-Provian ceasefire then?
Also, it seems odd that the AI will only accept missions it approves of, but will offer missions that are more to your advantage than its own. It should restrict its own offers far more, to its worst enemy or at most its own top two enemies.
What might allow more variation in diplomacy is if the envoys also had the option to spread dissent and hatred.. I'm sure that this is in keeping with the general theme of Rebellion? How about two new abilities for envoy vessels- Dissent, which could generally lower the liking for every foreign faction (including your own) and Hatred, which could inflame racial tensions, perhaps doubling and trebling the normal racial modifier? That might help.. then you could dominate an enemy gravwell but not bombard it and instead use it for underground propaganda... nice. After all, thats what envoys have always done. Also, stirring up trouble might have the ability to influence the negative modifier for adjacency, doubling or tripling it perhaps, though this modifier is still at too low a level, as I've repeatedly said. The base needs to be 3-5 times larger, at least. This could then be the factor Boshimi would like, that the AI doesn't want a ceasefire, it might also help shape the AI expansion into more sustainable empires, which in 'Sedition' is achieved largely via the map configuration.
I noted a spelling error in 1/2 level of 'Unwavering belief' the spelling of 'initiated' is inebriated into 'iniatiated,' you might want to correct that.
Not a very successful session for me, though I annihilated yet another Provian fleet, killed many Adjuns, gained a roid and have built up the starbases on my newly acquired planets. It would have been good consolidation but for the loss of the giant starbase in the plasma. I'm not sure that I can afford to keep unsupported starbases up, which would shorten my trade routes dramatically. Still, I built about ten new trade ports this session, thats something.
My Advent 42 credit rate 1232 supply, it claims 7 planets but I just took a roid for 8. Adjun 27 (2237) 11 planets, Lexmada 24 (1988) 7 planets, Raseaida 21 (1974) 7 planets, Provians 19 (1313) 4 planets, Illlus (1760) 4 planets, the other out have been eliminated. Adjuns, Rasaeida, Provains and Lexmada all have dead roids included in their total. My figure is temporarily low because my trade route is broken, it was up to 57 before I incresed supply and I have been building trade ports since, but the AI seem to have picked up as time has passed. I just worry that the AI supply/credit rate targets have been adjusted for the benefit of the Vicious AI, which will cripple any lesser AI, and I still consider that the Vicious AI underperforms vastly. With 5x the credits available for investment any economy AI should have a base level better than the player, even before its multiplier. Otherwise how can the non-Vicious economy-minded AI hope to compete?
Lastly, on the trade issue, the underlying issue is not that the AI needs extra help to set up 'proper' trade routes but that the existing system is stupid and this unnecessarily favours the human player because trade becomes such a huge factor after the early game. Also, the AI can do little about the human players trade, especially once it is based around starbases- with the possible exception of Advent trade starbases in plasma storms, of course!
What is stupid about the existing system is that it is based on trade ports rather than trade ships. What seems especially odd is the need to have a 'chain' of ports, I'm not sure how this can be rationalised, the trade ships are often seen making journeys of more than one well and in historical terms trade is encouraged by the proximity of ports, they don't have to be strung out in a chain. Tea clippers, sugar fleets.. the most lucrative voyages were point to point, not along a chain of ports. Also the 'South Sea Bubble' is a notorious example of equating simple distance with the value of trade. The 'chain' concept is quite simply a quick-fix to ensure that the AI doesn't count ports twice when calculating a multiplier. Even at this level it is still peculiar and senseless, you could just count the number of gravwells with trade to get the multiplier instead, which would then help the AI factions enormously as they were no longer penalised for being too stupid to build chains. Why make it harder for the AI, just to make such a simpler calculation easier for the CPU???
Basing trade off the ports does give you a straighforward 'credit rate,' but this isn't even a clever simplification. I consider it yet another example of the numbers telling you too much- you could just have the basic rate from planets plus the number of trade ships based in each gravwell. Sins doesn't have the incredibly overblown and very ahistorical 'pay as build' feature anyway, so the need to know the exact credit rate is not really there. What you lose is that the trade ships themselves are transformed into a fast-travelling spacefaring extension of the trade ports, except that they are quickly rebuilt for free, so that they become nothing more than an annoyance and a distraction for pirates. One of those little adjustments that would profit the AI enormously is if the trade ships were slower and were given a far far longer rebuild rate so that they were important to the player. Trade ships are vulnerable as they travel to the edges of gravwells and the player typically wins by concentrating defences at planets. Making trade ships important would also make lower levels of pirate raids more important without them being a direct threat to the player empire. This is what pirates are for.
Previously I have argued that there should be a basic system of trade goods, which would add depth, but perhaps this is too much complexity. However I would still recommend the following system- when you build a port, you designate which other port it will trade with- including any foreign ports where you have a treaty. Longer voyages and voyages between different types of gravwell are more valuable. Ports can be redesignated if required. However the pay-off is not a 'rate' but comes only upon a completed journey. This shouldn't take too much CPU power, even with many many ports the calculations necessary are far less than a single medium-sized fleet combat. However it adds depth to trade, you can have convoys, attacking trade becomes far more fun. It doesn't even require the screens away from combat that relations required and its still far less complicated.
Maybe this is too radical. But if we absolutely have to keep the current and somewhat rotten system, please make it that the trade ships are slower and can be destroyed without then being rebuilt at once, so that AI fleets and pirates gain some advantages from dominating the space around player centres of resistance. As it is, the effective loss if you destroy one is about 10-20 credits, and they are quite tough! If they took 4-5 minutes to rebuild the player might have to look after them more, and the AI will be hurt a lot less because their trade is less important.. Also please ensure that the bounty system is working, really the AI should make money off that too!
no one bothers with trade ships ... other than building two trade ports as an awesome defense against pirates because they chase the ships all over the gravity well!!!
Trade ships should be softer, slower, and meaningful to create. Right now, players just build trade and they get income rates that FAR outstrip the rates of crystals & metal etc ... In the original vanilla sins camping trade ships was an easy way to make some money (to an extent) now, those ships are meaningless. The research upgrades to bolster trade ship armor currently are meaningless as no one bothers trade ships.
My objectives for this session were to explore a means of turning Lexmada against Adjun, to bolster my defences and build up to my supply limit including a sixth capital, to get the final flat rate relations boost and finally to take the Provian dead roid nearest the pirates which seems to have an artifact on it- and I will also search my own holdings once I have scrying. With the dead roid in my hands it should be easier to maintain a starbase in the plasma storm. I have raided the roid before ,but it went wrong when some of my fleet were inexplicably not on hold position and my 3 Halcyons attempted an AI-style daylight attack over the minefield to bring their guns to bear on the starbase, lucky I caught this before any were too heavily damaged. It shows how rubbish my attack would be without clinging to the gravwell edge, why cant the AI use 'hold position' itself?
