AI Theory

[UesugiKenshin]

In 4X strategy/RTS games, the AI has an unusual task that is very different than the human players'.

Human players
This group has a diverse group of people who want a pleasurable journey in which part is a mystery that does not immediately make sense, but involves storytelling. Part is a logical puzzle in which clever intuitive players doing what makes sense , then are rewarded. Part is about an obstacle course as human beings like challenges and acquire a sense of accomplishment in traveling in ordinary and unique terrain and then overcoming obstacles, and then are rewarded. This is the map aspect of play. Then some like to fight and so a variety of units and in BotE's case... starships then allow various kinds of offensive and defensive strategies to "win" battles and paying attention to attrition and trying to have a range where winning results ij as little as 5% to 30 % and losing can be a range as well. Losing inspires reloading a save game and trying different tactics and strategy, so that in similar combats, players who are along a continuum from ignorant to having military prowess, can learn to win. Often only 20% can consistently win, but human players don't know this.

Even that number is high when random maps exist. How many strategy/RTS 4X games are played consistently on random maps. No typically. They most often are real world maps. Then players learn how to win on that map. Then try new maps.

Subsets of winning include the various achievements in the game, but usually have a "peace" or "wonder achievement" so winning is not just by military means. Most people dislike conquerors. Winning feels hollow when that is all there is.

AI
The AI exists to provoke the human player so induces conflict, and this ranges from economic competitors, research competitors, assymetric competitors, often a religious cometitor, a cultural competitors, diplomatic cometitors, and military competitors.

Human beings expect conflict...so the AI regularly induces conflict even when doing so accomplishes nothing. The AI cannot be doggedly trying to win as much of a game is micromanagement and doing the identical patterns in new mapregions to create profit and so facilitate the formation of a military and provide for defenses. Players would not enjoy being beaten 75% of the time.

Why? The AI is not emotional and does not care if it loses, but will use any way to wage war to increase its overall success as it is driven by competition.

This is why the AI turns evil by doing heinous actions that shock human players as they would feel bad making the same choices. The AI has no qualms so seldom will be duplitous and only seek alliances that temporarily will apply pressure to a weaker or stronger AI empire.

Human players do not intentionally provoke conflict with the AI ever 10-20 turns as doing so is pointless. Thus the AI has a "lull" that humans do not have.

Exploits

Humans
Human players will seek out any way to succeed that gives a higher chance of success even if doing so seems like cheating to some other human players. Winning is more important that following the rules.

AI
The AI cannot typically win in any real sense as it cannot feel satisfaction. It wins by causing conflict. It can do this poorly so difficulty levels are written in such ways as to first give the AI faster access to resources, or better economies, or faster training, spawning military units as the AI often poorly defends, etc. At higher difficulties these things happen AND the human player is penalized.

[UesugiKenshin]

Two of the most successful strategy/RTS 4X games are Rome 1 Total War and Medieval 2 Total War. They are played on an established map even if modded. They begin with starting infrastructure, an economy, a techtree that has been partially accomplished, thus units can be created or agents. They have an established military.

The empires initially face "barbarian" empires ie minor alien races in BotE. That is the immediate threat that must be eliminated as the human empires or the main empires (major races on BotE) are too strong and so fighting would cause attrition and a stalemate.

So the AI and human player know which to target first. And they can within 10-20 do so and reduce the barbarian empires and increase their fiefs and so enlarge their lands and so increase various kinds of income.

BotE is different. BotE's minor alien races can have significant defenses in sectors and whole fleets that the major races don't have for hundreds of turns...even 600-700 turns in some cases.

BotE also has celestial phenomena that can induce attrition.
BotE has ranges that limit expansion.
BotE has no established economy.
BotE has no established infrastructure.
BotE has no routine way to establish various types of income (farming, mining, trade). A player has to figure out ways to make income by mining and economic manipulation and trade is a tiny fraction that actually makes no sense as engaging in trade harms the players own economy. The AI never establishes trade routes. Why? Because doing so is not helpful.

BotE has no true military forces to any degree to allow safely taking a minor race sector without extremely high attrition.

See the problems?

[UesugiKenshin]

Civ 2 and Civ 3 are two of the most popular similar games. But few people play on random maps. The most popular follow a Rome and Medieval format as mods on established maps, economies, units, etc.

There is a pattern here.

BotE's AI and the human player lack the standard arsenal to allow playing and so just to begin, one has to create what say Medieval 2 begins with.

This is how to fix the AI in BotE as well as increase the enjoyment of human players. It's why BotE at the very least needs a save game editor, but hopefully a campaign editor, or some mechanism so an established "milieu" and an established map has established empires with established military forces made affordable by an established economy.