One question was why Provians even have this dead roid in the first place, they lost their home to culture why not this roid? Well it seems that both Provians and Lexmada, once they captured the desert planet, promptly spammed culture on it. Provians have 4 on theirs and Lexmada 5. This seems to have happened early enough to save the Lexmada home from Illus. It must be some sort of scripted behaviour, but it shows the limitations of the Vicious AI, confronted with a lack of available space it cannot make rational decisions. Along with the mines on Drone Hosts this has to be one of the biggest holes in the current AI. It really hampers the TEC and Advent, it mght be interesting to play a Vasari-only unlocked and see how that develops. However even Vasari have a special AI hindrance..
Vengeance is a cruel mistress, but this does not stop me withdrawing my ships from the plasma storm to my home. It just doesn't make sense to fight in the storm. Also, I build a factory on my newly acquired roid to construct 8 envoys to send to Lexmada through Rasaeida territory, as I am blocked from the star and it is hazardous to use that route. Then I scuttle the factory and build more Harmony. The lull in the fighting also allows me to examine the map and my relation options more thoroughly.
There are two Advent-TEC wars ongoing, Rasaeida against Provians and Illus against Lexmada. Rasaeida make a rubbish attack on the Provian desert but it does show that something is definitely wrong with TEC mines as the Rasaeida frigate fleet passes straight through a huge minefield and hardly any mines explode. However Rasaeida are hampered by their silly Drone Hosts and as they retreat a Provian fleet catches them. Fortunately for Rasaeida the Provian defences are on the wrong side of the planet, facing out into empty space, or their failure to destroy the inhibitor would have been more costly. Lexmada seem to have destroyed an Illus starbase at the ice with Ogrovs, but then they have been caught by the Illus fleet which is based around Halcyons and has hardly any Drone Hosts. Lexmada have another entire fleet though, and thats why my envoys are going to Lexmada! However I am at a loss how to make them dislike Adjuns, even though they are not very friendly to them, only neutral, I cannot work out how to get this lower. I had hoped that there might be Adjun envoys and was goimng to send a small anti-envoy fleet, but no. There seems to be no way to 'massively' bribe them to attack, either, the only hostile options for mission targets involve their only existing enemy Illus. Yet Adjuns routinely give me missions to attack Provians, who they have a ceasefire with.. I can only conclude that the AI will not accept missions that break a ceasefire but are happy to offer missions against their own partners. Lexmada even want me to attack Rasaeida, which is out of the question, but seems far more sensible as an AI plan since it is exactly the sort of unreasonable activity I want from them though am barred from offering! But why would Adjun offer me Provians as a target, whom I am fighting already? Also, my reward for completing/not completing Lexmada missions is puny, a 600 metal mission is 0.13 or -0.05. Even though Lexmada's attitude is vital I don't care, I need the metal more. Rasaeida missions have a more reasonable reward of 1.68 but they are too dangerous and I can afford not to bother. Even in the unlikely event that I lose a pact, so what? Losing a ceasefire is a strategic consideration, anything beyond that loses relevance quickly. As it is my second wave of envoys gets me Lexmada ship vision.
Heres a closer look at my relations with Adjun, my main opponents. My diplomatic preference is 0.2, and racial -1. Curiously I have 1.98 from missions because of the Provian war. Fleet strength is -0.01, which seems small as they are probably two levels up.. I still consider that this should work the other way, and that there should be a separate penalty for leading in the economic sphere. As it is I have a credit rate double theirs by skimping on fleet and with almost no penalty. Military actions costs me -6.83 and diplomatic actions a further -3, obviously we are not chummy. I get a flat 3 from research, the other modifers are all 0. From the Adjun side, their diplomatic preference is 0.15 and racial -0.15, they must have top-level research. I have 0.06 from missions, I must have given them an easy one sometime? Military actions is -7.5 and diplomatic actions -4. Resources is 0.4 (bribery!), trade 0.7 and research 3. It seems fairly clear that in 1.32 research and military and diplomatic actions are the main modifiers. I would want adjacency and fleet and economic strength all boosted so that the AI adjusted on the strategic level as well as tactically. Earlier I said a 3-5x boost for base adjacency but looking at this 5-10x might be more realistic.
At this point the huge Adjun fleet I had left floating in the plasma storm met another at the wormhole and attacked my ice, while yet a third Adjun fleet attacked the ex-Provian home. Show time. My entire fleet had finally been brought together and were sent to the ice, minus my newly built battleship which went straight from the slips on the volcanic to the ex-Provian terran. I noticed that the Adjuns were fairly easily able to drain my starbase of antimatter at the ice, whereas I had assumed this would only happen at the plasma. However their attack on the starbase and its assorted defences of turrets, repair and hangars was not so effective that any more than two repair had gone down before my fleet showed up, even without meteors. My fleet truly hammered them no contest, they lost three level ten Kortuls and twenty overseers, it was a wipe out. Their fleet seemed too defensive, it had 15 of the subverters as well but only 7 heavy cruisers and 10 assailants, just not enough firepower. The 8 carrier cruisers werent enough either. I did notice that the Vasari siege frigate was operating from further out than normal, must have been research. Yes thats the sort of thing, lets go further in that direction with the others and make the Vasari long range upgrade truly spectacular. At the ex-Provian Terran I had one battleship versus a Vasari fleet, it just sat behind the defences and got to level five. It was a Marauder-led fleet but just not big enough to trouble my defences, and the inhibitor meant that it too was annihilated. This gave me a break from Adjun attacks and I was able to reduce the Provian dead roid- completing Adjun and Illus missions in the process, ho hum. Even the Vasari, who seem able to negate the main offensive abilities from starbases, cannot damage a well-prepared defence. It might make a great deal of difference if they accompanied their attacks by building starbases though, like all humans do- the Advent bring Starfish and TEC Ogrovs when they attack gravwells, why don't Vasari use their own specialist anti-starbase attack?
To cover my attack I asked Raseida to attack the other Provian dead roid and they promptly obliged, taking on Provian and Adjun ships for me. This made me wonder whether the ability to order your allies around is not perhaps too strong in a game like this. I wouldn't consider it amiss if there was a massive relations penalty for each order you gave your ally and that it complied with. Otherwise the potential for abuse is just too great. I was effectively using their ships as a meatshield for mine... I also noticed that they made another attack on the desert with some scouts, 2 siege frigates and 9-10 Starfish, what made me laugh was that the defences were on the other side so the Starfish and scouts were mainly clearing the minefield that the entire Rasaeida frigate fleet had already passed through... The Provian starbase had control of the planet so the siege was ineffective and of course once the Starfish got near the broadcast centers they started taking too much damage. Sneak attacks like that just dont work in the game, the AI should stop trying them.
Finally got divination and not a single lousy artifact in my entire empire. Another note- why can't I have already searched an ex-homeworld as I will never find anything there? Or should you get a fixed extra benefit for occupying some other factions former home 'This was the original home world of the ..... and its possession gives you ....' as well as the undying emnity of that faction, of course. Anyway I exploded the 20000 health Provian starbase, swept their 150 mines, destroyed their turrets and hangars and took it all for myself, then sending my Illuminators to secure the plasma once more and build another starbase there as well as the one I was building at the dead roid. My six capitals bombarded the roid, whaddya know, a Planetary Organic AI, that will more than come in handy.
State of the galaxy- My Advent 46 cr 1360 supply 8 planets (it has not yet recorded that the dead roid changed hands, my cr will zoom once the plasma trade starbase is back) Adjun 27cr 2251 supply 11 planets, Lexmada 26cr 2000 supply, 8 planets, Rasaeida 22cr 2000 supply 8 planets, Illus 11cr 1692 supply, 4 planets, Provians 10 supply, 1453 supply, 4 planets. Provians are doomed, so are Illus. I have a firm alliance with Rasaeida and am very comfortable with Lexmada, Adjuns hate Rasaeida and are very shaky with Lexmada, there is a truce between Lexmada and Raseida. Seems like a winning position, we'll see. If the AI adapted to relative power as it should do there might still be a great deal of game left, as it is, possibly not.
This session I wanted to end the Provian War, perhaps slightly more ambitious than my last objective of taking a single dead roid, but then again the situation was more favourable. Lets consider my advantages at this stage, even in a Vicious game-
-I can go up to the penultimate supply stage, making the advantage that a Vicious AI gets that much more marginal. Simply being able to replace ships quickly just drags the game out once the AI has no advantage in numbers. It might be better to have rather less super-cheat on resources and research and instead at the higher level of AI restrict the player supply to the smallest level while the AI gets largest- this wouldn't break the games hardcoded limits yet would extend the AI advantage longer into the game. Even so, I am catching up to their Vicious research level, there is little that can be done about that, unless you want to give the AI combat bonuses, which I loathe, I'd much prefer adjustments to supply levels. If at setup you made this a faction tab rather than a game option it might also be a method to advantage new players against skilled players.
-The starbase system is yet another arbitrary Sins system that seems designed to punish the AI. Yes, you can justify a great deal in a far future environment (except advanced societies without telescopes perhaps), but it is normal to make game compromises that favour the AI, that help to disguise its limitations. A starbase will damage a fleet leaving a system and remove all its antimatter- by some unknown method not justified by lore- however it is not itself an inhibitor! So a human player if in trouble can destroy an inhibitor quickly, jump, and rest until a fleet is recovered. The AI cannot identify inhibitors and cannot judge the health of its fleet, so it can't take advantage of this peculiar loophole. It is very odd that starbases have magical powers throughout a gravwell yet cannot act as phase inhibitors as well, it smacks of a typical Sins compromise 'Dont make my inhibitors useless, they are a graphical element!.' Also, starbases are often less advanced on the tech tree than inhibitors... If a starbase can affect an entire gravwell it would seem normal to have its weapons range throughout the gravwell, but no, a sweet spot is also provided for the player to exploit the AI. The AIs reluctance to use starbases in any gravwell but its own must be considered an AI problem rather than a mechanics problem though... I have seen it deploy starbases in enemy gravwells before, just not in recent versions.
-The AI is incapable of assembling fleets, it does not understand good composition. Once its first crop of elite capitals is gone it tends to fall apart. I had an Adjun fleet of 2 carrier cruisers, 8 Overseers, 1 Skirmisher, 3 Flak, 9 Assailants and 9 subverters attack the ex-Provian terran with its starbase and full array of defenses. That size fleet will just be destroyed, its not that important for production costs but it does subtract from the main AI fleet, and the AI tends to employ 2-3 such fleets. Where this might have come from is pre-Entrenchment coding, where it was possible to defeat planetary defences with a much smaller fleet, but its survival into the Entrenchment era and beyond is an anachronism. More than just the supply points value of a fleet, the AI should have a better ratio of support to fighting ships, it can't be that hard for it to build Assailants/lrf, which are also the easiest ships for it to use.
-I wondered if the superweapons would be a threat at Vicious level but the AI hasn't managed to build any yet, unlike Cruel. It does seem to like planetary government on its starbases, while I haven't bothered on mine, this makes mine that much more effective. If the AI can build them, random attacks from Advent and Vasari superweapons don't seem much problem, they both require too much coordination for effective use, especially the Advent weapon.
-Even if the AI manage to ally against the player, if the player has at least one firm ally it is easy to arrange matters so that they do the brunt of the fighting for no reward. Vicious level even makes this far easier, as your ally does not run out of fleet and seems always eager to obey.
The battle between Rasaeida, Adjun and Provians at the other Provian dead roid raged on while I sent my fleet to the Provian desert and picked that up. It seemed that the Starfish must have been more effective than I had supposed as the starbase was gone (perhaps the earlier fleet attack had damaged it) and 2 of the culture centres had been replaced by much-needed labs, either that or the AI had made a very poor attempt at reorganisation. I took it easily, another 150 useless TEC mines out of the game, speeding up framerate now! Then I moved on the Provian dead roid where Rasaeida were fighting and took that as well. Provians made a last attack on the volcanic and also sent a dozen Ogrovs to their former home, but both attacks were easily cleaned up, the Ogrovs by the planetary defences alone. I starbased the wormhole near my ice in the face of another small Adjun fleet and sent some flak and Guardians to help as I built it up. I also began the nefarious process of starbasing the Rasaieda empire heh heh heh. The Raseida fleet were sent on to the last Provian roid, the one on the edge of the system and linking the two systems. My fleet arrived much later to avoid casualties, but in the only disappointment of the session, the crafty Rasaeidans managed to colonise the roid just after it was bombarded! It was an artifact roid as well, they'll regret that! Exit Provians, anyway. There was another skirmish between a small Adjun fleet and the defences at my artifact roid but I won that easily, the plasma storm is now secure! Fairly routine play.
Illus are still holding on in the other system though Lexmada are wearing them down. I was watching the Lexmada relations to Adjun closely as they seem the most important. The only significant modifiers were racial penalty -1.25 (not yet full research?) military action 2.96, diplomatic action -3.25, trade 2, research 3, AI bonus 2. However what I discovered was that when the Adjuns were no longer involved with the Rasaeida-Provian fighting and their attacks were solely on me, the beloved friends of Lexmada, the military actions modifier began to drop and it seems that Lexmada will eventually join the war against Adjuns. So while I couldn't directly affect the Adjun-Lexmada relationship, me ordering Raseida around had an effect.