The AI in BotE doesn't know how much it can spend until that happens, just as a human player has a quandry to way increase income to wage war.

Then the AI should prioritize against which empire is closest and so limits its expansion to enlarge its territory.

[UesugiKenshin]

That is why BotE's AI does random skirmishes to weaken the human player way too early. What the AI should do is prioritize its economy to develop a mechanism adding to defensive forces and then begin offensive units to take uncontested territory and then way later contested territory.

You could start with a huge map and two main empires and very few minor aliens and so the AI can expand with little opposition.

A smarter way is have two modders play at least 200 turns amd develop those empires without war. Whoever loses units first will likely end up weaker. You can't afford to buy them or replace them until the economy is established.

The idea of starting from scratch with a 1000 credits is inherently crippling.

The idea of minor alien races is compelling, but then the AI major empire has very limited expansion. I believe if you do both preset empires and many minor alien races, then those will be the immediate threat, not the human player on the opposits end of a huge galaxy.

[UesugiKenshin]

Diplomacy

In most 4X games, with some exceptions, the AI has a demeanor which is largely neutral and then is modified by history as turns take place.

BotE has attributes and the AI by proximity sees a competitor and that some have similar ethos and so may be potentially temporary allies. And that is a misnomer as the minor alien races are locked in their sector when "barbarians" are typically not. Thus real alliances can exist in a fashion and may hold.

I doubt any alliances are practical in BotE and are unlikely to persist.

The standard "carrot" that lures empires into trade and then increasing trust and treaties is PROFIT as trade income may be 3 X ordinary income. The only way to win is promoting trade so you get better military units.

This does not exist in BotE. Trade hurts you especially after turn 300 as the minor and major aliens have unlimited resources and the empires need resource routing badly.

[UesugiKenshin]

Standard 4X games can have lots of empires in the double digits.

BotE has six and sabotage induces asymmetric warfare that is actually fierce when compared to standard warfare with the limitations of range, speed, and exspensive starships and units. It's smarter to sabotage at little risk due to proximity and some have such MONUMENTAL amounts of intelligence points for missions that others can never acquire.

Sometime calculate how many vast resources are required to build a battleship. It's HUGE and that doesn't even account for construction time and complexity of resource routing. Even to make minor mid level troops takes far more than most games.

Even by turn 500-600, six empires and asymmetric warfare ruins the game.

[UesugiKenshin]

Wars begin over competition for land and resources. If you have 5 sectors with six resources, you have an infinite amount in BotE. Seriously. The central issue in BotE is not the infinite resources in storage,but the artificial cap of 125,000 per resource and commision reductions which is just thievery, and the lack of adequate resource routes. You are artificially creating problems that do not in fact exist. Why would a sector of 50 billion only have 4 resource routes??? And worse, the vast majority of my time in hundreds of hours was spent on resource routes being created and canceled and looking for ones that serves no purpose. You have infinity resources but crippled at 125,000 and then have to look for a spare resource slot elsewhere. It's maddening.

It's make work.

War in a galaxy makes no sense for the first 600 turns in BotE due to the plausible expansion amounts. But AI theory induces conflict because that is the expectation.

IF the empires are established, then conflict based in proximity, threat assessment, military force dispersion, diplomatic relations, length of peace, and the need to expand then are the impetus for war. That is why early conflict has no sound military basis to exist in BotE.

[UesugiKenshin]

It's easy to exploit the sectors to get fanatical morale to grow to the maximum amounts and then new growth rates can be HUGE and are easily generated and no issues exist from endless colony ships, starships, units, whatever. There is such abundance that war makes little sense.

The sectors have no mechanism to please the people as I guess it's pointless as there is maximal fanatical morale and no scarcity.

I'm showing you what exists. Obviously I adore the game, but between the language barrier, the documentation issues, the lack of a mechanism to mod on the Android, the random maps, the starting from scratch, the dearth of credits at the genesis, these are probably why it's not wildly popular when it should be with the smallest changes.

[UesugiKenshin]

BotE has an immense potential to story driven missions and learn about LOTS of major and minor aliens, but drops the ball. Then they become generic when they are diverse.

BotE has the potential for complicated unit composition and formations, but does not do so. Multiplayer scenarios with established militaries could be terrific, but don't exist.

We don't have even rudimentary hex editing primers for save games.

[UesugiKenshin]

Philosophically, BotE breaks every standard 4X rules, and succeeds to a tiny subset of 4X players because it's configured that way. When in my opinion, it has the potential of being one of the top games of all time that I have experienced in six decades of every kind of strategy game on any format. But it falters as tiny program changes aren't being remedied and modders lack the tools to fix it...on the Android platform.