I continued to get rewards for fighting my long-standing enemy Provians from Adjun and Illus, that really needs some more work. If the AI only offered missions against those factions it would accept from me as an offensive mission target, that would be sufficient.
At least the change in the Adjun-Lexmada relationship shows that the relations situation is not entirely static in the late game, though it would be good if Adjun had responded to its sliding influence with envoys, so I could shoot at them. The AI just needs to add strategic goals to its influence from tactical goals, and the strategic factors need to be on a par with the 2-3 level that the tactical factors seem to be on as well as the flat rate research and AI boost. Yarlen willing, it may be possible to get a fairly sophisticated relations environment. It shouldn't really be a problem on the same level as the combat AI, and though I'm still concerned that this behaviour would only appear at Vicious level and not before, it would take AI improvements elsewhere before that could be tested. Perhaps some of the limitations in the AI's use of fleets might not be exposed if it had only one opponent to deal with at a time, if Lexmada begin to attack I would expect the Adjun moves to become even more nonsensical.
State of the war: My Advent 48 cr 1675 supply 11 planets, Lexmada 28cr 1998 supply 7 planets, Adjun 27cr 2293 supply 11 planets, Rasaeida 24 cr 2000 supply 9 planets plus the roid they just took, Illus 9cr 947 supply, 3 planets. At least the big four remaining Vicious dominate their quarters of the map, which is more than happened during the Cruel game, and it will also be very hard for me to betray Rasaeida until the end, unlike the Cruel game where I took the option to have one system all to myself since I had no enemies in the other system. If I betrayed Raseaida now both they and Adjuns would be against me, 2 Vicious empires of 10+ planets. With the strategic factors that I would like made more influential, I would get a general -3 or far more for economic factors (considering I have double anyone else's base income and only 20% less fleet, and one level of fleet more should be a -2 penalty, with one level less a +2 advantage- this would make my economic penalty -10!) as well as a heavy penalty for adjacency with Rasaeida, this would make my situation more flimsy. Perhaps there should be another factor, some sort of jealousy, so that if one partner in an alliance does much better out of it in terms of planets gained it places a strain on relations? However I would still go for a straightforward relations penalty for every order as the best way to limit player abuse of an allied fleet.
Just as we foresaw... I haven't really had any experience of having to play the AI this far into the game, this is the first time I've known them have substantial empires. I had no particular objectives this session, there was no real pressure so I wanted to examine what was happening in the game while I filled the Rasaeida empire with starbases. I also decided to ask Rasaieda to take the Adjun dead roids in the Adjun system, my plan was to get the two equal-sized AIs deeply into a fight while preserving my own ships. One problem is that if Adjun fight Rasaeida this makes them more popular with Lexmada, but Lexmada seem fairly inept and slow to finish Illus so they might not be much help anyway.
However Adjun seemed not to fight much and when Raseaida took an ice in the top half of the system formerly owned by Illus I began to get concerned that I needed to move quicker. I used the wormhole to attack the Adjun ice and took out the huge starbase but before I had destroyed all the hangars and mines the entire Adjun fleet appeared and as I was in danger of being in the position of protracted combat I had hoped Rasaeida would be in, I opted to retreat. It was a good concentration by Adjun with their entire fleet, about 50% bigger than mine, appearing simultaneously from three directions, and the retreat cost me a few ships including my prized captured Subverters and my first capital ship losses of the game, though of course this was no penalty in terms of levels. Poor luck or has the AI some means to decide that ice planets are to be protected and dead roids are not valuable? Rasaeida had been very quick to starbase the dead roids but Adjun then took the one near the pirates on. Rasaeida also captured another ex-Illus roid on the other side of the star and at this point are doing far better from our alliance, though I have starbased everything they own in our system apart from their dead roids. I even wonder whether this is a good plan, as most of the starbases will need fleet support and that divides my fleet up, though I hope to catch the Rasaeida fleet in the other system and right at the end build up to have the star fully starbased to interdict them as they return.
At the end of this session I pushed over the phase lane between the systems and took the Adjun roid to extend my trade, using a trade starbase at the Rasaeida-owned roid. I also have a starbase in the asteroid belt beyond that and my fleet are there, though Adjun have a huge force at the Volcanic deterring an attack on the Volcanic or Terran. I keep urging Rasaeida to defend their dead roid but they aren't interested- perhaps there is an existing routine to make them ignore you more the more you ping? I wouldn't want to spring my surprise with Adjun still with a large empire, but I may have no choice if Raseida continues to expand into the other system. Also Rasaieda have built a superweapon on the ex-Illus roid, in a very vulnerable position but then a superweapon is a superweapon. I wonder if the AI has probems with tactical slots when building superweapons, so that it prefers newer acquisitions?
Some observations about the later game-
Both Rasaeida and Lexmada keep pinging and urging me to attack the pirate planet in the Adjun system, which seems a particularly odd request? If this is just random at present, why not link it to their offered missions?
The mission system seems to be falling apart later in the game. Even fifteen ship missions can be completed without effort due to the longer times and starbase attrition. Fifteen ships is a great many in the early game but not much later on, perhaps it should scale better, or at all? Also some of the rewards are derisory, Rasaeida wanted another envoy though they already have six of mine, the reward was 0.06 and the penalty -0.03. I was going to use the mission system more but I really can't be bothered its too much time off the main screen, I just ask for crystal from my main allies now and then, I get some crystal and it keeps my population happy. The numbers on the rewards and penalties need adjustments to make the mission system at all relevant.
The envoys seem to have been overboosted though, I'm at 20.00 toward both Raseida and Lexmada, with 12.00 and 8.00 from envoys, and there seem to be scarcely any envoys at my planets, maybe 3/4 total from each faction. Is this from the new envoy options? Perhaps the boosts from the new research items need recalibrating because +3 relations from a single envoy seems a lot, especially compared to the AI relations boost of +2 max and the mission rewards.
I've noticed a few times planets being described as 'invulnerable' in green at the bottom of the card, though it doesn't last long, what's going on there? It happens to both my planets and theirs, seems random.
Also, I have both seen Advent AI use the notorious mines in combat in an enemy gravwell, and the Vasari AI (at the dead roid) build a starbase in an attack on an enemy starbase. So both do happen. However, the offensive starbase build should be routine during a large Vasari fleet attack, not exceptional. With the Drone Hosts, I just wonder whether they can't carry 2 squadrons minimum and have at most one set of mines at any one time. A rule like that would seem to solve the associated difficulties?
State of the war- My Advent 64cr 1680 supply 12 planets, Rasaeida 35cr 1999 supply 13 planets, Lexmada 27cr 1998 supply 7 planets, Adjun 26cr 2286 supply 8 planets, Illus 2cr 54 supply 2 planets. So the Vicious AI does have some economy, 150+cr is a lot. I wonder whether I have made the right decision, my slow build up in the Rasaeida empire has given the AI more chance than I would normally allow but then Adjun were a bigger threat than I had faced before. If I can take out the main Raseaida planets quickly and put culture up they will never recover, but its difficult to make headway 'quickly' against a Vicious, my hope is to have their fleet far away because once attrition begins they have the advantage. I'm concerned that Adjun have been using their entire fleet together, this is good AI but very threatening in terms of this particular game.
Would it be possible to program some sort of random element into the AI, so that in a good position it would suddenly betray, like I plan to do to it..? I'm not sure how you would implement it, but I imagine that you could have a starting point if it has 95%(?) research and say controls 33%(?) of the planets then it might check every so often whether it will go for one of its allies?
Couple of comments based on reading your recent posts DesConnor.
1) I agree it seems envoys are too powerful now. Too easy to get to high pact status.
2) Too easy to get ceasefires. But at least with the AI relationship bonus reduced they are not as likely to all cease fire each other early game and they will start to engage each other. In 1.31 at cruel and vicious since they start with lots of techs (including some diplo) analyzing the diplomacy relationships showed they were getting to the 3.5-4.0 mark with respect to each other in a matter of minutes which depending on the map could often lead to fireworks for the human player.
3) In 1.32 the trick seems to be to research to the 4 diplomacy boost techs asap and then get up to the credit / metal boost tech in time to secure cease fires when you want them. It is so easy now to ninja colonize a planet and when the AI shows up to beat your ehad in you dump in some metal or credits and get a cease fire. Rinse and repeat till you are left with one enemy you want to fight and go from there.
4) Some AI behavior is random I believe. Sometimes I play and they never have planetary government even on cruel or vicious. Rarely do the Advent SBs use mass disorientation let alone meteor showers (thank God).
5) I still often see super-weapon deployment but I think in 1.31 super-weapon firing was non-stop and early. The firing may have more to do with how many actual wars are going on. In 1.31 since I often tended to ignore diplomacy for the first 45 minutes or so of a game, I would end up getting hit multiple times with novaliths or kosturas or deliverances (one game I got hit by two novaliths, one kostura and two deliverances in like 2 minutes of game time .. of course I was fighting 3 allied AIs). But in 1.32 I can secure early cease-fires, abuse the diplomacy system a bit and I haven't gotten hit with one single superweapon even though I hear and see them getting deployed.
Personally I would like to see the ceasefires a bit harder to get for both the human and the AI. I played last night a huge map random vs cruel Ais as a Vasari and like 40 minutes into the game I had 7 ceasefires and one enemy. In 1.31 I would have had minimum of three allied AIs coming at me with at least two working in one grav well at a time together non-stop with their majority fleets because they weren't fighting really anyone else.
The capacity for mischief with the diplomacy abuse smacks a little bit of Galactic Civilizations where the best strategy was to trade techs for low pop planets while going into debt. You never had to build a colony ship and though your econ was red you prevented all other empires from expanding beyond 3-4 worlds while you ended up with 40 planets and the game was over. That ruined that game for me unless I turned off tech trading.
Anyways in 1.32 I reiterate that locked teams is harder than unlocked. In 1.31 unlocked teams was MUCH harder than locked teams. Thus by transitive logic, 1.32 unlocked is much easier than 1.31 unlocked, It isn't rocket science imo.
Oh well just my opinion.
Cheers.
Ceasefires are the real issue, they are just too powerful in unlocked FFA, once either you or an AI faction has one there are so many options to reinforce it. A slightly higher initial threshold required might work better, perhaps 4.5 rather than the current 3.5. To go with this a slightly higher AI boost, as in 1.31, would be an improvement. Both the envoy and mission rewards need better balancing. However what the system also lacks at the moment are methods of splitting alliances, so that what you get is either AI that don't ally with each other, or AI that are easier to pick off very early using ceasfires but also capable of forming super-alliances that last forever. Your own alliances tend to leave you with eager puppets until they are the last faction left.
Another possible issue is that you get too much reward for simply fighting another faction, even if the conflict is remote from any other faction, the AI just sees a faction it hasn't a ceasefire with and rewards you accordingly for victories that are entirely in your interest. Perhaps the military actions points from destroying stuff should be dependent on the AI faction also having a significantly negative military actions score with regard to the faction in question, or some similar approach?
Generally, I would introduce more negative factors into the relations system. Adjacency does not work currently as the numbers are too small, and I suspect that the concept is not extended over enough gravwells. Hyping up adjacency might help a great deal as expanding empires begin to make contact over many points. The AI should then fight a nearer enemy and produce a more coherent empire. There could also be more negative factors for owning a large fleet and having a large economy, the latter being particularly unhelpful to the player against super-cheat AI. I would even consider a flat-rate negative factor as you worked your way up the supply levels. Economy is slightly more difficult to quantify but a player typically has at least twice the income of any AI and it is difficult to imagine any alliance lasting under those circumstances, the penalties should be appropriate. It can't be that a faction with no income should hate everyone though, so its tricky to make it relative.. perhaps there should be penalties for any rate above thirty credits, say a general -1 per 5 or fraction of 5 above? Also, how about simple penalties for owning too much of the galaxy?
The current relations system isn't really working but with a little tuning and getting the AI to recognise when the player has a large economic lead, it could be serviceable. The biggest obstacles are that it is only semi-moddable and also the wide variation in the types of games being run. When it isn't working we can only test the early stages and report that, the later stages aren't telling us much. I shouldn't expect to get much out of a Cruel game in this version but you might consider trying my map at Vicious level, either that or try a map like Gateway where the factions are also much closer and set up to clash early. I'm going to continue reporting on my Vicious game, but without any real pressure to defend it is becoming routine. Sins isn't the only strategy game that gets dull in single player because of the lack of a stop-the-leader concept, but with a dedicated relations expansion and all the numbers the game can check it should be very easy to implement one, there's no excuse not to.
Agreed ceasefires give the player too much leverage. Some AI relationship bonus is probably reasonable but there needs to be a higher threshold to get a ceasefire so there is more equilibrium time for various empires to fight each other and then pair off into alliances based on adjacency, competition, common hatreds, etc. Missions are broke imo. They offer little impact overall. Not enough dynamic alterations possible of alliances once they set into place.