It's marginalized on standard formats as it can't compete due to its dated presentation.

It could just take off on the Android, even now, with small practical changes.

Read the comments about BotE and what I am saying is not new. Players intuit these things. Some players recognize how great it is but are frustrated by sabotage.

[UesugiKenshin]

You have programs like Uciana, FreeOrion, and Endless Space, but in my opinion, many of these have more "eye candy" when BotE has a simple elegant interface that does more, and has more untapped potential to do amazing things. Like say every time a planet is colonized, then a percent chance that indigenous lifeforms are found. So any genuine hardcore science fiction fan would smile at the idea of discovering say a dolphin species across the galaxy. And then ethical issues would exist in terraforming.

Or say all human infections are caused by viruses, bacteria, fungus, or parasites. And while a virus is more like a biochemical robot (or say prions which are misfolded proteins that induce abberant protein formation), then the other 3 are lifeforms. So say BotE then reduces the population cap from say 50 billion to 49 billion as some are infectious and will affect population maximums.

Or some of these lifeforms will have an effect, so on say a D or O or L world, there is a building. And if you elect to create it, and it takes energy, then it positively raises a helpful sector attribute like sector income, but has a strong subtraction of morale. Why? It is meddling with a semi-intelligent alien mammalian or simian species. Some locals in the sectors are subversives and they mount protests that the Omega are exploiting the evolutionary process tantamount to slavery.

Say a edible alien goat is found on a K class planet. Is it wrong to eat it and raise it as livestock and therefore increases income? So then you end up helping income in the sector, but then a PETA subversive group forms or even lowers morale empire wide as some complain that they should only use food replicators.

Would bored colonists then start hunting another large mammalian species that then is akin to the buffalo that was crucial to the Sioux Native Americans. What if a protohumanoid species is found and to help them develop in that sector, you organize a livestock operation so instead of rarer hunts, you have organized animal husbandry? That affects two species and maybe the ordinary development of the sentient one.

BotE can do very interesting things, but isn't doing them. But modders could IF the process was facilitated by allowing the save game data to modify buildings, ships, troops, starting income, etc.

[UesugiKenshin]

In the early days before artificial heart valves, then pig heart valves were used. Imagine a future where they are grown by genetic experimentation but involving a semi-intelligent species in a sector. Then some scientists confide that that semi-intelligent species is on the cusp of sentience. If you make the building, it confers a benefit, but causes a morale outrage that might even affect diplomacy.

It's a shame to have diverse extraterrestrials and then not do anything with them that affects the storytelling of the game.

Ponder this. In a game like Medieval 2 Total War, players battle to take settlements from countries. BotE deals in sectors comprised of planets. In a 40x40 map (1600 sectors), is it ethical to invade a sector when you own 400 sectors and your sphere of influence is 700 sectors of spacetime? Why is that 1 sector, that is already inhabited, more important than an uninhabited one?

BotE could then raise all kinds of moral and ethical questions about whether it is personally or corporately wrong to infringe upon the natural rights of extraterrestrials in a sector to live in autonomy.

[UesugiKenshin]

If say the Terrans refuse to invade that 1 sector as it is ethically wrong to do so, and the Khaoran do invade to exploit the sector, do the Terrans have the right to liberate the 1 sector to free the local sentient extraterrestrials and proto-sentient and semi-intelligent species of that 1 sector?

It is ethical to sacrifice not only mining resources and troops and starships plus time to wage war when it's not the territory of the Terrans? Who made the Terrans the arbiters of what is ethical? If the Terrans liberate it, then who controls it? Have the Terrans then just become occupiers no different than the Khaorans?

Does the occupation then become a transition period so the local extraterestials then break away to a friendship or even just a trade status, or are they in effect member vassals?

[UesugiKenshin]

Say the Omega discover that certain planet classes confers a way to safely do cloning and so significantly raise the population cap. But then these clones live a life of ten years and then die a cruel painful death as cloning exacerabtes cancer that always enters systemic aggressive metastasis.

The clones never have a childhood, just live briefly as adults. It does however raise the research points by 25% in the sector cumulatively added to other research processes.

The cruel brief lives of clones then monumentally alter the scientific ability of the Omega allowing a far faster way to gain powerful starships.

Is that ethical as it certainly would preserve the Omega from Cartare domination?

[UesugiKenshin]

In every case, the AI will choose what is abberant and unethical as it then will win. Which is why alliances are always either temporary or incorrect as they will always be your enemies.

There are whole topics in GalCiv 2 where people discussed why you should play as an evil race as there is no benefit to being "good and right".