In 1.31 doing every mission offered still could not prevent AI from doubling up or worse on the human player in a short period of time. They simply crystallized into AI alliances way too fast. Now in 1.32 the human can leverage ceasefires too easily. I play Vasari and now I always build imperial labs first. Why? Not for ice colonization or eventual trade. But for diplomatic boosts. Then I build one military lab for kanraks and from there I get a 3rd eco lab for yes you guessed it more diplomatic boosts. Once I have the 4 diplo boosts + the metal giving boost I am usually set. The alternative is not to do this and let diplomacy ride but then in 1.32 you will pay a heavy price. Right now the more challenging position is to probably just turn off diplomacy and FFA.
I will give your map a try
I agree that the cease fire should be bumped up from 3.5 to 4.5, maybe even 5 at the high end. I play as vasari as well and the diplo boosts almost feel like a cheat. Maybe they boost too much?
Just finished two games on your map DesConnor (nice map btw though I detest plasma storms ). I played both at cruel (since I don't have the time to play large vicious games right now). Playing with unlocked teams was still formulaic via exploiting the diplomatic boosts and fighting one at a time. I quit once I had killed two enemies and after the rest of the dust settled there were only 2 other empires that were any threat. I quit since once I am max fleet Vasari the remaining AI were doomed at cruel difficulty since they were fighting each other and they never know how to put all their forces together. The second game was FFA. That was actually more fun since I got hit from not only my 3 jump neighbor but also from another AI across the asteroid link across systems early on. It actually forced me to drop one SB so my fleet wasn't pinned and could do some damage elsewhere. I also had to build more fleet earlier than I originally intended meaning I had to delay some eco development and logistics was a bit bottle-necked. In the end it took only a little more time to achieve the same rough end result but imo it was more fun. When I get enough time I will attempt a vicious FFA on that map. Cheers!
Wow, what a research. You should be paid to do this
Perhaps it is useful for people that dont have that much time, to make some brief summaries of the main points of your posts, by using bullets format. I suspect it will also increase the odds that Yarlen will do something with it!
Thanks for the map DesConnor. I had a fun cruel where while I was putting the finishing touches on stomping one of my neighbors, an enemy Advent brought the house via the backdoor wormhole from the other system. That was fun and unexpected
Yeah, though it is difficult to recommend unlocked at the Cruel level in 1.32 on any map, I wondered if proximity might help, I'm pleased you had some fun!
Other games tend to have many AARs or detailed commentaries, which have long been absent from this site, I wanted to do something about it as well as moan about it. However any FFA takes ages to play, it would be good to have a choice among several victory conditions. The multiplayer is rarely any more than 5v5s, which just seem a method of mixing good, bad and indifferent players in one game and are also too affected by lag and quits or drops to make for good commentaries.
A brief observation would be that 1.31 was better, which is where I started... However, I did want to play through some games to test how the relations system developed. The system does have quite a lot of potential if it isn't crippled by the starting values, but it needs extensive recalibration to achieve what it is capable of. The solution to the AI super-alliances wasn't to reduce the AI advantages, but to correct the real problem that there are too few negative modifiers other than military ones, making both ceasefires and alliances too effective. Stop-the-leader and stronger adjacency modifiers would help greatly. There, that's quite a concise summary?
I played some more of my Vicious game, I had reduced an Adjun terran when I managed to get Rasaeida to attack the adjacent volcanic and then I decided to betray them, to see what would happen as much as anything else, though it is easier to defend one system than two half-systems- especially for the AI.
Anyway my fifth column starbases proved to be expensive follies- my plan was to advance around the Rasaeida gas giant relieving each beleagured starbase in turn, but they went down rather faster than I could destroy the Rasaeida starbases- and this is with 40-50 squadrons of bombers against theirs, with only a handful against mine? Dunno why that was. The original plan was to send a Halcyon to the furthest ones to help, perhaps I should have tried that, as I didn't need to avoid splitting the fleet. However I had the income to starbase the stars, something else the AI is not good at, so a total disaster was averted.
I was made to regret not taking out the pirates in my system at the first opportunity as well, as they finally got tasty enough to destroy the starbase at my only artifact planet, ho hum. I had to use my last supply on a fleet to hold it while I began to refortify it.
Before the war Raseida envoys were getting a total +16 boost from my population, which seems rather high, though there were more of them than I had suspected, certainly half-a-dozen and possibly as many as nine. The AI semed to have one per planet, is that the most effective method now? After the war started relations with Raseaida then fell to -20... I had wondered whether Adjun would warm to me and they did, but insufficiently to get near ceasefire levels. Lexmada, having witnessed my insanely treacherous unprovoked attack on my gamelong closest allies, are now my dearest chums. The relations penalties for being transparently unprincipled are vastly underdone- if you betray an ally like that, why would anyone at all ally with you in that game ever? That modifier needs work as well, together we can stop these ruthless AI exploiters!
Anyway, Rasaeida found it difficult to disentangle themselves from Adjun, and though Adjun also gained my roid in their system, by the end of the session the only Rasaieda starbase left around the Rasaeida gas giant was at the volcanic. They also have the dead roids and a roid between Adjun and me, but the rest of their holdings are in the other system, basically the former Illus empire- which they seem to have taken instead of Lexmada, who did all the fighting for it. I have acquired one of their artifacts, the phase accelerator, and will pick up another in due course. Most of my new holdings aren't well-developed as yet. Adjun have recovered, when I could have put them out, but they are only as strong as they were in the past.
Sometimes I wonder how much unspent income the Vicious AIs have, it might be useful to know. I can spam starbases now, but I don't doubt that they have been able to for some time, yet not only do they very rarely use starbases when attacking, they have also failed to fill any of the available space in terms of plasma storms, asteroid belts and stars, when it is very much to their advantage to do so, especially on multistars. Another AI routine that could be improved.
State of the war: My Advent 57 credit rate 2000 supply 14 planets, Adjun 30cr 2300 supply 9 planets, Lexmada 28cr 1998 supply 7 planets, Rasaeida 16cr 1933 supply 8 planets. Lexmada are near to war with both Adjun and Rasaeida but not quite. I wish they'd stop pinging for joint attacks on the pirates though, even if they have nothing better to do!
Where are you getting the precise credit income rate / fleet supply / planet numbers? From the end-game stats page? If so can't you just go to the "Credits Available" statistic to see how much the Vicious are saving, if anything?
Shouldn't be too difficult.
I just had a weird game in locked teams FFA on the Sedition map with cruel. I was early on fighting for my life as I had snatched one of the TEC AIs 1-jump asteroids and caught a bunch of hammering due to that, meanwhile I start getting rocked by another Adevent AI on my own 1-jump starting asteroid. By the time I got my act together and launch a retaliation on the volcano world again 1 hop from the TEC homeworld. And what do I find? They hadn't even colonized it yet a good 25-30 minutes into the game. I snatched that planet up and proceed to annihilate the TEC HW while playing defense on the Advent. Never looked back.
Yeah I have given up with cruel AIs with unlocked teams in 1.32. You can almost immediately secure ceasefires with the all AIs except the first you are already in pitched battle with. From there you roll them 1 by 1. It isn't any different at vicious, it is just a question of whether you survive the single enemy on vicious who is on your doorstep.
The wormholes and the asteroid bridges are a nice feature in the Sedition map btw.
The AI can be very hit-and-miss when it comes to volcanics, and though they are not very valuable very early on, not colonising them often seems to become an error. If I made a map specifically for AI I might cluster ice and volcanics in the centre and use a two-roid start rather than one roid plus ice or volcanic. However that sort of approach tends to leave the AI too distant from the player. Wormholes are a feature that don't seem to be used enough, and they are very rarely useful on random maps- though I once won a multiplayer 2v2 where my ally had quit with a surprise siege frigate rush through wormholes, it was one of the proudest moments of my early playing days. Also, I like to use every single gravwell type, its peculiar that there aren't more- in three expansions, not even one extra terrain type?
My Vicious game is definitely in the twilight of its usefulness, as I am not experiencing any AI attacks at all. For a brief time I even had a ceasefire with Adjun, but this ended as Lexmada went to war with both them and Rasaeida, who are stil at war with Adjun from the days of their unfortunate alliance with me. So I own one entire system while my much smaller ally wages war on my behalf in the other one. This is the sort of situation stop-the-leader modifiers are implemented to prevent. The AI at present cannot even 'see' how many planets I have compared to its own and the total number in the game.
I also noted that Lexmada are still having problems with culture even at this stage, why they built several culture centres at the desert and none anywhere else is a mystery. Culture seems to be much more of a weak spot for the TEC AI than the Vasari AI, dunno why. I now have seven or eight superweapons bombing the other system but as ever with the Advent superweapon I'm not sure that it does much, it might help accelerate the spread of culture where you already have an advantage, but its certainly not a technology I'd rush to. Another aspect that needs a rework.
Also, I have two of the Advent siege capitals past level six and they are truly sad at sieging. The graphic is good, so at least you get a long time to watch it... The ships are so marginal that the ultimate being meh leaves them somewhat without purpose. It should one-shot planets, with maybe a small addition to the channel time. After all its not like the Vasari colony cap ultimate, where you get resources from it.
State of the war me 75 credit rate 2000 supply 20+1 planets, Lexmada 31cr 1993 supply 8 planets, Adjun 27cr 2300 supply 9 planets, Rasaeida 13cr, 640 supply, 4-1 planets. I just took the last Rasaeida roid in my system, the artifact one next to Adjun, it was the metal artifact. I also took out the pirates in my system- which stops my superweapons firing at them- and lost a capital carelessly in the process, no big deal. None of the pirate ships had ever returned, so it was just their basic defences. Those turrets are meaty though, I kept away and took them out with bombers. Having cleaned out well after well of Rasaeida's Advent homing mines it doesn't seem tricky as long as you have Guardians and plenty of strikecraft, I've lost perhaps six scouts to clear several fully-stocked gravwells.
Lastly I had a look at the credits available- I'm floating 1628, Lexmada have 303950, Adjun 603149 and Rasaeida 122482. Korsul Armada went out early with 128950 in the bank. Its not as if the AI was short of metal or crystal either, their reserves are as impressive as their banks. It makes their mission requests for 200 crystal seem extremely pointless. The missions definitely need scaling as the game progresses, and surely flush AI should not request credits or resources? Also, with the equivalent of 30 huge starbases in available funds, it makes the AIs inability to use starbases in attacks or to fill stars and plasma storms/junk/asteroid belts with starbases that much more of a drawback. Vicious AI is just not much of a threat later in the game, after all what use are such huge reserves? Instead of a general fleet supply toggle in the at-start options, one per faction would allow the AI to be set with more of an advantage throughout the game. After all, as it is a late-game Normal AI with credits to spare is as dangerous as a Vicious AI with credits to spare, no?
Finally got to play a vicious FFA on Sedition. That was a pretty rough start. I got lucky early when one of the enemy AIs stopped pounding me and I guess had to engae another AI encroaching its far border, leaving me to deal with my nearest neighbor.
Had no choice but to SB up since they are right on top of you from the get go. Was tough though since the enemy AI kept throwing Ogrovs at me. Tougher on this map since you can't easily get a good trade chain up. Finally turned the corner when I jumped his HW then phase stabilizers helped me to secure my borders with one full fleet.
Anyways once I finally I secured the majority of my system and had my 2400 (Vasari fleet) I felt like things were going to be easy. Silly me I jumped into a well in the other system that I didn't scout (was tired late at night) and ran smack into a full TEC fleet with 10 capitals ... four of them level 7 or higher Marzas. Did I take the smart approach and kite and bomb? No I didn't pay attention and four missile barrages later decided it was time for bed
Conclusion: vicious FFA much harder than vicious with diplomacy. Vicious edge is early but is quite an edge when they are on top of you like in Sedition.
Corollary: Late game vicious is not that impressive, except if they have multiple Marzas and you fall asleep at the switch ((
I'm just about finished my game, not quite. The Lexmada superweapon has taken the other pirates out so they have failed to return to base making some areas of space uninhabitable- I counted about 220 pirates around one neutral volcanic. The AI seems to have it in for the pirates, they have a limited lifespan once a TEC faction has a superweapon, why is this? It seems peculiar that surviving AI TEC factions should eliminate them easily and noone else does, why make the pirate defences stronger to everyoone other than TEC? Also the AI has little concept of how dangerous pirates are when massed around a neutral planet, Lexmada kept trying to have reinforcements transit pirate-dominated wells, leading to their inevitable destruction.
Lexmada and Rasaeida seem to have a truce but I destroyed the Rasaeida fleet remnants as I pushed into Adjun space. Lexmada came off worst in a battle with the Adjuns, who have 15 capitals, but my culture is depriving Adjuns of planets. I have just repositioned my fleet to take the Lexmada roid with their lone superweapon, building only a couple of advance starbases this time. The game is over really, but it would have been good to have Lexmada and Adjun patch up their differences and make some attempt to stop me, rather than have me order the ever-gullible Lexmadans to die for me. I'll see if Lexmada stop fighting Adjun once I attack, but usually an AI faced with a two-front war collapses quickly, their diplomats should always make it a priority to have only one enemy. I know mine do...
I noticed that the Raseida fleet has dwindled and it seems possible that they have no factories despite plenty of cash, while the lone Lexmada superweapon is at the vulnerable edge of their empire. With so much resources available, it might be useful to have the AI rationalise their holdings occasionally, even if this meant scuttling some of their structures. They should certainly never be in a position where they can't build ships!
I now have 74% culture in the galaxy- possibly there could be a further basic victory condition where Advent need culture to win, TEC need to liberate populations (planets) and Vasari need resources (total number of extractors)? It would seem a fairly easy condition to implement, just to prevent games having to be finished off completely.
My Advent 84 credit rate 2000 supply 24 planets, Lexmada 32cr 1248 supply 7 planets, Raseida 12cr 257 supply 3 planets, Adjun 7cr 2299 supply 3 planets. A count of starbases might be a more accurate representation of overall power, as I also have four at each star and more at neutral gravwells- yes I know that I can get this information, I will do next time!
I finished my Vicious game, it was quite tame as Adjun had the only other fleet but they were defeated by culture. Rasaeida and Lexmada fought to their last planets, somewhat pointlessly. Many strategy games are fairly dull to have to finish off, but the epic scale of Sins makes for epic dullness in an obviously won game. Perhaps another option to consider would be some form of time limit when playing against the AI, or the game is drawn between the survivors. I would still preferred the wonder-style victory however- the Dimensional Cascades! That would have obvious benefits for the AI, as otherwise time is definitely on the players side with limitless resources, another reason to prefer limited resource games.
Another aspect of this is that the AI may finally be starting to build superweapons and its capital ships to use ultimates, but the top game-winning techs have become so weak, stuff like Insurgency and the returning fleets, and this is a further handicap for the AI. This game certainly wasn't developed in an AI-friendly manner. Insurgency should be an AI's best friend. The 'space tax' on the returning fleets just makes it impossible for the AI to use, even human players have to calculate very carefully to get value out of the ability. Is it rocket science to have a small number of very strong but simple high-level techs, perhaps one or two per faction? Make the returning ships free, and not count to supply limits.. make Insurgency not only spawn guerrilla fleets but paralyse all structures in a grav well until the fleets are defeated, allow Advent a game-winning research as well, perhaps the nullification of hostile culture centres for peroids of the game. When there was all the fuss about the returning Vasari ships being the end of the game I was pleased, at least that was working, then it was just a matter of allowing the rival factions a similar ability.. but no. The decision instead was to nerf, which was stupid and got the game moving in the wrong direction.
Rasaeida did have some factories at the end, but too few military labs- they had about eight culture centres over three wells. Of course Adjun took a different approach with only a single culture centre and went out earlier, so Rasaeida would have been right in a time-limited game, or in a game where they had any chance to come back, eight was probably an overeaction, they could drop three and then at least have LRF. Though having said that I didn't encounter vast disciple fleets so even with a hoard of credits they simply weren't building ships for some reason? If the Advent AI has some preprogrammed urge to build many many culture centres once it has a superweapon, then it is just sad, a reflection of what the developers might have hoped the game to be like rather than how it is like. Games are not honed by wishful thinking.
Lexmada did try starbase attacks at the end, though without the support of their fleet, and only on planets I had just taken, for some reason. I just wish I knew why. Some RTS games make scripting available to the players, then the best AI can emerge competitively. I suspect the AI with the strategy of sending lone starbase constructors to planets they just lost wouldn't get very far in a competitive tournament! That would be a great game though wouldn't it, if we could play remotely by sending in AI scripts and merely watch our children slug it out... It might reveal much about the game though, as humans began to adapt the AI to conquer rather than merely to participate in a semi-authentic manner.
Those Revelation capitals Advent have are truly meh, its possibly the best graphic in the game when they use their ultimate, but the effect is so disappointing it takes the impression away. Advent have so few siege options that to have their siege capital so dire is extremly unhelpful. Without its ultimate it has no extra siege at all. As I wrote earlier I would make one use of the ultimate kill any planet, it does have to be channelled. An amusing element of the endgame was to watch a high level Adjun Marauder take on my four starbases at the star. The starbases were all culture/strikecraft, so perhaps they weren't typical, but even with sixteen squadrons of bombers they were unable to finish it off- all four of them! It just sat there for ages, phasing out as often as it could- of course it did no damage to the starbases but still.. and should Vasari surrender when they have no planets left anyway? Why would they do that?
I believe that single player Sins 1.32 could be reworked into a fairly strong game by comparison with many other RTS games, though as Darvin pointed out once the current level of AI is not high. It might take a Sins 2 to make best use of the engine though, as some of the concepts of the original game just seem too flawed, and all these flawed concepts seem to penalise the AI. Any weak research or ship category will hit the AI far harder than the player, who just knows to avoid some technologies or ships until a game is already won. Even the basics of the the combat system penalise the AI. The diplomacy system is no exception, and just makes an already severe AI problem far worse. If Sins 2 is going to be a mainly single player game then why not design the game around the AI in the first place? Its not as if a space game imposed any sort of constraints from reality...
That would be great. If the AI were moddable I would definitely be interested in putting lots and lots of effort into it.
Earlier in the post, it was mentioned that the AI 'styles' are lacking and the cheating levels annoying (paraphrasing a bit here). I thought about it, and wondered what it'd be like if there was different cheating the AI got (AI must have some kind of advantage to win. The goal is to make it less blatant and seem like you're on equal footing).
So:
Now, this is still cheating, obviously. But it makes sense that researcher cheats with research, an aggressor cheats with fleets, and an economist cheats with money. Even though it's cheating, I don't think I'd feel cheated so much this way. Fortifier could use some additional ideas, as I don't think it scales into late game (but then again, what good are defenses in the late game?).
On another note, the last time I played was when 1.32 was first released. I was happy with the AI, tbh. Yes, it's still pretty dumb, but.... idk, it seems to now have a knack for creating some really interesting scenarios (I'm struggling to come up with something specific, and I'll probably have to play a few to back up my claims). What I think I'm trying to say is that 1.32, imho, made the AI more fun to play against. Whether it actually made it harder or easier, I don't know, but I was enjoying the experience alot more.
While that would be an interesting variation on AI playstyles, I'd really rather see the devs instead work on opening up detailed AI modifications to players, who could (hopefully) then code things similar to what you said and more.
I could agree with that, but last I heard that wasnt going to happen.
And, btw, I'm not saying this is the only thing that needs fixing. Better auto-casting would be huge for the player and the opposing AI, but I think this gives the AI some character. If you have the AI on random, you'll be trying to figure out which 'type' of AI your facing... and that's akin, imho, to trying to ascertain a real opponent's strategy. Heck, we could stand to have a whole slew of AI types, especially with Rebellion and the rebel/loyalist paths.
I suppose that the devs could make an in-game AI wizard, ala the in-game map creator, but thats about as far as I think they are willing/can go.
There are many great features available to you once you register, including:
Sign in or Create Account
